FOREWORD
espite nearly 20 years of clinical experience with MRI, new instrumentation, new
D technologies, and new applications continue to emerge at a staggering rate.
Understanding the technological change—the new techniques—and implementing
them effectively in the clinical setting—poses a vexing challenge for many MR users,
physicians and technologists alike. In some instances, this has led to a reluctance by
radiologists to implement new techniques and has limited the range of examinations
offered, and, equally importantly, has resulted in the performance capabilities of current
MR systems not being fully utilized in clinical settings. The rate of change in technical
approaches is daunting as well. The abundance of new software capabilities and their
continual modification serve as a further barrier, discouraging many in the field from
using the most updated versions.
In this context, Current Protocols in Magnetic Resonance Imaging is a welcome
addition. The volume contains thorough descriptions of protocols and technical ap-
proaches to MRI applications throughout the body. The chapters are clearly organized,
with step-by-step details of scan protocols. Concise tables are provided highlighting
the key parameters, and succinct descriptions are provided of clinical issues which may
effect the scan protocols in individual situations. Dr. Haacke has assembled a most
talented group of experts to provide their current protocols as well as the rationale for
their implementation.
An intriguing feature of the current text is the commitment to ongoing quarterly
updates, so that the readers can be assured of continuing to have state-of-the-art
approaches in areas of clinical MR applications. This will allow MR users a reliable
way to understand and implement the latest approaches in MR imaging. This text
should prove to be a valuable reference that is found in scan control rooms around
the world.
Herbert Y. Kressel, MD
Radiologist-in-Chief
Beth Israel Deaconess Medical Center
Miriam H. Stoneman Professor of Radiology
Harvard Medical School
Boston, Massachusetts
Contributed by Herbert Y. Kressel
Current Protocols in Magnetic Resonance Imaging (2001)
Copyright © 2001 by John Wiley & Sons, Inc.
PREFACE
agnetic resonance imaging is a powerful noninvasive tool for imaging the human
M body. Over the past twenty years, since its inception, a variety of important core
procedures for imaging disease have been developed. Despite the fact that spin echo
sequences or their variants remain present in most protocols, numerous other methods
have come into being to address not only morphological issues but also functional aspects
of the human body. And there is no end in sight to potential advances in the field. New
methods are constantly coming onto the scene, but only a few make it into clinical practice.
Nevertheless, there is a plethora of material in the literature, and for this reason we felt it
time to tame this protocol beast and have a forum within which a hierarchy of protocols
could be established and regularly updated. Of course, this mandates a publication that is
able to metamorphose, a form of living text per se, in order both to build a complete
coverage of methodology in the field and to stay current.
There are a number of serious challenges to this endeavor. First, there are a wide variety
of field strengths that have been used to image humans, from 0.05 T all the way to 8.0 T
today, with the most common range being from 0.2 T to 1.5 T. Second, there are numerous
manufacturers of these systems, all with their own unique protocols. Third, as systems
become out of date, even among those offered by a given manufacturer, hardware and
often software change, and so what was available on one system might not be on the next
generation of the same system. These issues will take time to deal with, and we welcome
comments from the MR community on new or alternative protocols.
Each unit contains an introduction, the prescribed steps to run the scan, the scan
parameters, troubleshooting information, clinical comments, and references. This tem-
plate is, in some sense, repetitive for each unit, as many of the initial set-up steps for the
patient are the same for each protocol. However, rather than make the reader hunt for
commonalities, we chose to present each unit as a complete source in and of itself.
The present issue contains two sections, “established clinical protocols” and “educational
material.” We anticipate adding three further sections: one on “clinical research protocols”
(which may or may not eventually metamorphose into accepted protocols and then move
up to the first section), one on “quality assurance protocols,” and one on “animal
protocols” as a final section.
This publication is available in looseleaf format and is updated quarterly. We anticipate
it becoming available on the Web eventually. Subjects in this manual are organized by
sections and then chapters, with each chapter subdivided into units. Page numbering
reflects this modular arrangement; for example, page A5.1.7 refers to the clinical proto-
cols section, Chapter A5 (Miscellaneous Brain Pathology), Unit 1 (Multiple Sclerosis),
and page 7 of that particular unit.
The basic tenet of this manual is simple: to have a resource which can be accessed to run
any protocol from start to finish in magnetic resonance imaging, and provide you with
the understanding of what is being done, why it is being done, and any difficulties that
may arise during the process. On behalf of all the editorial board and contributors to
CPMRI, we hope you find this resource an aid in your everyday practice of clinical
magnetic resonance imaging.
Current Protocols Current Protocols
in Magnetic in Magnetic
Resonance Resonance
Imaging Imaging
i Contributed by E. Mark Haacke and Weili Lin iii
Current Protocols in Magnetic Resonance Imaging (2003) iii-iv
Copyright © 2003 by John Wiley & Sons, Inc. Supplement 8
 ACKNOWLEDGMENTS
We would like to thank Virginia Chanda, Scott Holmes, Allen Ranz, Susan Lieberman,
Tom Downey, Joseph White, Michael Gates, Mary Keith Trawick, Tom Cannon, and the
rest of the staff from the Current Protocols division of John Wiley & Sons for their
continuing efforts in the production process. We also wish to thank publishers Shawn
Morton and Susan King of John Wiley & Sons for their early involvement and encour-
agement, Marguerite Devers for her indexing, Jan Koonce of Washington University in
St. Louis, Rachel Martis-Laze and Diane Zaltsman of the MRI Institute for Biomedical
Research, Lisa Brownschidle of Wayne State University, and Lucy Rosenberg of Case
Western Reserve University for their secretarial support, Chia-Chi Chang and William G.
Sherwin for manuscript editing, and Azim Celik (at General Electric), Yiu-Cho Chung (at
Siemens), Michael R. Thompson (at Philips), and Yi Wang (at St. Francis Hospital) for
their technical discussions.
For the editors,
E. Mark Haacke, Editor-in-Chief
Weili Lin, Associate Editor-in-Chief
Preface
iv
Supplement 8 Current Protocols in Magnetic Resonance Imaging
Screening Forms for Patients for APPENDIX 1
MR Procedures and Individuals for the
MR Environment
The establishment of thorough and effective screening procedures for patients and other
individuals is one of the most critical components of a program that guards the safety of
all those preparing to undergo MR procedures or to enter the MR environment (Shellock
and Kanal, 1991, 1994, 1996; Sawyer-Glover and Shellock, 2000, 2001; Shellock, 2001;
Kanal et al. 2002). An important aspect of protecting patients and individuals from MR
system–related accidents and injuries involves an understanding of the risks associated
with the various implants, devices, accessories, and other objects that may cause problems
in this setting (Shellock and Kanal, 1991, 1994, 1996; Sawyer-Glover and Shellock, 2000,
2001; Shellock, 2001, 2002a,b, 2003; Kanal et al., 2002; Shellock and Crues, 2002). This
requires constant attention and diligence to obtain information and documentation about
these objects in order to provide the safest MR setting possible. In addition, because most
MR-related incidents have been due to deficiencies in screening methods and/or a lack
of properly controlled access to the MR environment (especially with regard to preventing
personal items and other potentially problematic objects from getting into the MR system
room), it is crucial to set up procedures and guidelines to prevent such incidents from
occurring.
MAGNETIC RESONANCE (MR) PROCEDURE SCREENING FOR PATIENTS
Certain aspects of screening patients for MR procedures may take place during the
scheduling process. This should be conducted by a healthcare worker that is specially
trained in MR safety (i.e., this person should be trained to understand the potential hazards
and issues associated with the MR environment and MR procedures and be familiar with
all of the information contained on the screening forms for patients and individuals).
During this time, it may be ascertained if the patient has any implant that may be
contraindicated for the MR procedure (e.g., a ferromagnetic aneurysm clip, pacemaker,
etc.) or if there is any condition that needs careful consideration (e.g., the patient is
pregnant, has a disability, etc.). Preliminary screening helps to prevent the scheduling of
patients who may be inappropriate candidates for MR examinations (Sawyer-Glover and
Shellock, 2000, 2001; Shellock, 2003).
After preliminary screening, every patient must undergo a comprehensive screening in
preparation for a magnetic resonance (MR) procedure (i.e., MR imaging, MR angiogra-
phy, functional MRI, MR spectroscopy). Comprehensive patient screening involves the
use of a printed form to document the screening procedure, a review of the information
on the screening form, and a verbal interview to verify the information on the form and
to allow discussion of any questions or concerns that the patient may have. An MR-safety
trained healthcare worker must conduct this aspect of patient screening.
A screening form for patients developed by Sawyer-Glover and Shellock (2000) was
recently revised in consideration of new information in the peer-reviewed literature. This
two-page form, entitled Magnetic Resonance (MR) Procedure Screening Form for
Patients, was also created in conjunction with the Medical, Scientific, and Technology
Advisory Board and the Corporate Advisory Board of the Institute for Magnetic Reso-
nance Safety, Education, and Research (IMRSER). A downloadable version of this form
may be obtained from the MR safety Web sites, http://www.IMRSER.org and
http://www.MRIsafety.com.
Screening Forms
Contributed by Frank G. Shellock A1.1
Current Protocols in Magnetic Resonance Imaging (2003) A1.1-A1.9
Copyright © 2003 by John Wiley & Sons, Inc. Supplement 11
 The first page of this screening form requests general patient-related information (e.g.,
name, age, sex, height, weight, etc.) as well as information regarding the reason for the
MR procedure and/or symptoms that may be present. Pertinent information about the
patient is required not only to ensure that the medical records are up-to-date, but also in
the event that the MR facility needs to contact the referring physician for additional
information regarding the examination or to verify the medical condition of the patient.
The form requests information regarding a prior surgery or operation to help determine
if there may be an implant or device present that could create a problem for the patient.
Information is also requested pertaining to prior diagnostic imaging studies that may be
helpful to review for assessment of the patient’s condition.
Next, important questions are posed in an effort to determine if there are possible problems
or issues that should be discussed with the patient prior to permitting entry to the MR
environment. For example, information is requested regarding any problem with a
previous MR examination, an injury to the eye involving a metallic object, or any injury
from a metallic object or foreign body. Questions are posed to obtain information about
current or recently taken medications as well as the presence of drug allergies. There are
also questions to assess past and present medical conditions that may affect the MR
procedure or the use of an MRI contrast agent in the patient.
At the bottom of the first page, there is a section for female patients that poses questions
that may impact MR procedures. For example, questions regarding the date of the last
menstrual period, pregnancy, or late menstrual period are included. A definite or possible
pregnancy must be identified prior to permitting the patient into the MR environment so
that the risks versus the benefits of the MR procedure can be considered and discussed
with the patient. MR procedures should only be performed in pregnant patients to address
important clinical questions. MR facilities should have a clearly defined procedure to
follow in the event that the patient has a confirmed or possible pregnancy.
Questions pertaining to the date of the last menstrual period and use of oral contraceptives,
hormonal therapy, and fertility medication are necessary for female patients undergoing
MR procedures that are performed to evaluate breast disease or for OB/GYN (obstetrics
gynecology) applications, as these may alter the tissue appearance on MR imaging. An
inquiry about breastfeeding is included in case the administration of MRI contrast media
is being considered for nursing mothers.
The second page of the form has a statement at the top that states: “WARNING: Certain
implants, devices, or objects may be hazardous to you and/or may interfere with the MR
procedure (i.e., MRI, MR angiography, functional MRI, MR spectroscopy). Do not enter
the MR system room or MR environment if you have any questions or concerns regarding
an implant, device, or object. Consult the MRI Technologist or Radiologist BEFORE
entering the MR system room. The MR system magnet is ALWAYS on.”
Next, there is a section that lists various implants, devices, and objects to identify anything
that could be hazardous to the patient undergoing the MR procedure or that may produce
an artifact that could interfere with the interpretation of the MR procedure. In general,
these items are arranged on the checklist in order of the relative safety hazard (e.g.,
aneurysm clip, cardiac pacemaker, implantable cardioverter defibrillator, electronic im-
plant, etc.), followed by items that may simply produce imaging artifacts that could be
problematic for the interpretation of the MR procedure. Additionally, questions are posed
to determine if the patient has a breathing problem, movement disorder, or claustrophobia
because these are known to present difficulties for MR procedures.
Figures of the human body are included on the second page of the screening form for the
patient as a means of showing the location of any object inside of or on the body. This
Screening Forms information is particularly useful so that the patient may indicate the approximate position
A1.2
Supplement 11 Current Protocols in Magnetic Resonance Imaging
of any object that may be hazardous or that could interfere with the interpretation of the
MR procedure by producing an artifact.
The second page of the screening form also has an “Important Instructions” section that
states: “Before entering the MR environment or MR system room, you must remove all
metallic objects including hearing aids, dentures, partial plates, keys, beeper, cell phone,
eyeglasses, hair pins, barrettes, jewelry, body piercing jewelry, watch, safety pins,
paperclips, money clip, credit cards, bank cards, magnetic strip cards, coins, pens, pocket
knife, nail clipper, tools, clothing with metal fasteners, and clothing with metallic threads.
Please consult the MRI Technologist or Radiologist if you have any questions or concerns
BEFORE you enter the MR system room.”
Finally, there is a statement on the “Magnetic Resonance (MR) Procedure Screening Form
for Patients” that indicates hearing protection is “advised or required” to prevent possible
problems or hazards related to acoustic noise. In general, this should not be an option for
a patient undergoing an MR procedure on a high-field-strength (e.g., ≥1.0-Tesla) MR
system. By comparison, it may not be necessary for the use of hearing protection by
patients undergoing MR procedures on low-field-strength MR systems.
It should be noted that undergoing previous MR procedures without incidents does not
guarantee a safe subsequent MR examination. Various factors (e.g., the static magnetic
field strength of the MR system, the orientation of the patient, the orientation of a metallic
implant or object, etc.) can substantially change the scenario. Thus, a written screening
form must be completed each time a patient prepares to undergo an MR procedure. This
is not an inconsequential matter because a surgical intervention or accident involving a
metallic foreign body may have occurred that could impact the safety of an MR procedure
or of entering the MR environment.
With the use of any type of written questionnaire, limitations exist related to incomplete
or incorrect answers provided by the patient. For example, there may be difficulties
associated with patients that are impaired with respect to their vision, language fluency,
or level of literacy. Therefore, an appropriate accompanying family member or other
individual (e.g., referring physician) should be involved in the screening process to verify
any information that may impact patient safety. Versions of this form should also be
available in other languages, as needed (i.e., specific to the demographics of the MR
facility; Sawyer-Glover and Shellock, 2000; Kanal et al. 2002; Shellock, 2003).
In the event that the patient is comatose or unable to communicate, the written screening
form should be completed by the most qualified individual (e.g., physician, family
member, etc.) that has knowledge about the medical history and present condition of the
patient. If the screening information is inadequate, it is advisable to look for surgical scars
on the patient and/or to obtain plain films of the skull and/or chest to search for implants
that are known to be particularly hazardous in the MR environment (e.g., aneurysm clips,
cardiac pacemakers, etc.).
Following completion of the “Magnetic Resonance (MR) Procedure Screening Form for
Patients,” an MR-safety trained healthcare worker should review the form’s content. Next,
a verbal interview should be conducted by the MR-safety trained healthcare worker to
verify the information on the form and to allow discussion of any questions or concerns
that the patient may have before undergoing the MR procedure. This allows for clarifica-
tion or confirmation of the answers to the questions posed to the patient so that there is
no miscommunication regarding important MR safety issues. In addition, because the
patient may not be fully aware of the medical terminology used for a particular implant
or device, it is imperative that this particular information on the form be discussed during
the verbal interview.
Screening Forms
A1.3
Current Protocols in Magnetic Resonance Imaging Supplement 11
 After the comprehensive screening procedure is completed, any patient that is transferred
by a stretcher, gurney, or wheelchair to the MR system room should be checked thoroughly
and systematically for metal objects under the sheets or blankets such as ferromagnetic
oxygen tanks, monitors, or other objects that could pose a hazard.
MAGNETIC RESONANCE (MR) ENVIRONMENT SCREENING FOR
INDIVIDUALS
Before any “non-patient” individual (e.g., MRI technologist, MR support person, family
member, visitor, allied health professional, physician, maintenance worker, custodial
worker, fire fighter, security officer, etc.) is allowed into the MR environment, he or she
must be screened by an MR-safety trained healthcare worker. Proper screening for
individuals involves the use of a printed form to document the screening procedure, a
review of the information on the form, and a verbal interview to verify the information
on the form and to allow discussion of any questions or concerns that the individual may
have before permitting entry into the MR environment.
In general, magnetic resonance (MR) screening forms were developed with patients in
mind and, therefore, pose many questions that are inappropriate or confusing to other
individuals that may need to enter the MR environment. Therefore, a screening form was
recently created specifically for individuals that need to enter the MR environment and/or
an MR system room. This form, entitled “Magnetic Resonance (MR) Environment
Screening Form for Individuals,” was developed in conjunction with the Medical,
Scientific, and Technology Advisory Board and the Corporate Advisory Board of the
Institute for Magnetic Resonance Safety, Education, and Research (IMRSER). A down-
loadable version of this form may be obtained from the MR safety Web sites,
http://www.IMRSER.org and http://www.MRIsafety.com.
At the top of this form, the following statement is displayed: “The MR system has a very
strong magnetic field that may be hazardous to individuals entering the MR environment
or MR system room if they have certain metallic, electronic, magnetic, or mechanical
implants, devices, or objects. Therefore, all individuals are required to fill out this form
BEFORE entering the MR environment or MR system room. Be advised, the MR system
magnet is ALWAYS on.”
The “Magnetic Resonance (MR) Environment Screening Form for Individuals” requests
general information (e.g., name, age, address, etc.) and poses important questions to
determine if there are possible problems or issues that should be discussed with the
individual prior to permitting entry to the MR environment. A warning statement is also
provided on the form, as follows: “WARNING: Certain implants, devices, or objects may
be hazardous to you in the MR environment or MR system room. Do not enter the MR
environment or MR system room if you have any questions or concerns regarding an
implant, device, or object.” In addition, there is a section that lists various implants, devices,
and objects to identify the presence of anything that could be hazardous to an individual
in the MR environment (e.g., an aneurysm clip, cardiac pacemaker, implantable
cardioverter defibrillator (ICD), electronic or magnetically activated device, metallic
foreign body, etc.).
Finally, there is an “Important Instructions” section on the form that states: “Remove all
metallic objects before entering the MR environment or MR system room including
hearing aids, beeper, cell phone, keys, eyeglasses, hair pins, barrettes, jewelry (including
body piercing jewelry), watch, safety pins, paperclips, money clip, credit cards, bank
cards, magnetic strip cards, coins, pens, pocket knife, nail clipper, steel-toed boots/shoes,
and tools. Loose metallic objects are especially prohibited in the MR system room and
Screening Forms
A1.4
Supplement 11 Current Protocols in Magnetic Resonance Imaging
MR environment. Please consult the MRI Technologist or Radiologist if you have any
questions or concerns BEFORE you enter the MR system room.”
The proper use of this written form along with thorough verbal screening of the individual
by an MR-safety trained healthcare worker should prevent accidents and injuries in the
MR environment (Sawyer-Glover and Shellock, 2000; Kanal et al. 2002; Shellock, 2003).
ACKNOWLEDGMENT
Portions of this text were adapted with permission from Sawyer-Glover and Shellock
(2001) The screening forms, “Magnetic Resonance (MR) Procedure Screening Form For
Patients” and “Magnetic Resonance (MR) Environment Screening Form for Individuals,”
were developed in conjunction with the Medical, Scientific, and Technology Advisory
Board and the Corporate Advisory Board of the Institute for Magnetic Resonance Safety,
Education, and Research (IMRSER) (Shellock, 2003).
LITERATURE CITED
Kanal, E., Borgstede, J.P., Barkovich, A.J., Bell, C., Bradley, W.G., Felmlee, J.P., Froelich, J.W., Kaminski,
E.M., Keeler, E.K., Lester, J.W., Scoumis, E.A., Zaremba, L.A., Zinninger, M.D. 2002. American college
of radiology white paper on MR safety. Am. J. Roentgen. 178:1335-1347.
Sawyer-Glover, A. and Shellock, F.G. 2000. Pre-MRI procedure screening: Recommendations and safety
considerations for biomedical implants and devices. J. Magn. Reson. Imag. 12:92-106.
Sawyer-Glover, A. and Shellock, F.G. 2001. Pre-magnetic resonance procedure screening. In Magnetic
Resonance Procedures: Health Effects and Safety (F.G. Shellock, Ed.). CRC Press, LLC, Boca Raton,
Fla.
Shellock, F.G. 2001. New recommendations for screening patients for suspected orbital foreign bodies.
Signals, No. 36, Issue 4. pp. 8-9.
Shellock, F.G. 2002a. Biomedical implants and devices: Assessment of magnetic field interactions with a
3.0-Tesla MR system. J. Magn. Reson. Imaging 16:721-732.
Shellock, F.G. 2002b. MR safety update 2002: Implants and devices. J. Mag. Res. Imaging 16:485-496.
Shellock, F.G. 2003. Reference Manual for Magnetic Resonance Safety. Edition. Amirsys, Salt Lake City, Utah.
Shellock, F.G. and Crues, J.V. 2002. Commentary. MR safety and the American College of Radiology White
Paper. Am. J. Roentgen. 178:1349-1352.
Shellock, F.G. and Kanal, E. 1991. Policies, guidelines, and recommendations for MR imaging safety and
patient management. J. Magn. Reson. Imaging 1:97-101.
Shellock, F.G. and Kanal, E. 1994. SMRI Report. Policies, guidelines and recommendations for MR imaging
safety and patient management. Questionnaire for screening patients before MR procedures. J. Magn.
Reson. Imaging 4:749-751.
Shellock, F.G. and Kanal, E. 1996. Magnetic Resonance: Bioeffects, Safety, and Patient Management. Second
Edition, Lippincott-Raven Press, New York.
INTERNET RESOURCES
http://www.MRIsafety.com
http://www.IMRSER.org
Contributed by Frank G. Shellock
Keck School of Medicine
University of Southern California and
Institute for Magnetic Resonance Safety, Education, and Research
Los Angeles, California
Screening Forms
A1.5
 MAGNETIC RESONANCE (MR) PROCEDURE SCREENING FORM FOR PATIENTS

Date _____/_____/_____ Patient Number ______________________

Name ______________________________________________ Age ________ Height ________ Weight ________

Last name First name Middle Initial

Date of Birth _____/_____/_____ Male Ì Female Ì Body Part to be Examined _________________________

month day year

Address ____________________________________________ Telephone (home) (_____) _____-________

City ____________________________________________ Telephone (work) (_____) _____-________

State _______________________ Zip Code ___________

Reason for MRI and/or Symptoms ___________________________________________________________________________

Referring Physician __________________________________ Telephone (_____) _____-______________

1. Have you had prior surgery or an operation (e.g., arthroscopy, endoscopy, etc.) of any kind? Ì No Ì Yes
If yes, please indicate the date and type of surgery:
Date _____/_____/_____ Type of surgery _________________________________________________________
Date _____/_____/_____ Type of surgery _________________________________________________________
2. Have you had a prior diagnostic imaging study or examination (MRI, CT, Ultrasound, X-ray, etc.)? ÌNo Ì Yes
If yes, please list: Body part Date Facility
MRI _________________________ _____/_____/_____ ____________________________________
CT/CAT Scan _________________________ _____/_____/_____ ____________________________________
X-Ray _________________________ _____/_____/_____ ____________________________________
Ultrasound _________________________ _____/_____/_____ ____________________________________
Nuclear Medicine _________________________ _____/_____/_____ ____________________________________
Other__________ _________________________ _____/_____/_____ ____________________________________

3. Have you experienced any problem related to a previous MRI examination or MR procedure? Ì No Ì Yes
If yes, please describe: ________________________________________________________
4. Have you had an injury to the eye involving a metallic object or fragment (e.g., metallic slivers,
shavings, foreign body, etc.)? Ì No Ì Yes
If yes, please describe: ________________________________________________________
5. Have you ever been injured by a metallic object or foreign body (e.g., BB, bullet, shrapnel, etc.)? Ì No Ì Yes
If yes, please describe: ________________________________________________________
6. Are you currently taking or have you recently taken any medication or drug? Ì No Ì Yes
If yes, please list:_____________________________________________________________
7. Are you allergic to any medication? Ì No Ì Yes
If yes, please list:_____________________________________________________________
8. Do you have a history of asthma, allergic reaction, respiratory disease, or reaction to a contrast
medium or dye used for an MRI, CT, or X-ray examination? Ì No Ì Yes
9. Do you have anemia or any disease(s) that affects your blood, a history of renal (kidney)
disease, or seizures? Ì No Ì Yes
If yes, please describe: ________________________________________________________

For female patients:

10. Date of last menstrual period:_____/_____/_____ Post menopausal? Ì No Ì Yes
11. Are you pregnant or experiencing a late menstrual period? Ì No Ì Yes
12. Are you taking oral contraceptives or receiving hormonal treatment? Ì No Ì Yes
13. Are you taking any type of fertility medication or having fertility treatments? Ì No Ì Yes
If yes, please describe: ________________________________________________________
14. Are you currently breastfeeding? Ì No Ì Yes
 WARNING: Certain implants, devices, or objects may be hazardous to you and/or may interfere with the MAGNETIC RESONANCE (MR) ENVIRONMENT SCREENING FORM FOR INDIVIDUALS*
MR procedure (i.e., MRI, MR angiography, functional MRI, MR spectroscopy). Do not enter the MR system room
or MR environment if you have any question or concern regarding an implant, device, or object. Consult the MRI
Technologist or Radiologist BEFORE entering the MR system room. The MR system magnet is ALWAYS on. The MR system has a very strong magnetic field that may be hazardous to individuals entering the
MR environment or MR system room if they have certain metallic, electronic, magnetic, or mechanical
implants, devices, or objects. Therefore, all individuals are required to fill out this form BEFORE entering
the MR environment or MR system room. Be advised, the MR system magnet is ALWAYS on.
Please indicate if you have any of the following:
Ì Yes Ì No Aneurysm clip(s) Please mark on the figure(s) below *NOTE: If you are a patient preparing to undergo an MR examination, you are required to fill out a different form.
Ì Yes Ì No Cardiac pacemaker the location of any implant or metal
Ì Yes Ì No Implanted cardioverter defibrillator (ICD) inside of or on your body.
Ì Yes Ì No Electronic implant or device Date _____/_____/_____ Name ____________________________________________________ Age _______
Ì Yes Ì No Magnetically-activated implant or device month day year Last Name First Name Middle Initial
Ì Yes Ì No Neurostimulation system
Ì Yes Ì No Spinal cord stimulator Address __________________________________________ Telephone (home) (_____) _____-________
Ì Yes Ì No Internal electrodes or wires
City __________________________________________ Telephone (work) (_____) _____-________
Ì Yes Ì No Bone growth/bone fusion stimulator
Ì Yes Ì No Cochlear, otologic, or other ear implant
State ____________________ Zip Code ___________
Ì Yes Ì No Insulin or other infusion pump
Ì Yes Ì No Implanted drug infusion device
1. Have you had prior surgery or an operation (e.g., arthroscopy, endoscopy, etc.) of any kind? Ì No Ì Yes
Ì Yes Ì No Any type of prosthesis (eye, penile, etc.)
If yes, please indicate date and type of surgery: Date ____/____/____ Type of surgery________________
Ì Yes Ì No Heart valve prosthesis
2. Have you had an injury to the eye involving a metallic object (e.g., metallic slivers, foreign body)? Ì No Ì Yes
Ì Yes Ì No Eyelid spring or wire
If yes, please describe: _____________________________________________________________________
Ì Yes Ì No Artificial or prosthetic limb
3. Have you ever been injured by a metallic object or foreign body (e.g., BB, bullet, shrapnel, etc.)? Ì No Ì Yes
Ì Yes Ì No Metallic stent, filter, or coil
If yes, please describe: _____________________________________________________________________
Ì Yes Ì No Shunt (spinal or intraventricular)
4. Are you pregnant or suspect that you are pregnant? Ì No Ì Yes
Ì Yes Ì No Vascular access port and/or catheter
Ì Yes Ì No Radiation seeds or implants
Ì Yes Ì No Swan-Ganz or thermodilution catheter WARNING: Certain implants, devices, or objects may be hazardous to you in the MR environment or
Ì Yes Ì No Medication patch (Nicotine, Nitroglycerine) MR system room. Do not enter the MR environment or MR system room if you have any question or concern
Ì Yes Ì No Any metallic fragment or foreign body regarding an implant, device, or object.
Ì Yes Ì No Wire mesh implant IMPORTANT INSTRUCTIONS i l d i bj
Ì Yes Ì No Tissue expander (e.g., breast)
Ì Yes Ì No Surgical staples, clips, or metallic sutures Before entering the MR environment or MR system Please indicate if you have any of the following:
Ì Yes Ì No Joint replacement (hip, knee, etc.) room, you must remove all metallic objects including Ì Yes Ì No Aneurysm clip(s) IMPORTANT INSTRUCTIONS
Ì Yes Ì No Bone/joint pin, screw, nail, wire, plate, etc. hearing aids, dentures, partial plates, keys, beeper, cell Ì Yes Ì No Cardiac pacemaker
Ì Yes Ì No IUD, diaphragm, or pessary phone, eyeglasses, hair pins, barrettes, jewelry, body Ì Yes Ì No Implanted cardioverter defibrillator (ICD)
piercing jewelry, watch, safety pins, paperclips, money Ì Yes Ì No Electronic implant or device Remove all metallic objects before entering the MR
Ì Yes Ì No Dentures or partial plates
clip, credit cards, bank cards, magnetic strip cards, Ì Yes Ì No Magnetically-activated implant or device environment or MR system room including hearing
Ì Yes Ì No Tattoo or permanent makeup
Ì Yes Ì No Body piercing jewelry coins, pens, pocket knife, nail clipper, tools, clothing Ì Yes Ì No Neurostimulation system aids, beeper, cell phone, keys, eyeglasses, hair pins,
Ì Yes Ì No Hearing aid with metal fasteners, & clothing with metallic threads. Ì Yes Ì No Spinal cord stimulator barrettes, jewelry (including body piercing jewelry),
(Remove before entering MR system room) Ì Yes Ì No Cochlear implant or implanted hearing aid watch, safety pins, paperclips, money clip, credit
Ì Yes Ì No Other implant _______________________ Please consult the MRI Technologist or Radiologist if Ì Yes Ì No Insulin or infusion pump cards, bank cards, magnetic strip cards, coins, pens,
Ì Yes Ì No Breathing problem or motion disorder you have any question or concern BEFORE you enter Ì Yes Ì No Implanted drug infusion device pocket knife, nail clipper, steel-toed boots/shoes, and
Ì Yes Ì No Claustrophobia the MR system room. Ì Yes Ì No Any type of prosthesis or implant tools. Loose metallic objects are especially prohibited
Ì Yes Ì No Artificial or prosthetic limb in the MR system room and MR environment.
NOTE: You may be advised or required to wear earplugs or other hearing protection during Ì Yes Ì No Any metallic fragment or foreign body
the MR procedure to prevent possible problems or hazards related to acoustic noise. Ì Yes Ì No Any external or internal metallic object
Ì Yes Ì No Hearing aid Please consult the MRI Technologist or Radiologist if
(Remove before entering the MR system room) you have any question or concern BEFORE you enter
I attest that the above information is correct to the best of my knowledge. I read and understand the contents of this form and had the
opportunity to ask questions regarding the information on this form and regarding the MR procedure that I am about to undergo. Ì Yes Ì No Other implant______________________ the MR system room.

Signature of Person Completing Form: _______________________________________ Date _____/_____/_____ I attest that the above information is correct to the best of my knowledge. I have read and understand the entire contents of this
Signature form and have had the opportunity to ask questions regarding the information on this form.

Form Completed By: Ì Patient Ì Relative Ì Nurse ____________________________________ ____________________________ Signature of Person Completing Form: ______________________________________ Date _____/_____/_____
Print name Relationship to patient Signature

Form Information Reviewed By: ____________________________________________ ____________________________________ Form Information Reviewed By: _____________________________________ ______________________________________
Print name Signature Print name Signature

Ì MRI Technologist Ì Nurse Ì Radiologist Ì Other___________________________________________ Ì MRI Technologist Ì Radiologist Ì Other ______________________________
Imaging Parameters Used in the Sequence
Tables
The purpose of this appendix is to elaborate on some of the parameters used in the
sequence tables included in the individual units. This may help the reader to understand
those imaging parameters when it is necessary to translate them into parameters for a
particular scanner, which will vary depending upon the manufacturer. The reader is also
referred to UNIT A7.4 because some additional explanation is provided there.
Cine gradient echo sequence: In this sequence, the repeat time (TR) listed is the effective
repeat time—i.e., temporal resolution.
Delay time (TD): Sometimes this is referred to as the delay time after the R wave.
Display matrix (Dx, Dy): This corresponds to the size of an image. The reader should not
confuse this with “number of data points collected.”
Echo train length (ETL): We define this as the number of k-space lines collected after
the π/2-pulse in fast spin echo sequences. This parameter belongs to a subgroup of the
“number of lines per segment” collected in k-space.
Fast spin echo: This sequence is also called a turbo spin echo sequence.
Flip angle (FA): Generally speaking, the flip angle is the angle whose value is displayed
on the MR console. For some spin echo sequences, the system displays the angle of the
refocusing pulse instead of the π/2 pulse.
Multi-phase: This is the terminology used by GE; it is similar to the number of
repetitions, but is usually used when the sequence is a perfusion or MR angiography scan
and it is desired to continually collect data after the injection of a contrast agent.
Number of acquisitions (Nacq): This is the number of times that a sequence is repeated,
but the same number of images is produced whether Nacq = 1 or some larger integer.
Increasing the number of acquisitions increases the signal-to-noise ratio by the square
root of the number of acquisitions. One should not confuse this parameter with number
of repetitions. When this parameter is less than one, it means partial Fourier is used in the
sequence. To GE users, this parameter is equivalent to the “number of excitations”
parameter which is called NEX. In Marconi’s system, this term is referred to as the
“number of signal averages.”
Number of acquisitions on a GE system: Unfortunately, GE uses this term to refer to
the number of repetitions of a scan when there is insufficient time for short TR sequences
to collect all the slices of interest. For example, if only 10 slices can be run with a given
TR and TE and 30 slices are required to cover the region of interest (ROI), GE scanners
will run this scan three times but appropriately shift the slices to ensure the full coverage
of the ROI.
Number of cardiac phases: This is the number of phases (or frames or images) of one
slice that can be obtained in one heartbeat. This number is usually calculated as (R-to-R
interval) multiplied by 85%, divided by TR.
Number of data points collected (Nx, Ny): This is the acquisition matrix and it is the
one that the reader needs to choose on the system before running the scan. Nx is the number
of points collected in the read direction and Ny is the number of phase-encoding steps. If
Current Protocols in Magnetic Resonance Imaging (2004) A2.1-A2.2
Copyright © 2004 by John Wiley & Sons, Inc.
APPENDIX 2 the sequence swaps the read and phase encoding directions, then Nx will be the number
of phase-encoding steps and Ny will be the sampling points in the read direction. This
setting is not the same as the “display matrix” (see above) which represents the number
of pixels used to display an image.
Number of excitations (NEX): See Nacq above.
Number of lines per segment: This is the number of k -space lines that can be collected
in one TR for a 2-D slice or a 3-D slab.
Number of repetitions: This is the number of times that a sequence is repeated. The
number of images is proportional to this setting. Increasing the number of repetitions will
increase the scan time but will not increase the signal-to-noise ratio. In Siemens’
terminology, this is called the “number of measures.”
Number of slices: This is also referred to as the “number of partition encoding steps” if
a 3-D sequence is performed.
Read direction: This is also called the frequency direction.
Receiver bandwidth (RBW): This value has been quoted as plus or minus a frequency
(e.g., ±12 kHz) for GE’s scanners. This means that, in GE’s scanners, the sampling time
along the readout direction is Nx/(2 × |receiver bandwidth|). In Siemens Sonota systems,
in order to change the echo time, sometimes it is necessary to change the receiver
bandwidth.
Rectangular fields of view: We usually give the true values of “fields of view (FOVx,
FOVy)” in this series. The aspect ratio for a rectangular field of view is equal to FOVy/
FOVx.
Slab thickness: This is the “number of slices” multiplied by “slice thickness” (Δz).
Swap read and phase encoding: In this series, we assume that the default setting of the
read direction for transverse images is left-right. Thus, in order to avoid ghosting from
the eyes in transverse images in the brain, one would usually swap read and phase-encod-
ing directions. (However, this is not done for an echo planar sequence.) The default setting
of the read direction in the other two orientations is superior-inferior; however, readers
are strongly advised to check with the technicians at their own site for this setting.
Zerofill interpolation process (ZIP): This is an image reconstruction process. An image
can be zero-filled and interpolated to any matrix size for display. For example, if an image
is collected at an acquisition matrix (Nx, Ny) of 256 by 256, then “ZIP 512” means that
this image now has a display matrix (Dx, Dy) of 512 by 512 and thus the in-plane resolution
of this image is doubled. Under the same explanation, “ZIP 2” means that the through-
plane resolution of an image is doubled. This process is accomplished by filling k-space
with zeros out to the desired value and then Fourier-transforming the data.
Imaging Imaging
Parameters Used Parameters Used
in the Sequence in the Sequence
Tables Tables
A2.1 A2.2
Supplement 12 Supplement 12 Current Protocols in Magnetic Resonance Imaging
MR Safety Guide
GENERAL INFORMATION
1. Before anyone (i.e., staff, subject, visitor) enters the magnet room, a screening form
must be completed by the subject and reviewed by the investigators or the technolo-
gists who will be performing the MRI exam.
2. Before entering the magnet room, all pockets are to be emptied of: watches, pagers,
wallets, pens, pencils, hair clips, jewelry, keys, coins, and any other possible projec-
tiles.
APPENDIX 3 MRI ORIENTATION AND TRAINING PROGRAM
To ensure a safe and knowledgeable working environment, all potential MRI operators
must complete formal training as outlined below. Upon satisfactory completion of this
orientation program, only those deemed “Qualified Operators” by the MRI Steering
Committee will be assigned a 1 year safety certificate.
Each MRI operator will complete an orientation in the MR facility via a walk-through
and demonstration of the equipment and safety procedures. All orientation sessions will
be supervised and documented on the form found on the facing page. The orientation and
training will include:

3. It is recommended that the person being scanned change into MR-compatible Nursing procedures
clothing. Screening/safety
Consent forms
Pajama bottoms and gowns are provided for this purpose.
Monitoring equipment
4. The magnet room door should be kept closed at all times except when entering or Emergencies
exiting the room. Code procedures
Keeping the door shut tight will reduce RF noise during imaging. Scanning procedures
Coil selection/setup
5. All persons scanned must wear ear protection. If they decline, the exam must be Positioning
canceled. Pulse sequence selection
6. Never scan with an unplugged surface (or body) coil in the bore of the magnet. Imaging options selection
Scanner startup/shutdown
7. Use of any research equipment, coils, or supplies not supplied by the MRI facility Record keeping/documentation
must be approved by an MR Steering Committee or appropriate official. Archiving/deletion of data
8. All investigators and technologists must clean up after themselves. Return all equip- Filming procedures (if applicable)
ment to its proper place. The room should be kept neat and tidy. Do not lay coils or Processor startup/shutdown
phantoms on the floors. Laser camera startup/shutdown
9. Please report all malfunctioning or damaged equipment to the technologist(s) imme- Filming software
diately. Developing films.

10. All studies should be archived immediately after the exam is complete. Data may
remain on the system for up to 7 days (or some agreed upon time) unless disk space
becomes critical, in which case, the oldest studies will be removed first. If a study
requires extra time on the disks, log it in the “DO NOT REMOVE” logbooks, which
are located in each control room.
11. A physician must be available to cover MR exams where patients are being scanned
or a contrast agent is being administered. For normal volunteers, no medical coverage
is necessary, however, the investigators and technologists must be familiar with code
procedures and calling 911 if a medical emergency should arise.
12. Investigators and technologists should never leave their subjects unattended in the
magnet. If the investigator or technologist must leave the MR area, then he/she must
get another qualified person to cover.
13. Subjects should never be left lying on the scan table outside of the magnet. Due to
the lack of side rails on the table, there is a risk of the subject falling off the table.

MR Safety Guide MR Safety Guide

Contributed by Glenn Jay Foster A3.1 A3.2
Current Protocols in Magnetic Resonance Imaging (2003) A3.1-A3.27
Copyright © 2003 by John Wiley & Sons, Inc. Supplement 11 Supplement 11 Current Protocols in Magnetic Resonance Imaging
TOPIC REVIEWED BY DATE

Nursing procedures
• Screening/safety

• Review screening form

• Review consent form

• Review permission form

• Monitoring equipment
ECG (electrocardiogram)

Blood pressure

Pulse oximeter

• Emergencies

• Code procedures

TOPIC REVIEWED BY DATE

Scanning procedures
• Coil selection/set up

• Positioning

• Pulse sequence selection

• Imaging options selection

• Scanner startup/shutdown

• Record keeping/documentation

• Archiving/deletion of data

TOPIC REVIEWED BY DATE

Filming procedures
• Processor startup/shutdown

• Laser camera startup/shutdown

• Filming software

• Developing films

Documentation of Approval

_____________________ is approved to operate the________________________

(Investigator’s name) (Name and location of magnet)

Approval given by________________________________ Date:_______________

MRI SAFETY STUDY GUIDE
Section I: The Magnetic Environment
Magnetic field
It is important to remember when working around a superconducting magnet that the
magnetic field is always on. Under normal working conditions, the field is never turned
off. Therefore, it is important to be aware of safety issues regarding ferrous objects that
can act as projectiles and watching out for patients who may have contraindicated devices
implanted in their bodies.
There are two units used to describe magnetic field strength. They are Tesla and Gauss:
1 Tesla equals 10,000 Gauss. The strength of the magnetic fields of a 1.5T (15,000 Gauss)
system is ∼30,000 times that of the earth’s magnetic field. The 5-Gauss line defines the
area at which the magnetic field becomes dangerous to patients with cardiac pacemakers.
For many 1.5T systems the 5-Gauss line is located at the foot of each scan table as well
as at the same distance from the back of the magnet.
Keep doors closed
The doors leading to the magnet room should be closed at all times except when entering
or exiting the room. This will prevent people who do not belong in the room from
wandering into the room. This also can dramatically reduce RF noise appearing in the
images (the room is not properly RF screened).
Consent and screening procedures
Consent forms
Every research patient will need to sign a consent form. Research often involves the use
of non-FDA (Food and Drug Administration)-approved MRI sequences on patients and
volunteers. These sequences are important to the advancement of the science of MRI. The
research patients and volunteers must be informed that these non-approved sequences are
being performed and consent must be given. Copies of the consent forms for patients and
volunteers to sign every time they have a research exam should be readily available. In
addition, the consent form must be signed by the Principal Investigator or his/her designee.
Only one consent form is needed per exam, but the research patient must sign a new
consent form every time he/she returns for another exam. Offer the volunteer or signee a
copy of the consent form. See the MRI research technologists or the scanning investigator
for a copy of the consent form.
Screening forms
As stated earlier, the magnets have a very strong magnetic field surrounding them, which
has the potential to attract certain types of metal. The magnetic field can also interfere
with the normal operation of electronic devices. For these reasons, a detailed health history
for every person that enters the magnet room is necessary. This includes all staff members,
investigators, patients, and volunteers. The repercussions associated with a patient,
volunteer, or staff member being injured because of negligence on the part of the scanning
investigator could be severe and could cause clinical scanning or research to be halted at
this facility.
Screening forms have been designed and must be completed by every person entering the
magnet room. Examples of the screening forms can be found in APPENDIX 1. In the event
that a staff member has an accident or surgery where a metallic foreign object or electronic
device is implanted into their body, the staff member would be restricted from going into
the magnet room until the metallic/electronic object can be cleared for safety purposes.
Current Protocols in Magnetic Resonance Imaging
It is up to the staff member to be aware of such circumstances and to report any such
events to their direct supervisor.
From this point on, patients and volunteers will be referred to as “subjects.” However, it
should be noted that the word “patient” refers to any person scanned in the clinical or
research facility who is under the care of a physician; that physician being one who is an
investigator in the clinical or research study.
Subjects who return for other MRI exams must fill out a new screening form for each
visit. Every screening form must be signed by the subject and the investigator or
technologist who is performing the scan.
Keep in mind that all subjects who are providing information regarding their health history
must be conscious and coherent. Any gaps in memory or lack of information about a
surgical procedure is grounds for canceling the subject, unless a family member can
provide a detailed history. If there is ever any question about the past health history of a
subject regarding metal in their body, it is required that the MRI exam be put on hold until
the question can be investigated thoroughly.
An in-depth explanation of contraindications to MRI will follow in the next section.
Section II: Contraindications for MRI
There are several types of contraindications that would prevent a subject from having an
MRI scan. Metallic implants and foreign bodies as well as the physical condition of the
subject will be discussed in this section. All subjects are required to remove any clothing
that has metal on it. Gowns and pajama bottoms are provided for the subject to change
into. All subjects and staff members must empty their pockets of any loose metallic
objects (e.g., hair pins, safety pins, coins, keys, ID badges, wallets, credit cards, banking
cards, lighters, pocket knives, scissors, stethoscopes, hemostats, etc.) before entering the
magnet room.
Surgical implants
There are hundreds of metallic implants that can be surgically placed into a person’s body
for various reasons. Some of these implants are ferrous and may be attracted to the
magnetic field. Some may be electronic in nature, in which case, the magnetic field can
interrupt the normal operations of the device. By placing an electronic device in the
magnetic field, a current may be induced in the conducting wires of the device, which
could possibly burn the patient. There are many metallic implants that are non-ferrous
and may be compatible for MRI such as orthopedic screws, rods, and plates. It is suggested
that a waiting period of at least 2 weeks after surgery is necessary for the tissues around
the implant to take hold of it to prevent any potential movement of the implant. Although
the 2-week period is generally observed, in some more emergent instances, a subject with
a non-ferrous implant may be scanned as soon as 1 day after the implant is in place. There
are also some ferrous implants (e.g., heart valves, venous blood clot filters) that are
compatible for MRI. Typically, the waiting period for these implants is between 2 and 6
weeks. The bottom line is that the waiting period decision should be left to a radiologist
who is familiar with the implant and its magnetic properties. To prevent injury to the
subject, it is extremely important that the scanning investigator be familiar with the
difference between compatible and contraindicated implants and devices.
MR Safety Guide MR Safety Guide
A3.5 A3.6
Supplement 10 Supplement 10 Current Protocols in Magnetic Resonance Imaging
Accidental metallic foreign bodies
Occasionally, a subject may have been injured by a piece of metal that punctured his/her
body in some way, shape, or form. Common causes of this type of injury are bullets,
buckshot, pellets, or BBs. Other frequent causes are metal slivers that fly off by those who
grind, sand, or cut metal. These metal slivers often fly into the eyes, hands, or face (and
it is likely that the people working in these areas are completely unaware because these
metal slivers are so fine). Also, people who have been involved in wartime activity may
have pieces of shrapnel or other metal fragments in their body. Any of these circumstances
must be investigated thoroughly to prevent injury to the subject.
Checklist of tested implants, devices, and metallic foreign bodies
There is a small handbook entitled “Pocket Guide to MR Procedures and Metallic Objects:
Update” (Shellock, 2001). This book lists hundreds of surgical implants and metallic
foreign bodies that have been tested by leading MRI safety authorities in magnetic
environments for evidence of deflection and torquing of the metallic objects. This book
should be used to investigate any questionable implant or foreign body.
Procedure to clear metallic implants and foreign bodies
The subject should list all of his/her surgical or accident history on the screening form.
In addition to this, the investigator should requestion the subject about his/her history
even if the subject has written that he/she has not had any surgery. It is not uncommon
for a subject to conceal or forget about a procedure or accident that may have happened
some time ago. He/she also may be uncomfortable about writing it down for others to
see. Further questioning the subject and explaining the importance of honesty can
sometimes provide additional information to the investigator.
If the subject has had a surgical implant or an accident involving metal, find out the
following information:
1. What was the procedure? What was the nature of the accident?
2. What kind of implant is it? What does it do? What is it used for?
3. When was the procedure done? What year?
4. Do you know for sure that it is metal?
5. Who was the doctor/surgeon who performed the procedure? Is he or she still in
practice?
6. At what hospital was the procedure performed?
7. If it was an accident, did you have any X-rays done at that time and was the metal
removed?
Once all of the answers to these questions are obtained, proceed with the following:
1. Provide the above information to the MRI technologist. The technologists have been
educated as to what may or may not be scanned and in many instances will be able
to assist.
2. If the technologist does not know of the implant or thinks the subject may have to be
canceled, the principal investigator and a radiologist must be consulted to obtain more
information. If the subject does not know if the implant is metal, the radiologist may
suggest X-rays be done to rule out metal. X-rays may not be performed without the
permission of the principal investigator.
Current Protocols in Magnetic Resonance Imaging
3. If the radiologist does not know of the implant, contact the surgeon who placed the
implant and request a copy of the operating room report, which should describe the
model and name of the implant. Attach this report to the screening form of the subject
for permanent documentation.
4. The final responsibility of canceling or proceeding with the exam lies with the
principal investigator, who should make an informed decision based on the informa-
tion provided by the MRI technologists and radiologists.
5. If the subject is cleared, a written permission signed by the principal investigator for
the subject to undergo the MRI exam must be provided to the MRI research staff.
Attach this permission form to the screening form of the subject for permanent
documentation.
IMPORTANT NOTE: Any person (subject or staff) who has a history of working with
metal as an occupation or hobby should have X-rays of their orbits to rule out any metallic
foreign body before they enter the magnet room. A case of a patient being blinded by a
metal sliver piercing the optic nerve was a former metal worker who had no knowledge
of a piece of metal existing in his eyes. Typically, if metal workers get a sliver of metal
in their eyes, it is removed in an emergency room by a physician. However, without X-rays,
there is no way of knowing if the entire piece of metal was removed. Usually, X-rays will
be ordered at the time the metal is taken out. If a copy of the report from those X-rays can
be obtained and the subject has no more metal in his/her eyes since the time of the X-rays,
then use the original X-ray report to clear the subject for the MRI exam.
Emergency removal of the subject from the MRI scanner
If the investigator has placed a subject in the scanner and upon looking at the first set of
images notices a metallic artifact present, the investigator must follow the proper proce-
dure for removing the subject from the scanner and the magnet room.
1. Inform the subject that he or she is going to be removed from the magnet. Instruct
him/her to remain perfectly still and to not sit up at any time.
2. Pull the table out of the scanner very slowly.
3. Move a gurney into the magnet room and place it next to the table.
4. Have the subject slide, without sitting up, onto the gurney.
5. Slowly pull the gurney straight away from the magnet without turning.
6. Once the doorway is reached, slowly turn the gurney and move it out through the
doorway.
7. Once the subject is safely outside of the room, he/she may sit up.
NOTE: This procedure should also be used if the subject informs the technologist of a
contraindicated metallic implant in his/her body after already being placed in the magnet
(or if being taken out for any other reason as well, e.g., he/she pushes the button or
squeezes the ball requesting to come out).
Pregnant subjects
It is the policy of the MRI department to not scan any pregnant subjects for research
purposes. In the clinical environment, pregnant patients are only scanned in emergency
situations. With this in mind, and realizing that research is not done on an emergency
basis, pregnant subjects must wait until after they give birth in order to participate in a
research project.
MR Safety Guide MR Safety Guide
A3.7 A3.8
Supplement 10 Supplement 10 Current Protocols in Magnetic Resonance Imaging
If a subject believes she may be pregnant, it is up to the principal investigator to decide
if the subject should undergo a pregnancy test. If the principal investigator deems the
pregnancy test necessary, all arrangements and financial responsibility will be taken care
of by the principal investigator or his/her designee.
If the pregnancy test is negative and the subject is to undergo the MRI, a copy of the
pregnancy test report will be needed to attach to the screening form of the subject for
permanent documentation.
Contrast agents used in subjects who are breastfeeding
In the case where a subject is breastfeeding her child, the mother must be informed that
her breastmilk must be expressed with a breast pump and thrown away for 48 hr following
the injection of gadolinium contrast agent. It is important that she be aware of this in order
to store enough milk to feed the child during the 48 hr after the contrast agent injection.
Pregnant staff
It is the policy of the MRI department that all pregnant staff members be restricted from
the magnet room when radiofrequency pulses are on. Any pregnant ancillary staff member
(e.g., nurses, coordinators, assistants) who does not need to be in the magnet room should
stay out of the room unless there is an emergency with a subject. Pregnant staff members,
such as MRI technologists, who must enter the room on a regular basis should only stay
in the room as long as necessary, i.e., positioning subjects, emergencies, etc. The goal is
to keep the pregnant staff member out of the magnet room as much as possible unless it
is part of the job description to be in the room or if there is an emergency with the subject.
If an MRI staff member is pregnant, she should inform her supervisor immediately.
Ancillary staff members should let the MRI technologists know if they are pregnant.
Radiofrequency and specific absorption rate
MRI employs radiofrequency (RF) pulses to disturb the alignment of protons in the
nucleus of hydrogen atoms in the body. These RF pulses deposit heat into the tissues of
the body. This heat deposition is termed specific absorption rate (SAR). SAR is measured
in watts per kilogram and is a function of several variables, including: (1) the type of RF
pulse used (90° or 180°); (2) the number of RF pulses in a sequence; (3) the pulse width;
(4) the TR; (5) the weight of the patient; and (6) the type of coil used. The FDA and IEC
(International Electrotechnical Commission) have developed guidelines to regulate the
acceptable amount of deposited heat. Currently all manufacturers of MRI equipment are
permitted to submit their pulsing sequences to the FDA for SAR review.
Conditions in the examination room
The examination room should have an ambient temperature of 21°C ±3°C with a relative
humidity of 50% to 70%.
Specific absorption rates
1. The FDA (http://www.fda.gov/cdrh/ode/magdev.html) does not require an approval
of an investigational device exemption (IDE) if the field strength is not >4 Tesla or
if SAR is less than or equal to:
a. 4 W/kg averaged over the whole body for any period of 15 min; or
b. 3 W/kg averaged over the head for any period of 10 min; or
c. 8 W/kg in any gram of tissue in the head or torso, or 12 W/kg in any gram of tissue
in the extremities, for any period of 5 min.
Current Protocols in Magnetic Resonance Imaging
2. IEC 60601-2-33 normal mode; i.e., when the field strength is ≤2 Tesla (see IEC
60601-2-33 Medical Electrical Equipment—Part 2: Particular requirements for the
safety of magnetic resonance equipment for medical diagnosis, 2002, 2nd Ed.,
International Electrotechnical Commission). The SAR are limited to:
a. 2 W/kg averaged over the whole body for any period of 6 min; or
b. 3.2 W/kg averaged over the head for any period of 6 min; or
c. 10 W/kg in any gram of tissue in the head or trunk, or 20 W/kg in any gram of
tissue in the extremities, for any period of 6 min; or
d. three times the appropriate long-term SAR level for any period of 10 sec.
3. IEC 60601-2-33 first controlled mode; i.e., when the field strength is >2 Tesla but is
not >4 Tesla. The SAR are limited to:
a. 4 W/kg averaged over the whole body for any period of 6 min; or
b. 3.2 W/kg averaged over the head for any period of 6 min; or
c. 10 W/kg in any gram of tissue in the head or trunk, or 20 W/kg in any gram of
tissue in the extremities, for any period of 6 min; or
d. three times the appropriate long-term SAR level for any period of 10 sec.
Acoustic noise
1. The FDA regulates the peak unweighted sound pressure level to be not >140 dB or
the A-weighted r.m.s. sound pressure level to be not >99 dBA with hearing protection
in place.
2. IEC 60601-2-33 regulates that the peak unweighted sound pressure level is <140 dB.
NOTE: It should be noted that subjects with thermoregulatory illnesses such as a fever,
or diseases in which the patient is unable to sweat, may be compromised by heat deposition
in MRI. Extreme care should be taken with these subjects to keep them cool during the
exam. Choose sequences that do not result in high amounts of heat deposition. The eyes
are also particularly susceptible to heat deposition.
Section III: Emergency Procedures
CPR requirements
It is a common requirement of MRI departments that all investigators, technologists,
students, or other staff who will be conducting MRI experiments on humans are certified
in cardiopulmonary resuscitation (CPR).
Crash carts
There is usually one crash cart located in any MRI area.
Calling 911
In order to get emergency help, call 911 and follow the instructions of the dispatcher. Tell
them the address right away. Remember to speak clearly and slowly so there is no
misinterpretation of information. Some sites may have their own internal procedures.
Code procedures
In order to know the status of the subject at all times, it is strongly recommended that the
pulse oximeter be placed on the finger of every subject that goes into the magnet. This
will provide the heart rate and oxygen saturation for the subject while he/she is in the
scanner. If the subject should become unresponsive, begin the code procedure as listed
MR Safety Guide MR Safety Guide below.
A3.9 A3.10
Supplement 10 Supplement 10 Current Protocols in Magnetic Resonance Imaging
 1. Immediately remove the subject from the bore of the magnet.
2. Try to arouse the subject by shaking him/her gently.
3. If there is no response, feel for a pulse at the carotid artery in the neck. At the same
time, listen for breathing near his/her mouth and nose while looking at his/her chest
to determine breathing.
4. If the subject is not breathing or does not have a pulse, call 911 immediately, also call
security.
5. Return to the subject immediately and begin rescue breathing and/or CPR appropriate
to the condition of the patient.
6. If more than one investigator is present, have him/her call 911 and security for help
and then round up as many medical personnel as possible.
7. The second person should also bring the stretcher into the magnet room to enable
transport of the subject out of the room. Make sure the stretcher is made of aluminum,
which is MR compatible.
8. DO NOT BRING THE CRASH CART INTO THE MAGNET ROOM.
9. If the subject is transported outside of the magnet room, proceed to the holding area
and bring the crash cart. The crash cart is equipped with an ambu bag that may be
used to breathe the patient if trained to do so.
10. Once emergency personnel arrive, DO NOT let them into the magnet room. Secure
the subject out of the magnet room without endangering any personnel who have not
been cleared to enter the room. Among other items, paramedics always have stetho-
scopes, scissors, and hemostats on them, which will turn into deadly projectiles if
brought near the magnet. In addition, their health history is unknown and they may
have contraindicated implants in them.
11. Inform the emergency personnel of the incident and stay close by to assist them with
any questions they may have or items they may need.
12. After the subject is taken away by the paramedics, fill out an incident report.
Quench
The term “quench” is used to describe the rapid boil off of the cryogens that keep the
magnet cooled and in a superconducting state. Cryogens are supercooled liquid helium
and nitrogen. Without cryogens, the magnet loses its magnetic field. Usually a quench is
undesirable and is due to a malfunction within the system. In rare instances, a quench
may be necessary to free someone from the magnet if they have been accidentally struck
by a projectile ferrous object and pinned to the magnet. In each control room, there are
boxes on the wall that enclose quench buttons that should be pushed in the event that the
magnetic field must be manually run down.
When a quench occurs, either spontaneously or manually, evacuate from the magnet room
immediately to avoid being overcome by helium gas, should the room not vent properly.
If manual quenching of the magnet is to be done, make sure the door to the scan room is
left open to avoid a vacuum forming, which may seal the door shut. If the magnet quenches
spontaneously, and the door does not open, break the window between the control room
and the magnet room in order to get the subject and possibly yourself out of the room.
Remember to stay low so that helium or nitrogen is not breathed in.
Current Protocols in Magnetic Resonance Imaging
Projectile injury
If a subject or staff member becomes pinned to the magnet by a ferromagnetic object,
evaluate the situation quickly before taking any action. If the person is unconscious,
bleeding profusely, at risk of losing a limb or extremity, or in severe pain, manually quench
the magnet to bring down the field in order to release the object and the person. If the
person is responsive and able to communicate his/her well being, leave him/her in the
position until a service engineer can respond and ramp the magnet down slowly to avoid
a full quench. If the latter is chosen, and the person then loses consciousness, or his/her
condition worsens, immediately quench the magnet manually. Keep in mind that the
cryogens are expensive to replace, so evaluate the situation carefully but never put cost
above the life or well being of the subject.
Once released, secure the subject out of the room and call 911 for emergency medical
help.
Responsible parties
Any time a patient is scanned, or any contrast agent is administered to a human in the
MRI area, a physician must be available to cover in the event of a medical emergency. If
the principal investigator is not a medical doctor, arrangements must be made to have a
medical doctor available to respond for emergency purposes. If a designee of the principal
investigator is present with the subject, the designee must know how to reach the principal
investigator, or a medically responsible party, immediately in the case of an emergency.
This also means when a subject is scanned, the responsible physician may not be out of
town without arranging for medical coverage in the event of an emergency. The MRI
technologists will assume that these arrangements have been made before the subject is
scanned and will not be responsible for medical treatment of the patient other than proper
emergency procedures in the event of an emergency or adverse event.
Section IV: The Scanner and Related Equipment
Bringing computers up and shutting computers down
Bringing the MRI computer up, or booting the computer, is a fairly simple procedure.
Make sure that the information from the local service representative for the specific
system is available.
From this point on, in order to use the scanners, the scan time on the research magnets
must be pre-approved.
Table controls and table stop buttons
Each of the magnets is equipped with a table that moves into and out of the scanner by using
the plasma controls on the front of the magnet gantry. Each scanner should have buttons or
pedals on the scan table or around the magnet in order to manually release the scan table. For
example, in the Siemens Vision, there are two orange arrows that indicate “up/down” and
“in/out.” On both sides of each table are red table release buttons. Normal position for the
buttons is in the “out” position. To remove the table rapidly, push one of the red buttons “in”
to release the table. The table may then be pulled out of the magnet bore manually.
Stereo/headphones/earplugs
All subjects are required to wear ear protection while undergoing an MRI exam. Earplugs
will be provided. Headphones are also provided, which hook into the stereo system so the
subject may enjoy music during the exam. The headphones are also part of the noise
cancellation system that helps drown out the knocking noise of the gradients. For the
high-field systems, if the subject refuses all hearing protection, the scan cannot be
performed.
MR Safety Guide MR Safety Guide
A3.11 A3.12
Supplement 10 Supplement 10 Current Protocols in Magnetic Resonance Imaging
Communicating with the subject while in the scanner
It is important to maintain voice contact with the subject throughout the exam. The
researcher or the technologist should routinely establish contact between each sequence.
Speak to the subject while depressing the “speaking” button on the keyboard. In order to
hear the subject, the “hearing” button on the keyboard must be depressed.
Patient alarm
Every subject should be given the patient alarm ball to hold in his/her hand during the
exam. The subject should be instructed to squeeze the ball if:
a. he/she needs to speak with the investigator or technologist in between sequences;
b. he/she wants to come out of the scanner immediately;
c. something is hurting.
Because the scanner cannot be put in a pause mode, stopping a scan to speak to the subject
will require one to start the scan over again from the beginning. For this reason, it is wise
to advise the subject to squeeze the ball only in situations of pain, injury, or claustrophobia.
If the investigator or technologist is communicating with the subject routinely between
sequences, the subject will be less likely to squeeze the ball in the middle of a sequence
to ask a non-emergent question.
In vivo monitoring system
With an in vivo system, one may be able to do the following:
1. obtain an ECG trace from the subject;
2. obtain blood oxygen saturation percentages from the subject;
3. obtain blood pressures from the subject.
ECG and cardiac gating
Currently, there are several types of studies where cardiac gating is desirable. Cardiac
gating functions to allow imaging in areas of the body where there is considerable motion.
For example, when imaging the heart, cardiac gating is used to instruct the computer to
image all of the slices at the same point in the heart cycle every time the heart beats. This
gives the appearance that the heart motion is frozen resulting in images with the
appearance of little or no motion artifact.
Special attention must be given when attaching the electrode and leads to the chest of a subject.
Because the lead wires are placed in a magnetic field, it is possible to induce an undesirable
current in those wires, which may burn the subject. It is imperative that the lead wires and
the main ECG cable have no loops in them when placed on the subject. The main cable should
not touch the sides of the magnet or the skin of the subject as it is run out of the magnet bore.
The cable should be run straight out of the bore with no loops and should not cross over the
body of the subject at any point. If the cable must be obtained from the left side of the subject
to the plug-in port on the right side of the table, run the cable down the left side of the subject,
and then across the foot of the table. A washcloth, sheet, or towel must be placed between the
skin and any wire that makes contact with the skin.
Blood pressure monitor
The blood pressure monitor is able to measure the blood pressure of the subject noninva-
sively at prescribed intervals throughout the exam. The researcher may set a time interval
at which the monitor will automatically inflate the cuff on the arm of the subject while in
the scanner. An updated blood-pressure reading is displayed with each measurement
interval.
Current Protocols in Magnetic Resonance Imaging
The blood-pressure cuff may be placed on either arm but care should be taken not to place
it on an arm that has an i.v. placed in it. Also, it is not uncommon for women who have
undergone a mastectomy to have poor lymph circulation in the arm on the side of the
mastectomy. Because of this, these subjects cannot usually withstand pressure placed on
the arm of the same side as the mastectomy. For example, if the woman has had her right
breast removed, place the blood-pressure cuff on her left arm. If she has had both breasts
removed, ask her on which arm she would prefer to have the cuff placed.
Blood oxygen saturation monitor (pulse oximeter)
The pulse oximeter, when placed on the fingertip of the subject, will display the heart rate
of the subject and percentage of oxygen in the blood. The pulse oximeter may be placed
on any finger of the subject, however, the index or middle finger has been found to work
the best. Place a piece of adhesive tape around the clip when it is on the finger to hold it
securely in place.
The corrugated cable running from the finger clip to the pulse oximeter contains fragile
fiber optic wires. It is important that this cable is not stepped on or crushed; be careful
with it. Damage to the fiber optics or a break in the corrugated cover could cause the pulse
oximeter to malfunction or produce RF artifacts.
NOTE: For safety purposes, it is strongly recommended that the pulse oximeter be placed
on all subjects who are having an MRI exam. If the subject should fall asleep in the scanner
and become unresponsive when the investigator or technologist attempts to speak to
him/her, the investigator or technologist will know whether or not the subject is well or
in distress based on the readouts from the pulse oximeter. This will save the investigator
or technologist from having to stop the experiment to go into the magnet room to check
on the subject. Also, if the subject should have heart failure or breathing difficulty, the
investigator or technologist will know immediately based on the pulse oximeter readout.
Removing subjects from the MRI scanner
If a subject requests to be removed from the magnet at any time, the investigator or
technologist should do so promptly. Whether due to pain, illness, or claustrophobia, the
investigator or technologist must never keep the subject in the MRI scanner against his/her
will. If a subject asks to be brought out, communicate with him/her to determine the
problem. Ask the subject if he/she can continue, if not, remove the subject immediately.
Starting/stopping a scan
To start a scan, simply click on the “measurement” (or its equivalent) button on the control
screen. To stop a scan, click on the “stop” (or its equivalent) button on the control screen.
Oxygen/suction/room air supplies
Some MRI facilities are equipped with oxygen, suction, and room air channels. These
may be mounted on a wall or hung from a ceiling to the left of the scanner. Oxygen is
marked by the green hose or regulator, suction is white, and room air is yellow.
Section V: Data Acquisition and Management
Responsibility for acquired data, archiving, and deletion of data
All investigators and technologists are responsible for the data acquired. Data must be
archived immediately after the exam is complete in order to prevent loss by removal off
the MRI system disks.
MR Safety Guide MR Safety Guide
A3.13 A3.14
Supplement 10 Supplement 10 Current Protocols in Magnetic Resonance Imaging
Data may stay on the MRI system disks for up to 7 days. It is requested that if the
investigator or technologist does not need the data or is finished processing it, then it
should be removed. If the disks become full and deletion is necessary, the oldest studies
will be removed first.
If an investigator or technologist needs his/her data to stay on the disks for longer than 7
days, he/she should fill out the “DO NOT REMOVE” logbooks, which reside in each scan
control room. Also, if an investigator or technologist restores a study to the MRI system,
he/she should log this in the “DO NOT REMOVE” logbook so that the study is not
removed before the data is used.
Investigators or technologists may only remove one’s own data. All other MRI data may
only be removed from the system disks by the MRI chief technologists.

MR Safety Guide

A3.15
Current Protocols in Magnetic Resonance Imaging Supplement 10
MRI OPERATOR EXAM 6. You have just positioned a subject inside of the magnet for a knee scan when she
suddenly remembers she had a brain aneurysm repaired 10 years ago. You ...
Name: a. Rapidly pull the table out of the scanner, immediately lower it, and rush her out
Date: of the room as fast as possible
b. Continue with the exam because you are not scanning her head
Department: c. Call 911
d. Very slowly, pull the table out of the scanner and then have the subject slowly slide
1. Mr. Jones is feeling a little claustrophobic during his exam and would like his wife onto a stretcher so that you can slowly cross her through the magnetic field lines
to sit with him during his MRI exam. Mrs. Jones had a pacemaker implanted 5 years and out of the room
ago, therefore, it is permissible to allow her into the room as long as she stays at least
10 feet from the front of the magnet. 7. Which of the following represent the current FDA recommendations regarding
a. True specific absorption rates for an MR scanner that does not have a field strength >4
Tesla?
b. False
a. Not to exceed 4 W/kg averaged over the whole body for any period of 15 min
2. For subject safety purposes, which of the following items are contraindicated for an b. Not to exceed 3 W/kg averaged over the head for any period of 10 min
MRI scan. c. Not to exceed 8 W/kg in any gram of tissue in the head or torso, or 12 W/kg in any
a. Orthodontic braces gram of tissue in the extremities, for any period of 5 min
b. Orthopedic screws in the ankle placed 6 months ago d. All of the above
c. Aneurysm clips
8. If a subject is unresponsive when you try to communicate with him in between
d. Pacer wires (no pacemaker) left in the chest after open heart surgery 3 years ago scanning sequences, you should ...
e. c and d only
a. Continue with the scan and assume that he is sleeping
f. All of the above
b. Call 911
3. You walk into the scan room and find a maintenance employee pinned to the magnet c. Defibrillate him
by a floor buffer. He is unresponsive but has a weak pulse. You immediately ... d. Pull him out of the scanner or at the very least go into the room to see if he is O.K.
a. Throw cold water on him
9. To avoid the confusion that a patient is well or not, it is a good idea to use the pulse
b. Use smelling salts to try to rouse him oximeter on everyone that is placed into the magnet.
c. Call his supervisor and tell him that the maintenance employee is sleeping on the
job a. True
d. Call for help then attempt to free the employee; if unsuccessful, press the quench b. False
button to eliminate the magnetic field
10. Your subject has informed you that he has worked as a metal grinder for 10 years but
4. While scanning a subject for a heart study, he suddenly complains of a stinging, he has never had any metal shavings fly into his eyes.
burning sensation at the location of one of the electrodes placed on his chest for a. Proceed with the scan since he has never had any metallic foreign bodies in his
cardiac gating. You .... eyes
a. Tell him to remain quiet until the sequence has finished b. Do not proceed with the scan and at the discretion of the principle investigator,
b. Turn the radio up louder so that you cannot hear his complaint send the subject for X-rays of his orbital area to rule out possible metallic foreign
bodies
c. Tell him the stinging will subside in a few hours
d. Immediately remove him from the magnet and inspect the area of the leads for 11. Your subject has informed you that she had a metal sliver enter her eye 10 years ago
loose or crossed wires but that the doctor got it all out. You ...

5. When positioning the ECG monitoring cable on a subject for a heart study, it is OK a. Proceed with the scan since the metal sliver was removed
to run the cable across the subject’s body to plug it into the connector box. b. Do not proceed with the scan and at the discretion of the principle investigator,
send the subject for X-rays of the orbital area to rule out possible metallic foreign
a. True
body
b. False
MR Safety Guide MR Safety Guide

A3.17 A3.18
Current Protocols in Magnetic Resonance Imaging Supplement 10 Supplement 10 Current Protocols in Magnetic Resonance Imaging
12. While performing a brain MRI on a 24-year-old woman, she begins to complain of 18. Cryogens are ...
some stinging around her eyelids. Upon examining her you find out that she has
a. Industrial-strength cleaners
permanent tattoo eyeliner. You ...
b. Sedatives
a. Terminate the exam, take her out of the scanner, apply cold compresses to her eyes,
c. Liquefied gases used to cool a superconducting magnet and that are deadly to breathe
and seek the proper medical attention
d. Sewer gases
b. Take her out of the scanner, apply ice-cold compresses to her eyes to lower the
temperature in the area, and then continue with the scan if she feels up to it
19. A quench is ...
c. Inform the subject to shut her eyes and the pain will go away
a. A relief of thirst for a hardworking technologist
13. Your subject has had a heart attack while undergoing an MRI exam. You ... b. A rain cloud that occurs inside of the magnet room due to overuse of the magnet
by physicists
a. Call 911 and immediately bring the crash cart into the magnet room to begin
resuscitation measures c. A rapid, usually undesired, release of the cryogens, which causes a rapid decline
of the magnetic field
b. Call 911 and bring the subject out of the room on a stretcher to begin resuscitation
measures d. The German word for magnet
c. Call a hospital code team and bring the subject out of the room on a stretcher to
20. The first thing that you should do if a quench occurs is ...
begin resuscitation measures
d. Call a hospital code team and immediately bring the crash cart into the magnet a. Call 911
room to begin resuscitation measures b. Call the service engineers
c. Immediately remove the subject from the scanner
14. During the middle of a measurement, you hear the subject attempting to speak to you
d. Call the EPA (Environmental Protection Agency)
but you can not hear her clearly. You ...
a. Ignore her and continue the measurement 21. If during a quench the cryogens do not vent out of the room properly, a vacuum may
b. Stop the measurement and speak to her form, making it difficult to open the door of the magnet room to get the subject out.
If you are unable to open the door, you should ...
c. Stop the measurement and yell at her for interrupting
d. Terminate the entire exam a. Sit and wait
b. Use a blow torch to burn through the door
15. Your subject tells you she had surgery 20 years ago on her brain but she can not c. Break the window between the control room and the magnet room
remember exactly what it was for. You ...
d. Panic
a. Proceed with the scan because she seems credible
b. Cancel the scan for today until further investigation into the matter can occur 22. If your subject becomes short of breath while in the MR scanner, you should
immediately grab the oxygen tank off of the crash cart and administer oxygen.
c. Tell the subject of the risks involved due to her lack of history and allow her to
decide if she wants to have the MRI a. True
d. Proceed with the test only after having the subject signs a waiver that releases us b. False
from responsibility of injury
23. It is acceptable to let someone go into the magnet without emptying their pockets of
16. Which of the following are contraindicated for an MRI exam? metal objects if they are only going to be in there for <30 min.
a. Cardiac pacemaker a. True
b. Swan-Ganz catheter b. False
c. Aneurysm clips
24. Which of the following objects are approved for entry into the MR scanner room?
d. All of the above
a. Stethoscope
17. It is acceptable to loop an insulated surface coil cable around a subject’s arm or leg b. Hemostats
to help keep the cable from getting caught in the table.
c. Employee I.D. badges
a. True d. Small paper clips and coins
b. False e. None of the above
MR Safety Guide MR Safety Guide

A3.19 A3.20
Current Protocols in Magnetic Resonance Imaging Supplement 10 Supplement 10 Current Protocols in Magnetic Resonance Imaging
 32. Any time a patient is scanned for research purposes, a medical doctor or other
25. The system stop key will cut power to ... qualified medical personnel associated with the study must be present or in the near
a. All computer cabinets except for the gradient chiller cabinet vicinity, and available at all times.
b. The lights a. True
c. The fire alarms b. False
d. The automatic sprinklers
33. An in vivo MR–compatible monitoring system is capable of measuring which of the
26. If you get a message that the gradient chiller system is malfunctioning, you should following?
immediately ... a. ECG and heart rate
a. Continue to scan because the chiller is on automatic controls b. Blood pressure
b. Run from the room screaming c. Blood oxygen saturation
c. Stop scanning to avoid overheating the gradient coils d. All of the above
d. Call 911
34. It is acceptable to leave a subject lying on the scan table outside of the magnet bore
27. If the subject sets off the patient alarm during the scan you should ... while you leave the room.

a. Page the MRI technologists a. True

b. Tell the subject to be patient b. False
c. Communicate with the subject to check on his/her status
35. All subjects undergoing an MRI should be given ear protection in the form of earplugs
d. Turn off the alarm and keep scanning with or without the noise cancellation headphones.
28. Which of the following are contraindications for MRI? a. True
b. False
a. Cochlear implants
b. Neurostimulators 36. Every person that enters the magnet room should have completed a screening form
c. Patient-controlled anesthesia devices regarding metal in the body.
d. All of the above a. True
29. Pregnant women should never be scanned in our research facility. b. False

a. True 37. It is mandatory that subjects remove all metal, metal jewelry, clothing with metal on
b. False it, and pocket contents before having an MRI scan.
a. True
30. Pregnant staff members are not permitted to enter the magnet room.
b. False
a. True
b. False 38. Metallic objects in the body will produce artifacts on the images that look like ...
a. A black signal void with bright edges
31. Which of the following is the correct method of investigating a metallic implant in a
subject? b. A green streak
c. A line through the image
a. Take the subject’s word that his/her implant is non-ferrous.
d. Pink swirls
b. Call the surgeon responsible for placing the implant and get a detailed written
operating room dictation explaining where the implant was placed, what it is, what 39. If caught in the magnet room during a quench, where should you place your body to
it is made of, the manufacturer, and the model name and number avoid inhalation of the cryogens?
c. Ask the subject’s family member
a. On top of the scan table
d. Call a physician referral service
b. Perpendicular to the main magnetic field
c. As close to the floor as possible
MR Safety Guide MR Safety Guide d. Pressed up against the window between the scan room and the control room

A3.21 A3.22
Current Protocols in Magnetic Resonance Imaging Supplement 10 Supplement 10 Current Protocols in Magnetic Resonance Imaging
40. The simplest way to stop a scan is to ... 46. Under normal circumstances, when the MRI scan is not an emergency, how long must
subjects who have had non-ferrous metal, e.g., orthopedic screws or plates, surgically
a. Click the mouse on the “stop” (or its equivalent) button on the menu of the scan
implanted wait before undergoing an MRI scan?
console
b. Quench the magnet a. 2 to 6 weeks
c. Turn the key off b. 5 years
d. Speak the word “stop” into the microphone c. 1 year
d. Forever
41. Which of the following is the correct reason to shut the MRI system down (not the
magnetic field)? 47. How long must subjects who have had ferrous metal devices that are not contraindi-
cated for MRI, i.e., heart valves, venous blood clot filters, surgically implanted wait
a. Building utilities are to be shut off
before undergoing an MRI scan?
b. The scanner applications have locked up
a. 2 to 6 weeks
c. No one is going to be using the system for several hours
b. 5 years
d. All of the above
c. 1 year
42. When should the doors to the magnet rooms be shut? d. Forever
a. When measurements are running
48. After completing a scan, you should ...
b. Whenever the system is not in use
a. Leave the room clean and tidy
c. Whenever there is the potential that someone not associated with the study will
wander into the room b. Leave the room as you found it
d. All of the above c. Leave the phantoms on the floor
d. Call the technologists to clean up the room
43. If anyone asks you to take them into the magnet room to look at it, what do you do?
49. Ferromagnetic objects are dangerous to take near the magnet because ...
a. Take them right into the magnet room
b. Insist that they empty their pockets and then take them into the magnet room a. They can become projectiles and harm the subject
c. Tell them absolutely not, under any circumstances, will you let them into the b. They will burn the subject
magnet room c. They become very heavy
d. Screen them for metal in their bodies as if they were patients, and if they are d. There are too many north poles and the subject gets dizzy
cleared, have them remove all loose metal from hair, pockets, and clothing, and
then take them into the magnet room 50. The leads for cardiac gating must be handled as follows:
a. Wound together and routed down the center of the magnet
44. All subjects must complete consent forms of some type before having MRI scans in
the research facility. b. Run in parallel across the subject’s chest
c. Placed on the shin of the subject
a. True
d. Coiled in a loop and placed across the subject’s chest
b. False
e. None of the above
45. A volunteer had an MRI scan 6 months ago and they are now participating in the
research program again. What documents do they need to complete? 51. Echo-planar imaging and other sequences can be quite loud. You should ...

a. None, the ones they completed before are still good a. Tell the subject to grin and bear it
b. Screening form only b. Have the subject cup his hands over his ears
c. Consent form only c. Use earplugs with or without earphones to dampen the sound
d. Consent and screening forms d. Have the subject sing during the scan to dampen the sound

MR Safety Guide MR Safety Guide

A3.23 A3.24
Current Protocols in Magnetic Resonance Imaging Supplement 10 Supplement 10 Current Protocols in Magnetic Resonance Imaging
52. Consent forms must be signed by all subjects when non-FDA sequences are used. 59. Which of the following sequences tend to have the worst problems with RF power
deposition?
a. True
b. False a. Low flip angle RF spoiled gradient echo sequences
b. Spin echo, FLAIR (fluid attenuated inversion recovery), and turbo-spin echo
53. While scanning a subject for a heart study, he suddenly complains of chest pain. You ... sequences
a. Tell him to remain quiet until the sequence is finished c. TRUE FISP (true fast imaging with steady state free precession) sequences
b. Ignore his complaints d. Both b and c
c. Immediately remove him from the scanner, get him out of the room, and call 911
d. Give him a drink of water and put him back into the scanner to continue the exam

54. If ferrous devices and devices containing motors or questionable ferrous parts are
needed in the magnet room, they should be secured when brought into the magnet
room.
a. True
b. False

55. The 5-Gauss line is ...

a. A pixel line in the matrix
b. The magnetic field line territory at which the magnetic field can become harmful
c. A catheter inserted into the femoral artery
d. A geometric theorem

56. Where is the 5-Gauss line located in most MR facilities?

a. At the front door of the building
b. In the street in front of the building
c. Between the foot of the table and the RF window/screen, which is between the
control room and the magnet room
d. At the opening of the magnet bore

57. When a subject is in the bore of the magnet, it is okay for the investigator or
technologist to leave the control room ...
a. To check the weather outside
b. Whenever he/she feels like it
c. When he/she wants to get himself/herself a soda
d. If, and only if, he/she gets someone else to monitor the subject while gone

58. Potential subjects who have magnetically controlled artificial body parts or prosthe-
ses, such as artificial limbs or eyes, can not undergo an MRI exam.
a. True
b. False

MR Safety Guide MR Safety Guide

A3.25 A3.26
Current Protocols in Magnetic Resonance Imaging Supplement 10 Supplement 10 Current Protocols in Magnetic Resonance Imaging
ANSWERS TO THE EXAM Cerebral Infarct/Intracranial UNIT A1.1
1. b 21. c 41. d Cerebrovascular Disease
2. e 22. b 42. d
3. d 23. b 43. d When imaging patients for intracranial cerebral vascular disease, the goals are (1) to as-
4. d 24. e 44. a sess the degree of parenchymal injury and identify intraparenchymal hemorrhage; (2) to
5. b 25. a 45. d determine if there are areas of altered perfusion that may be at risk for future injury; and
6. d 26. c 46. a (3) to assess the intracranial arteries (patency as well as direction of flow). The corre-
7. d 27. c 47. a
sponding sequences are (1) T2 fast spin echo (FSE) and fluid attenuated inverson recov-
8. d 28. d 48. a
ery (FLAIR) to detect acute, subacute, and chronic injury, as well as diffusion-weighted
9. a 29. a 49. a
imaging to detect hyperacute injury and to determine age of injury, and gradient-echo
10. b 30. b 50. b
imaging to detect intracranial hemorrhages; (2) perfusion-weighted imaging; followed
11. b 31. b 51. c
by (3) 3-D time-of-flight MR angiography (TOF-MRA) to examine vascular patency,
12. a 32. a 52. a
and phase-contrast MRA (PC-MRA) to provide information regarding the direction of
13. b 33. d 53. c
14. b 34. b 54. a flow in major vessels. The Basic Protocol can be used to evaluate stable patients with
15. b 35. a 55. b acute, subacute, or chronic cerebrovascular symptoms. For a more tailored protocol in
16. d 36. a 56. c cases of hyperacute strokes or cerebrovascular symptoms in unstable patients, see the
17. b 37. a 57. d Alternate Protocol.
18. c 38. a 58. a
19. c 39. c 59. d STANDARD IMAGING FOR CEREBRAL INFARCT/INTRACRANIAL BASIC
20. c 40. a CEREBROVASCULAR DISEASE PROTOCOL
In order to perform the recommended perfusion and diffusion studies, a scanner with
ACKNOWLEDGEMENT echo planar capabilities is required (Table A1.1.1). MR angiography and the standard
The authors would like to thank Professor E. Mark Haacke for his original input and anatomical MR imaging sequences included in the protocol do not require these faster
continuing contributions, and Dr. Yu-Chung N. Cheng for his review of the material. gradients. The parameters given here are optimized for a 1.5 Tesla (T) General Electric
(GE) LX system with version 8.35 software and may need modification for different
software versions, field strengths, or manufacturers.
Contributed by Glenn Jay Foster Data processing is required for the following three sequences described in this protocol,
Washington University
St Louis, Missouri
(1) 3-D TOF MRA, (2) diffusion-weighted imaging, and (3) perfusion-weighted imag-
ing. For optimal assessment of the volumetric 3-D TOF MRA data, separate volumes
of interest (VOI) of the left anterior, right anterior, and posterior circulation should be
created. Images from 15◦ rotations about the vertical axis should be saved. An additional
anterior circulation volume of interest should include both anterior cerebral arteries and
the region of the anterior communicating artery for visual inspection. Images from 15◦
rotations about the horizontal axis should also be saved. The authors recommend maxi-
mum intensity projection (MIP) over volume and surface reconstructions for assessment
of vascular patency, as algorithms in many software packages for surface and volume
reconstructions may smooth out and therefore obscure regions of arterial narrowing. All
of the authors’ MRA data is processed on a General Electric Medical Systems (GEMS)

Table A1.1.1 Equipment Requirements for Cerebral Vascular Assessment

Coil type Quadrature head coil
Gradient coil strength 25 mT/m (or whatever the system permits)
Slew rate ≥120 T/m/sec
Cardiac gating Not necessary
Peripheral gating Not necessary
Respirator If required by patient
Oxygen If required by patient
Motion cushions Useful
Use of contrast agents Yes Intracranial
MR Safety Guide Arterial Disease

A3.27 Contributed by P. Ellen Grant, Pamela W. Schaefer, and R. Gilberto Gonzalez A1.1.1
Current Protocols in Magnetic Resonance Imaging (2007) A1.1.1-A1.1.13
Current Protocols in Magnetic Resonance Imaging Supplement 11
C 2007 by John Wiley & Sons, Inc.
Copyright  Supplement 14
 Advantage Windows Workstation version 3.1P using standard GEMS Advantage Win-
dows software or FuncTool version 1.9M.
Diffusion imaging uses a spin echo (SE) echo planar sequence with additional bipolar
gradients around the 180◦ RF (radio frequency) pulse to increase the sensitivity to
molecular motion. This includes blood flow and water diffusion. Water diffusion is
highly anisotropic in the brain due to the highly ordered, tightly packed, and anisotropic
alignment of white matter tracts. When the bipolar gradient is oriented parallel to the
white matter tracts, water diffuses readily and signal loss occurs. If the gradients are
oriented perpendicular to the tracts in a direction of restricted water diffusion, no signal
loss occurs. Thus, the calculation of an apparent diffusion coefficient must balance these
anisotropic effects by measuring signal changes in ≥3 directions. In fact, a full description
would require the measurement of the full diffusion tensor but this is not usually used
clinically. The raw data from the diffusion-weighted sequence must be processed to
provide both the trace diffusion-weighted image (commonly referred to as the DWI)
and the apparent diffusion coefficient (ADC) map. Most diffusion sequences, including
the GE diffusion sequence, process the data automatically to provide the DWI and the
spin echo (SE) echo-planar (EP) T2 sequence but most do not automatically provide
the ADC map. The standard GE diffusion sequence can be processed on an Advantage
Windows Workstation version 3.1P using FuncTool version 1.9M to obtain ADC maps.
With the authors’ custom diffusion sequence, both DWI and ADC maps are automatically
obtained.
The perfusion of blood through the brain parenchyma can be monitored by tracking the
first passage of a rapid, well defined bolus of intravenously administered MR contrast
agent through the brain using echo planar imaging. There is a magnetic susceptibility
induced signal reduction as the bolus of paramagnetic contrast agent, confined to the
vascular compartment, passes through the brain. The time series data from the perfusion
sequence must be processed in order to obtain relative cerebral blood volume (rCBV),
time to peak (TTP), relative cerebral blood flow (rCBF), and mean transit time (MTT)
maps. The most accurate method for obtaining these parameters is to convert plots
of signal intensity versus time to plots of R2 * = (1/T2 * ) versus time. A filter is
used to decrease high frequency noise prior to conversion of signal intensity to R2 * .
Next, to determine the bolus dispersion in the smaller arterial feeders and tissue, the
bolus dispersion that occurs in the larger feeding artery is removed by deconvolving the
tissue curve with the arterial input curve, typically the middle cerebral artery. Although
the deconvolution step is not needed for the calculation of rCBV or TTP, the authors
typically calculate all parameters after the deconvolution. In order to reduce errors from
motion artifact, time points with excessive motion are deleted prior to processing. This
perfusion data analysis is custom software. However, perfusion analysis software is also
available with FuncTool version 1.9M that will give similar results.
For perfusion-weighted imaging in adults, the use of a double dose of contrast agent
injected by an MR compatible injection pump, such as a MEDRAD injector, at a rate
of 5 ml/sec through an 18-G angiocatheter is recommended. If an injection pump is not
available, perfusion sequences can still be performed with a hand injector using a 3-way
stopcock to inject a contrast bolus followed rapidly by a saline flush. The contrast agent
should be injected ∼10 sec after the start of the scan to provide enough images prior to
contrast agent arrival to determine the precontrast baseline and to provide enough images
after the contrast agent arrival to see the entire first pass of the bolus.
Cerebral Infarct/ NOTE: Be sure that technologists and nurses have immediate access to any emergency
Intracranial equipment that may be relevant to a given study, or that may be needed for a particular
Cerebrovascular patient, such as crash carts or oxygen.
Disease
A1.1.2
Supplement 14 Current Protocols in Magnetic Resonance Imaging
Materials
Normal saline (0.9% NaCl), sterile
Gadolinium-based MR contrast agent (e.g., Magnevist, Omniscan, or Prohance)
Set up patient and equipment
1. Interview (screen) the patient to ensure that he or she has no contraindications such
as cardiac pacemakers or other implants containing ferromagnetic materials. Also be
sure to find out if the patient has any health conditions that may require the presence
of special emergency equipment during the scanning procedure, or necessitate any
other precautions. Do not forget to ask if the patient has any drug allergies, and
document them.
Generally, standard screening forms are used for all patients scanned in a magnetic
resonance system.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(1996) for discussion of what implants may be safely scanned using magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member, who
can sit in the room during the scan and comfort the patient as needed. This companion
must be screened as well to ensure the absence of loose metal objects on the body or
clothing.
2. If the procedure is a research protocol, have the patient sign any necessary consent
form.
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating
and image artifacts.
5. Inform the patient about what will occur during the procedure, what he or she will
experience while in the magnet, and how to behave, including the following:
a. If earphones or headphones are used to protect the ears from the loud sounds
produced by the gradients, the patient will be asked to wear these, but will be able
to communicate with you at any time during the imaging.
b. The patient will be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this works).
c. For good results the patient should not talk, and should avoid or minimize swal-
lowing or other movement, during each scan—i.e., as long as the banging sounds
continue. Between scans, talking and swallowing are allowed in most cases, but
should be avoided when comparative positional studies are being performed; the
patient will be informed when this is the case.
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
6. If an 18-G i.v. is not already in place, an i.v. with an 18-G angiocatheter should be
started.
7. Have the patient mount onto the table. Either before or right after the patient lies
down, set up any triggering devices or other monitoring equipment that is to be
used.
8. Connect the MR compatible injection pump that is cleared of air and loaded with a
double dose of contrast agent and with saline flush, to the patient. Do a small test Intracranial
Arterial Disease
A1.1.3
Current Protocols in Magnetic Resonance Imaging Supplement 14
 injection to ensure that the i.v. is working properly so that contrast agent can be Table A1.1.2 Rapid Sagittal T1 Scout
injected when the perfusion sequence is performed. Patient position Supine
9. Center the patient in a head or neck coil at the region where the key information Scan type Gradient echo
is desired. Make sure that the head and neck are constrained to prevent motion, Imaging plane (orientation) Sagittal
especially if high-resolution scans are to be run. Central slice or volume center Nasion
Echo time (TE ) Minimum (at least 4.1 msec)
Generally the patient’s head is fixed so that the head is horizontal (not tilted) and the
Receiver bandwidth (RBW) 15.6 kHz
neck and head lie along the axis of the patient table; other positions may be appropriate
depending on the needs at hand. Repeat time (TR ) 100 msec
Flip angle (FA) 60◦
Most scanners have a special neck coil for MRA; otherwise, a head coil should be used
Fields of view (FOVx , FOVy ) 240 mm, 240 mm
and the patient placed as far in as possible so that the bifurcation of the common carotid
artery into the internal and external carotid arteries can be imaged. Resolution (x, y) 0.94 mm, 1.25 mm
Number of data points collected (Nx , Ny ) 256, 192
10. If needed, place a pillow or other support under the knees to make the patient more Display matrix (Dx , Dy ) 256, 256
comfortable. Slice thickness (z) 7 mm
11. Use the centering light focused on the nasion to position the patient and put him or Number of slices 7
her into the center of the magnet. Slice gap 2 mm
Number of excitations (NEX) 1
Once this step has been performed, so long as the patient does not move on the table,
Number of acquisitions (Nacq ) 1
the table itself can be moved and then replaced in the same position as before without
jeopardizing the positioning of one scan relative to another. Read direction Superior–inferior
Scan time 16 sec
12. If the patient is unable to hold still, provide an appropriate sedative.
13. Program the 5 ml/sec injection rate for both contrast agent and saline, contrast agent Table A1.1.3 Transverse Diffusion
dose required (double dose by weight), saline flush required (typically 40 ml), and Patient position Supine
10-sec time delay. Start the saline running to keep the vein open and arm the injection Scan type Single shot SE-EPI, diffusion
pump. Do not inject the contrast agent until the perfusion sequence. Imaging plane (orientation) Transverse
Pulse sequence database (PSD) epi2NV
Sequence 1: Rapid sagittal T1 scout
Central slice or volume center Nasion
14. To determine the patient’s position, a fast sagittal scout scan is performed using
Echo time (TE ) 101 msec
the imaging sequence in Table A1.1.2. (Some medical centers may prefer a rapid
Repeat time (TR ) 7500 msec
three-plane scout instead of this multislice sagittal scout.)
Flip angle (FA) 90◦
Sequence 2: Transverse diffusion Fields of view (FOVx , FOVy ) 220 mm, 220 mm
15. From the sagittal scout, select the image through the center of the brain to set up the Resolution (x, y) 1.72 mm, 1.72 mm
locations for the transverse diffusion sequence. The transverse images should begin Number of data points collected (Nx , Ny ) 128, 128
at the foramen magnum and end at the top of the brain. Set up imaging parameters Display matrix (Dx , Dy ) 128, 128
as shown in Table A1.1.3. Slice thickness (z) 5 mm
Number of slices 24
16. Warn the patient that the sequence results in loud beeping (banging) noises and begin Slice gap 1 mm
the scan.
Number of excitations (NEX) 3
Data processing and viewing for sequence 2 Number of acquisitions (Nacq ) 1
17. Most diffusion sequences will perform a minimum of 4 sequences: three with orthog- Read direction Right–left
onal diffusion gradient directions and one with a minimal or no diffusion gradient. Saturation pulses Fat saturation (automatic with EPI)
Control variables (CV) Ramp sampling = 1,
Often, a fifth set of images is also provided, which combines the 3 orthogonal gradient
burst sampling = 0
images to produce a set of images whose signal intensity is not affected by the diffusion
direction. This is the DWI. In order to obtain apparent diffusion coefficient (ADC) maps, b-value 1000 sec/mm2
the images may need further processing. On a GE system, the ADC maps can be obtained Scan time 2 min, 38 sec
by processing on the Advantage Window Workstation version 3.1P using FuncTool version
1.9M.

Sequence 3: Transverse T2 *-weighted gradient echo

Cerebral Infarct/ 18. To detect hemorrhages, this gradient echo sequence should be performed. From the
Intracranial
Cerebrovascular sagittal scout, select the image through the center of the brain to set up the locations Intracranial
Disease for the transverse T2 * -weighted gradient echo sequence. The locations should be Arterial Disease

A1.1.4 A1.1.5
Supplement 14 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 14
 Table A1.1.4 Transverse T2 -Weighted Gradient Echo Table A1.1.6 Transverse FLAIR
Patient position Supine Patient position Supine
Scan type 2-D gradient echo Scan type Inversion recovery fast spin echo
Imaging plane (orientation) Transverse Imaging plane (orientation) Transverse
Variable bandwidth Yes Variable bandwidth Yes
Central slice or volume center Nasion Central slice or volume center Nasion
Echo time (TE ) 25 msec Echo time (TE ) 140 msec
Receiver bandwidth (RBW) 15.63 kHz Receiver bandwidth (RBW) 20.83 kHz
Repeat time (TR ) 750 msec Echo train length (ETL) 14
Flip angle (FA) 20◦ Repeat time (TR ) 10,000 msec
Fields of view (FOVx , FOVy ) 240 mm, 240 mm Inversion time (TI ) 2200 msec
Resolution (x, y) 0.94 mm, 1.25 mm Flip angle (FA) 180◦
Number of data points collected (Nx , Ny ) 256, 192 Fields of view (FOVx , FOVy ) 240 mm, 240 mm
Display matrix (Dx , Dy ) 256, 256 Resolution (x, y) 0.94 mm, 1.25 mm
Slice thickness (z) 5 mm Number of data points collected (Nx , Ny ) 256, 192
Number of slices 23 Display matrix (Dx , Dy ) 256, 256
Slice gap 1 mm Slice thickness (z) 5 mm
Number of excitations (NEX) 2 Number of slices 14
Number of acquisitions (Nacq ) 1 Slice gap 1 mm
Read direction Anterior–posterior Number of excitations (NEX) 1
Flow compensation Yes Number of acquisitions (Nacq ) 2
Scan time 3 min, 41 sec Read direction Anterior–posterior
Scan time 5 min, 20 sec
Table A1.1.5 Transverse T2 -Weighted FSE
Patient position Supine the same as those chosen for the transverse diffusion sequence. Set up the imaging
Scan type Fast spin echo parameters as shown in Table A1.1.4.
Imaging plane (orientation) Transverse
Central slice or volume center Nasion 19. Warn the patient that the sequence is starting and begin the scan.
Variable bandwidth Yes Sequence 4: Transverse T2 -weighted FSE
Pulse sequence database (PSD) FSE-XL
20. From the sagittal scout, select the image through the center of the brain to set up
Echo time (TE ) 102 msec
the locations for the transverse T2 -weighted FSE sequence. The locations should be
Receiver bandwidth (RBW) 11.36 kHz
the same as those chosen for the transverse diffusion sequence. Set up the imaging
Echo train length (ETL) 12
parameters as shown in Table A1.1.5.
Repeat time (TR ) 6000 msec
Flip angle (FA) 90◦ 21. Warn the patient that the sequence is starting and begin the scan.
Fields of view (FOVx , FOVy ) 240 mm, 240 mm
Sequence 5: Transverse FLAIR
Resolution (x, y) 0.94 mm, 1.25 mm
22. From the sagittal scout, select the image through the center of the brain to set up the
Number of data points collected (Nx , Ny ) 256, 192
locations for the transverse FLAIR sequence. The locations should be the same as
Display matrix (Dx , Dy ) 256, 256
those chosen for the transverse diffusion sequence. Set up the imaging parameters
Slice thickness (z) 5 mm
as shown in Table A1.1.6.
Number of slices 28
Slice gap 1 mm 23. Warn the patient that the sequence is starting and begin the scan.
Number of excitations (NEX) 2
Sequence 6: Transverse 3-D TOF MRA
Number of acquisitions (Nacq ) 1
Read direction Anterior–posterior
24. From the sagittal scout, select the image through the center of the brain to set up the
locations for the transverse 3-D TOF MRA sequence. Set up the imaging parameters
Flow compensation Yes
as shown in Table A1.1.7. Two slabs are posted on the midline sagittal image,
Extended dynamic range (EDR) Yes
centered at the top of the sella turcica.
Scan time 2 min, 48 sec
Cerebral Infarct/
25. Warn the patient that the sequence is starting and begin the scan.
Intracranial
Cerebrovascular Intracranial
Disease Arterial Disease

A1.1.6 A1.1.7
Supplement 14 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 14
 Table A1.1.7 Transverse 3-D TOF MRA Table A1.1.8 Transverse 2-D Phase Contrast
Patient position Supine Patient position Supine
Scan type 3-D short-TR gradient echo Scan type 2-D Vascular phase contrast gradient
Imaging plane (orientation) Transverse echo
Variable bandwidth Yes Imaging plane (orientation) Transverse
Central slice or volume center Nasion Central slice or volume center Nasion
Echo time (TE ) 3.2 msec Echo time (TE ) 6.7 msec
Receiver bandwidth (RBW) 15.6 kHz Repeat time (TR ) 33 msec
Repeat time (TR ) 33 msec Flip angle (FA) 20◦
Flip angle (FA) 20◦ Fields of view (FOVx , FOVy ) 220 mm, 165 mm
Fields of view (FOVx , FOVy ) 160 mm, 160 mm Resolution (x, y) 0.86 mm, 1.03 mm
Resolution (x, y) 0.63 mm, 0.84 mm Number of data points collected (Nx , Ny ) 256, 160
Number of data points collected (Nx , Ny ) 256, 192 Display matrix (Dx , Dy ) 256, 256
Display matrix (Dx , Dy ) 256, 256 Slice thickness (z) 15 mm
Slice thickness (z) 1.4 mm Number of slices 3
Number of slices per slab 32 Slice gap 0 mm
Number of slabs 2 Number of excitations (NEX) 4
Slab overlap 4 slices Number of acquisitions (Nacq ) 3
Slice gap 0 mm Read direction Anterior–posterior
Number of excitations (NEX) 1 Flow compensation Yes
Number of acquisitions (Nacq ) 1 Slice series Sequential
Read direction Anterior–posterior Vascular options Flow reconstruction type = phase
Flow compensation Yes difference
ZIP 512 Yes Velocity encoding = 80 cm/sec
ZIP 2 Yes Acquired flow directions = all
Extended dynamic range (EDR) Yes Collapse = on
Saturation pulses Superior Flow analysis = off
Magnetization transfer Yes Additional flow images = R/L, A/P,
Vascular options Reprojections = 19◦ S/I, MAG
Collapse = on Scan time 2 min, 6 sec
Ramp pulse = inferior to superior
Scan time 6 min, 32 sec Sequence 8: Transverse perfusion
29. Using the transverse FSE T2 -weighted images, find the slice location where the
middle cerebral arteries are best seen. Using the imaging parameters as shown in
Data processing and viewing for sequence 6 Table A1.1.9, set up the transverse images starting at the same location that the MCA’s
26. Once the transverse partitions from the 3-D TOF MRA are obtained, volumes of (middle cerebral artery) were identified on the FSE T2 -weighted images. Typically,
interest (VOI) that select out the posterior circulation, the right anterior circulation, only 10 or 11 slices can be obtained in the given TR interval. The slice thickness and
and the left anterior circulation should be selected, and the maximum intensity gap can be modified so that the perfusion study covers any diffusion abnormality.
projections (MIPs) of these 3 VOIs rotated through 360◦ about the vertical axis, 30. Check that the injection pump is ready to inject the contrast agent and that the
saving images at 15◦ intervals. A fourth VOI that selects out both anterior circulations, injection of the contrast agent is set to start 10 sec after the scan begins.
should be rotated through 360◦ about the horizontal axis, saving images at 15◦
intervals. 31. Warn the patient that the sequence is starting and that part of the way through the
scan the i.v. injection will occur. Warn the patient that a cool sensation may be felt
Sequence 7: Transverse 2-D phase contrast during the injection. Begin the scan.
27. From the sagittal scout, select the image through the center of the brain to set up the
Two doses of contrast agent (0.2 mmol/kg) are given here.
locations for the transverse 2-D PC MRA sequence. Set up the imaging parameters
as shown in Table A1.1.8. Three slices are posted on the midline sagittal image, Process data and view for sequence 8
centered at the top of the sella turcica. 32. Process the perfusion images.
Cerebral Infarct/ 28. Warn the patient that the sequence is starting and begin the scan. The perfusion images can be processed on an Advantage Windows Workstation version
Intracranial 3.1P using FuncTool version 1.9M to give maps of relative cerebral blood volume (rCBV),
Cerebrovascular relative cerebral blood flow (rCBF), and mean transit time (MTT). Intracranial
Disease Arterial Disease

A1.1.8 A1.1.9
Supplement 14 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 14
 Table A1.1.9 Transverse Perfusion Sequence 11: Transverse T2 -weighted FSE
4. Run transverse T2 -weighted FSE sequence (see Basic Protocol, sequence 4, steps 20
Patient position Supine
and 21).
Scan type Single shot SE-EPI
Imaging plane (orientation) Transverse Sequence 12: Transverse perfusion
5. Run transverse perfusion sequence (see Basic Protocol, sequence 8, steps 29 and 32).
Variable bandwidth Yes
Pulse sequence database (PSD) epi2perf Sequence 13: 2-D phase contrast MRA
Central slice or volume center Nasion 6. Run transverse 2-D phase contrast MRA sequence (see Basic Protocol, sequence 7,
steps 27 and 28).
Echo time (TE ) 65 msec
Repeat time (TR ) 1520 msec COMMENTARY
Flip angle (FA) 90◦
Background Information the infarct will increase in size, although the
Fields of view (FOVx , FOVy ) 220 mm, 220 mm Strokes are the third leading cause of death mechanism of this interval growth is not yet
Resolution (x, y) 1.72 mm, 1.72 mm and the major cause of adult long-term dis- clear.
ability. The concept of “Brain Attack” has The 3-D TOF MRA is useful for determin-
Number of data points collected (Nx , Ny ) 128, 128
emphasized the importance of rapid treatment ing the patency of major vessels and the degree
Display matrix (Dx , Dy ) 128, 128 in stroke management. To enable rapid treat- of intracranial atherosclerotic disease. The
Slice thickness (z) 5 or 6 mm ment, rapid detection of an acute stroke is es- 2-D PC MRA gives a lower resolution evalua-
sential. With the high sensitivity of diffusion- tion of vessel patency but is significantly faster
Number of slices 11
weighted imaging to acute stroke (Gonzalez and therefore more appropriate when time is
Slice gap 1 mm et al., 1999), it has become the mainstay of an issue. It is adequate to detect vessel cut-off.
Number of excitations (NEX) 1 stroke evaluation. In animal models, diffusion- The 2-D PC MRA also provides information
weighted imaging can detect acute ischemic regarding the direction of flow in major ves-
Number of acquisitions (Nacq ) 1
injury as early as 10 min after arterial occlu- sels that can help determine routes of collateral
Read direction Left–right sion (Kucharczyk et al., 1991). In the authors’ flow in cases with vascular occlusion.
Slice location Inferior slice to include MCA clinical experience, diffusion-weighted imag- The standard anatomical MR sequences,
ing can detect hyperacute infarcts in humans namely the FLAIR and T2 -weighted FSE se-
Multi-phase 46 phases per location, minimum
as early as 30 min after a witnessed event, well quences, allow us to assess the condition of
delay between acquisitions
before the standard T2 -weighted or FLAIR the brain parenchyma and determine if there
Slice series Interleaved imaging becomes abnormal. In the subacute is evidence of old ischemic events. The T2 * -
Control variables (CV) Ramp sampling = on, stages, when there is mixture of chronic and weighted gradient echo sequence is probably
burst sampling = off recent events, diffusion-weighted imaging is as accurate as computed tomography (CT) at
helpful in determining which of many bright determining if there is evidence of blood prod-
Scan time 1 min, 11 sec
areas on T2 -weighted images is the more re- ucts in the setting of acute stroke (Perl et al.,
cent lesion. The DWI, when combined with 1999).
the ADC map, can also help determine the age
ALTERNATE RAPID IMAGING FOR ACUTE CEREBRAL INFARCT of the lesion (Warach et al., 1995; Burdette Critical Parameters and
PROTOCOL et al., 1998; 1999).
When an acute arterial stroke is suspected, MR imaging with diffusion and perfusion Troubleshooting
imaging provides information that is necessary for rapid, optimal patient treatment. If Perfusion-weighted imaging provides com- For diffusion-weighted imaging, the “b-
the patient is unstable, if the patient is an intra-arterial thrombolysis candidate, or if the plementary information about blood volume in value” is a critical parameter in determin-
patient must be squeezed into an already over-booked MR schedule, imaging is kept and blood flow through the tissue. This infor- ing ADC values for different tissues, partic-
mation may be helpful in assessing the risk ularly when the b-value is between 1000 and
to a minimum. The choice of T2 -weighted sequence for sequence 3 depends on user
for infarct growth and the degree of hemo- 2000 sec/mm2 (Le Bihan et al., 1991; Mulkern
preference. Arguments for T2 -weighted FSE sequence include shorter scan time with
dynamic compromise (Sorensen et al., 1996; et al., 1999. The human studies on stroke imag-
good spatial resolution, for FLAIR, increased sensitivity to cortical and periventricular Rordorf et al., 1998; Sorensen et al., 1999). ing in the literature have all been done with
hyperintensity, and for gradient echo, increased sensitivity to blood products. In most clinical settings where imaging is per- b-values close to 1000 sec/mm2 and the
formed at least a few hours after the event, the protocol described here uses a b-value of
Set up patient and equipment
area of decreased rCBV best predicts the fi- 1000 sec/mm2 . At this b-value, there is a good
1. Use the same equipment and perform the same equipment and patient set up as in nal infarct size (Sorensen et al., 1999). When signal to noise ratio (SNR) and the ADC values
Basic Protocol, steps 1 to 13. the areas of decreased relative cerebral blood of gray and white matter are essentially equal.
flow and elevated mean transit time are signif- With increasing b-values, the SNR drops and
Sequence 9: Rapid sagittal scout
icantly larger than the diffusion abnormality, ADC values of gray and white matter start
2. Run rapid sagittal scout (see Basic Protocol, sequence 1, step 14). the ipsilateral ICA (internal carotid artery) or to diverge. Many systems now have the ca-
Cerebral Infarct/
Intracranial MCA is typically occluded. When such areas pability of doing diffusion-weighted imaging
Cerebrovascular
Sequence 10: Transverse diffusion Intracranial
of “mismatch” between relative blood flow and at b-values >1000 sec/mm2 but extreme cau-
Disease 3. Run transverse diffusion sequence (see Basic Protocol, sequence 2, steps 15 to 17). diffusion exist, there is an increased risk that tion must be used when interpreting these Arterial Disease

A1.1.10 A1.1.11
Supplement 14 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 14
 results. The TE is also an important parame-
ter in diffusion-weighted imaging. To optimize
SNR, the shortest possible TE is used. TR has
no effect on image contrast. With increasing or
decreasing TR , the number of slices obtained
increases or decreases.
For perfusion-weighted imaging, a spin
echo echo-planar sequence or a gradient echo
echo-planar sequence can be used. In this pro-
tocol, a spin echo echo-planar sequence is cho-
sen because it has a greater contribution from
small vessels (Boxerman et al., 1995). The
trade-off is a decreased change in signal with
the bolus of contrast compared to the gradient
echo sequence. In order to improve the signal
change, a double dose of contrast agent is used
for the spin echo sequence that results in a sig-
nal drop of at least 10%. In perfusion weighted
imaging, increasing TR increases the number
of slices available but at the cost of increas-
ing the temporal spacing between points on
the signal intensity versus time curve. Short
spacing in time enables better characteriza-
tion of the bolus. The TR chosen in this se-
quence is what the authors determined to be
the best compromise between these two fac-
tors. The TE is also an important parameter to
optimize. Varying the TE not only alters the
SNR but also changes the contribution of dif-
ferent vessel sizes on relative blood volume
and relative blood flow calculations (Boxer-
man et al., 1995). The authors have chosen a
TE of 65 msec to best optimize SNR and the
contribution of small diameter vessels.
Important basic parameters in MR angiog-
raphy include TR , flip angle, and TE . The TR
is chosen to be long enough to avoid satu-
ration of flowing blood but short enough to
improve contrast by saturation of background
tissue. TE is typically minimized to minimize
flow related dephasing. The flip angle must be
optimized to be short enough to avoid satura-
tion of small vessels but long enough to give
sufficient suppression of background tissue.
Most vendors have ramped flip angles where
the flip angle is gradually increased in the di-
rection of the flowing blood to minimize satu-
ration. Interpolation, performed by zero filling
the acquisition matrix (number of data points
collected) in k-space, is helpful in reducing
partial volume averaging effects. Typically, the
authors interpolate in both the slice select and
the phase encoding directions. In cases with
high-grade stenosis, turbulence may result in
Cerebral Infarct/ an over estimation of the degree of vascular
Intracranial narrowing. Decreasing the TR and flip angle
Cerebrovascular can minimize these effects at the cost of de-
Disease
creased background suppression. Other alter-
A1.1.12
Supplement 14
natives include contrast agent administration 1999. Multicomponent apparent diffusion coef- W.A., Look, R.B., Finklestein, S.P., Rosen,
or black blood techniques (see Table A7.4.10). ficients in human brain. N.M.R. Biomed. 12:51- B.R., and Koroshetz, W.J. 1996. Hyperacute
62. stroke: Evaluation with combined multisec-
Perl, J., 2nd, Tkach, J.A., Porras-Jimenez, M., tion diffusion-weighted and hemodynamically
Anticipated Results weighted echo-planar MR imaging. Radiology
Lieber, M., Obuchowski, N., Ross, J.S., Ding,
Diffusion-weighted imaging should allow 199:391-401.
X.P., Ruggieri, P.M., Shearer, D.M., Khajavi,
detection of ischemic necrosis as early as K., and Masaryk, T.J. 1999. Hemorrhage de- Sorensen, A.G., Copen, W.A., Ostergaard, L.,
30 min and at least as long as 10 days after tected using MR imaging in the setting of acute Buonanno, F.S., Gonzalez, R.G., Rordorf, G.,
the ischemic event. Perfusion-weighted imag- stroke: An in vivo model. Am. J. Neuroradiol. Rosen, B.R., Schwamm, L.H., Weisskoff, R.M.,
ing provides complementary information on 20:1863-1870. and Koroshetz, W.J. 1999. Hyperacute stroke:
Rordorf, G., Koroshetz, W.J., Copen, W.A., Simultaneous measurement of relative cerebral
the relative blood volume and blood flow that
Cramer, S.C., Schaefer, P.W., Budzik, R.F. blood volume, relative cerebral blood flow, and
helps determine areas of hemodynamic com- mean tissue transit time. Radiology 210:519-
Jr., Schwamm, L.H., Buonanno, F., Sorensen,
promise. The MRA allows assessment of the 527.
A.G., and Gonzalez, G. 1998. Regional is-
degree of underlying cerebral vascular disease chemia and ischemic injury in patients with Warach, S., Gaa, J., Siewert, B., Wielopolski, P.,
and can detect areas of vascular occlusion. acute middle cerebral artery stroke as de- and Edelman, R.R. 1995. Acute human stroke
fined by early diffusion-weighted and perfusion- studied by whole brain echo planar diffusion-
weighted MRI. Stroke 29:939-943. weighted magnetic resonance imaging. Ann.
Time Considerations Neurol. 37:231-241.
The full assessment of patients at risk for Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
an ischemic event often includes neck MR an- Philadelphia.
giography as well as the Basic Protocol de- Contributed by P. Ellen Grant, Pamela W.
Sorensen, A.G., Buonanno, F.S., Gonzalez, R.G.,
scribed above. This results in a study of at Schwamm, L.H., Lev, M.H., Huang-Hellinger, Schaefer, and R. Gilberto Gonzalez
least 45 min in length. When time is crucial Massachusetts General Hospital
F.R., Reese, T.G., Weisskoff, R.M., Davis, T.L.,
for optimal patient treatment, as in candidates Suwanwela, N., Can, U., Moreira, J.A., Copen, Boston, Massachusetts
for intra-arterial thrombolysis, the Alternate
Protocol is used to minimize imaging time yet
provides the critical information required for
decisions on urgent patient management.
Literature Cited
Boxerman, J.L., Hamberg, L.M., Rosen, B.R., and
Weisskoff, R.M. 1995. MR contrast due to
intravascular magnetic susceptibility perturba-
tions. Magn. Reson. Med. 34:555-566.
Burdette, J.H., Ricci, P.E., Petitti, N., and Elster,
A.D. 1998. Cerebral infarction: Time course of
signal intensity changes on diffusion-weighted
MR images. Am. J. Roentgenol. 171:791-795.
Burdette, J.H., Elster, A.D., and Ricci, P.E. 1999.
Acute cerebral infarction: Quantification of
spin-density and T2 shine-through phenomena
on diffusion-weighted MR images. Radiology
212:333-339.
Gonzalez, R.G., Schaefer, P.W., Buonanno, F.S.,
Schwamm, L.H., Budzik, R.F., Rordorf, G.,
Wang, B., Sorensen, A.G., and Koroshetz, W.J.
1999. Diffusion-weighted MR imaging: Di-
agnostic accuracy in patients imaged within
6 hours of stroke symptom onset. Radiology
210:155-162.
Kucharczyk, J., Mintorovitch, J., Asgari, H.S.,
and Moseley, M. 1991. Diffusion/perfusion MR
imaging of acute cerebral ischemia. Magn.
Reson. Med. 19:311-315.
Le Bihan, D., Moonen, C.T., van Zijl, P.C., Pekar,
J., and DesPres, D. 1991. Measuring random
microscopic motion of water in tissues with MR
imaging: A cat brain study. J. Comput. Assisted
Tomogr. 15:19-25.
Mulkern, R.V., Gudbjartsson, H., Westin, C.F.,
Zengingonul, H.P., Gartner, W., Guttmann, C.R.,
Robertson, R.L., Kyriakos, W., Schwartz, R., Intracranial
Holtzman, D., Jolesz, F.A., and Maier, S.E. Arterial Disease
A1.1.13
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 14
Rule Out (R/O) Intracranial Aneurysm
When imaging patients for intracranial aneurysm, the goals are: (1) to assess the contour
of the intracranial arteries, particularly in the regions of the ACOM (anterior communi-
cating artery), PCOM (posterior communicating artery), ICA (internal carotid artery)
bifurcation, MCA (middle cerebral artery) trifurcation, basilar tip, and PICA (posterior
inferior cerebellar artery); (2) to assess the anatomy of the Circle of Willis and direction
of flow, and; (3) to determine if there is evidence of a recent subarachnoid bleed. The 3-D
time-of-flight (TOF) magnetic resonance angiography (MRA) is used for assessing
contour and anatomy of the major intracranial arteries. The 2-D phase contrast (PC)
sequence is used to determine the direction of flow. A fluid-attenuated inversion recovery
(FLAIR) sequence is used to search for evidence of a recent subarachnoid bleed. If
vasospasm is a clinical concern, diffusion- and perfusion-weighted imaging are added to
the imaging assessment. The following Basic Protocol can be used for the evaluation of
stable patients, and requires 20 min, 18 sec to complete.
STANDARD IMAGING FOR ANEURYSM DETECTION
In order to perform the optional sequences for perfusion and diffusion studies, a scanner
with echo-planar capabilities is required (Table A1.2.1). MR angiography and the standard
anatomical MR imaging sequences included in the protocol do not require these faster
gradients. The parameters given here are optimized for a 1.5T General Electric (GE) LX
system with software version 8.35 and may require modification if using different
software versions, field strengths, or manufacturers.
Data processing is required for the 3-D TOF MRA sequence described in this protocol.
For optimal assessment of the volumetric 3-D TOF MRA data, separate volumes of
interest (VOI) for the left anterior, right anterior, and posterior circulation should be
created. Images from 15°-rotations about the vertical axis should be saved. An additional
anterior circulation volume of interest should include both anterior cerebral arteries and
the region of the anterior communicating artery for visual inspection. Images from
15°-rotations about the horizontal axis should also be saved. Although maximum intensity
projections (MIP) are superior to volume and surface reconstructions for assessment of
vascular patency, surface reconstructions can be helpful in visualizing aneurysms. All of
our MRA data are processed on a GE Advantage Windows Workstation version 3.1P using
standard GE Advantage Windows software or FuncTool version 1.9M. In the hands of an
experienced technologist, the postprocessing will require ∼5 min.
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.
Table A1.2.1 Equipment Requirements for Cerebral Vascular
Assessment
Coil type Quadrature head or neurovascular coil
Gradient coil strength 22 mT/m (or whatever the system permits)
Slew rate ≥120 T/m/sec
Cardiac gating Not necessary
Peripheral gating Not necessary
Respirator If required by patient
Oxygen If required by patient
Motion cushions Useful
Use of contrast agents Yes
Contributed by P. Ellen Grant, R. Gilberto Gonzalez, and Pamela W. Schaefer
Current Protocols in Magnetic Resonance Imaging (2002) A1.2.1-A1.2.12
Copyright © 2002 by John Wiley & Sons, Inc.
UNIT A1.2 Set up patient and equipment
1. Interview (screen) the patient to ensure that he or she has no counterindications such
as cardiac pacemakers or other implants containing ferromagnetic materials. Also be
sure to find out if the patient has any health conditions that may require the presence
of special emergency equipment during the scanning procedure, or necessitate any
other precautions. Do not forget to ask if the patient has any drug allergies and
document them.
Generally, standard screening forms are used for all patients scanned in a magnetic
resonance system (see APPENDIX 1).
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(1996) for discussion of what implants may be safely scanned using magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
BASIC 2. If the procedure is a research protocol, have the patient sign any necessary consent
PROTOCOL
forms.
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating
and image artifacts.
5. Inform the patient about what will occur during the procedure, what he or she will
experience while in the magnet, and how to behave, including the following:
a. If earphones or headphones are used to protect the ears from the loud sounds
produced by the gradients, the patient will be asked to wear these, but will be able
to communicate with you at any time during the imaging.
b. The patient will be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this works).
c. For good results the patient should not talk, and should avoid or minimize
swallowing or other movement, during each scan—i.e., as long as the banging
sounds continue. Between scans, talking and swallowing are allowed in most
cases, but should be avoided when comparative positional studies are being
performed; the patient will be informed when this is the case.
d. Nevertheless, the patient may call out at any time if he or she feels that it is
necessary.
6. Have the patient mount onto the table. Either before or right after the patient lies
down, set up any triggering devices or other monitoring equipment that is to be used.
7. Center the patient in a head or neurovascular coil at the region where the key
information is desired. Make sure that the head and neck are constrained to prevent
motion, especially if high-resolution scans are to be run.
Generally, the patient’s head is fixed so that the head is horizontal (not tilted) and the neck
and head lie along the axis of the patient table; other positions may be appropriate
depending on the needs at hand.
Most scanners have a special neurovascular coil for MRA; otherwise, a head coil should
Rule Out (R/O)
Intracranial Intracranial be used and the patient is placed as far in as possible so that the bifurcation of the main
Arterial Disease Aneurysm carotid artery into the internal and external carotid arteries can be imaged.
A1.2.1 A1.2.2
Supplement 6 Supplement 6 Current Protocols in Magnetic Resonance Imaging
 Table A1.2.2 Rapid Sagittal T1-Weighted Scout Table A1.2.4 Transverse 3-D TOF MRA

Patient position Supine Patient position Supine

Scan type 2-D gradient echo Scan type 3-D vascular TOF short TR gradient
Imaging plane (orientation) Sagittal echo
Central slice or volume center Nasion Imaging plane (orientation) Transverse
Echo time (TE) Minimum (at least 4.1 msec) Variable bandwidth Yes
Receiver bandwidth (RBW) ±15.6 kHz Central slice or volume center Nasion
Repeat time (TR) 100 msec Echo time (TE) 3.2 msec
Flip angle (FA) 60° Receiver bandwidth (RBW) ±15.6 kHz
Fields of view (FOVx, FOVy) 240 mm, 240 mm Repeat time (TR) 33 msec
Resolution (Δx, Δy) 0.94 mm, 1.88 mm Flip angle (FA) 20°
Number of data points collected (Nx, Ny) 256, 128 Fields of view (FOVx, FOVy) 220 mm, 165 mm
Display matrix (Dx, Dy) 256, 256 Resolution (Δx, Δy) 0.86 mm, 0.86 mm
Slice thickness (Δz) 7 mm Number of data points collected (Nx, Ny) 256, 192
Number of slices 7 Display matrix (Dx, Dy) 256, 256
Slice gap 2 mm Slice thickness (Δz) 1.4 mm
Number of excitations (NEX) 1 Number of slices per slab 32
Number of acquisitions (Nacq) 1 Number of slabs 2
Read direction Anterior–posterior Slab overlap 4 slices
Scan time 16 sec Slice gap 0 mm
Number of excitations (NEX) 1
Number of acquisitions (Nacq) 1
Read direction Anterior–posterior
Table A1.2.3 Transverse FLAIR
Flow compensation Yes
Patient position Supine ZIP 512 Yes
Scan type Inversion recovery, fast-spin echo ZIP 2 Yes
Imaging plane (orientation) Transverse Saturation pulses Superior
Variable bandwidth Yes Extended dynamic range (EDR) Yes
Central slice or volume center Nasion Magnetization transfer Yes
Echo time (TE) 140 msec Vascular options Reprojections = 19°
Collapse = on
Receiver bandwidth (RBW) ±20.83 kHz
Ramp pulse = inferior to superior
Echo train length (ETL) 12 (not selectable on GE’s system)
Scan time 6 min, 32 sec
Repeat time (TR) 10,000 msec
Inversion time (TI) 2200 msec
Flip angle (FA) 180°
Fields of view (FOVx, FOVy) 220 mm, 165 mm 8. If needed, place a pillow or other support under the knees to make the patient more
Resolution (Δx, Δy) 0.86 mm, 0.86 mm comfortable.
Number of data points collected (Nx, Ny) 256, 192
Display matrix (Dx, Dy) 256, 256 9. Place the patient such that the laser light is centered over the patient’s nasion, and put
Slice thickness (Δz) 5 mm him or her into the center of the magnet.
Number of slices 12 Once this step has been performed, so long as the patient does not move on the table, the
Slice gap 1 table itself can be moved and then replaced in the same position as before without
Number of excitations (NEX) 1 jeopardizing the positioning of one scan relative to another.
Number of acquisitions (Nacq) 2
10. If the patient is unable to hold still, provide an appropriate sedative.
Read direction Anterior–posterior
Scan time 5 min, 20 sec
Sequence 1: Rapid sagittal T1-weighted scout
11. To validate the patient’s position, perform a fast sagittal scout scan using the imaging
parameters in Table A1.2.2. (Some medical centers may prefer a rapid three-plane
scout instead of this multi-slice sagittal scout.)
Rule Out (R/O)
Intracranial Intracranial
Arterial Disease Aneurysm

A1.2.3 A1.2.4
Current Protocols in Magnetic Resonance Imaging Supplement 6 Supplement 6 Current Protocols in Magnetic Resonance Imaging
 Table A1.2.5 Transverse 2-D Phase Contrast Table A1.2.6 Transverse T2-Weighted FSE

Patient position Supine Patient position Supine

Scan type 2-D vascular phase contrast gradient echo Scan type Fast-spin echo
Imaging plane (orientation) Transverse Imaging plane (orientation) Transverse
Central slice or volume center Nasion Variable bandwidth Yes
Echo time (TE) 6 msec Pulse sequence database (PSD) FSE-XL
Repeat time (TR) 20 msec Central slice or volume center Nasion
Flip angle (FA) 20° Echo time (TE) 102 msec
Fields of view (FOVx, FOVy) 220 mm, 165 mm Receiver bandwidth (RBW) ±11.36 kHz
Resolution (Δx, Δy) 0.86 mm, 1.29 mm Echo train length (ETL) 12
Number of data points collected (Nx, Ny) 256, 128 Repeat time (TR) 6000 msec
Display matrix (Dx, Dy) 256, 256 Flip angle (FA) 90°
Slice thickness (Δz) 10 mm Fields of view (FOVx, FOVy) 220 mm, 165 mm
Number of slices 3 Resolution (Δx, Δy) 0.86 mm, 0.86 mm
Slice gap 0 mm Number of data points collected (Nx, Ny) 256, 192
Number of excitations (NEX) 4 Display matrix (Dx, Dy) 256, 256
Number of acquisitions (Nacq) 3 Slice thickness (Δz) 5 mm
Read direction Anterior–posterior Number of slices 28
Flow compensation Yes Slice gap 1 mm
Slice series sequential Number of excitations (NEX) 2
Vascular options Flow reconstruction type = phase difference Number of acquisitions (Nacq) 1
Velocity encoding = 80 cm/sec Read direction Anterior–posterior
Acquired flow directions = all Flow compensation Yes
Collapse = on Extended dynamic range (EDR) Yes
Flow analysis = off
Scan time 2 min, 48 sec
Additional flow images = R/L, A/P, S/I, MAG
Scan time 1 min, 33 sec

Sequence 2: Transverse FLAIR Sequence 4 (optional): Transverse 2-D phase contrast

12. From the sagittal scout, select the image through the center of the brain to set up the
locations for the transverse FLAIR sequence. The locations should be the same as
those chosen for the transverse diffusion sequence (if it will be run at a latter time).
Set up the imaging parameters according to Table A1.2.3.
13. Warn the patient that this sequence is starting and begin the scan.
Sequence 3: Transverse 3-D TOF MRA
14. From the sagittal scout, select the image through the center of the brain to set up the
locations for the transverse 3-D TOF MRA. Set up the imaging parameters according
to Table A1.2.4.
Two slabs are posted on the midline sagittal image, centered at the top of the sella tursica.
15. Warn the patient that this sequence is starting and begin the scan.
Data processing and viewing for sequence 3
16. Once the transverse partitions from the 3-D TOF MRA have been obtained, select
three volumes of interest (VOI) that select out the posterior circulation, the right
anterior circulation and the left anterior circulation, and rotate the maximum intensity
projections or surface reconstructions of these three VOIs through 360° about the
vertical axis, saving images at 15°-intervals. Rotate a fourth VOI, that selects out both
anterior circulations, through 360° about the horizontal axis, saving images at
15°-intervals.
Flow direction information may be helpful in determining which vessels contribute flow
to the aneurysm.
17. From the sagittal scout, select the image through the center of the brain to set up the
locations for the transverse 2-D PC MRA. Set up the imaging parameters according
to Table A1.2.5.
Three slices are posted on the midline sagittal image, centered at the top of the sella tursica.
18. Warn the patient that this sequence is starting and begin the scan.
Sequence 5: Transverse T2-weighted fast-spin echo (FSE)
19. From the sagittal scout, select the image through the center of the brain to set up the
locations for the transverse T2-weighted FSE sequence. The locations should be the
same as those chosen for the transverse diffusion sequence (if it will be used). Set up
the imaging parameters according to Table A1.2.6.
20. Warn the patient that this sequence is starting and begin the scan.
Rule Out (R/O)
Intracranial Intracranial
Arterial Disease Aneurysm

A1.2.5 A1.2.6
Current Protocols in Magnetic Resonance Imaging Supplement 6 Supplement 6 Current Protocols in Magnetic Resonance Imaging
ANEURYSM SCREEN AND ASSESSMENT FOR VASOSPASM/INFARCTION ALTERNATE Table A1.2.7 Transverse Diffusion
PROTOCOL
When there is concern for vasospasm and infarction, the imaging sequences in the Basic
Patient position Supine
Protocol should be performed with the addition of the diffusion and perfusion sequences
Scan type Single-shot spin-echo-EPI (echo
described below. planar imaging), diffusion
Materials Imaging plane (orientation) Transverse
Pulse sequence database (PSD) epi2NV
Normal saline (0.9% NaCl), sterile
Central slice or volume center Nasion
Gadolinim-DTPA contrast agent (e.g., Magnevist, Omniscan, or Prohance)
Echo time (TE) 99.3 msec
Repeat time (TR) 6000 msec
Set up equipment and patient
Flip angle (FA) 90°
1. Repeat Basic Protocol, steps 1 to 5.
Fields of view (FOVx, FOVy) 220 mm, 220 mm
2. Establish an intravenous (i.v.) line from which the contrast agent can be injected, and Resolution (Δx, Δy) 1.72 mm, 1.72 mm
attach this line securely to the patient so that movement into or out of the magnet will Number of data points collected (Nx, Ny) 128, 128
not pull at the patient’s arm. Display matrix (Dx, Dy) 128, 128
Slice thickness (Δz) 5 mm
It is preferable to insert the line prior to imaging and to leave the patient in the magnet,
with no intervening motion, between the scans run before contrast agent injection and those Number of slices 23
run after injection. Slice gap 1 mm
Number of excitations (NEX) 3
3. Connect the MR compatible injection pump that is cleared of air and loaded with a Number of acquisitions (Nacq) 1
double dose of contrast agent and with saline flush, to the patient. Do a small test Read direction Right–left
injection to ensure that the i.v. is working properly. Saturation pulses Fat saturation (automatic with EPI)
4. Program the 5 ml/sec injection rate for both contrast agent and saline, contrast agent Control variables (CV) Ramp sampling = 1, burst sampling = 0
dose required (double dose by weight), saline flush required (typically 40 ml) and b-value 1000 sec/min2
10-sec time delay. Start the saline running to keep the vein open, and arm the injection Scan time 2 min, 38 sec
pump. Do not inject the contrast agent at this time.
Table A1.2.8 Transverse Perfusion
5. Repeat Basic Protocol, steps 6 to 10.
Patient position Supine
Sequences 1 to 5 Scan type Single-shot spin-echo-EPI
6. Perform sequences 1 to 5 as described in the Basic Protocol. Pulse sequence database (PSD) epi2perf
Imaging plane (orientation) Transverse
Sequence 6: Transverse diffusion Variable bandwidth Yes
7. From the sagittal scout, select the image through the center of the brain to set up the Central slice or volume center Nasion
locations for the transverse diffusion-weighted sequence. The transverse images Echo time (TE) 65 msec
should begin at the foramen magnum and end at the top of the brain. Set up imaging Receiver bandwidth (RBW) ∼200 kHz (not selectable on GE’s system)
parameters according to Table A1.2.7. Repeat time (TR) 1500 msec
8. Warn the patient that this sequence results in loud beeping noises and begin the scan. Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 220 mm, 220 mm
Data processing and viewing for sequence 6 Resolution (Δx, Δy) 1.72 mm, 1.72 mm
9. Process the diffusion-weighted images. Number of data points collected (Nx, Ny) 128, 128
Display matrix (Dx, Dy) 128, 128
Most diffusion-weighted sequences will perform a minimum of four sequences: three with Slice thickness (Δz) 5 or 6 mm
orthogonal diffusion gradient directions and one with a minimal or no diffusion gradient.
Number of slices 11
Often, a fifth set of images is also provided that combines the three orthogonal gradient
images to produce a set of images whose signal intensity is not affected by the diffusion Slice gap 1 mm
direction. This is the diffusion-weighted trace image (DWI). In order to obtain Apparent Number of excitations (NEX) 1
Diffusion Coefficient (ADC) maps the images may need further processing. On a GE Number of acquisitions (Nacq) 1
system, the ADC maps can be obtained by processing on the Advantage Window Worksta- Read direction Left–right
tion version 3.1P using FuncTool version 1.9M. Slice location Inferior slice to include MCA
Multi-phase Yes, 46 phases per location, minimum delay
between acquisitions
Rule Out (R/O) Slice series Interleaved
Intracranial Intracranial
Arterial Disease Aneurysm Control variables (CV) Ramp sampling = on, burst sampling = off
Scan time 1 min, 11 sec
A1.2.7 A1.2.8
Current Protocols in Magnetic Resonance Imaging Supplement 6 Supplement 6 Current Protocols in Magnetic Resonance Imaging
Sequence 7: Transverse perfusion
10. Using the transverse FSE T2-weighted images, find the slice location where the
middle cerebral arteries are best seen. Using the imaging parameters as shown in A B
Table A1.2.8, set up the transverse images starting at the same location as the MCA
were identified on the T2-weighted FSE images. Typically, only 10 or 11 slices can
be obtained in the given TR interval. The slice thickness and gap can be modified so
that the perfusion study covers any diffusion abnormality.
11. Check that the injection pump is ready to inject the contrast agent and that the
injection is set to start 10 sec after the scan begins.
12. Warn the patient that this sequence is starting and that part of the way through the
scan the iv injection will occur. Warn the patient that a cool sensation may be felt
during the injection. Begin the scan.
A dose of 0.2 mmol/kg (which is equivalent to two usual doses) of contrast agent is usually
given.

Data processing and viewing for sequence 7

13. Process the perfusion images on an Advantage Windows Workstation version 3.1P
using FuncTool version 1.9M to give maps of relative cerebral blood volume (rCBV),
relative cerebral blood flow (rCBF) and mean transit time (MTT).

COMMENTARY
Background Information
Subarachnoid hemorrhage (SAH) due to
rupture of intracranial aneurysms is a major
cause of cerebrovascular disease related death
(Whisnant et al., 1982; Ingall et al., 1989).
Between 1% and 7.9% of the population have
unruptured aneurysms (Locksley, 1969;
Rosenorn et al., 1988; Atkinson et al., 1989;
Ingall et al., 1989; Inagawa and Hirano, 1990;
Iwata et al., 1991; Nakagawa and Hashi, 1994;
Rinkel et al., 1998) with annual risks of rupture
estimated to be 0.5% (International Study of
Unruptured Intracranial Aneurysms Investiga-
tors, 1998) to 2.3% (Yasui et al., 1997). Mor-
tality rates with SAH have been reported to be
as high as 52% to 67% (Winn et al., 1977;
Whisnant et al., 1982; Juvela et al., 1993; Yasui
et al., 1997; International Study of Unruptured
Intracranial Aneurysms Investigators, 1998).
With the advent of MR (and computed to-
mography; CT) angiography, there has been
increasing interest in screening for the presence
of aneurysms, particularly in patients with
higher risks such as patients with polycystic
kidney disease, connective tissue disorders, or
two first-degree relatives with intracranial
aneurysms. The sensitivity of 3-D TOF MRA
at 1.5T for detecting aneurysms has been re-
ported to be 75% dropping to 58% for aneu-
rysms <5 mm (Korogi et al., 1997). When all
image sequences are closely scrutinized includ-
ing the 3-D TOF MRA partition data and MIP,
sensitivity has been reported to be 81% with
100% specificity in the setting of acute SAH
(Wilcock et al., 1996).
In most medical centers, MR (or CT)
angiography has replaced catheter angiography
in screening patients at risk for intracranial
aneurysms and to follow known aneurysms. In
patients with acute subarachnoid hemorrhage,
it can also be used prior to catheter angiography
to plan the optimal projections on catheter
angiography and to help visualize the surgical
approach. CT angiography is beginning to pro-
vide strong competition in this area, especially
for presurgical planning where bony landmarks
are often helpful to the surgeon.
With the location of aneurysms being most
common in the ACOM, MCA bifurcation,
PCOM, carotid siphon, basilar tip, and PICA
origin, these regions are the main target areas
for MR angiography. MR angiography is not
sensitive at detecting aneurysms in distal
branches and therefore if distal aneurysms,
such as mycotic aneurysms, are suspected,
catheter angiogram is still the study of choice.
FLAIR imaging is highly sensitive in the
detection of subarachnoid hemorrhage in the
acute stage (Fig. A1.2.1; Singer et al., 1998;
Bakshi et al., 1999) but artifacts from flowing
CSF (cerebrospinal fluid) that is not suppressed
may be difficult to differentiate from true
subarachnoid blood since both are hyperin-
tense. Some authors suggest using proton den-
Figure A1.2.1 Transverse FLAIR images at the level of the basal cisterns (A) and level of the basal ganglia
(B). Increased signal is seen throughout the subarachnoid spaces including the anterior interhemispheric
fissure, basal cisterns and bilateral sylvian fissures. High signal material is also seen layering in the posteriorly
in the trigones of lateral ventricles in (B). The high signal is due to acute subarachnoid and intraventricular
blood.
sity–weighted images in conjunction with representations are difficult to interpret, surface
FLAIR images to help differentiate flow arti- reconstructions may improve aneurysm visu-
facts from blood since all cases of subarachnoid alization. Also, multiplanar (including curved)
blood seen on CT were also seen on the proton reformations may help in optimizing aneurysm
density–weighted sequence (Wiesmann et al., visualization, particularly visualization of the
1999). The authors find correlation with the aneurysm neck. If there is a large and/or par-
T2-weighted FSE sequence helpful in distin- tially thrombosed aneurysm, a post contrast
guishing flow artifacts but still obtain noncon- 3-D TOF MRA may be helpful. With the post-
trast CT studies prior to MRI to rule out acute contrast MRA, the TR can be decreased and the
SAH. flip angle increased compared to the pre-con-
In patients with vasospasm, those that were trast study to improve visualization.
symptomatic showed areas of increased signal With FLAIR sequences, it is important to
on diffusion-weighted imaging consistent with make sure that the TR and T1 values have not
areas of acute ischemia. In these same patients, been modified and that vendor-specific recom-
perfusion-weighted imaging showed larger ar- mendations for optimal CSF suppression have
eas of increased transit time in a distribution been followed. For example, in GE FLAIR
that correlated with clinical symptoms (Ror- sequences, at least 2 interleaved acquisitions
dorf, 1999). This study indicated that there was are required for optimal CSF suppression.
a role for diffusion- and perfusion-weighted
imaging in the diagnosis and evaluation of Anticipated Results
patients with symptomatic vasospasm. The 3-D TOF MRA data should allow
accurate 3-D visualization of aneurysms >3
Critical Parameters and mm in size as well as visualization of the
Troubleshooting aneurysm neck and relationship to other in-
Important basic parameters in MR angiog- tracranial structures (Fig. A1.2.2). When
raphy are discussed in UNIT A1.1. If the MIP MIPs are confusing, multiplanar reforma-
Rule Out (R/O)
Intracranial Intracranial
Arterial Disease Aneurysm

A1.2.9 A1.2.10
Current Protocols in Magnetic Resonance Imaging Supplement 6 Supplement 6 Current Protocols in Magnetic Resonance Imaging

A 339:1725-1733.
B Aneurysms and Subarachnoid Hemorrhage: A
Figure A1.2.2 MIP of the right anterior circulation, ACOM and left A2 segment (A) from the same
patient shown in Figure A1.2.1. A small aneurysm arising from the right ACOM/A2 junction, is seen
pointing inferiorly and to the right (arrow). (B) MIP of the left anterior circulation in a different patient
shows a small aneurysm (arrow) arising from the left ICA just above the origin of the ophthalmic
artery.
tions may be helpful to differentiate an in-
fundibulum from an aneurysm and may help in
the assessment of the aneurysm neck. Partially
thrombosed aneurysms may have regions of
intrinsically bright T1-weighted signal due to
methemaglobin but these can be distin-
guished from flow related enhancement by
viewing the partition data.
Literature Cited
Atkinson, J.L.D., Sindt, T.M., Houser, O.W., and
Whisnant, J.P. 1989. Angiographic frequency of
anterior circulation intracranial aneurysms. J.
Neurosurg. 70:551-555.
Current Protocols in Magnetic Resonance Imaging
Bakshi, R., Kamran, S., Kinkel, P.R., Bates, V.E.,
Mechtler, L.L., Janardhan, V., Belani, S.L., and
Kinkel, W.R. 1999. Fluid-attenuated inversion-
recovery MR imaging in acute and subacute
cerebral intraventricular hemorrhage. Am. J.
Neuroradiol. 20:629-636.
Inagawa, T. and Hirano, A. 1990. Autopsy study of
unruptured incidental intracranial aneurysms.
Surg. Neurol. 34:361-365.
Ingall, T.J., Whisnant, J.P., Wiebers, D.O., and
O’Fallon, W.M. 1989. Has there been a decline
in subarachnoid hemorrhage mortality? Stroke
20:718-724.
International Study of Unruptured Intracranial
Aneurysms Investigators. 1998. Unruptured in-
tracranial aneurysms: Risk of rupture and risks
of surgical intervention. N. Engl. J. Med.
Iwata, K., Misu, N., Terada, K., Kawai, S., Momose,
M., and Nakagawa, H. 1991. Screening for un-
ruptured asymptomatic intracranial aneurysms
in patients undergoing coronary angiography. J.
Neurosurg. 75:52-55.
Juvela, S., Porras, M., and Heiskanen, O. 1993.
Natural history of unruptured intra-cranial aneu-
rysms: A long-term follow-up study. J. Neuro-
surg. 79:174-182.
Korogi, Y., Takahashi, M., Mabuchi, N., Watabe, T.,
Shiokawa, Y., Shiga, H., O’Uchi, T., Nakagawa,
T., Miki, H., Horikawa, Y., Fujiwara, S., and
Furuse, M.1997. MR angiography of intracranial
aneurysms: A comparison of 0.5 T and 1.5 T
Comput. Med. Imaging Graph. 21:111-116.
Locksley, H.B. 1969. Natural history of subarach-
noid hemorrhage, intracranial aneurysms and ar-
teriovenous malformations. In Intracranial
Cooperative Study (A.L. Sahs, G.E. Perret, and
H.R. Locksley, eds.) pp. 37-108. Lippincott,
Philadelphia.
Nakagawa, T. and Hashi, K. 1994. The incidence
and treatment of asymptomatic, unruptured cere-
bral aneurysms. J. Neurosurg. 80:217-223.
Rinkel, G.J.E., Djibuti, M., Algra, A., and van Gijn,
J. 1998. Prevalence and risk of ruptured of intrac-
ranial aneurysms: A systematic review. Stroke
29:251-256.
Rordorf, G., Koroshotz, W.J., Copen, W.A., Gon-
zalez, G., Yamada, K., Schaefer, P.W.,
Schwamm, L.H., Ogilvy, C.S., and Sorensen,
A.G. 1999. Diffusion-weighted MR imaging:
Diagnostic accuracy in patients within 6 hours of
stroke symptom onset. Radiology 210:155-162.
Rosenorn, J., Eskesen, V., and Schmidt, K. 1988.
Unruptured intracranial aneurysms: An assess-
ment of the annual risk of rupture based on
Rule Out (R/O)
Intracranial Intracranial
Arterial Disease Aneurysm
A1.2.11 A1.2.12
Supplement 6 Supplement 6
epidemiological and clinical data. Br. J. Neuro-
surg. 2:369-378.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Singer, M.B., Atlas, S.W., and Drayer, B.P. 1998.
Subarachnoid space disease: Diagnosis with
fluid-attenuated inversion-recovery MR imaging
and comparison with gadolinium-enhanced
spin-echo MR imaging-blinded reader study.
Radiology 208:417-422.
Whisnant, J.P., Phillips, L.H., and Sundt, T.M. 1982.
Aneurysmal subarachnoid hemor-rhage: Timing
of surgery and mortality. Mayo Clin. Proc.
57:471-475.
Wiesmann, M., Mayer, T.E., Medele, R., and Bruck-
mann, H. 1999. Diagnosis of acute subarachnoid
hemorrhage at 1.5 Tesla using proton-density
weighted FSE and MRI sequences. Radiologe
39:860-865.
Wilcock, D., Jaspan, T., Holland, I., Cherryman, G.,
and Worthington, B. 1996. Comparison of mag-
netic resonance angiography with conventional
angiography in the detection of intracranial
aneurysms in patients presenting with subarach-
noid haemorrhage. Clin. Radiol. 51:330-334.
Winn, H.R., Richardson, A.E., and Jane, J.A. 1977.
The long-term prognosis in untreated cerebral
aneurysms, I: The incidence of late hemorrhage
in cerebral aneurysms: 10-year evaluation of 364
patients. Ann. Neurol. 1:358-370.
Yasui, N., Suzuki, A., Nishimura, H., Suzuki, K.,
and Abe, T. 1997. Long-term follow-up study of
unruptured intracranial aneurysms. Neurosur-
gery 40:1155-1160.
Contributed by P. Ellen Grant, R. Gilberto
Gonzalez, and Pamela W. Schaefer
Massachusetts General Hospital
Boston, Massachusetts
Current Protocols in Magnetic Resonance Imaging
Extracranial Carotid Artery Disease
Magnetic resonance angiography (MRA) is a noninvasive means by which to study the
integrity of the vascular system (Potchen et al., 1992; Anderson et al., 1993).
This unit presents two MRA techniques for imaging the carotid arteries, time-of-flight
MRA (see Basic Protocol 1) and contrast-enhanced MRA (see Alternate Protocol). As a
complement to MRA, following either of the preceding two procedures, high-resolution
spin-echo scanning can be used to study the vessel wall in more detail (see Basic Protocol
2). The parameters given here are derived from experience at 1.5 T and may need to be
altered slightly depending on the field strength and the equipment manufacturer. In
particular, optimal TE may vary with different field strength.
IMAGING THE CAROTID ARTERIES BY TIME-OF-FLIGHT MRA
Magnetic resonance angiography (MRA) scans can be run at a range of field strengths
with varying degrees of success. The high signal-to-noise ratio (SNR) obtained at a
high-field strength makes it possible to run the MRA scans either faster or with higher
resolution than at lower fields. (Low-field strength is generally ≤0.2 T, mid-field strength
is between 0.5 T and 1 T, and high-field strength is ≥1.5 T.) At low-field strengths,
contrast-enhanced MRA (see Alternate Protocol) may give better results than conven-
tional time-of-flight (TOF) MRA. The sequences described herein are based on the
authors’ experience with a Siemens 1.5 T Vision scanner, but are expected to be equally
applicable to machines from other manufacturers.
The following three sequences comprise the preferred protocol. In cases where a patient
requires a faster procedure (e.g., for trauma patients or those intolerant to a long scanning
session), either the two-dimensional (2-D) sequence (sequence 2) or the three-dimen-
sional (3-D) sequence (sequence 3) may be omitted. Because the 3-D sequence offers
higher through-plane resolution (1 mm) than the 2-D sequence, some radiologists and
researchers have suggested that 3-D techniques are more accurate than 2-D techniques in
detecting carotid stenosis. However, in older patients, for whom flow can be very slow,
the 2-D sequence may give better results. For patients with tortuous vessels, the contrast-
enhanced method (see Alternate Protocol) may be most appropriate. If desired, this
procedure may be followed by high-resolution spin-echo scanning (see Basic Protocol 2)
for a more detailed view of vessel walls. The entire Basic Protocol 1 takes ∼40 min.
Table A1.3.1 lists the hardware necessary to perform the procedure, along with appropri-
ate parameters. The available gradient strength will depend on the scanner, and the echo
Table A1.3.1 Equipment Parameters for Time-of-Flight MRA
Coil type Circularly polarized neck coil or
neuro-vascular coil (or phased-array neck
coil, if available)
Gradient coil strength 25 mT/m (or whatever the system permits)
Cardiac gating No
Peripheral gating For safety only
Respiratory gating No
Respirator If required by patient
Oxygen If required by patient
Motion cushions Useful
Use of contrast agents Only use in the Alternate Protocol
Contributed by E. Mark Haacke, Daniel Kido, and Karen Tong
Current Protocols in Magnetic Resonance Imaging (2002) A1.3.1-A1.3.13
Copyright © 2002 by John Wiley & Sons, Inc.
UNIT A1.3 times given below may be varied accordingly (the lower the gradient strength, the longer
the echo time for a particular scan).
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.
Set up patient and equipment
1. Interview (screen) the patient to ensure that he or she has no contraindications such
as cardiac pacemakers or other implants containing ferromagnetic materials that may
be problematic for patient safety or good image acquisition. Also be sure to find out
if the patient has any health conditions that may require the presence of special
emergency equipment during the scanning procedure, or necessitate any other pre-
cautions.
BASIC
PROTOCOL 1 Generally, standard screening forms (APPENDIX 1) are used for all patients scanned in a
magnetic resonance system.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(2001) for a discussion of what implants may be safely scanned using magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing,
as well as other items as described above.
2. If the procedure is a research protocol, have the patient sign any necessary consent
forms.
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating
and image artifacts.
5. Inform the patient about what will occur during the procedure, what he or she will
experience while in the magnet, and how to behave, including the following:
a. If earplugs, earphones, or headphones are used to protect the ears from the loud
sounds produced by the gradients, the patient will be asked to wear one of these
devices, but will be able to communicate with you at any time during the imaging.
b. The patient may be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this works).
c. For good results, the patient should not talk, and should avoid or minimize
swallowing or other movement during each scan—i.e., as long as the banging
sounds continue. Between scans, talking and swallowing are allowed in most
cases, but should be avoided when comparative positional studies are being
performed; the patient will be informed when this is the case.
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
6. Have the patient mount onto the table. Either before or right after the patient lies
down, set up any triggering devices or other monitoring equipment that is to be used.
Extracranial
Intracranial Carotid Artery
Arterial Disease Disease
A1.3.1 A1.3.2
Supplement 7 Supplement 7 Current Protocols in Magnetic Resonance Imaging
 7. Center the patient in a neck coil, or a neuro-vascular coil, at the region where the key Table A1.3.2 Primary Clinical Imaging Parameters for Sequence 1 (Pilot Scan)
information is desired. Make sure that the head and neck are constrained to prevent
motion, especially if high-resolution scans are to be run. Patient position Supine
Scan type 3-D gradient echo
Generally, the patient’s head and neck are fixed so that they are horizontal (not tilted) and
lie along the axis of the patient table; other positions may be appropriate depending on Imaging plane (orientation) Sagittal/coronal/transverse (many
the imaging needs. scanners are now capable of triple
plane orthogonal scouts)
Most scanners have a special neck coil or a neuro-vascular coil for MRA; otherwise, a Central slice or volume center Laser light centered on edge of jaw
head coil and/or cervical spine coil should be used and the patient placed so that the
Echo time (TE) As short as possible (e.g., 6 msec)
bifurcation of the main carotid artery into the internal and external carotid arteries can be
imaged. Repeat time (TR) As short as possible (e.g., 15 msec)
Flip angle (FA) 15°
8. If needed, place a pillow or other support under the knees to make the patient more Fields of view (FOVx, FOVy) 300 mm, 300 mm
comfortable. Resolution (Δx, Δy) 1 mm, 1 mm
9. Use the centering light to position the patient at the angle of the mandible and put the Number of data points collected (Nx, Ny) 300, 300
patient into the center of the magnet. Slice thickness (Δz) 8 mm
Once this step has been performed, so long as the patient does not move on the table, the Number of slices 3
table itself can be moved and then returned to the same position as before without Slice gap Not applicable
jeopardizing the positioning of one scan relative to another. Number of acquisitions (Nacq) 1
Swap read and phase encoding Standard default
10. If the patient is unable to hold still, provide an appropriate sedative.
Slice locations Not applicable
Sequence 1: Rapid three-plane positioning pilot Saturation pulses Not applicable
11. To validate the patient’s position, run the system’s pilot (or scout) scan to ensure Scan time 16 sec
correct location of the neck in three dimensions using the imaging sequence given in
Table A1.3.2 or similar parameters. Refer to manufacturer specifications for standard
default settings and swap as necessary. Table A1.3.3 Primary Clinical Imaging Parameters for Sequence 2 (2-D TOF)
This sequence usually consists of three orthogonal planes to allow localization. The images
are often also used later to determine where to place the saturation pulses and to set up Patient position Supine
total coverage of the volume of interest. Scan type 2-D gradient echo
Imaging plane (orientation) Transverse
Sequence 2: 2-D time-of-flight MRA Central slice or volume center Centered on the carotid bifurcation
12. Bring the sequence for a rapid 2-D transverse scan up onto the console. Set the Echo time (TE) As short as possible (usually <7 msec)
imaging parameters as shown in Table A1.3.3. Repeat time (TR) 25 msec
13. Use the pilot (scout) images run earlier (sequence 1) to locate the vessel positions in Flip angle (FA) 35°
three dimensions and ensure coverage of the region of interest. Specifically, make Fields of view (FOVx, FOVy) 256 mm, 160–200 mm
sure the carotid bifurcation is in the center of the field of view. Use the pilot images Resolution (Δx, Δy) 1 mm, 1–1.25 mm or 0.5 mm, 1–1.25
to set up either a saturation pulse caudally to suppress signal from the incoming mm
arterial flow or cephalically (cranially) to suppress signal from the incoming venous Number of data points collected (Nx, Ny) 256, 160 or 512, 160
flow (see Fig. A1.3.1). Display matrix (Dx, Dy) 512, 512
Slice thickness (Δz) 2–3 mm
See Potchen et al. (1992), Anderson et al. (1993), and Haacke et al. (1999) for details of
Number of slices 50–70, or as many as needed to cover
setting up a saturation pulse.
the region of interest
The system usually applies this as a traveling saturation pulse to continue to saturate Slice gap –0.33 mm (overlapping)
venous flow near the plane being imaged but not so close as to interfere with application Number of acquisitions (Nacq) 1
of the usual slice selection process. Read direction Anterior–posterior
14. Let the patient know you are ready, and begin the scan. Slice locations Carotid bifurcation or stenosis
Saturation pulses Yes; cephalically (cranially) thick slab
Data processing and viewing for sequence 2 to saturate veins or caudally thick slab
15. If a better view of the images is necessary, modify the window and level settings and to suppress arteries
obtain a close-up view of a region by zooming in on the data (images). For example, Slice series Ascending or descending, but not
if a stenosis appears in the image that needs to be measured, zoom in on the image interleaveda
by at least a factor of 2 and then use the distance function in the manufacturer’s Scan time ∼5 min
software to determine the width of the stenosis. This can then be compared to a more aThis will help eliminate reconstruction artifacts.
normal portion of the vessel, if possible, either distal to the stenosis (NASCET Extracranial
Intracranial Carotid Artery
criteria) or proximal to the stenosis. Arterial Disease Disease

A1.3.3 A1.3.4
Current Protocols in Magnetic Resonance Imaging Supplement 7 Supplement 7 Current Protocols in Magnetic Resonance Imaging
 Table A1.3.4 Primary Clinical Imaging Parameters for Sequence 3 (3-D TOF)a

Patient position Supine

saturation Scan type 3-D gradient echo
pulse Imaging plane (orientation) Transverse
Central slice or volume center Centered on the carotid bifurcation
Echo time (TE) As short as possible; but using 7 msec
or 2.5 msec to 3.0 msec ensures that
3-D bright dark
excited blood blood water and fat are out of phase
slab Repeat time (TR) 20–40 msec
Flip angle (FA) 15°–25°
Fields of view (FOVx, FOVy) 256 mm, 160 mm
Resolution (Δx, Δy) 1 mm, 0.83 mm
Number of data points collected (Nx, Ny) 512, 192
Display matrix (Dx, Dy) 512, 512
artery vein
Slice thickness (Δz) 1 mm
Number of slices 64
Slab thickness 64 mm
Figure A1.3.1 Schematic of the carotid artery and jugular vein. The images resulting from the 3-D Slice gap −0.38 mm (overlapping)
TOF scan (sequence 3; Fig. A1.3.2) show dramatically reduced signal from the veins. Number of acquisitions (Nacq) 1
Read direction Anterior–posterior
Slice locations Carotid bifurcation or stenosis
Saturation pulses Yes; 50 to 80 mm thick slabs
A B cephalically (cranially) to saturate
veins or caudally to suppress arteries
Fat suppression Yes
Scan time ∼4–8 min
aNeed RF spoiled pulse.

This zoom feature is only useful if it is done by interpolation of the images, not by pixel
duplication. Ideally, this should be done in the in-plane view, but if this is not available or
the data cannot be manipulated using multiplanar reformatting, a through-plane view
through the center of the vessel can be substituted.
16. Once the scan is complete, process the data as needed to produce a maximum intensity
projection (MIP) image (UNIT B7.3) of the vessels at different angles.
The resulting projections can then be viewed as a cine or movie loop to allow the physician
to obtain a more global perspective of the vasculature. If there are questions of accuracy
associated with this processing, go back and view the pristine data by reexamining the
original images or by slicing the original data in a transverse mode (or the view that gives
the best cross-sectional information) via a multiplanar reformatting to better view the
cross-section of the vessel of interest.

C D Sequence 3: 3-D time-of-flight MRA (Fig. A1.3.2)

17. Review the pilot scans and ensure that the saturation pulse is correctly placed above
the slab of interest.
18. Run the 3-D scan using parameters as close as possible to those in Table A1.3.4 to
obtain high signal from the arteries.
This sequence is usually performed in a transverse mode to benefit from a maximum inflow
effect.
Figure A1.3.2 Transverse and coronal planes of a human neck. (A) A transverse image across the neck 19. Process the data from sequence 3 similar to that for sequence 2 (see steps 15 and 16).
without a saturation pulse shows a cross-section of both the carotid artery and the jugular vein (arrows). (B)
After a saturation pulse, the signal from the jugular vein is dramatically suppressed. (C, D) Coronally reformatted Extracranial
images obtained from the same 3-D transverse data sets used to obtain the images shown in (A) and (B), Carotid Artery
Disease
respectively. Note that the internal jugular veins (arrows) have vanished in (B) and (D).
A1.3.5 A1.3.6
Current Protocols in Magnetic Resonance Imaging Supplement 7 Supplement 7 Current Protocols in Magnetic Resonance Imaging
IMAGING THE CAROTID ARTERY BY CONTRAST-ENHANCED ALTERNATE Table A1.3.5 Primary Clinical Imaging Parameters for Sequence 4 (Timing Bolus
TIME-OF-FLIGHT MRA PROTOCOL Scan)

The set of imaging sequences presented in Basic Protocol 1 often suffices to reveal the Patient position Supine
necessary angiographic information. However, the accuracy of non-contrast-enhanced Scan type 2-D gradient echo
MRA is only slightly better than that of an ultrasound. For slow-flow cases, to obtain Imaging plane (orientation) Transverse
better coverage or higher resolution, or where speed is of the essence (such as for elderly Central slice or volume center Centered on the carotid bifurcation
patients, trauma patients, or patients who are uncooperative), an alternative procedure Echo time (TE) 2.3 msec
using a T1-reducing contrast agent is preferred (Prince et al., 1997; Wilman et al., 1998). Repeat time (TR) 4.5 msec
This can be considered a supplementary approach, but in the near future, it is anticipated Flip angle (FA) 8°
that this contrast-enhanced (CE) sequence will eventually replace sequences 2 and 3 Fields of view (FOVx, FOVy) 350 mm, 262 mm
described in Basic Protocol 1. If desired, this procedure may be followed by high-reso- Resolution (Δx, Δy) 1.37 mm, 1.36 mm
lution spin-echo scanning (see Basic Protocol 2) to obtain a more detailed view of vessel Number of data points collected (Nx, Ny) 256, 192
walls. The entire Alternate Protocol takes ∼25 min. Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 10 mm
Materials
Number of slices 1
Gadolinium-based MR contrast agent (e.g., Magnevist, Omniscan, or Prohance) Slice gap N/A
Normal saline (0.9% NaCl), sterile Number of acquisitions (Nacq) 1
Power injector Number of repetitions 60
Read direction Anterior–posterior
Set up patient and equipment
Slice locations Carotid bifurcation or stenosis
1. Set up patient as in Basic Protocol 1, steps 1 to 6.
Scan time ∼52 sec
2. Establish an i.v. line from which the contrast agent can be injected, and attach this
line securely to the patient so that movement into or out of the magnet will not pull
at the patient’s arm. Table A1.3.6 Primary Clinical Imaging Parameters for Sequence 5 (3-D CE TOF)
It is preferable to insert the line prior to imaging and to leave the patient in the magnet, so
that there is no intervening motion between the scans run before contrast agent injection Patient position Supine
and those run after injection. Scan type 3-D gradient echo
Imaging plane (orientation) Coronal
3. Repeat Basic Protocol 1, steps 7 to 10. Depending on the coverage required, use a Central slice or volume center Centered on the carotid bifurcation
neck coil, a neuro-vascular coil, a phased-array head and neck coil, or the body coil. Echo time (TE) As short as possible (usually 2–3 msec)
Repeat time (TR) As short as possible (usually 4–8 msec)
Run pilot scan Flip angle (FA) 10°–45°
4. Repeat a rapid three-plane positioning pilot scan by running sequence 1 (see Basic Fields of view (FOVx, FOVy) 256 mm, 160 mm
Protocol 1). Resolution (Δx, Δy) 1 mm, 1 mm
Number of data points collected (Nx, Ny) 256, 160
Sequence 4: Short TR gradient echo timing bolus scan
Display matrix (Dx, Dy) 256, 256
The timing of the contrast bolus can be critical. Image acquisition should occur during
Slice thickness (Δz) 2–3 mm
the first pass of the bolus, coinciding with arterial enhancement before arrival into adjacent
Number of slices 16–24, depending on the coverage
veins. Ideally, the filling of central k-space should occur during peak arterial enhancement. needed
Circulation times can vary widely among patients although bolus arrival times in the Slab thickness 32–72 mm, depending on the coverage
carotid arteries generally occur 15 to 20 sec after initiation of the bolus. This can be needed
calculated by using a timing run with a small test bolus of contrast agent measured by a Slice gap 0 mm
single slice fast 2-D gradient echo sequence that acquires an image approximately every Number of acquisitions (Nacq) 1
second over the duration of the circulating bolus. Some newer scanners are also capable Read direction Cranial/caudal
of real-time “fluoroscopic MR imaging,” which allows the contrast-enhanced MRA Slice locations Centered along long axis of vessel of
sequence to be manually triggered immediately after the fast 2-D sequence upon obser- interest to include vessel at all levels
vation of peak arterial enhancement. For both injections, a power injector should be used required
to deliver the solutions at a rate of 2 to 3 ml/sec. Scan time ∼20–30 sec

5. Prepare the injector for a dose of 1 to 2 ml of contrast agent to be followed by a 15

to 20 ml bolus of saline flush.
Extracranial
Intracranial Carotid Artery
Arterial Disease Disease

A1.3.7 A1.3.8
Current Protocols in Magnetic Resonance Imaging Supplement 7 Supplement 7 Current Protocols in Magnetic Resonance Imaging
 6. Start sequence 4 according to the parameters in Table A1.3.5 prior to injection so that Table A1.3.7 Primary Clinical Imaging Parameters for Sequence 6 (2-D SE)
it runs continuously for ∼50 sec.
Patient position Supine
7. After 15 sec, start the injection so that the same data are collected prior to injection Scan type Spin echo
for baseline purposes. Imaging plane (orientation) Transverse
8. Find the image in which there is a maximum signal. Some systems come with Central slice or volume center Centered on the abnormality of
interest (commonly at the skull base or
software to allow a region-of-interest plotted over all the sequence acquisitions.
carotid bifurcation)
Obtain the time when the signal is at maximum. This time indicates the scan delay
Echo time (TE) 15 msec
time.
Repeat time (TR) 500–700 msec
A detailed estimation of the scan delay time is described in UNIT A10.1, Basic Protocol 2. Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 256 mm, 200 mm (or as small as
Sequence 5: Short TR 3-D gradient echo post-contrast scan appropriate without significant
9. Let the patient know that you are ready to start the main scan and instruct the patient aliasing)
to remain as motionless as possible. Resolution (Δx, Δy) 0.5 mm, 1 mm
Number of data points collected (Nx, Ny) 512, 200
10. Prepare the injector with a dose of 0.1 to 0.2 mmol/kg of contrast agent to be followed Display matrix (Dx, Dy) 512, 512
by a 15- to 20-ml saline solution flush. Slice thickness (Δz) 3 mm
11. Inject the contrast agent and flush the saline solution. Wait for the scan delay time. Number of slices 10–20, if more slices are needed to
cover the region of interest, then the
12. After obtaining the scan delay time, immediately start the scan according to the scan can be repeated as necessary
parameters given in Table A1.3.6 so that the center of k-space corresponds to the peak Slice gap 0 mm
signal from the analysis in step 8. Number of acquisitions (Nacq) 1
Read direction Anterior–posterior
The center of k-space should be used at the peak of the bolus. Preferably a centric ordered
elliptic k-space coverage will be used. Slice locations Centered about the abnormality
Saturation pulses Yes; caudal and cephalic (cranial)
Waiting too long to start the scan will result in high signal from veins as well as arteries. thick slabs to saturate both arteries and
veins using parallel saturation bands
13. If a more detailed scan is needed to better visualize the vessels and time permits,
Fat suppression Yes (for carotid dissection)
immediately collect a second, higher resolution scan by changing Nxfrom 256 to 512
Slice series Interleaved
and Nyfrom 160 to 320.
Scan time ∼2 min, 30 sec
See Haacke and Masaryk (1989) for a discussion of this high-resolution scan.
The patient can then be removed from the magnet and the data analyzed.
COMMENTARY
IMAGING BLOOD VESSEL WALLS BY SPIN-ECHO MRA BASIC Background Information technique—based, like ultrasound, on the de-
PROTOCOL 2 Atherosclerosis of the carotid artery is an tection of flow through the artery. The related
The inclusion of some sequences to study blood vessel walls after the vessels have been
imaged initially by time-of-flight MRA (see Basic Protocol 1) and an abnormality (e.g., important medical condition because it is asso- imaging technique of contrast-enhanced MRA,
stenosis) has been discovered is likely to become commonplace in the future. The simplest ciated with strokes, the third-greatest cause of on the other hand, does not depend on the
death in the United States. Obtaining detailed presence of significant blood flow and may, in
approach for studying vessel-wall disease is to collect high-resolution images using a
information about carotid artery stenosis (nar- the future, prove to be the most accurate means
spin-echo (SE) or turbo-spin-echo sequence (Yuan et al., 1996; Raynaud et al., 1998). rowing) can be critical to the determination of of assessing the vessel lumen. Both of these
Although this is not yet a common procedure, it is anticipated that this will become part appropriate treatment, e.g., data from the North potentially complementary methods are pre-
of a single overall imaging protocol for evaluating carotid artery disease. The entire Basic American Symptomatic Carotid Endarterec- sented in this unit (see Basic Protocol 1; see
Protocol 2 takes ∼25 min. tomy Trial (NASCET) indicated that in patients Alternate Protocol). Common indications for
with symptomatic carotid stenosis of >70%, the performing carotid MRA include transient is-
Set up patient and equipment incidence of stroke was lower in patients treated chemic attacks, stroke, neck injuries, suspicion
1. Repeat Basic Protocol 1, steps 1 to 10. with surgery than in those receiving medical of carotid dissection, neck tumors, tumor inva-
care only. The traditional means of imaging sion of the carotid artery, carotid bruit on clini-
Sequence 6: High-resolution spin-echo scan stenosis has been the invasive technique of cal exam, or abnormal ultrasound screening
2. Run a spin-echo sequence using the parameters described in Table A1.3.7 to obtain carotid angiography. More recently, ultrasound results prior to carotid bypass surgery. Other
vessel-wall information and general morphological information about the vessels and has been suggested as a non-invasive alterna- methods for imaging blood vessels include ul-
surrounding tissue. tive, but numerous studies have demonstrated trasound and computed tomography angiogra-
Extracranial that it is less accurate. Magnetic resonance phy (CTA). Ultrasound provides good imaging
Intracranial Carotid Artery angiography (MRA) is another non-invasive results only for surface vessels, but can also
Arterial Disease Disease

A1.3.9 A1.3.10
Current Protocols in Magnetic Resonance Imaging Supplement 7 Supplement 7 Current Protocols in Magnetic Resonance Imaging
give information about vessel-wall thickness.
CTA does about as good a job as MRA but
requires the use of ionizing radiation and a
contrast agent.
Critical Parameters and
Troubleshooting
Magnetic resonance angiography without
contrast agents works on the principle of de-
tecting enhanced signal from blood due to the
inflow of fresh spins (see UNIT B7.3 for more
technical details). For a given repeat time (TR),
the choice of flip angle determines not only the
available signal from the blood but also the
available contrast between the blood and sur-
rounding tissue. The faster the flow, the larger
the flip angle that can be used to optimize the
signal from the blood and suppress the signal
from background tissue. For small vessels, high
resolution is key to creating a successful MR
angiogram. This is also true for larger vessels
if it is important to quantify the degree of
stenosis. The actual application of the se-
quences presented in this unit may require some
modifications depending on the age of the pa-
tient, the exact region being evaluated, and the
time available for scanning. For example, new-
borns have very slow arterial flow, whereas in
children aged 2 to 10, flow is very fast, which
makes it possible to increase the slab thickness
without loss of vascular information. Elderly
patients often have tortuous vessels and slow
flow, possibly requiring contrast-enhanced
MRA. Patients with tight stenoses and rapid
flow (up to 4 m/sec) may demonstrate vessels
with signal dropout, where contrast-enhanced
imaging may also help. Artifacts in carotid
artery imaging pose significant problems for
the interpretation of results. Three common
classes of artifacts are discussed below; for
further information, any one of a number of
standard texts, such as Potchen et al. (1992),
may be consulted.
Saturation artifacts
As a patient ages, blood vessels can become
tortuous and blood flow significantly slows
relative to younger people. The winding of the
vessels can lead to retrograde flow in those
patients relative to the presumed upward direc-
tion and this can cause saturation on the 2-D
TOF sequence (see UNIT B7.3). This is less of a
problem with 3-D imaging, where a lower flip
angle is used along with a spatially variable
pulse to try and avoid this difficulty. Still, even
3-D images can suffer if the flow is too slow.
Also, a shorter echo time (TE) can often give
Current Protocols in Magnetic Resonance Imaging
better results because it has better flow com-
pensation and reduced acceleration effects.
A
Signal loss artifacts with gradient echo
imaging
Signal loss in the vessels themselves can be
caused by insufficient velocity compensation.
This can be improved upon by running a veloc-
ity compensated sequence with a shorter echo
time. Significant signal loss throughout the
image or in a large local region in gradient-echo
imaging can be caused by the presence of small
amounts of metals, especially those containing
iron or nickel. Even iron particulates in clothing
can lead to giant gaps in the image. Removing
all clothing, jewelry, and mascara can help
avoid this problem.
Artifacts in post-contrast scanning
Two major problems can occur. If the 3-D
post-contrast scan is started too soon, only the
edges of the arteries will enhance. If it is started
too late, the veins may be as bright as the
arteries making the viewing of the data in the
MIP form very difficult. Ideally, when the tim-
ing is perfect, the arteries will be much brighter
B
than the veins and only the edges of the veins
will appear if at all.
Anticipated Results
The goal in studying the carotid vessels is
to diagnose the degree of narrowing and, if
possible, the damage to the vessel wall itself.
Magnetic resonance imaging offers the neces-
sary contrast to separate the vessels from other
tissue. Conventional 3-D TOF scans are gener-
ally effective for visualizing the carotid arteries
(see Fig. A1.3.2). With a resolution of 1 mm in
all three directions and possibly higher for
in-plane resolution, vessels are well delineated.
With the better modern gradients allowing
shorter echo times and less signal loss from
dephasing, stenoses can be seen clearly. In
principle, it is also possible to quantify the
vessel lumen (i.e., the extent of stenosis), al-
though approved software offering this feature
is not yet widely available from equipment
manufacturers. Using MRA, the carotid arter-
ies can be imaged with a cross-sectional reso-
lution of ∼0.5 mm by 0.5 mm, with a slice Figure A1.3.3 Images from a contrast-enhanced scan immediately after injection of the contrast
thickness of 1 mm. This is sufficient to obtain agent. Data was acquired coronally so that all vessels could be seen in projection mode. The
10 to 20 pixels across the vessel lumen. This in-plane resolution is just high enough to show the diseased vessels. (A) A major stenosis is present
resolution then limits how accurately one can on the right carotid and some abnormality on the left carotid. (B) A magnified view of the right carotid
quantify the stenosis. The use of contrast agents shows the latter problem more clearly.
to enhance resolution is now commonplace.
Extracranial
These agents reduce the T1 of the blood, leading Intracranial Carotid Artery
to a larger signal from blood relative to all other Arterial Disease Disease
A1.3.11 A1.3.12
Supplement 7 Supplement 7 Current Protocols in Magnetic Resonance Imaging
tissues independent of flow, so that the blood
will be bright even in cases of slow or ob-
structed flow. Immediately post injection, only
the arterial blood appears bright; after ∼30 sec,
the signal from venous blood begins to enhance
as well. The advantage of using a contrast agent
is the increased coverage and the rapid imaging
time (see Fig. A1.3.3). In-phase resolution is
sufficient to visualize severe stenoses easily.
Acknowledgments
The authors would like to thank Glenn Fos-
ter, R.T., for reading this material and providing
helpful comments. The authors would also like
to thank Yingjian Yu for checking sequence
parameters.
Literature Cited
Anderson, C.M., Edelman, R.R., and Tarshi, P.A.
(eds.) 1993. Clinical Magnetic Resonance
Angiography. Raven Press, New York.
Haacke, E.M. and Masaryk, T.J. 1989. The salient
f ea t u re s o f MR a ngio graphy. Radiology
173:611-612.
Haacke, E.M., Brown, R.W., Venkatesan, R., and
Thompson, M.R. 1999. Magnetic Resonance
Imaging: Physical Principles and Sequence De-
sign. John Wiley & Sons, New York.
Potchen, J.E., Haacke, E.M., Siebert, J.E., and
Gottschalk, A. (eds.) 1992. Magnetic Resonance
Angiography. C.V. Mosby, St. Louis.
Prince, M.R., Chenevert, T.L., Foo, T.K., Londy,
F.J., Ward, J.S., and Maki, J.H. 1997. Contract
enhanced abdominal MR angiography: Optimi-
zation of imaging delay time by automating the
detection of contrast material arrival in the aorta.
Radiology 203:109-114.
Raynaud, J.-S., Bridal, S.L., Toussaint, J.-F., Fornès,
P., Lebon, V., Berger, G., and Leroy-Willig, A.
1998. Characterization of atherosclerotic plaque
components by high resolution quantitative MR
and US imaging. J. Magn. Reson. Imaging
8:622-629.
Shellock, F.G. 2001. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Current Protocols in Magnetic Resonance Imaging
Wilman, A.H., Riederer, S.J., Huston, J., 3rd., Wald,
J.T., and Debbins, J.P. 1998. Arterial phase ca-
rotid and vertebral artery imaging in 3-D con-
trast-enhanced MR angiography by combining
fluoroscopic triggering with an elliptical centric
acquisition order. Magn. Reson. Med. 40:24-35.
Yuan, C., Skinner, M.P., Kaneki, E., Mitsumori,
L.M., Hayes, C.E., Raines, E.W., Nelson, J.A.,
and Ross, R. 1996. Magnetic resonance imaging
to study lesions of atherosclerosis in the hyper-
lipidemic rabbit aorta. Magn. Reson. Imaging
14:93-102.
Internet Resources
http://www.mravvh.com
Vascular Visible Human Project, providing an over-
view of MRA imaging for different parts of the body.
http://www.mrisafety.com
Covers a number of important patient management
issues related to MR imaging, including recom-
mended safety procedures, a list of metallic implants
that have been tested for MR compatibility, and a
list of other sources on MR safety.
Key References
Potchen, E.J., Haacke, E.M, Siebert, J.E., and
Gottschalk, A. 1993. Magnetic Resonance
Angiography: Concepts and Applications. C.V.
Mosby, St. Louis.
Contains detailed descriptions of the technical as-
pects of magnetic resonance angiography, as well
as numerous clinical examples.
Shellock, F.G. 2001. Magnetic Resonance Proce-
dures: Health effects and safety. CRC Press, Fla..
Covers a number of important patient management
issues related to MR imaging, including recom-
mended safety procedures, a list of metallic implants
that have been tested for MR compatibility, and a
list of other sources on MR safety.
Contributed by E. Mark Haacke
The MRI Institute for Biomedical Research
St. Louis, Missouri
Daniel Kido and Karen Tong
Loma Linda University
Loma Linda, California
Intracranial
Arterial Disease
A1.3.13
Supplement 7
High-Resolution Multi-Contrast MRI of the UNIT A1.4
Carotid Artery Wall for Evaluation of
Atherosclerotic Plaques
MRI can be used for noninvasive detection and characterization of atherosclerotic plaques BASIC
in extracranial carotid arteries (Yuan et al., 2001a). MRI provides information about PROTOCOL
plaque geometry, distribution, and tissue composition. Once atherosclerotic disease is
detected, follow-up scans may be helpful in the evaluation of disease progression and
effect of treatment.
This unit presents a protocol for multi-contrast, high-resolution imaging of the wall of
extracranial carotid arteries. Identification of basic plaque components (fibrous tissue,
lipid-rich necrotic core, calcification) and incidental changes (a rupture or hemorrhage)
can be accomplished by comparison of black-blood images with basic contrast weight-
ings—T1-weighted, T2-weighted, and proton density (PD)–weighted—in conjunction
with bright-blood images obtained using time-of-flight MR angiography (TOF MRA).
This Basic Protocol summarizes the authors’ experience at 1.5 T with the use of several
models of GE scanners (Signa Echospeed, LX, and CVi) and a Philips Gyroscan NT
scanner. Use of the protocol with different field strengths may require special tests in
order to optimize sequence parameters because of changes in relaxation times and the
signal-to-noise ratio (SNR).
The entire protocol consists of seven imaging sequences. The first two sequences are
used to locate the bilateral carotid bifurcations. Sequence 3 (black-blood MRA of the
carotid bifurcation) allows one to precisely determine the position of the bifurcation,
which is then used as a landmark for all subsequent transverse sequences 4 to 7. The
side of the bifurcation chosen for this purpose is referred to below as the index side. The
index side is recorded for use in follow-up examinations, if indicated. Sequence 3 is also
helpful to visualize plaque distribution along the artery. If necessary, this sequence can
be repeated for the contralateral side. Four transverse scans (sequences 4 to 7) provide
black-blood images with basic contrast weightings (T1-weighted, PD-weighted, and
T2-weighted) and bright-blood (3-D TOF) images, which are necessary for plaque
characterization. In order to facilitate further comparisons, coverage of sequences 4
through 7 is prescribed so that slices obtained from different sequences are positioned
at the same distance from the carotid bifurcation of the index side. The Basic Protocol
takes ∼30 to 40 min depending on coverage (number of slices). The entire examination,
including patient setup, will take ∼1 hr.
Table A1.4.1 lists the hardware necessary to perform the procedure, along with the
appropriate parameters. The available gradient strength will depend on the scanner, and
the echo times given in other tables below may be varied accordingly (the smaller the
gradient strength, the longer the echo time for a particular scan).
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.
Set up patient and equipment
1. Interview (screen) the patient to ensure that he or she has no contraindications such
as cardiac pacemakers or other implants containing ferromagnetic materials that may
be problematic for patient safety or good image acquisition. Also be sure to find out Intracranial
Arterial Disease
Contributed by Vasily L. Yarnykh and Chun Yuan A1.4.1
Current Protocols in Magnetic Resonance Imaging (2003) A1.4.1-A1.4.17
Copyright © 2003 by John Wiley & Sons, Inc. Supplement 11
 Table A1.4.1 Equipment Parameters for High-Resolution
Multicontrast MRI of the Carotid Artery Wall

Coil type Phased-array bilateral surface carotid coil

Gradient coil strength
Cardiac gating
Peripheral gating
Respiratory gating
Respirator
Oxygen
Motion cushions
Use of contrast agents
(Fig. A1.4.1)
25 mT/m (or whatever the system permits)
No
Yes (can be omitted if a double
inversion-recovery, DIR, method is used
for blood suppression in sequences 3–6)
No
If required by patient
If required by patient
Useful
No

if the patient has any health conditions that may require the presence of special
emergency equipment during the scanning procedure, or necessitate any other pre-
cautions.
Generally, standard screening forms (see APPENDIX 1) are used for all patients scanned in
a magnetic resonance system.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(2001) for discussion of what implants may be safely scanned using magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing,
as well as other items as described above.
2. If the procedure is a research protocol, have the patient sign any necessary consent
forms.
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating
and image artifacts.
5. Inform the patient of what will occur during the procedure, what he or she will
experience while in the magnet, and how to behave, including the following:
a. If earplugs, earphones, or headphones are used to protect the ears from the loud
sounds produced by the gradients, the patient will be asked to wear one of these
devices, but will be able to communicate with you at any time during the imaging.
b. The patient may be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this works).
c. For good results, the patient should not talk, and should avoid or minimize
swallowing or other movement during each scan—i.e., as long as the banging
sounds continue. Between scans, talking and swallowing are allowed in most
cases, but should be avoided when comparative positional studies are being
MRI of
performed; the patient will be informed when this is the case.
the Carotid
Artery for d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
Atherosclerotic
Plaques
Figure A1.4.1 Photo of the bilateral phased-array carotid coil with head holder (Pathway Medical).
6. Set up the carotid phased-array coil with head holder (Fig. A1.4.1) on the moving
table at approximately the same position as for a standard head coil. Have the patient
mount onto the table in a supine position with head placed on the head holder. After
the patient lies down, attach a peripheral gating lead to the patient’s finger and check
the appearance of the appropriate triggering signal according to manufacturer’s
guidelines.
The need for peripheral gating depends on the particular setup of sequences 3 to 6. Gating
improves quality of blood suppression in sequences 5 and 6, if inflow saturation pulses are
employed. However, gating can be omitted if a double inversion-recovery (DIR) method is
used for blood suppression in sequences 3 through 6, and if the software of the scanner
allows non-gated acquisition with DIR. See annotations in corresponding sequences for
details.
7. Place bilateral receiver panels of the carotid phased-array coil (Fig. A1.4.1) on the
left and right sides of the patient’s neck. Fix the coil panels using setscrews. If
necessary, adjust the coil panels for the comfort of the patient.
Generally, the head of the patient is fixed so that the head is horizontal (not tilted) and the
neck and head lie along the axis of the table.
If a special carotid phased-array coil is unavailable, other small surface coils (like a coil
for temporomandibular joints scanning, TMJ coil) can be used if they can provide sufficient
signal-to-noise ratio.
8. If needed, place a pillow or other support under the knees to make the patient more
comfortable.
9. Use the centering light to position the patient at the angle of the mandible and put
him or her into the center of the magnet.
Intracranial
Arterial Disease

A1.4.2 A1.4.3
Supplement 11 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 11
 Table A1.4.2 Primary Clinical Imaging Parameters for Sequence 1 (Pilot Scan) Table A1.4.4 Primary Clinical Imaging Parameters for Sequence 3 (Oblique 2-D
Black-Blood MRA)a
Patient position Supine
Scan type 2-D gradient echo Patient position Supine
Imaging plane (orientation) Sagittal/coronal/transverse (many Scan type 2-D FSE (fast spin echo) with DIR
scanners are capable of triple plane preparation
orthogonal scouts) Imaging plane (orientation) Oblique-sagittal
Central slice or volume center Laser light centered on edge of jaw Variable bandwidth Yes
Echo time (TE) As short as possible (∼2–6 msec) Central slice or volume center Along line connecting internal and
Repeat time (TR) As short as possible (∼10–15 msec) external carotid arteries
Flip angle (FA) 15° Echo time (TE) As short as possible (∼7–10 msec)
Fields of view (FOVx, FOVy) 250 mm, 250 mm Receiver bandwidth (RBW) 31 kHz or close
Resolution (Δx, Δy) 0.98 mm, 1.95 mm Echo train length (ETL) 16
Number of data points collected (Nx, Ny) 256, 128 Repeat time (TR) 2 R-to-R intervalsa or 1600 msecb
Display matrix (Dx, Dy) 256, 256 Inversion time (TI) 550 msecb
Slice thickness (Δz) 5–7 mm Fields of view (FOVx, FOVy) 160 mm, 160 mm
Number of slices 3–22 (1–12 sagittal, 1–5 coronal, 1–5 Resolution (Δx, Δy) 0.63 mm, 0.63 mm
transverse) Number of data points collected (Nx, Ny) 256, 256
Slice gap 0–2 mm Display matrix (Dx, Dy) 512, 512
Number of acquisitions (Nacq) 1 Slice thickness (Δz) 2 mm
Scan time ∼5–40 sec Number of slices 5–7
Slice gap 0 mm
Number of acquisitions (Nacq) 2
Table A1.4.3 Primary Clinical Imaging Parameters for Sequence 2 (2-D TOF) Read direction Superior–inferior
Slice locations Carotid bifurcation
Patient position Supine Flow compensation No
Scan type 2-D spoiled gradient echo ZIP 512 Yes
Imaging plane (orientation) Transverse Saturation pulses No
Variable bandwidth Yes Fat suppression Yes
Central slice or volume center Close to the carotid bifurcations Slice series Sequential
Echo time (TE) As short as possible (<5 msec) Scan time ∼50–60 sec/slice
Receiver bandwidth (RBW) ±16 kHz or close aDepending on the specifications of the scanner, this sequence may require peripheral gating. Alternatively,
Repeat time (TR) As short as possible (∼10–20 msec) the sequence can be run without gating (if allowed by a scanner) or with gating simulator.
Flip angle (FA) 30° bT for a non-gated sequence. The delay between DIR pulses and an acquisition sequence (if available for
R
Fields of view (FOVx, FOVy) 160 mm, 120 mm control) should be set to 550 msec.
Resolution (Δx, Δy) 0.63 mm, 0.94 mm
Number of data points collected (Nx, Ny) 256, 128 Once this step has been performed, so long as the patient does not move on the table, the
Display matrix (Dx, Dy) 256, 256 table itself can be moved and then returned to the same position as before without
Slice thickness (Δz) 2–3 mm jeopardizing the positioning of one scan relative to another.
Number of slices 30–50 or as many as needed to 10. If the patient is unable to hold still, provide an appropriate sedative.
visualize carotid bifurcations
Slice gap 0 mm Sequence 1: Rapid three-plane positioning pilot
Number of acquisitions (Nacq) 1 11. Run the system’s pilot (or scout) scan to ensure correct location of the neck in
Read direction Left–right three dimensions, using the imaging sequence given in Table A1.4.2 or similar
Slice locations Carotid bifurcation parameters.
Flow compensation Yes
Saturation pulses Yes; superior to saturate veins This sequence usually consists of three orthogonal planes to allow localization. The images
Slice series Sequential, ascending are often also used later to determine where to place the saturation pulses and to set up
total coverage of the volume of interest.
Scan time ∼0.5–2 min
If allowed by a scanner, the authors recommend using a group of slices for each plane, i.e.,
MRI of 8 to 12 sagittal, 3 to 5 coronal, and 3 to 5 transverse slices. Lateral sagittal images can be
the Carotid used to locate carotid bifurcations. If the bifurcations are visible on scout images, a fewer
Artery for number of slices can be used for the next sequence (2-D TOF MRA).
Atherosclerotic Intracranial
Plaques Arterial Disease

A1.4.4 A1.4.5
Supplement 11 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 11
 Table A1.4.5 Alternate Imaging Parameters for Sequence 3 (Oblique 2-D
Black-Blood MRA) with the Use of a Multislice DIR Methoda
A B
Patient position Supine
Scan type 2-D FSE with DIR preparation
Imaging plane (orientation) Oblique-sagittal
Variable bandwidth Yes
Pulse sequence database (PSD) “fse-m”a
Central slice or volume center Along line connecting internal and
external carotid arteries
Echo time (TE) As short as possible (∼7–10 msec)
Receiver bandwidth (RBW) ±31 kHz or close
Echo train length (ETL) 16
Repeat time (TR) 1800 msecb
Inversion time (TI) 270 msecb
Fields of view (FOVx, FOVy) 160 mm, 160 mm
Resolution (Δx, Δy) 0.63 mm, 0.63 mm
Number of data points collected (Nx, Ny) 256, 256
Display matrix (Dx, Dy) 512, 512
Slice thickness (Δz) 2 mm
Number of slices 6b Figure A1.4.2 Setting scan geometry. (A) Position of slices for sequence 3 (oblique sagittal black-blood MRA) using
Slice gap 0 mm transverse 2-D TOF MRA (sequence 2) as a localizer. 2-D TOF MRA image is also used to determine coordinates of the
Number of acquisitions (Nacq) 2 FOV (field of view) center (asterisk) for all subsequent transverse sequences. (B) Finding an inferior-superior coordinate
of the bifurcation (arrow) on the oblique black-blood image obtained in sequence 3.
Number of repetitions 2
Read direction Superior–inferior
This sequence uses sequential slice acquisition. On most scanners, acquired slices are
Slice locations Carotid bifurcation
reconstructed immediately, and the operator may see recently obtained images. When both
Flow compensation No left and right bifurcations are passed and 5 to 10 images are obtained above the bifurca-
ZIP 512 Yes tions, acquisition can be stopped to save scan time.
Saturation pulses No
Fat suppression Yes Sequence 3: Oblique sagittal 2-D black-blood MRA of the carotid bifurcation
Slice series Interleaved 15. Set up parameters according to Table A1.4.4.
Scan time ∼2 min/6 slices This sequence uses a double inversion–recovery (DIR) method (Edelman et al., 1991) to
aA special pulse sequence database (PSD) is required. Currently, it is installed on GE imagers in the authors’ obtain images with heavily suppressed blood signal. In the specifications of the scanners,
and several other institutions. DIR is often referred to as “blood suppression” or “black blood” option. Sequence 3 can
bExact correspondence between T , T , and a number of slices is required. The total number of slices (6)
R I
be run in either non-gated or gated mode. In most commercially available scanner software
corresponds to acquisition of two interleaved groups by 3 slices. The parameters are given for acquisition versions, gating is a default option for black-blood imaging. In the authors’ experience,
of three slices per TR. See Yarnykh and Yuan (2003) for details. gating does not improve blood suppression or image quality for this sequence, and therefore
a custom-designed non-gated sequence is used. In patients with arrhythmias or an
For older models of MR scanners where three-plane scans may be unavailable, use a rapid abnormally fast or slow heart rate, non-gated acquisition is preferable. In such situations,
gradient-echo scan in a sagittal plane. Prescribe 12 to 16 slices to ensure coverage of the a modified (non-gated) pulse sequence or gating simulator should be used.
neck in the left-right direction. Scan time for sequence 3 can be considerably reduced using a new multi-slice DIR
technique (Yarnykh and Yuan, 2003). If a multi-slice DIR pulse sequence is available, set
Sequence 2: 2-D time-of-flight MRA up parameters according to Table A1.4.5.
12. Set up the imaging parameters according to Table A1.4.3. 16. Review the images obtained in sequence 2. Find the carotid bifurcation on the index
13. Use the scout images obtained earlier to locate the carotid arteries and ensure side. Locate the first slice above the bifurcation, where internal and external carotid
coverage of the region of interest. Prescribe slices graphically to cover long enough arteries are seen separately (Fig. A1.4.2A). Use this slice as a locator in the graphic
segments of the carotid arteries, including both bifurcations. Use the pilot images to display. Place a slice group (slab) to be imaged along the line connecting the internal
set up a superior saturation band for suppression of venous flow signal. and external carotid artery (Fig. A1.4.2A).

14. Start scan and ensure the appearance of both left and right bifurcations on the obtained 17. Instruct the patient to refrain from swallowing during the scan and run the scan.

MRI of
images. If, occasionally, bifurcations are not found, repeat this scan with superior or 18. If indicated by a local protocol, repeat steps 16 and 17 for the contralateral side.
the Carotid inferior shift of a slice group (slab) until bifurcations are found. The index side can be marked by a physician or chosen during the examination based on
Artery for
Atherosclerotic the appearance of stenosis on MRA images (sequence 2). Black-blood MRA is also useful Intracranial
Plaques for estimation of plaque distribution along the artery. For purposes of surgical planning, Arterial Disease

A1.4.6 A1.4.7
Supplement 11 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 11
 Table A1.4.6 Primary Clinical Imaging Parameters for Sequence 4 (Transverse
T1-Weighted Black-Blood FSE)a
A B
Patient position Supine
Scan type 2-D FSE with DIR preparation
Imaging plane (orientation) Transverse
Variable bandwidth Yes
Central slice or volume center Carotid bifurcation or stenosis
Echo time (TE) As short as possible (∼7–11 msec)
Receiver bandwidth (RBW) ±21 kHz or close (or ±31 kHz if TR = 2
R-to-R intervals)b
Echo train length (ETL) 10 (or 12 if TR = 2 R-to-R intervals)b
Repeat time (TR) 800 msecc or 1 R-to-R intervala or 2
R-to-R intervalsb
Inversion time (TI) 330 msecc
Fields of view (FOVx, FOVy) 160 mm, 120 mm
Resolution (Δx, Δy) 0.63 mm, 0.63 mm
Number of data points collected (Nx, Ny) 256, 192
Display matrix (Dx, Dy) 512, 512
Slice thickness (Δz) 2 mm
Number of slices 8–16
Slice gap 0 mm
Number of acquisitions (Nacq) 2
Read direction Left–right
Figure A1.4.3 Oblique sagittal 2-D black-blood MRA images (sequence 3) processed by multi- Slice locations Precise position relative to carotid
planar reformation (MPR; A) and minimal-intensity projection (MinIP; B). Images were obtained at bifurcation
the left side from a patient with moderate atherosclerotic disease. Transverse images of this patient Flow compensation No
are shown in Figure A1.4.4. Abbreviations correspond to common (CCA), internal (ICA), end
ZIP 512 Yes
external (ECA) carotid arteries, jugular vein (JV), and plaque (P). Reformatted view (A) shows a
fragment of plaque (P), whereas minimal-intensity projection (B) confirms that the lumen is almost Saturation pulses No
unobstructed. Fat suppression Yes
Slice series Sequential
an additional scan with sequence 3 can be obtained in the oblique plane orthogonal to the Scan time ∼30–60 sec/slice
plaque surface at the point of maximal thickness. Such a scan can be prescribed after aDepending on the specifications of the scanner, this sequence may require peripheral gating. Alternatively,

obtaining transverse images (sequences 4 through 7). the sequence can be run without gating (if allowed by a scanner) or with gating simulator.
bSpecific parameter setting for GE scanners, if the sequence is gated to actual patient’s heart rate with T
R

Image processing and viewing for sequences 2 and 3 = 2 R-to-R intervals.

19. Find bilateral carotid bifurcations in images obtained from sequence 2. Find the
approximate position of the central point at the line connecting the bifurcations using
the cursor (Fig. A1.4.2A). Document “left-right” and “anterior-posterior” coordinates
of the central point for future use.
These coordinates will be used to ensure identical positioning of the center of the field of
view (FOV) in the next sequences.
20. Review the images obtained from sequence 3 for the index side (Fig. A1.4.2B). If
necessary, adjust window level to clearly see a dark lumen of the carotid artery. Zoom
images ∼2×.
21. Find slice(s) containing the carotid bifurcation (Fig. A1.4.2B). If several slices dissect
the bifurcation, find the slice with the lowest (i.e., most inferior) position of the
bifurcation using the cursor. Document the inferior-superior coordinate of the bifur-
cation for future use.
MRI of
the Carotid The carotid bifurcation on the index side is used as the internal reference for prescription
Artery for
Atherosclerotic of transverse multi-contrast images in the next sequences. Use of the above coordinates
Plaques facilitates registration between images with different contrast weightings. If follow-up
cT for a non-gated sequence. The delay between DIR pulses and an acquisition sequence (if available for
R
control) should be set to 330 msec.
examinations of a patient are expected, in each examination, transverse images should be
obtained at the same distance from the bifurcation on the index side.
22. If indicated, perform multi-planar reformation of images obtained from sequence 3.
Load images into the multi-planar reformation tool. Rotate reformatted plane to
clearly visualize the extent of the plaque into the lumen and its relationship to the
common, internal and external carotid arteries (Fig. A1.4.3A). If also indicated,
prepare a minimal intensity projection (min IP) for the image subset crossing the
carotid arteries (Fig. A1.4.3B).
This step can be done after the examination or during the acquisition of subsequent
sequences. A particular algorithm of image processing depends on available software tools
and clinical requirements. Generally, multi-planar views may help to estimate the distri-
bution of the plaque along the artery. A minimal intensity projection may allow the
physician to estimate the degree of lumen obstruction. However, care must be taken near
adjacent jugular veins, which may interfere with the carotid arteries on minimal intensity Intracranial
projections. Arterial Disease

A1.4.8 A1.4.9
Supplement 11 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 11
 Table A1.4.7 Primary Clinical Imaging Parameters for Sequences 5 and 6 Table A1.4.8 Alternate Imaging Parameters for Sequences 5 and 6 (Transverse
(Transverse PD- and T2-Weighted FSE)a PD- and T2-Weighted FSE) with the Use of Multislice DIR Methoda

Patient position Supine Patient position Supine

Scan type 2-D FSE Scan type 2-D FSE with DIR preparation
Imaging plane (orientation) Transverse Imaging plane (orientation) Transverse
Variable bandwidth Yes Variable bandwidth Yes
Central slice or volume center Carotid bifurcation or stenosis Pulse sequence database (PSD) “fse-m”a
Echo time (TE) ∼7–12 msec for sequence 4, Central slice or volume center Carotid bifurcation or stenosis
PD-weighted; ∼50–60 msec for Echo time (TE) ∼7–12 msec for sequence 4,
sequence 5, T2-weighted PD-weighted; ∼50–60 msec for
Receiver bandwidth (RBW) ±31 kHz or close sequence 5, T2-weighted
Echo train length (ETL) 8 Receiver bandwidth (RBW) ±31 kHz or close
Repeat time (TR) 3–4 R-to-R intervalsa (∼2500–3000 Echo train length (ETL) 12
msec) Repeat time (TR) 2500 msec (3000 msec)b
Fields of view (FOVx, FOVy) 160 mm, 120 mm Inversion time (TI) 270 msec (230 msec)b
Resolution (Δx, Δy) 0.63 mm, 0.63 mm Fields of view (FOVx, FOVy) 160 mm, 120 mm
Number of data points collected (Nx, Ny) 256, 192 Resolution (Δx, Δy) 0.63 mm, 0.63 mm
Display matrix (Dx, Dy) 512, 512 Number of data points collected (Nx, Ny) 256, 192
Slice thickness (Δz) 2 mm Display matrix (Dx, Dy) 512, 512
Number of slices 16–18 Slice thickness (Δz) 2 mm
Slice gap 0 mm Number of slices 8 (12)b
Number of acquisitions (Nacq) 2 Slice gap 0 mm
Read direction Left–right Number of acquisitions (Nacq) 2
Slice locations Precise position relative to carotid Number of repetitions 2
bifurcation Read direction Left–right
Flow compensation No Slice locations Precise position relative to carotid
ZIP 512 Yes bifurcation
Saturation pulses Yes; superior and inferior 50–60 mm Flow compensation No
slabs ZIP 512 Yes
Fat suppression Yes Saturation pulses No
Scan time ∼2.5–3 min Fat suppression Yes
aPeripheral gating is required.
Slice series Interleaved
Scan time ∼3 min/8 slices (∼4 min/12 slices)b
Sequence 4: Transverse T1-weighted black-blood fast-spin echo (FSE) aA special pulse sequence database (PSD) is required. Currently, it is installed on GE scanners in the

23. Set up sequence parameters according to Table A1.4.6.
As in sequence 3, this sequence uses a DIR blood-suppression method. See annotations in
step 15 for details. In the authors’ standard protocol, sequence 3 is run without peripheral
gating (using a custom-designed pulse sequence) to guarantee identical TR, and, therefore,
reproducible tissue contrast. Sequence 3 can potentially be run with gating, although a fast
or slow heart rate may compromise the results. If a non-gated sequence is not available,
24. Prescribe a slice group (slab) at the specified distance from bifurcation (or exactly at
one may also use a gating simulator programmed for the rate of 75 beats/min, i.e., TR =
800 msec. A particular problem in the setup of sequence 4 occurs on GE scanners with the
standard commercial software, which does not permit black-blood acquisition with TR = 1
R-to-R interval (only 2 R-to-R can be selected). The problem can be resolved by using an
external gating simulator programmed for the appropriate heart rate, i.e., 150 beats/min,
which results in TR = 800 msec for 2 R-to-R intervals. Alternatively, the sequence can be
gated to the heart rate of the patient with TR = 2 R-to-R intervals, but proper T1-weighting
may not be achieved. This should be taken into account when interpreting multi-contrast
images. At the same time, this standard sequence provides effective blood suppression and
can be used for lumen and wall identification.
MRI of
the Carotid
Artery for
Atherosclerotic
Plaques
authors’ and several other institutions.
bAn exact correspondence between T , T , and a number of slices is required. The total number of slices
R I
(8 or 12) corresponds to acquisition of two interleaved groups by 4 or 6 slices, respectively. Two validated
combinations of parameters are given: for the acquisition of 4 slices per TR and for the acquisition of 6
slices per TR (in brackets). See Yarnykh and Yuan (2003) for details.
the bifurcation) with the center of FOV between the left and right bifurcations (Fig.
A1.4.2A). Use the coordinates determined in steps 19 and 21 to precisely position
the slices.
Graphic displays for the prescription of scan geometry may not provide sufficient accuracy
for this step. The authors recommend using exact numbers as coordinates for the slice
group. Specifically, slices may be centered at the bifurcation or at the specified distance
from the bifurcation based on the appearance of atherosclerotic plaque. This procedure
enables the use of the bifurcation as a landmark for further comparison between images
with different contrast weightings and, possibly, at different time periods in follow-up
studies.
Intracranial
Arterial Disease

A1.4.10 A1.4.11
Supplement 11 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 11
 Table A1.4.9 Primary Clinical Imaging Parameters for Sequence 7 (3-D TOF Note that PD- and T2-weighted sequences 5 and 6 utilize identical parameters except for
MRA) the echo time (TE). When setting up parameters according to Table A1.4.7 or Table A1.4.8,
use proper TE as indicated for sequences 5 and 6.
Patient position Supine
Scan type 3-D spoiled gradient echo 27. Prescribe slices according to the above recommendations (see step 24).
Imaging plane (orientation) Transverse 28. Place inferior and superior saturation bands 2 to 3 cm from the edges of the slice
Variable bandwidth Yes group.
Central slice or volume center Carotid bifurcation or stenosis
Echo time (TE) 3.0–3.5 msec This step is not needed if a multislice DIR sequence is used (see Table A1.4.8).
Receiver bandwidth (RBW) ±12.5 kHz or close 29. Instruct the patient to refrain from swallowing during the scan and begin the scan.
Repeat time (TR) 20–25 msec
Flip angle (FA) 25° Sequence 6: Transverse T2-weighted FSE
Fields of view (FOVx, FOVy) 160 mm, 120 mm 30. Repeat steps 26 to 29. Use proper echo time (TE) for T2-weighted imaging from Table
Resolution (Δx, Δy) 0.63 mm, 0.63 mm A1.4.7.
Number of data points collected (Nx, Ny) 256, 192
Operators may copy the entire set of parameters from sequence 5 and change only TE
Display matrix (Dx, Dy) 512, 512
according to Table A1.4.7. As with sequence 5, sequence 6 can be run using multislice
Slice thickness (Δz) 2 mm DIR-FSE method. Refer to Table A1.4.8 in this case.
Number of slices 24
Slab thickness 48 mm Sequence 7: 3-D time-of-flight MRA
Number of slabs 1 31. Prior to scanning, ensure proper position of the imaged slab based on the above
Slice gap 0 recommendations (see step 24). Run sequence 7 according to Table A1.4.9.
Number of acquisitions (Nacq) 2
Read direction Left–right More details about 3-D TOF MRA of carotid arteries can be found in UNIT A1.3. The
sequence parameters (Table A1.4.9) presented here are slightly different, as the primary
Slice locations Precise position relative to carotid
bifurcation goal of this sequence is to provide complementary information about the tissue composition
of the atherosclerotic plaque rather than to visualize blood flow.
Flow compensation Yes
ZIP 512 Yes
ZIP 2 Yes COMMENTARY
Saturation pulses Yes; superior 50–80 mm thick slab to Background Information wall. Unlike the imaging modalities focused on
saturate veins Carotid atherosclerosis is thought to be one the assessment of stenosis—X ray and CT
Fat suppression No of the major contributors to stroke, which is (computed tomography)–based angiography,
Scan time ∼3.5 min among the leading causes of death and disabil- and Doppler ultrasound—MRI has the unique
ity throughout the world. Traditionally, the de- potential to characterize various tissue compo-
gree of lumen stenosis is used as a marker for nents of atherosclerotic plaque. As a method for
Manual prescription of scan geometry may differ between manufacturers. It may require vulnerable plaques, which pose an increased plaque characterization, MRI considerably out-
entering the position of the center of a slice group (Philips) or the position of the first (or risk for a thromboembolic event resulting in performs alternative techniques—high-resolu-
last) slice (GE). In the latter case, the starting position (first slice) can be calculated using brain ischemia. However, two clinical trials (the tion CT and B-mode ultrasound—which have
the position of the bifurcation (see step 21), number of slices, slice thickness, and slice gap. North American Symptomatic Carotid Endar- limited dynamic range for soft tissues (CT) or
For example, if there are 16 contiguous slices with a thickness of 2 mm and it is necessary terectomy Trial and the Asymptomatic Carotid are prone to anatomic restrictions (B-mode
to center the slice group at the bifurcation, the starting inferior position will be 15 mm Atherosclerosis Study) demonstrated that lu- ultrasound). A number of studies (Toussaint et
below the bifurcation. In this case, the eighth slice will be exactly at the bifurcation. men narrowing alone is a poor indicator of al., 1996; Hatsukami et al., 2000; Yuan et al.,
25. Instruct the patient to refrain from swallowing during the scan and begin the scan. vulnerability, predicting only 1 out of 4 strokes 2001b) have shown that MRI is capable of
in symptomatic patients and 1 out of 10 in identifying clinically relevant plaque constitu-
asymptomatic patients. Many histological ents and morphological features in vivo, such
Sequence 5: Transverse PD-weighted FSE
studies (for details see Yuan et al., 2001a) re- as the lipid core, intraplaque hemorrhage, ne-
26. Set up sequence parameters according to Table A1.4.7. Use proper echo time (TE) for vealed the specific morphological factors of crosis, calcified tissue, and ruptured fibrous
PD-weighted imaging from Table A1.4.7. plaque instability. Some of these features are cap. Due to the complexity of tissue compo-
The sequence in Table A1.4.7 uses gated FSE acquisition with saturation of inflowing the erosion and rupture of the fibrous cap over- nents, multiple contrast weightings are needed
arterial and venous blood. This technique provides a fast scan with a large coverage but lying the lipid core, the presence of a large for plaque characterization (Toussaint et al.,
may suffer from residual flow artifacts. Improved flow suppression with little penalty of necrotic core or intraplaque hemorrhage, and 1996; Hatsukami et al., 2000; Yuan et al.,
reduced coverage (or increased scan time) can be obtained using a new multislice DIR-FSE the appearance of calcium nodules on or near 2001b). Good overall agreement between
MRI of technique (Yarnykh and Yuan, 2003). If this pulse sequence is available, use Table A1.4.8 the lumen surface. multi-contrast MRI and histology was docu-
the Carotid to set up parameters. MRI is a noninvasive imaging technique, mented for the classification and staging of
Artery for which can visualize the arterial lumen and pro- atherosclerotic lesions (Cai et al., 2002). Addi-
Atherosclerotic Intracranial
Plaques vide detailed information about the arterial tionally, MRI allows accurate measurements of Arterial Disease

A1.4.12 A1.4.13
Supplement 11 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 11
 plaque size (Yuan et al., 1998; Kang et al.,
2000), providing a means of monitoring the
progression of atherosclerosis over time, as
well as control of treatment.
High-resolution MRI of atherosclerotic
plaques is a relatively new area, not commonly
acknowledged as a routine diagnostic proce-
dure. Recent successful applications of this
method in relatively large groups of patients
became possible due to the use of a specialized
phased-array coil with improved SNR (Hayes
et al., 1996; see also Fig. A1.4.1) and an opti-
mized imaging protocol (Yuan et al., 2001a).
Further improvement of image quality was
achieved by using a new time-efficient black-
blood imaging method, multislice DIR
(Yarnykh and Yuan, 2003), although its avail-
ability is not critical for the Basic Protocol. In
its current state, MRI of the atherosclerotic
plaque requires the same scan time and efforts
as many other MR protocols. Based on previous
experience, the main goals of MRI in studies
of carotid atherosclerosis are as follows: (1)
detection and staging of the atherosclerotic dis-
ease; (2) monitoring of disease progression and
effectiveness of therapy; and (3) detection of
vulnerable atherosclerotic plaques, which are
prone to disruption resulting in brain embo-
lisms. The Basic Protocol provides comprehen-
sive information for the assessment of all these
aspects.
Critical Parameters and
Troubleshooting
Image quality is critical for the identification
of the tiny details of plaque structures. The
following problems and artifacts are of major
concern for high-resolution imaging of the ca-
rotid artery wall.
Motion artifacts
Swallowing is the major cause of motion
artifacts in carotid imaging. It is especially
important to instruct the patient to refrain from
swallowing during each scan. For long single-
slice DIR acquisition (sequences 3 and 4), some
slices may be compromised due to motion,
while others will be of good quality. In such
situations, acquisition can be repeated for se-
lected locations with little loss of overall ex-
amination time.
Wrap-around artifacts
MRI of The requirement of high resolution imposes
the Carotid
limitations on the FOV size. The authors use a
Artery for
Atherosclerotic rectangular FOV for all transverse images to
Plaques save scan time. Anterior and posterior parts of
A1.4.14
Supplement 11
the neck may frequently appear outside the
FOV and cause wrap-around artifacts (see, for
example, Fig. A1.4.4). These are not critical A B
unless they overlap with the carotid arteries.
Wrap-around artifacts may create problems in
patients with large necks. In this case, an in-
creased FOVy or a no-phase-wrap option (over-
sampling in the phase-encoding direction) can
be used while keeping the same resolution,
although scan time will be lengthened.
Low SNR
MR signal observed with a surface coil de-
creases with an increase of distance from the
coil. If the area of interest (i.e., carotid arteries)
appears far from the receiver panels of the coil
(Fig. A1.4.1), image SNR may be insufficient.
Another reason for a signal drop can be elec-
tromagnetic coupling between coil arrays, if
receiver panels are improperly aligned. This
may happen if panels are close to each other at
the anterior surface of the neck and form an
obtuse angle. Ideally, receiver panels should be
parallel when placed at the lateral surfaces of
the neck. If signal is low, try to readjust the coil Figure A1.4.4 Transverse multi-contrast images (A) obtained from a patient with moderate atherosclerotic disease
position. This problem can be difficult to avoid of the left carotid artery. Circular arrows indicate minor wrap-around artifacts. Magnified views of the carotid arteries
in patients with a thick neck and deep carotid (B) show atherosclerotic plaque overlaying the left internal carotid artery (solid arrows). This lesion contains a large
arteries. In such cases, SNR can be improved fragment of loose fibrous matrix corresponding to a hyperintense area on the T2-weighted image. In the right carotid
by a reasonable increase of the slice thickness bifurcation (open arrows), black-blood images display normal wall and a minor residual blood signal in the carotid
(e.g., to 3 to 4 mm). bulb. This incompletely suppressed signal originates from recirculating blood and has minimal intensity on the
T1-weighted DIR image. Note the absence of flow enhancement in the right carotid bulb on the 3-D TOF image, which
Incomplete flow suppression in black-blood shows a tissue-like signal (open arrow). Without black-blood images, 3-D TOF MRA may result in misinterpretation of
this region as plaque.
imaging
Effective blood suppression is critical for the
weighted images obtained by using DIR (se- used, the optimal values of the delay are given
evaluation of plaque components at the blood-
quence 4). Usually, these artifacts disappear or in Tables A1.4.4 and A1.4.6. The thickness of
wall interface. Sometimes, residual blood sig-
have dramatically reduced area and intensity on an inverted slice should be 2 to 2.5 times larger
nal can be observed in the lumen, causing the
T1-weighted DIR images. than the thickness of an imaged slice. On
lumen boundary to be obscured. Also, unsup-
Rarely, incomplete blood suppression may Philips scanners, make sure that the thickness
pressed blood may result in the so-called
occur in DIR imaging (Fig. A1.4.4). This may of an inverted slice is set to the minimal per-
plaque-mimicking artifact (Steinman et al.,
happen if the time between DIR pulses and mitted value (currently 5 mm).
1998). This artifact arises from recirculating
acquisition is insufficient for outflow of blood An alternative solution for black-blood im-
blood and occurs mostly in the carotid bulb
from an imaged slice, particularly, if blood aging with long TR(PD and T2-weighted) is
(Steinman et al., 1998). Effectiveness of flow
recirculates or flows very slowly. If the se- based on recently developed multislice DIR
suppression depends on the suppression tech-
quence is used with cardiac gating, flow sup- methods (Parker et al., 2002; Song et al., 2002;
nique used. The best results can be obtained
pression may be compromised for very fast Yarnykh and Yuan, 2003), which improve the
using the double inversion-recovery (DIR)
heart rates. A non-gated acquisition is prefer- time-efficiency of DIR and provide almost the
method (Edelman et al., 1991), which is utilized
able in this case (see also sequences 3 and 4). same quality of blood suppression. One such
in sequences 3 and 4. However, DIR is very
The critical parameters determining perform- multislice DIR technique was developed in the
time-consuming, since it requires single-slice
ance of DIR are the thickness of the inverted authors’ institution (Yarnykh and Yuan, 2003),
acquisition. Due to scan-time limitations, DIR
slice and the delay between a double-inversion and it is used in the Basic Protocol (see Tables
cannot be used in all dark-blood sequences
pulse pair and a readout sequence. The delay is A1.4.5 and A1.4.8).
(sequences 3 through 6). The inflow saturation
an analog of the inversion time (TI) for inflow-
method (used in sequences 5 and 6) is less
ing blood. Usually, these parameters are auto- Loss of flow enhancement in 3-D TOF MRA
effective, but it does not lengthen scan times. If
matically calculated by a scanner, but they can Sequence 7 (3-D TOF MRA) may suffer
present, plaque-mimicking artifacts on PD/T2-
also be available for manual control (e.g., on from specific artifacts resulting in a reduced
weighted images (sequences 5 and 6) can be Intracranial
Philips scanners). If a non-gated sequence is signal in the vessel lumen. Due to the absence
easily recognized by comparing them with T1- Arterial Disease
A1.4.15
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 11
 of flow enhancement, such areas can be erro-
neously interpreted as tissue (parts of the
plaque) or thrombosis (Fig. A1.4.4). Common
reasons for these artifacts are due to the satura-
tion of slow flow and flow-dependent
dephasing. These issues are discussed in detail
in UNITS A1.3 & B7.3. In the Basic Protocol, 3-D
TOF is used as an additional contrast weighting
(T1-weighted gradient-echo contrast), which is
interpreted in conjunction with black-blood im-
ages. If complementary high-quality black-
blood images are available, signal-loss artifacts
in the lumen area on TOF images can be reliably
distinguished from the plaque (Fig. A1.4.4).
Anticipated Results
Examination with the Basic Protocol pro-
vides comprehensive characterization of the
atherosclerotic lesion (Figs. A1.4.3 and
A1.4.4). Usually, both lumen and plaque are
clearly visible on all dark-blood images.
Oblique sagittal black-blood MRA depicts the
distribution of the plaque along the artery and
lumen narrowing in a stenotic segment (Fig.
A1.4.3). Transverse images obtained with DIR
blood suppression allow excellent contrast be-
tween wall and lumen (Fig. A1.4.4). Bright-
blood MRA (3-D TOF) is generally unable to
depict the lumen-wall boundary as conspicu-
ously as the black-blood images (Fig. A1.4.4).
However, TOF images provide helpful comple-
mentary contrast features. Slice-by-slice com-
parison of images with four basic contrast
weightings (T1, T2, PD, and gradient-echo
TOF) offers identification of basic plaque com-
ponents, such as fibrous tissue, lipid core, ne-
crosis, calcifications, and hemorrhages. The
typical signal features of plaque tissues can be
found in a review by Yuan et al. (2001a). The
main plaque constituent is fibrous tissue with
signal characteristics similar to muscle. Early
plaques are usually composed of fibrous tissue
and display uniform signal intensity. During
evolution, plaques acquire lipids and calcifica-
tions, and also may undergo necrosis or inci-
dental changes due to hemorrhages and rup-
tures. Advanced plaques frequently show areas
of variable signal. Specifically, hyperintensity
on T1-weighted images may be associated with
either an intraplaque hemorrhage or the lipid
core; both are destabilizing factors making the
plaque prone to rupture. TOF images may help
to distinguish a recent hemorrhage and the lipid
MRI of core, since a recent hemorrhage (in contrast to
the Carotid the lipid core) frequently shows increased sig-
Artery for
Atherosclerotic nal on TOF images (Yuan et al., 2001b). At the
Plaques same time, signal characteristics of a hemor-
A1.4.16
Supplement 11
rhage are stage-dependent. Hyperintensity on
T2- and PD-weighted images in combination
with iso-/hypointense signal on T1-weighted
images indicate a so-called loose matrix, which
is a less integrated fibrous tissue with an ele-
vated liquid phase. This may be a result of an
old hemorrhage and necrosis. Calcified regions
can be identified as dark areas in all images.
Bright-blood TOF images are especially help-
ful to assign calcificates close to the lumen
surface (juxtalumenal calcium) or to identify a
ruptured fibrous cap (Hatsukami et al., 2000).
Plaque rupture is an important clinical condi-
tion, which is associated with neurological
symptoms (stroke and transient ischemic at-
tack). By comparing black- and bright-blood
images, plaque rupture can be identified as a
dark/bright tear originating at the lumen bound-
ary.
Finally, black-blood images can be used for
quantitative morphologic analysis, such as lu-
men and wall area and volume measurements
(Yuan et al., 1998; Kang, 2000). Precise mor-
phometry of the atherosclerotic plaque is now
becoming an important tool for studies of dis-
ease progression and treatment.
Acknowledgments
The authors would like to thank Dr. Bao-
cheng Chu for helpful comments; Drs. Jeffrey
Maki and Gregory J. Wilson for assistance in
implementing this protocol for a Philips scan-
ner; and Marina S. Ferguson for reviewing the
manuscript.
Literature Cited
Cai, J.M., Hatsukami, T.S., Ferguson, M.S., Small,
R., Polissar, N.L., and Yuan, C. 2002. Classifica-
tion of human carotid atherosclerotic lesions
with in vivo multicontrast magnetic resonance
imaging. Circulation 106:1368-1373.
Edelman, R.R., Chien, D., and Kim, D. 1991. Fast
selective black blood MR imaging. Radiology
181:655-660.
Hatsukami, T.S., Ross, R., Polissar, N.L., and Yuan,
C. 2000. Identification of fibrous cap thickness
and cap rupture in human atherosclerotic carotid
plaque in vivo with high resolution magnetic
resonance imaging. Circulation 102:959-964.
Hayes, C.E., Mathis, C.M., and Yuan, C. 1996. Sur-
face coil phased arrays for high-resolution imag-
ing of the carotid arteries. J. Magn. Reson. Im-
aging 6:109-112.
Kang, X.J., Polissar, N.L., Han, C., Lin, E., and
Yuan, C. 2000. Analysis of the measurement
precision of arterial lumen and wall areas using
high resolution magnetic resonance imaging.
Magn. Reson. Med. 44:968-972.
Current Protocols in Magnetic Resonance Imaging
Parker, D.L., Goodrich, K.C., Masiker, M., Tsuruda,
J.S., and Katzman, G.L. 2002. Improved effi-
ciency in double-inversion fast spin-echo imag-
ing. Magn. Reson. Med. 47:1017-1021.
Shellock, F.G. 2001. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Song, H.K., Wright, A.C., Wolf, R.L., and Wehrli,
F.W. 2002. Multislice double inversion pulse
sequence for efficient black-blood MRI. Magn.
Reson. Med. 47:616-620.
Steinman, D.A. and Rutt, B.K. 1998. On the nature
and reduction of plaque-mimicking flow arti-
facts in black blood MRI of the carotid bifurca-
tion. Magn. Reson. Med. 39:635-641.
Toussaint, J.F., LaMuraglia, G.M., Southern, J.F.,
Fuster, V., and Kantor, H.L. 1996. Magnetic reso-
n ance images lip id , fib rou s, calcified,
hemorhagic, and thrombotic components of hu-
man atherosclerosis in vivo. Circulation 94:932-
938.
Yarnykh, V.L. and Yuan, C. 2003. Multislice double
inversion-recovery black-blood imaging with si-
multaneous slice reinversion. J. Magn. Reson.
Imaging 17:478-483.
Yuan, C., Beach, K.W., Smith, L.H., and Hatsukami,
T.S. 1998. Measurement of atherosclerotic ca-
rotid plaque size in vivo using high resolution
magnetic resonance imaging. Circulation
98:2666-2671.
Yuan, C., Mitsumori, L.M., Beach, K.W., and Ma-
ravilla, K.R. 2001a. Carotid atherosclerotic
plaque: Noninvasive MR characterization and
identification of vulnerable lesions. Radiology
221:285-299.
Yuan, C., Mitsumori, L.M., Ferguson, M.S., Polis-
sar, N.L., Echelard, D., Ortiz, G., Small, R.,
Davies, J.W., Kerwin, W.S., and Hatsukami, T.S.
2001b. In vivo accuracy of multispectral mag-
netic resonance imaging for identifying lipid-
rich necrotic cores and intraplaque hemorrhage
in advanced human carotid plaques. Circulation
104:2051-2056.
Current Protocols in Magnetic Resonance Imaging
Key References
Shellock, F.G. 2001. Magnetic Resonance Proce-
dures: Health Effects and Safety. CRC Press,
Boca Raton, Fla.
Covers a number of important patient management
issues related to MR imaging, including recom-
mended safety procedures, a list of metallic implants
that have been tested for MR compatibility, and a
list of other sources on MR safety.
Yarnykh and Yuan, 2003. See above.
Describes technical principles and demonstrates
advantages of a multislice DIR method, which is
preferable for performing sequences 3, 5, and 6 in
the Basic Protocol. The method produces excellent
results and it is easy for implementation.
Yuan et al., 2001a. See above.
Comprehensive review of clinical and technical as-
pects of carotid plaque MRI.
Internet Resources
http://vil.rad.washington.edu
Web site of the Vascular Imaging Lab (University of
Washington) contains example images and useful
references on MRI of the carotid atherosclerotic
plaque.
http://www.mrisafety.com
Covers a number of important patient management
issues related to MR imaging, including recom-
mended safety procedures, a list of metallic implants
that have been tested for MR compatibility, and a
list of other sources on MR safety.
Contributed by Vasily L. Yarnykh and
Chun Yuan
University of Washington
Seattle, Washington
Intracranial
Arterial Disease
A1.4.17
Supplement 11
 Set up patient and equipment
Rule Out (R/O) Arteriovenous Malformation UNIT A1.5
1. Interview (screen) the patient to ensure that he or she has no contraindications such
as cardiac pacemakers or other implants containing ferromagnetic materials. Also be
When imaging patients for intracranial arteriovenous malformations, the goals are (1) to BASIC
sure to find out if the patient has any health conditions that may require the presence
determine if an arteriovenous malformation is present; (2) to assess size, shape, and lo- PROTOCOL
of special emergency equipment during the scanning procedure, or necessitate any
cation of the nidus; (3) to determine potential arterial feeders and venous drainage routes other precautions. Do not forget to ask if the patient has any drug allergies and
(including screening for aneurysms on arterial feeders); and (4) to determine the state document them.
of the surrounding brain parenchyma, including the identification of intracranial hemor-
rhage. A post-contrast 3-D spoiled gradient echo (SPGR) sequence is used for sequences Generally, standard screening forms are used for all patients scanned in a magnetic
resonance system (see APPENDIX 1).
1 to 3 and the standard MR imaging sequences are used to evaluate the brain parenchyma.
The following Basic Protocol can be used for the evaluation of stable patients. The presence of any ferromagnetic metals may be a health hazard to the patient when
he or she is inside the magnet, and will also affect the imaging. If in doubt as to the
In order to perform the optional sequences with perfusion and diffusion studies, a scanner exact composition of the items, it is best to exclude patients with any metal implants; see
with echoplanar capabilities is required (Table A1.5.1). The standard anatomical MR Shellock (1996) for discussion of what implants may be safely scanned using magnetic
imaging sequences included in the protocols do not require these faster gradients. The resonance.
parameters given here are optimized for a 1.5T GE LX system with 11.0 software and Patients may be accompanied into the magnet room by a friend or family member, who can
may need modification for different software versions, field strengths, or manufacturers. sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular 2. If the procedure is a research protocol, have the patient sign any necessary consent
patient, such as crash carts or oxygen. Reactions to contrast agents are rare, but the form.
resources are necessary. 3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
Materials
Gadolinium-based MR contrast agent (e.g., Magnevist, Omniscan, Prohance) 4. Have the patient wash off any mascara and other makeup to avoid local tissue heating
Normal saline (0.9% NaCl), sterile and image artifacts.
5. Inform the patient about what will occur during the procedure, what he or she will
experience while in the magnet, and how to behave, including the following:
Table A1.5.1 Equipment Requirements for Cerebral Vascular Assessment a. If earphones or headphones are used to protect the ears from the loud sounds
Type of system LX EchoSpeed produced by the gradients, the patient will be asked to wear these, but will be
Field strength 1.5 T able to communicate with the technologist at any/all times during the imaging.
Software level 11 b. The patient will be given a safety squeeze-bulb or similar equipment to request
Magnet type CX K4 assistance at any time (demonstrate how this works).
Polarity Positive c. For optimum results the patient should not talk, and should avoid or minimize
Active gradient shielding No swallowing or other movement, during each scan—i.e., as long as the banging
Passive gradient shielding No sounds continue. Between scans, talking and swallowing are allowed in most
Gradient strength (amplitude) 2.4 mT/cm cases, but should be avoided when comparative positional studies are being
Slew rate 12 mgauss/cm
performed; the patient will be informed when this is the case.
SAR (average) 2 W/kg d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
SAR (max) 8 W/kg 6. Have the patient mount onto the table. Either before or right after the patient lies
Superior conductive shimming 45 cm down, set up any triggering devices or other monitoring equipment that is to be used.
Resistive shimming No
7. Establish an intravenous line from which the contrast agent can be injected, and
Active magnetic shielding No
attach this line securely to the patient so that movement into or out of the magnet
Maximum noise level 100 dB
will not pull at the patient’s arm.
Type of body coil High pass
TPS recon system memory 386 It is preferable to insert the line prior to imaging and to leave the patient in the magnet, so
that there is no intervening motion between the scans run before contrast agent injection
Array processor (TPS) Reflex 100 and those run after injection.
Computer LX Octane
Weight limit 350 lb (130.6 kg)
Rule Out (R/O)
Intracranial Arteriovenous
Arterial Disease Malformation
Contributed by L. Kirkland Conrad, Steven Thibodeau, and Ellen Grant A1.5.1 A1.5.2
Current Protocols in Magnetic Resonance Imaging (2005) A1.5.1-A1.5.9
C 2005 by John Wiley & Sons, Inc.
Copyright  Supplement 13 Supplement 13 Current Protocols in Magnetic Resonance Imaging
 8. Center the patient in a head or neck coil at the region where the key information Table A1.5.3 Transverse T2 -Weighted FSE
is desired. Make sure that the head and neck are constrained to prevent motion, Patient position Supine
especially if high-resolution scans are to be run. Scan type Fast spin echo XL
Generally the patient’s head is fixed so that the head is horizontal (not tilted) and the Imaging plane (orientation) Transverse
neck and head lie along the axis of the patient table; other positions may be appropriate Variable bandwidth Yes
depending on the needs at hand.
Pulse sequence database (PSD) FSE-XL
Most scanners have a special neck coil for MRA; otherwise, a head coil should be used Central slice or volume center Nasion
and the patient placed as far in as possible so that the bifurcation of the main carotid
Echo time (TE ) 102 msec
artery into the internal and external carotid arteries can be imaged.
Receiver bandwidth (RBW) 11.36 kHz
9. If needed, place a pillow or other support under the knees to make the patient more Echo train length (ETL) 12
comfortable. Repeat time (TR ) 6000 msec
10. Use the centering light to position the patient and put him or her into the center of Flip angle (FA) 90◦
the magnet. Fields of view (FOVx , FOVy ) 220 mm, 165 mm
Resolution (x, y) 0.86 mm, 0.86 mm
Once this step has been performed, so long as the patient does not move on the table,
the table itself can be moved and then replaced in the same position as before without Number of data points collected (Nx , Ny ) 256, 192
jeopardizing the positioning of one scan relative to another. Display matrix (Dx , Dy ) 320, 256
Slice thickness (z) 5 mm
11. If the patient is unable to hold still, provide an appropriate sedative.
Number of slices 24
12. Have a single dose of 0.1 to 0.2 mmol/kg of contrast agent drawn up and ready Slice gap 1 mm
for injection. Have an angiocatheter or butterfly, alcohol swabs, and tourniquet Number of excitations (NEX) 2
ready. Number of acquisitions (Nacq ) 1
Read direction Anterior–posterior
Sequence 1: Rapid sagittal T1 -weighted scout Flow compensation Yes
13. To determine the patient’s position, perform a fast sagittal scout scan using the Extended dynamic range (EDR) Yes
imaging sequence in Table A1.5.2. Spatial saturation Inferior
Some centers may prefer a rapid three-plane scout instead of this multislice sagittal Scan time 3 min, 19 sec
scout.

Sequence 2: Transverse T2 -weighted fast spin echo (FSE)

Table A1.5.2 Rapid Sagittal T1 -Weighted Scout
14. From the sagittal scout, select the image through the center of the brain to set up the
Patient position Supine locations for the transverse T2 -weighted fast spin echo (FSE) sequence. Set up the
Scan type Gradient echo imaging parameters as shown in Table A1.5.3.
Imaging plane (orientation) Sagittal The locations should be the same as those chosen for the transverse diffusion.
Central slice or volume center Nasion
Echo time (TE ) Minimum (at least 8.2 msec) 15. Warn the patient that this sequence is starting and begin the scan.
Receiver bandwidth (RBW) 15.6 kHz
Sequence 3: Transverse fluid-attenuated inversion recovery (FLAIR)
Repeat time (TR ) 100 msec
16. From the sagittal scout, select the image through the center of the brain to set up the
Flip angle (FA) 60◦
locations for the transverse fluid-attenuated inversion recovery (FLAIR). Set up the
Fields of view (FOVx , FOVy ) 240 mm, 240 mm imaging parameters as shown in Table A1.5.4.
Resolution (x, y) 0.94 mm, 1.87 mm
The locations should be the same as those chosen for the FSE sequence.
Number of data points collected (Nx , Ny ) 256, 128
Display matrix (Dx , Dy ) 128, 128 17. Warn the patient that this sequence is starting and begin the scan.
Slice thickness (z) 7 mm
Number of slices 7 Sequence 4: Transverse T2 -weighted gradient echo
Slice gap 2 mm 18. To detect hemorrhage, from the sagittal scout select the image through the center
Number of excitations (NEX) 1 of the brain to set up the locations for the transverse gradient echo T2 -weighted
Number of acquisitions (Nacq ) 1 sequence. Set up the imaging parameters as shown in Table A1.5.5.
Read direction Anterior–posterior The locations should be the same as those chosen for the FSE sequence.
Rule Out (R/O)
Scan time 16 sec Intracranial Arteriovenous 19. Warn the patient that this sequence is starting and begin the scan.
Arterial Disease Malformation

A1.5.3 A1.5.4
Current Protocols in Magnetic Resonance Imaging Supplement 13 Supplement 13 Current Protocols in Magnetic Resonance Imaging
 Table A1.5.4 Transverse FLAIR Table A1.5.6 Transverse T1 -Weighted Spin Echo

Patient Position Supine Patient position Supine

Scan type Inversion recovery, FSE Scan type Spin echo
Variable bandwidth Yes Imaging plane (orientation) Transverse
Imaging plane (orientation) Transverse Variable bandwidth Yes
Central slice or volume center Nasion Central slice or volume center Nasion
Echo time (TE ) 120 msec Echo time (TE ) Minimum full (14 msec)
Receiver bandwidth (RBW) 20.83 kHz Receiver bandwidth (RBW) 15.6 kHz
Repeat time (TR ) 10,000 msec Repeat time (TR ) 400 msec
Inversion time (TI ) 2200 msec Flip angle (FA) 90◦
Flip angle (FA) 90◦ Fields of view (FOVx , FOVy ) 220 mm, 165 mm
Fields of view (FOVx , FOVy ) 220 mm, 220 mm Resolution (x, y) 0.86 mm, 1.14 mm
Resolution (x, y) 0.86 mm, 1.14 mm Number of data points collected (Nx , Ny ) 256, 144
Number of data points collected (Nx , Ny ) 256, 192 Display matrix (Dx , Dy ) 256, 192
Display matrix (Dx , Dy ) 256, 256 Slice thickness (z) 5 mm
Slice thickness (z) 5 mm Number of slices 24
Number of slices 24 Slice gap 1 mm
Slice gap 1 mm Number of excitations (NEX) 1
Number of excitations (NEX) 1 Number of acquisitions (Nacq ) 2
Number of acquisitions (Nacq ) 2 Read direction Anterior–posterior
Read direction Anterior–posterior Flow compensation No
Saturation pulses Inferior Extended dynamic range (EDR) Yes
Scan time 4 min, 44 sec Saturation pulses Inferior
Scan time 2 min, 8 sec

Table A1.5.5 Transverse T2 -Weighted Gradient Echo

Sequence 5: Transverse T1 -weighted spin echo
Patient position Supine 20. From the sagittal scout, select the image through the center of the brain to set up
Scan type Gradient echo the locations for the transverse T1 -weighted spin echo sequence. Set up the imaging
Imaging plane (orientation) Transverse parameters as shown in Table A1.5.6.
Variable bandwidth Yes The locations should be the same as those chosen for the transverse diffusion.
Central slice or volume center Nasion
Echo time (TE ) 25 msec 21. Warn the patient that this sequence is starting and begin the scan.
Receiver bandwidth (RBW) 15.63 kHz Sequence 6: Transverse post-contrast 3-D spoiled gradient echo
Repeat time (TR ) 750 msec 22. Let the patient know that you are going to bring them out of the scanner and give
Flip angle (FA) 20◦ them the contrast injection.
Fields of view (FOVx , FOVy ) 220 mm, 165 mm
23. Bring the patient out of the scanner and inject the single dose of contrast agent
Resolution (x, y) 0.86 mm, 1.14 mm
intravenously.
Number of data points collected (Nx , Ny ) 256, 144
Display matrix (Dx , Dy ) 256, 192 24. From the sagittal scout, select the image through the center of the brain to set up the
Slice thickness (z) 5 mm locations for the transverse 3-D spoiled gradient echo sequence. Set up the imaging
Number of slices 24 parameters as shown in Table A1.5.7.
Slice gap 1 mm The locations should be the same as those chosen for the FSE sequence.
Number of excitations (NEX) 2
25. Warn the patient that this sequence is starting and begin the scan.
Number of acquisitions (Nacq ) 1
Read direction Anterior–posterior Data processing and viewing for sequence 6
Flow compensation Yes 26. Obtain maximum intensity projections (MIP) or surface reconstructions of the nidus,
Scan time 3 min, 41 sec feeding vessels, and draining veins, by using appropriate software.
Rule Out (R/O)
Intracranial Arteriovenous On a GE system, one can use the Advantage Window Workstation with FuncTool version
Arterial Disease Malformation 1.9M.
A1.5.5 A1.5.6
Current Protocols in Magnetic Resonance Imaging Supplement 13 Supplement 13 Current Protocols in Magnetic Resonance Imaging
 Table A1.5.7 Transverse Post-Contrast 3-D Spoiled Gradient Echo
Patient position
Scan type
Imaging plane (orientation)
Variable bandwidth
Central slice or volume center
Echo time (TE )
Receiver bandwidth (RBW)
Repeat time (TR )
Flip angle (FA)
Fields of view (FOVx , FOVy )
Resolution (x, y)
Number of data points collected (Nx , Ny )
Display matrix (Dx , Dy )
Slice thickness (z)
Number of slices
Slice gap
Number of excitations (NEX)
Number of acquisitions (Nacq )
Read direction
Flow compensation
Extended dynamic range (EDR)
Saturation pulses
Scan time
COMMENTARY
Background Information
Catheter angiography remains the gold
standard, particularly in the detection of small
arteriovenous malformations, determining the
feeding arteries, and detecting aneurysms
and venous restriction. A 3-D spoiled gra-
dient echo sequence without an inferior sat-
uration pulse is similar to an MRA with-
out magnetization transfer and has the ad-
ditional advantage that whole-brain coverage
can be obtained in a reasonable time (Fig.
A1.5.1). A post-contrast 3-D spoiled gradi-
ent echo sequence (SPGR) often plays a sig-
nificant role in treatment planning since it
allows more accurate determination of the
nidus shape and is therefore more tailored
for treatment planning (Kondziolka et al.,
1994; Schlemmer et al., 1994). Post-contrast
3-D SPGR is superior to noncontrast MRA
in that it allows better visualization of the
slower velocity draining veins (Yanno et al.,
1997). Standard MRI in combination with
post-contrast 3-D SPGR is also helpful in de-
termining the degree of parenchymal involve-
ment (Nussel et al., 1991). MIP reconstruction
Current Protocols in Magnetic Resonance Imaging
Supine
3-D spoiled gradient echo
Transverse
Yes
Nasion
Minimum full (6 msec)
15.6 kHz
31 msec
30◦
240 mm, 180 mm
0.93 mm, 0.93 mm
256, 192
256, 256
1.5 mm
24
Figure A1.5.1 Post-contrast (A) transverse and (B) coronal 3-D gradient echo sequence demon-
0 strating an arteriovenous malformaton (AVM) located in the right frontal lobe. The AVM is supplied
1 by the right anterior and middle cerebral arteries, and is drained by enlarged cortical veins that
1 drain into the superior saggital sinus.
Anterior–posterior
Yes
Yes
Inferior
12 min, 45 sec
of the post-contrast 3-D SPGR data can prove
helpful in visualizing the complexity of the ar-
terial and venous vessels in relation to the AVM
nidus (Fig. A1.5.2).
It is important to remember that other
vascular malformations can look similar to
arteriovenous malformations if the images are
not closely evaluated. If a number of enlarged
vessels are identified, but there is no definite
nidus, a dural arteriovenous fistula is more Figure A1.5.2 MIP images from the 3-D spoiled gradient echo data in (A) transverse and (B)
likely than an arteriovenous malformation. coronal orientations. The feeding arteries from the anterior and middle cerebral arteries are demon-
Also, if a number of small veins are seen to strated, as well as the enlarged cortical veins draining the AVM into the superior saggital sinus.
drain into a larger anomalous vein, without
arterial feeders, a developmental venous restriction, and therefore catheter angiography
Anticipated Results
anomaly should be considered. Post-contrast 3-D SPGR should allow may still be required. Standard imaging will
accurate assessment of the size, shape, and allow assessment of the underlying brain
Critical Parameters and location of AVM nidus with the caveat that parenchyma and determine if hemorrhage has
Troubleshooting a very small nidus can be missed on MR occurred.
For a discussion of critical parameters and imaging. MRA can help in the assessment
troubleshooting issues, see UNIT A1.2 & A1.3. of potential arterial feeders and venous
In the evaluation of arteriovenous malfor- drainage routes and may help to identify Literature Cited
mations, the inferior saturation band is not aneurysms on feeding arteries or venous out- Kondziolka, D., Lunsford, L.D., Kanal, E., and Ta-
applied when acquiring the 3-D SPGR post- lagala, L. 1994. Stereotactic magnetic resonance
Rule Out (R/O) flow restriction. However, catheter angiogram
contrast, to allow inflow effect to improve ves- angiography for targeting in arteriovenous mal-
Intracranial Arteriovenous is still the gold standard for the detection
Arterial Disease Malformation formation radiosurgery. Neurosurgery 35:585-
sel conspicuity. of small AVMs, aneurysms, and venous 590; discussion 590-591.
A1.5.7 A1.5.8
Supplement 13 Supplement 13 Current Protocols in Magnetic Resonance Imaging
Nussel, F., Wegmuller, H., and Huber, P. 1991. Com-
parison of magnetic resonance angiography,
magnetic resonance imaging and conventional
angiography in cerebral arteriovenous malfor-
mation. Neuroradiology 33:56-61.
Schlemmer, H.P., Hess, T., Debus, J., Knopp, M.V.,
Schad, L.R., and Engenhart, R. 1994. TOF-MR
angiography in radiotherapy treated cerebral ar-
teriovenous malformations. Radiologe 34:447-
453.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Yano, T., Kodama, T., Suzuki, Y., and Watan-
abe, K. 1997. Gadolinium-enhanced 3D time-of-
flight MR angiography. Experimental and clini-
cal evaluation. Acta Radiol. 38:47-54.
Contributed by L. Kirkland Conrad,
Steven Thibodeau, and Ellen Grant
Massachusetts General Hospital
Boston, Massachusetts
Rule Out (R/O) Vasculitis UNIT A1.6
When imaging patients for vasculitis, the goals are (1) to determine if there is evidence
of acute or subacute cerebral injury and (2) to assess the contour of the major intracranial
arteries. An additional but still experimental goal is (3) to determine if there are areas
of altered perfusion that suggest active small vessel disease. Standard MR images and
diffusion-weighted imaging are used to detect and determine the age of parenchymal
lesions. The 3-D TOF MRA helps evaluate the large and medium vessels. Perfusion-
weighted imaging may detect regions of altered relative blood flow and blood volume.
The following Basic Protocol can be used for the evaluation of stable patients; an Alternate
Protocol is provided for evaluation of unstable patients.
STANDARD IMAGING FOR VASCULITIS BASIC
PROTOCOL
In order to perform the optional sequences with perfusion and diffusion studies, a scanner
with echoplanar capabilities is required (Table A1.6.1). MR angiography and the standard
anatomical MR imaging sequences included in the protocols do not require these faster
gradients. The parameters given here are optimized for a 1.5T GE LX system with 11.0
software and may need modification for different software versions, field strengths, or
manufacturers.

Data processing is required for the 3-D TOF MRA sequences described in this unit. For
optimal assessment of the volumetric 3-D TOF MRA data, separate volumes of interest
(VOI) of the left anterior, right anterior, and posterior circulation should be created.
Images from 15◦ rotations about the vertical axis should be saved. An additional anterior
circulation volume of interest should include both anterior cerebral arteries and the region
of the anterior communicating artery for visual inspection. Images from 15◦ rotations
about the horizontal axis should be saved.

Table A1.6.1 Equipment Requirements for Cerebral Vascular Assessment

Type of system LX EchoSpeed

Field strength 1.5 T
Software level 11
Magnet type CX K4
Polarity Positive
Active gradient shielding No
Passive gradient shielding No
Gradient strength (amplitude) 2.4 mT/cm
Slew rate 12 mgauss/cm
SAR (average) 2 W/kg
SAR (max) 8 W/kg
Superior conductive shimming 45 cm
Resistive shimming No
Active magnetic shielding No
Maximum noise level 100 db
Type of body coil High pass
TPS recon system memory 386
Array processor (TPS) Reflex 100
Computer LX Octane
Weight limit 350 lb (130.6 kg)
Intracranial Intracranial
Arterial Disease Arterial Disease

A1.5.9 Contributed by Steven Thibodeau, Ellen Grant, and Pamela W. Schaefer A1.6.1
Current Protocols in Magnetic Resonance Imaging (2005) A1.6.1-A1.6.11
Current Protocols in Magnetic Resonance Imaging Supplement 13
C 2005 by John Wiley & Sons, Inc.
Copyright  Supplement 13
 Although maximum intensity projections are used. All of the authors’ MRA data is
processed on a General Electric (GE) Advantage Windows Workstation version 3.1P
using standard GE Advantage Windows software or FuncTool version 1.9M. In the hands
of an experienced technologist, the post-processing will require ∼5 min.
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen. Reactions to contrast agents are rare, but the
resources are necessary.
Materials
Gadolinium-DTPA contrast agent (e.g., Manevist, Omniscan, Prohance)
Normal saline (0.9% NaCl), sterile
Injection pump
Set up patient and equipment
1. Interview (screen) the patient to ensure that he or she has no counterindications such
as cardiac pacemakers or other implants containing ferromagnetic materials. Also be
sure to find out if the patient has any health conditions that may require the presence
of special emergency equipment during the scanning procedure, or necessitate any
other precautions. Do not forget to ask if the patient has any drug allergies and
document them.
Generally, standard screening forms are used for all patients scanned in a magnetic
resonance system (see APPENDIX 1).
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(1996) for discussion of what implants may be safely scanned using magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
2. If the procedure is a research protocol, have the patient sign any necessary consent
forms.
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating
and image artifacts.
5. Inform the patient about what will occur during the procedure, what he or she will
experience while in the magnet, and how to behave, including the following:
a. If earphones or headphones are used to protect the ears from the loud sounds
produced by the gradients, the patient will be asked to wear these, but will be
able to communicate with the technologist at any time during the imaging.
b. The patient will be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this works).
c. For optimum results the patient should not talk, and should avoid or minimize
swallowing or other movement, during each scan—i.e., as long as the banging
sounds continue. Between scans, talking and swallowing are allowed in most
cases, but should be avoided when comparative positional studies are being
Rule Out (R/O) performed; the patient will be informed when this is the case.
Vasculitis
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
A1.6.2
Supplement 13 Current Protocols in Magnetic Resonance Imaging
6. Have the patient mount onto the table. Either before or right after the patient lies
down, set up any triggering devices or other monitoring equipment that is to be used.
7. Establish an intravenous line using an 18-G angiocatheter through which the contrast
agent can be injected, and attach this line securely to the patient so that movement
into or out of the magnet will not pull at the patient’s arm.
It is preferable to insert the line prior to imaging and to leave the patient in the magnet, so
that there is no intervening motion between the scans run before contrast agent injection
and those run after injection.
8. Center the patient in a head or neck coil at the region where the key information
is desired. Make sure that the head and neck are constrained to prevent motion,
especially if high-resolution scans are to be run.
Generally, the patient’s head is fixed so that the head is horizontal (not tilted) and the
neck and head lie along the axis of the patient table; other positions may be appropriate
depending on the needs at hand.
Most scanners have a special neck coil for MRA; otherwise, a head coil should be used
and the patient placed as far in as possible so that the bifurcation of the main carotid
artery into the internal and external carotid arteries can be imaged.
9. If needed, place a pillow or other support under the knees to make the patient more
comfortable.
10. Place the patient such that the laser light is centered over the patient’s nasion, and
put him or her into the center of the magnet.
Once this step has been performed, so long as the patient does not move on the table,
the table itself can be moved and then replaced in the same position as before without
jeopardizing the positioning of one scan relative to another.
11. If the patient is unable to hold still, provide an appropriate sedative.
12. Connect the MR-compatible injection pump, cleared of air, and loaded with a double
dose of contrast agent and saline flush, to the patient. Do a small test injection of
normal saline to ensure that the i.v. is working properly.
13. Program the 5 ml/sec injection rate for both contrast and saline, the contrast dose
required (double dose by weight), the saline flush required (typically 40 ml), and a
10-sec time delay. Start the saline running at TKVO (to keep vein open) and arm the
injection pump.
Sequence 1: Rapid sagittal T1 -weighted scout
14. To validate the patient’s position, perform a fast sagittal scout scan using the imaging
sequence in Table A1.6.2.
Some centers may prefer a rapid three-plane scout instead of this multislice sagittal scout.
Sequence 2: Transverse diffusion
15. From the sagittal scout, select the image through the center of the brain to set up
the locations for the transverse diffusion. The transverse images should begin at the
foramen magnum and end at the top of the brain. Set up imaging parameters as shown
in Table A1.6.3.
16. Warn the patient that this sequence results in loud beeping noises and begin the scan.
Most diffusion sequences will perform a minimum of four sequences: three with orthogonal
diffusion gradient directions and one with a minimal or no diffusion gradient. Often, a
fifth set of images is also provided which combines the three orthogonal gradient images
to produce a set of images whose signal intensity is not affected by the diffusion direction. Intracranial
This is diffusion-weighted imaging (DWI). In order to obtain apparent diffusion coefficient Arterial Disease
A1.6.3
Current Protocols in Magnetic Resonance Imaging Supplement 13
 Table A1.6.2 Rapid Sagittal T1 -Weighted Scout Table A1.6.4 Transverse FLAIR

Patient position Supine Patient position Supine

Scan type Gradient echo Scan type Inversion recovery, FSE
Imaging plane (orientation) Sagittal Imaging plane (orientation) Transverse
Central slice or volume center Nasion Variable bandwidth Yes
Echo time (TE ) Minimum (at least 4.1 msec) Central slice or volume center Nasion
Receiver bandwidth (RBW) 15.6 kHz Echo time (TE ) 120 msec
Repeat time (TR ) 100 msec Receiver bandwidth (RBW) 20.83 kHz
Flip angle (FA) 60◦ Repeat time (TR ) 10,000 msec
Fields of view (FOVx , FOVy ) 240 mm, 240 mm Inversion time (TI ) 2200 msec
Flip angle (FA) 90◦
Resolution (x, y) 0.94 mm, 1.87 mm
Fields of view (FOVx , FOVy ) 220 mm, 220 mm
Number of data points collected (Nx , Ny ) 256, 128
Resolution (x, y) 0.86 mm, 1.14 mm
Display matrix (Dx , Dy ) 128, 128
Number of data points collected (Nx , Ny ) 256, 192
Slice thickness (z) 7 mm
Display matrix (Dx , Dy ) 256, 192
Number of slices 7
Slice thickness (z) 5 mm
Slice gap 2 mm
Number of slices 24
Number of excitations (NEX) 1
Slice gap 1 mm
Number of acquisitions (Nacq ) 1 Number of excitations (NEX) 1
Read direction Anterior–posterior Number of acquisitions (Nacq ) 2
Scan time 16 sec Read direction Anterior–posterior
Saturation pulses Inferior
Scan time 4 min, 44 sec
Table A1.6.3 Transverse Diffusion

Patient position Supine Table A1.6.5 Transverse T2 -Weighted FSE

Scan type SE-EPI Patient position Supine
Imaging plane (orientation) Transverse Scan type Fast spin echo
Central slice or volume center Nasion Imaging plane (orientation) Transverse
Echo time (TE ) Minimum (80.9 msec) Variable bandwidth Yes
Echo train length (ETL) or shots (SH) 1 SH Pulse sequence database (PSD) FSE-XL
Repeat time (TR ) 5000 msec Central slice or volume center Nasion
Flip angle (FA) 90◦ Echo time (TE ) 102 msec
Fields of view (FOVx , FOVy ) 220 mm, 220 mm Receiver bandwidth (RBW) 11.36 kHz
Resolution (x, y) 1.72 mm, 1.72 mm Echo train length (ETL) or shots (SH) 12
Number of data points collected (Nx , Ny ) 128, 128 Repeat time (TR ) 6000 msec
Display matrix (Dx , Dy ) 128, 128 Flip angle (FA) 90◦
Slice thickness (z) 5 mm Fields of view (FOVx , FOVy ) 220 mm, 165 mm
Resolution (x, y) 0.86 mm, 0.86 mm
Number of slices 24
Number of data points collected (Nx , Ny ) 256, 192
Slice gap 1 mm
Display matrix (Dx , Dy ) 320, 256
Number of excitations (NEX) 3
Slice thickness (z) 5 mm
Number of acquisitions (Nacq ) 1
Number of slices 24
Read direction Right–left
Slice gap 1 mm
Saturation pulses Fat saturation (automatic with EPI)
Number of excitations (NEX) 2
Control variables (CV) Ramp sampling = 1,
Number of acquisitions (Nacq ) 1
burst sampling = 0
Read direction Anterior–posterior
b-value 1000 sec/min2
Flow compensation Yes
Direction 6 sec
Extended dynamic range (EDR) Yes
Scan time 3 min, 5 sec
Saturation pulses Inferior
Rule Out (R/O) Scan time 3 min, 19 sec Intracranial
Vasculitis Arterial Disease

A1.6.4 A1.6.5
Supplement 13 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 13
 (ADC) maps, the images may need further processing. On a GE system, the ADC maps
can be obtained by processing on the Advantage Window Workstation using FuncTool
version 1.9M.
Sequence 3: Transverse FLAIR
17. From the sagittal scout, select the image through the center of the brain to set up the
locations for the transverse FLAIR. The locations should be the same as those chosen
for the transverse diffusion. Set up the imaging parameters as shown in Table A1.6.4.
18. Warn the patient that this sequence is starting and begin the scan.
Sequence 5: Transverse 3-D TOF MRA
21. From the sagittal scout, select the image through the center of the brain to set up the
locations for the transverse 3-D TOF MRA. Set up the imaging parameters as shown
in Table A1.6.6. Two slabs are posted on the midline sagittal image, centered at the
top of the sella tursica.
22. Warn the patient that this sequence is starting and begin the scan.
This sequence can also be performed after Sequence 7 (perfusion sequence) to optimize
visualization of small vessels.

Sequence 4: Transverse T2 -weighted FSE Data processing and viewing for Sequence 5
19. From the sagittal scout, select the image through the center of the brain to set up the 23. Once the axial partitions from the 3-D TOF are obtained, select volumes of interest
locations for the transverse T2 -weighted FSE. The locations should be the same as (VOI) that select out the posterior circulation, the right anterior circulation, and the
those chosen for the transverse diffusion. Set up the imaging parameters as shown in left anterior circulation. Rotate the maximum intensity projections (MIPs) or surface
Table A1.6.5. reconstructions of these three VOIs through 360◦ about the vertical, saving images at
15◦ intervals. Rotate a fourth VOI that selects out both anterior circulations through
20. Warn the patient that this sequence is starting and begin the scan. 360◦ about the horizontal, saving images at 15◦ intervals.
Sequence 6: Transverse T1 -weighted spin echo
Table A1.6.6 Transverse 3-D TOF MRA
24. From the sagittal scout, select the image through the center of the brain to set up the
Patient position Supine locations for the transverse T1 -weighted spin echo. The locations should be the same
Scan type 3-D vascular TOF SPGR as those chosen for the transverse diffusion. Set up the imaging parameters as shown
Imaging plane (orientation) Transverse in Table A1.6.7.
Variable bandwidth Yes
25. Warn the patient that this sequence is starting and begin the scan.
Central slice or volume center Nasion
Echo time (TE ) 6.9 msec
Receiver bandwidth (RBW) 10.42 kHz
Table A1.6.7 Transverse T1 -Weighted Spin Echo
Repeat time (TR ) 50 msec
Flip angle (FA) 20◦ Patient position Supine
Scan type Spin echo
Fields of view (FOVx , FOVy ) 180 mm, 180 mm
Imaging plane (orientation) Transverse
Resolution (x, y) 0.357 mm, 0.468 mm
Variable bandwidth Yes
Number of data points collected (Nx , Ny ) 384, 224
Central slice or volume center Nasion
Display matrix (Dx , Dy ) 384, 224
Echo time (TE ) Minimum full (14 msec)
Slice thickness (z) 1.4 mm
Receiver bandwidth (RBW) 15.6 kHz
Number of slices/slab 40
Repeat time (TR ) 400 msec
Number of slabs 2
Flip angle (FA) 90◦
Slab overlap 4 Fields of view (FOVx , FOVy ) 220 mm, 165 mm
Slice gap 0 Resolution (x, y) 0.86 mm, 1.14 mm
Number of excitations (NEX) 1 Number of data points collected (Nx , Ny ) 256, 144
Number of acquisitions (Nacq ) 1 Display matrix (Dx , Dy ) 256, 192
Read direction Anterior–posterior Slice thickness (z) 5 mm
Flow compensation Yes Number of slices 24
ZIP 512/ZIP 1024 ZIP 512 Slice gap 1 mm
ZIP 2/ZIP 4 ZIP 2 Number of excitations (NEX) 1
Extended dynamic range (EDR) Yes Number of acquisitions (Nacq ) 2
Saturation pulses Inferior Read direction Anterior–posterior
Magnetization transfer Yes Flow compensation No
Vascular options Reprojections = 19, Extended dynamic range (EDR) Yes
collapse = on, Saturation pulses Inferior
ramp pulse = inferior to superior Scan time 2 min, 8 sec
Scan time 6 min, 32 sec
Rule Out (R/O) Intracranial
Vasculitis Arterial Disease

A1.6.6 A1.6.7
Supplement 13 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 13
 Table A1.6.8 Transverse Perfusion Table A1.6.9 Transverse Post-Contrast T1 -Weighted Spin Echo

Patient position Supine Patient position Supine

Scan type GRE-EPI Scan type Spin echo
Imaging plane (orientation) Transverse Imaging plane (orientation) Transverse
Variable bandwidth Yes Variable bandwidth Yes
Central slice or volume center Nasion Central slice or volume center Nasion
Echo time (TE ) 54 msec Echo time (TE ) Minimum full (20 msec)
Echo train length (ETL) 1 Receiver bandwidth (RBW) 15.6 kHz
Repeat time (TR ) 1500 msec Repeat time (TR ) 400 msec
Flip angle (FA) 35◦ Flip angle (FA) 90◦
Fields of view (FOVx , FOVy ) 220 mm, 220 mm Fields of view (FOVx , FOVy ) 220 mm, 165 mm
Resolution (x, y) 1.72 mm, 1.72 mm Resolution (x, y) 0.86 mm, 1.14 mm
Number of data points collected (Nx , Ny ) 128, 128 Number of data points collected (Nx , Ny ) 256, 144
Display matrix (Dx , Dy ) 128, 128 Display matrix (Dx , Dy ) 256, 192
Slice thickness (z) 5 mm Slice thickness (z) 5 mm
Number of slices 24 Number of slices 24
Slice gap 1 mm Slice gap 1 mm
Number of excitations (NEX) 1 Number of excitations (NEX) 1
Number of acquisitions (Nacq ) 1 Number of acquisitions (Nacq ) 2
Read direction Right–left Read direction Anterior–posterior
Flow compensation Yes
Slice location Inferior slice to include MCA
Extended dynamic range (EDR) Yes
Multiphase Yes, 46 phases per location,
minimum delay between acquisition Saturation pulses Inferior
Slice series Interleaved Scan time 2 min, 8 sec
Control variables (CV) Ramp sampling = on,
burst sampling = off Table A1.6.10 Coronal Post-Contrast T1 -Weighted Spin Echo
Scan time 1 min, 10 sec Patient position Supine
Scan type Spin echo
Imaging plane (orientation) Coronal
Sequence 7: Transverse perfusion Variable bandwidth Yes
26. Using the transverse T2 -weighted FSE images, find the slice location where the Central slice or volume center Nasion
middle cerebral arteries are best seen. Using the imaging parameters as shown in Echo time (TE ) Minimum full (20 msec)
Table A1.6.8, set up the transverse images starting at the same location as the MCA Receiver bandwidth (RBW) 15.6 kHz
were identified on the FSE T2 . Repeat time (TR ) 500 msec
Flip angle (FA) 90◦
Typically only ten or eleven slices can be obtained in the given TR interval. The slice thick-
ness and gap can be modified so that the perfusion study covers any diffusion abnormality. Fields of view (FOVx , FOVy ) 220 mm, 165 mm
Resolution (x, y) 0.86 mm, 1.14 mm
27. Check that the injection pump is ready to inject and that the injection is set to start Number of data points collected (Nx , Ny ) 256, 144
after a 10-sec delay. Display matrix (Dx , Dy ) 256, 192
28. Warn the patient that this sequence is starting and that part way through the scan, the Slice thickness (z) 5 mm
i.v. injection will occur. Begin the scan. Number of slices 26
Slice gap 1 mm
29. Process the perfusion images using appropriate workstation and software to give
Number of excitations (NEX) 1
maps of relative cerebral blood volume (rCBV), relative cerebral blood flow (rCBF),
Number of acquisitions (Nacq ) 2
and mean transit time (MTT).
Read direction Superior-inferior
On a GE system, one can use an Advantage Windows Workstation with FuncTool version Flow compensation Yes
1.9M.
Extended dynamic range (EDR) Yes
Saturation pulses Inferior
Rule Out (R/O) Scan time 2 min, 40 sec Intracranial
Vasculitis Arterial Disease

A1.6.8 A1.6.9
Supplement 13 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 13
 Sequence 8: Transverse post-contrast T1 and the flip angle increased compared to the nial vasculitis: The roles of MR and angiog-
precontrast study to improve visualization. raphy. A.J.N.R. Am. J. Neuroradiol. 15:317-
30. From the sagittal scout, select the image through the center of the brain to set up the 330.
locations for the coronal post-contrast T1 . The locations should be the same as those Anticipated Results Jager, H.R., Albrecht, T., Curati-Alasonatti, W.L.,
chosen for the transverse diffusion. Set up the imaging parameters as shown in Table This combination of sequences should al- Williams, E.J., and Haskard, D.O. 1999. MRI
A1.6.9. low the detection and distinction of acute from in neuro-Behcet’s syndrome: Comparison of
chronic parenchymal injury. The perfusion- conventional spin-echo and FLAIR pulse se-
31. Warn the patient that this sequence is starting and begin the scan. quences. Neuroradiology 41:750-758.
weighted sequence should allow regions with
altered hemodynamics to be identified. These Joelson, E., Ruthrauff, B., Ali, F., Lindeman, N., and
Sequence 9: Coronal post-contrast T1 Sharp, F.R. 2000. Multifocal dural enhancement
regions of altered hemodynamics may indicate
32. From the sagittal scout, select the image through the center of the brain to set up associated with temporal arteritis. Arch. Neurol.
acute small vessel involvement prior to the de- 57:119-122.
the locations for the coronal post-contrast T1 . The locations should be the same as velopment of acute infarction. In the rare cases
those chosen for the transverse diffusion. Set up the imaging parameters as shown in Pomper, M.G., Miller, T.J., Stone, J.H., Tidmore,
that medium or large vessels are involved, W.C., and Hellmann, D.B. 1999. CNS vasculitis
Table A1.6.10. the MRA may help identify areas of involve- in autoimmune disease: MR imaging findings
ment. The post-contrast studies should allow and correlation with angiography. A.J.N.R. Am.
33. Warn the patient that this sequence is starting and begin the scan. J. Neuroradiol. 20:75-85.
detection of dural involvement and also allow
detection of regions of blood brain barrier Shellock, F.G. 1996. Pocket Guide to MR Proce-
ALTERNATE RAPID ASSESSMENT FOR ALTERED PERFUSION AND ACUTE ISCHEMIC breakdown due to either the vasculitis or re- dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
PROTOCOL INJURY lated ischemic lesions.
Yuh, W.T., Ueda, T., Maley, J.E., Quets, J.P., White,
To evaluate an unstable patient, imaging must be kept to a minimum. In such cases, the Literature Cited M., Hahn, P.Y., and Otake, S. 1999a. Diagnosis
following alternate protocol should be followed. Cloft, H.J., Phillips, C.D., Dix, J.E., McNulty, B.C., of microvasculopathy in CNS vasculitis: Value
Zagardo, M.T., and Kallmes, D.F. 1999. Correla- of perfusion and diffusion imaging. J. Magn. Re-
tion of angiography and MR imaging in cerebral son. Imaging 10:310-313.
Set up patient and equipment vasculitis. Acta Radiol. 40:83-87. Yuh, W.T., Ueda, T., and Maley, J.E. 1999b. Perfu-
1. Perform equipment and patient setup as in steps 1 to 11 of the Basic Protocol. Duna, G.F. and Calabrese, L.H. 1995. Limita- sion and diffusion imaging: A potential tool for
tions of invasive modalities in the diagnosis of improved diagnosis of CNS vasculitis. A.J.N.R.
2. Draw up a syringe with a single dose of contrast by weight (typically 20 ml). primary angiitis of the central nervous system. Am. J. Neuroradiol. 20:87-89.
J. Rheumatol. 22:662-667.
3. Perform Sequences 1 to 4 as described in the Basic Protocol. Greenan, T.J., Grossman, R.I., and Goldberg, H.I.
1992. Cerebral vasculitis: MR imaging and an- Contributed by Steven Thibodeau,
4. Perform Sequence 8 as described in the Basic Protocol. giographic correlation. Radiology 182:65-72. Ellen Grant, and Pamela W. Schaefer
Harris, K.G., Tran, D.D., Sickels, W.J., Cornell, Massachusetts General Hospital
S.H., and Yuh, W.T. 1994. Diagnosing intracra- Boston, Massachusetts
COMMENTARY
Background Information reversible, blood-brain barrier breakdown due
Reports on the sensitivity of standard MR to subacute ischemic events, and dural pathol-
images versus catheter angiography vary and ogy. The presence of dural sinus pathology
one or both can be normal. Standard MR or brain stem atrophy in addition to multifo-
studies and catheter angiography actually ap- cal parenchymal involvement raises the possi-
pear to provide complimentary information bility of Neurological Becet’s disease. Dural
(Greenan et al., 1992; Harris et al., 1994; Duna enhancement has also been reported in tem-
and Calabrese, 1995; Cloft et al., 1999; Pom- poral arteritis (Joelson et al., 2000). FLAIR
per et al., 1999) Recently, perfusion-weighted images are also included as they may im-
imaging has been shown to detect regions prove detection of subtle subcortical lesions
of altered hemodynamics, presumably related when compared to T2 FSE images (Jager et al.,
to microvascular involvement, that were not 1999).
identified on catheter angiography (Yuh et al.,
1999a,b). Diffusion-weighted imaging may Critical Parameters and
also add additional information by identifying Troubleshooting
regions of acute necrotic change in a nonvas- With FLAIR sequences, it is important to
cular territory (Yuh et al., 1999b). Although in make sure that the TR and TI values have not
most cases, 3-D TOF angiography is unhelp- been modified and that vendor-specific recom-
ful because vascular involvement is limited mendations for optimal CSF suppression have
to small vessels, 3-D TOF angiography may been followed. For example in GE FLAIR se-
be helpful when medium to large vessels are quences, at least two interleaved acquisitions
involved, such as in vasculitis due to fungal in- are required for optimal CSF suppression.
vasion. Post-contrast studies are routinely ob- Important basic parameters in MR angiog-
Rule Out (R/O) tained to assess for blood-brain barrier break- raphy are as discussion in UNIT A1.1. With the Intracranial
Vasculitis Arterial Disease
down due to vasculitic involvement that may be post-contrast MRA, the TR can be decreased
A1.6.10 A1.6.11
Supplement 13 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 13
 Materials
Rule Out (R/O) Migraine UNIT A1.7
Gadolinium-DTPA contrast agent (e.g., Magnevist, Omniscan, Prohance)
When imaging patients with headaches, the goals are (1) to rule out a more ominous
etiology for their headaches and, when atypical or complicated migraine is suspected, Set up patient and equipment
(2) to assess the degree of hemodynamic and parenchymal involvement. The standard 1. Interview (screen) the patient to ensure that he or she has no counterindications such
images allow other etiologies such as mass lesions to be excluded and an MR venogram as cardiac pacemakers or other implants containing ferromagnetic materials. Also be
can rule out venous sinus thrombosis. Contrast may be helpful in differentiating multiple sure to find out if the patient has any health conditions that may require the presence
small metastases, demylinating lesions with inflammatory components, and subcortical of special emergency equipment during the scanning procedure, or necessitate any
infarcts from the nonspecific foci of increased T2 associated with migraines that do other precautions. Do not forget to ask if the patient has any drug allergies and
not enhance. Perfusion-weighted imaging allows detection assessment of hemodynamic document them.
compromise. Diffusion-weighted imaging, in combination with the standard MR images, Generally standard screening forms are used for all patients scanned in a magnetic
allows assessment of parenchymal involvement. The following Basic Protocol can be used resonance system (see APPENDIX 1).
for the evaluation of patients. The Alternate Protocol is used for the assessment of cerebral
perfusion in patients with visual auras. The presence of any ferromagnetic metals may be a health hazard to the patient when
he or she is inside the magnet, and will also affect the imaging. If in doubt as to the
exact composition of the items, it is best to exclude patients with any metal implants; see
STANDARD IMAGING FOR MIGRAINE BASIC Shellock (1996) for discussion of what implants may be safely scanned using magnetic
resonance.
PROTOCOL
In order to perform the optional sequences, perfusion and diffusion studies, a scanner
Patients may be accompanied into the magnet room by a friend or family member, who can
with echoplanar capabilities is required (Table A1.7.1). The standard anatomical MR sit in the room during the scan and comfort the patient as needed. This companion must
imaging sequences included in the protocols do not require these faster gradients. The be screened as well to ensure the absence of loose metal objects on the body or clothing.
parameters given here are optimized for a 1.5T GE LX system with 11.0 software and
may need modification for different software versions, field strengths, or manufacturers. 2. If the procedure is a research protocol, have the patient sign any necessary consent
form.
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular 3. Have the patient remove all jewelry and change into a gown to eliminate any metal
patient, such as crash carts or oxygen. Reactions to contrast agents are rare, but the that might be found in clothing.
resources are necessary. 4. Have the patient wash off any mascara and other makeup to avoid local tissue heating
and image artifacts.
Table A1.7.1 Equipment Requirements for Cerebral Vascular Assessment 5. Inform the patient about what will occur during the procedure, what he or she will
Type of system LX EchoSpeed experience while in the magnet, and how to behave, including the following:
Field strength 1.5 T
a. If earphones or headphones are used to protect the ears from the loud sounds
Software level 11
produced by the gradients, the patient will be asked to wear these, but will be
Magnet type CX K4 able to communicate with the technologist at any time during the imaging.
Polarity Positive
b. The patient will be given a safety squeeze-bulb or similar equipment to request
Active gradient shielding No
assistance at any time (demonstrate how this works).
Passive gradient shielding No
c. For good results, the patient should not talk, and should avoid or minimize
Gradient strength (amplitude) 2.4 mT/cm
swallowing or other movement, during each scan—i.e., as long as the banging
Slew rate 12 mgauss/cm
sounds continue. Between scans, talking and swallowing are allowed in most
SAR (average) 2 W/kg cases, but should be avoided when comparative positional studies are being
SAR (max) 8 W/kg performed; the patient will be informed when this is the case.
Superior conductive shimming 45 cm
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
Resistive shimming No
Active magnetic shielding No 6. Have the patient mount onto the table. Either before or right after the patient lies
Maximum noise level 100 db down, set up any triggering devices or other monitoring equipment that is to be used.
Type of body coil High pass 7. Center the patient in a head or neck coil at the region where the key information
TPS recon system memory 386 is desired. Make sure that the head and neck are constrained to prevent motion,
Array processor (TPS) Reflex 100 especially if high-resolution scans are to be run.
Computer LX Octane Generally the patient’s head is fixed so that the head is horizontal (not tilted) and the
Weight limit 350 lb (130.6 kg) neck and head lie along the axis of the patient table; other positions may be appropriate
Intracranial Rule Out (R/O)
Arterial Disease Migraine depending on the needs at hand.

Contributed by Steven Thibodeau and Ellen Grant A1.7.1 A1.7.2

Current Protocols in Magnetic Resonance Imaging (2005) A1.7.1-A1.7.11
C 2005 by John Wiley & Sons, Inc.
Copyright  Supplement 13 Supplement 13 Current Protocols in Magnetic Resonance Imaging
 Table A1.7.2 Rapid Sagittal T1 -Weighted Scout Table A1.7.3 Transverse FLAIR

Patient position Supine Patient position Supine

Scan type Gradient echo Scan type Inversion recovery FSE
Imaging plane (orientation) Sagittal Imaging plane (orientation) Transverse
Central slice or volume center Nasion Variable bandwidth Yes
Echo time (TE ) Minimum (at least 4.1msec) Central slice or volume center Nasion
Receiver bandwidth (RBW) 15.6 kHz Echo time (TE ) 120 msec
Repeat time (TR ) 100 msec Receiver bandwidth (RBW) 20.83 kHz
Flip angle (FA) 60◦ Repeat time (TR ) 10,000 msec
Fields of view (FOVx , FOVy ) 240 mm, 240 mm Inversion time (TI ) 2200 msec
Flip angle (FA) 90◦
Resolution (x, y) 0.94 mm, 1.87 mm
Fields of view (FOVx , FOVy ) 220 mm, 220 mm
Number of data points collected (Nx , Ny ) 256, 128
Resolution (x, y) 0.86 mm, 1.14 mm
Display matrix (Dx , Dy ) 128, 128
Number of data points collected (Nx , Ny ) 256, 192
Slice thickness (z) 7 mm
Display matrix (Dx , Dy ) 256, 192
Number of slices 7
Slice thickness (z) 5 mm
Slice gap 2 mm
Number of slices 24
Number of excitations (NEX) 1
Slice gap 1 mm
Number of acquisitions (Nacq ) 1 Number of excitations (NEX) 1
Read direction Anterior–posterior Number of acquisitions (Nacq ) 2
Scan time 16 sec Read direction Anterior–posterior
Saturation pulses Inferior
8. If needed, place a pillow or other support under the knees to make the patient more Scan time 4 min, 44 sec
comfortable.
9. Use the centering light to position the patient and put him or her into the center of Table A1.7.4 Transverse T2 -Weighted FSE
the magnet. Patient position Supine
Once this step has been performed, so long as the patient does not move on the table, Scan type Fast spin echo XL
the table itself can be moved and then replaced in the same position as before without Imaging plane (orientation) Transverse
jeopardizing the positioning of one scan relative to another. Variable bandwidth Yes
Pulse sequence database (PSD) FSE-XL
10. If the patient is unable to hold still, provide an appropriate sedative.
Central slice or volume center Nasion
Sequence 1: Rapid sagittal T1 -weighted scout Echo time (TE ) 102 msec
11. To determine the patient’s position, a fast sagittal scout scan is performed using the Receiver bandwidth (RBW) 11.36 kHz
imaging sequence in Table A1.7.2. Echo train length (ETL) or shots (SH) 12
Some centers may prefer a rapid three-plane scout instead of this multislice sagittal scout. Repeat time (TR ) 6000 msec
Flip angle (FA) 90◦
Sequence 2: Transverse FLAIR Fields of view (FOVx , FOVy ) 220 mm, 165 mm
12. From the sagittal scout, select the image through the center of the brain to set up Resolution (x, y) 0.86 mm, 0.86 mm
the locations for the transverse FLAIR. The locations should be the same as those Number of data points collected (Nx , Ny ) 256, 192
chosen for the transverse diffusion. Set up the imaging parameters as shown in Table Display matrix (Dx , Dy ) 320, 256
A1.7.3. Slice thickness (z) 5 mm
13. Warn the patient that this sequence is starting and begin the scan. Number of slices 24
Slice gap 1 mm
Sequence 3: Transverse T2 -weighted FSE Number of excitations (NEX) 2
14. From the sagittal scout, select the image through the center of the brain to set up Number of acquisitions (Nacq ) 1
the locations for the transverse T2 FSE. The locations should be the same as those Read direction Anterior–posterior
chosen for the transverse diffusion. Set up the imaging parameters as shown in Table Flow compensation Yes
A1.7.4. Extended dynamic range (EDR) Yes
15. Warn the patient that this sequence is starting and begin the scan. Saturation pulses Inferior
Intracranial Rule Out (R/O)
Arterial Disease Migraine Scan time 3 min, 19 sec

A1.7.3 A1.7.4
Current Protocols in Magnetic Resonance Imaging Supplement 13 Supplement 13 Current Protocols in Magnetic Resonance Imaging
 Table A1.7.5 Transverse Diffusion Table A1.7.6 Coronal 2-D TOF MR Venogram

Patient position Supine Patient position Supine

Scan type SE-EPI Scan type Vascular TOF SPGR
Imaging plane (orientation) Transverse Imaging plane (orientation) Coronal
Central slice or volume center Nasion Variable bandwidth Yes
Echo time (TE ) Minimum (80.9 msec) Central slice or volume center Nasion
Echo train length (ETL) or shots (SH) 1 SH Echo time (TE ) Minimum
Repeat time (TR ) 5000 msec Receiver bandwidth (RBW) 15.6 kHz
Flip angle (FA) 90◦ Repeat time (TR ) 40 msec
Fields of view (FOVx , FOVy ) 220 mm, 220 mm Flip angle (FA) 60◦
Fields of view (FOVx , FOVy ) 220 mm, 165 mm
Resolution (x, y) 1.72 mm, 1.72 mm
Resolution (x, y) 0.86, 1.14
Number of data points collected (Nx , Ny ) 128, 128
Number of data points collected (Nx , Ny ) 256, 144
Display matrix (Dx , Dy ) 128, 128
Display matrix (Dx , Dy ) 256, 192
Slice thickness (z) 5 mm
Slice thickness (z) 1.5 mm
Number of slices 24
Number of slices Variable (depends on area of
Slice gap 1 mm coverage)
Number of excitations (NEX) 3 Slice gap 0
Number of acquisitions (Nacq ) 1 Number of excitations (NEX) 1
Read direction Right–left Number of acquisitions (Nacq ) Variable (depends on area of
Saturation pulses Fat saturation (automatic with EPI) coverage)
Control variables (CV) Ramp sampling = 1, Read direction Superior–Inferior
burst sampling = 0 Saturation pulses Inferior
b-value 1000 sec/min2 Scan time 12 sec/slice
Direction 6
Scan time 3 min, 5 sec
Sequence 6: Transverse T1 -weighted spin echo (optional)
20. From the sagittal scout, select the image through the center of the brain to set up the
Sequence 4: Transverse diffusion locations for the transverse T1 . The locations should be the same as those chosen for
16. From the sagittal scout, select the image through the center of the brain to set up the transverse diffusion. Set up the imaging parameters as shown in Table A1.7.7.
the locations for the transverse diffusion. The transverse images should begin at the
21. Warn the patient that this sequence is starting and begin the scan.
foramen magnum and end at the top of the brain. Set up imaging parameters as shown
in Table A1.7.5. Sequence 7: Transverse post-contrast T1 -weighted spin echo (optional)
17. Warn the patient that this sequence results in loud beeping noises and begin the scan. 22. Using a small-gauge butterfly, inject a single dose of contrast by weight (usually
20 ml) intravenously.
Most diffusion sequences will perform a minimum of four sequences: three with orthogonal
diffusion gradient directions and one with a minimal or no diffusion gradient. Often, a fifth 23. From the sagittal scout, select the image through the center of the brain to set up
set of images is also provided, which combines the three orthogonal gradient images to the locations for the post-contrast transverse T1 . The locations should be the same
produce a set of images whose signal intensity is not affected by the diffusion direction. This as those chosen for the transverse precontrast T1 . Set up the imaging parameters as
is the DWI. In order to obtain apparent diffusion coefficient (ADC) maps, the images may shown in Table A1.7.8.
need further processing. On a GE system, the ADC maps can be obtained by processing on
the Advantage Window Workstation using FuncTool version 3.1P using FuncTool version 24. Warn the patient that this sequence is starting and begin the scan.
1.9M.
Sequence 8: Coronal post-contrast T1 -weighted spin echo (optional)
Sequence 5: Coronal 2-D TOF MR venogram (optional) Contrast may be helpful in differentiating multiple small metastases, demylinating lesions
18. From the sagittal scout, select the image through the center of the brain to set up the with inflammatory components, and subcortical infarcts from the nonspecific foci of
locations for the coronal TOF MR venogram. The entire brain should be included. increased T2 associated with migraines that do not enhance.
Set up the imaging parameters as shown in Table A1.7.6.
25. From the sagittal scout, select the image through the center of the brain to set up the
19. Warn the patient that this sequence is starting and begin the scan. locations for the coronal post-contrast T1 . Set up the imaging parameters as shown
in Table A1.7.9.
26. Warn the patient that this sequence is starting and begin the scan.
Intracranial Rule Out (R/O)
Arterial Disease Migraine

A1.7.5 A1.7.6
Current Protocols in Magnetic Resonance Imaging Supplement 13 Supplement 13 Current Protocols in Magnetic Resonance Imaging
 Table A1.7.7 Transverse T1 -Weighted Spin Echo Table A1.7.9 Coronal Post-Contrast T1 -Weighted Spin Echo
Patient position Supine Patient position Supine
Scan type Spin echo Scan type Spin echo
Imaging plane (orientation) Transverse Imaging plane (orientation) Coronal
Variable bandwidth Yes Variable bandwidth Yes
Central slice or volume center Nasion Central slice or volume center Nasion
Echo time (TE ) Minimum full (14 msec) Echo time (TE ) Minimum full (20 msec)
Receiver bandwidth (RBW) 15.6 kHz Receiver bandwidth (RBW) 15.6 kHz
Repeat time (TR ) 400 msec Repeat time (TR ) 500 msec
Flip angle (FA) 90◦ Flip angle (FA) 90◦
Fields of view (FOVx , FOVy ) 220 mm, 165 mm Fields of view (FOVx , FOVy ) 220 mm, 165 mm
Resolution (x, y) 0.86 mm, 1.14 mm Resolution (x, y) 0.86 mm, 1.14 mm
Number of data points collected (Nx , Ny ) 256, 144 Number of data points collected (Nx , Ny ) 256, 144
Display matrix (Dx , Dy ) 256, 192 Display matrix (Dx , Dy ) 256, 192
Slice thickness (z) 5 mm Slice thickness (z) 5 mm
Number of slices 24 Number of slices 26
Slice gap 1 mm Slice gap 1 mm
Number of excitations (NEX) 1 Number of excitations (NEX) 1
Number of acquisitions (Nacq ) 2 Number of acquisitions (Nacq ) 2
Read direction Anterior–posterior Read direction Superior–inferior
Flow compensation No Flow compensation Yes
Extended dynamic range (EDR) Yes Extended dynamic range (EDR) Yes
Saturation pulses Inferior Saturation pulses Inferior
Scan time 2 min, 8 sec Scan time 2 min, 40 sec

Table A1.7.8 Transverse Post-Contrast T1 -Weighted Spin Echo

ALTERNATE ASSESSMENT OF CEREBRAL PERFUSION IN PATIENTS WITH AURAS
Patient position Supine PROTOCOL
Occasionally perfusion weighted imaging may be helpful in patients with migraine to as-
Scan type Spin echo
sess for associated perfusion changes. During the aura and the first few hours of headache
Imaging plane (orientation) Transverse
transient hypoperfusion has been documented whereas others have noted hyperfusion but
Variable bandwidth Yes
this may occur at a later phase. Although still controversial, perfusion differences in the
Central slice or volume center Nasion setting of migraine do not typically respect vascular territories which help differentiate
Echo time (TE ) Minimum full (20 msec) these changes from stroke related changes.
Receiver bandwidth (RBW) 15.6 kHz
Repeat time (TR ) 400 msec Materials
Flip angle (FA) 90◦ Gadolinium-based MR contrast agent (e.g., Magnevist, Omniscan, Prohance)
Fields of view (FOVx , FOVy ) 220 mm, 165 mm Normal saline (0.9% NaCl), sterile
Resolution (x, y) 0.86 mm, 1.14 mm 18-G angiocatheter and MR-compatible injection pump
Number of data points collected (Nx , Ny ) 256, 144
Display matrix (Dx , Dy ) 256, 192 Set up patient and equipment
Slice thickness (z) 5 mm 1. Set up patient and equipment as described in Basic Protocol, steps 1 to 10.
Number of slices 24 2. Establish an intravenous line using an 18-G angiocatheter through which the contrast
Slice gap 1 mm agent can be injected. Attach this line securely to the patient so that movement into
Number of excitations (NEX) 1 or out of the magnet will not pull at the patients arm.
Number of acquisitions (Nacq ) 2
It is preferable to insert the line prior to imaging and to leave the patient in the magnet, so
Read direction Anterior–posterior that there is no intervening motion between the scans run before contrast agent injection
Flow compensation Yes and those run after injection.
Extended dynamic range (EDR) Yes
Saturation pulses Inferior
3. Connect the MR-compatible injection pump, cleared of air, and loaded with a double
Intracranial Rule Out (R/O) dose of contrast agent and saline flush, to the patient. Do a small test injection of
Scan time 2 min, 8 sec Arterial Disease Migraine
normal saline to ensure that the i.v. is working properly.
A1.7.7 A1.7.8
Current Protocols in Magnetic Resonance Imaging Supplement 13 Supplement 13 Current Protocols in Magnetic Resonance Imaging
 Table A1.7.10 Transverse Perfusion Data processing and viewing for Sequence 6
Patient position Supine 9. Process the perfusion images using an appropriate workstation and software to give
Scan type GRE-EPI maps of relative cerebral blood volume (rCBV), relative cerebral blood flow (rCBF),
Imaging plane (orientation) Transverse and mean transit time (MTT).
Variable bandwidth Yes On a GE system, one can use an Advantage Windows Workstation with FuncTool version
Central slice or volume center Nasion 1.9M.
Echo time (TE ) 54 msec Sequences 7 and 8
Echo train length (ETL) or shots (SH) 1 10. Perform sequences 7 and 8 as described in the Basic Protocol.
Repeat time (TR ) 1500 msec
Flip angle (FA) 35◦
Fields of view (FOVx , FOVy ) 220 mm, 220 mm COMMENTARY
Resolution (x, y) 1.72 mm, 1.72 mm Background Information Critical Parameters and
Number of data points collected (Nx , Ny ) 128, 128 The incidence of small foci of increased T2 Troubleshooting
Display matrix (Dx , Dy ) 128, 128 signal in patients with a history of migraine For a discussion of critical parameters and
Slice thickness (z) 5 mm varies from study to study. Reported incidence troubleshooting, see UNITS A1.1 & A1.2, as well
rates vary from 13% to 39.6% for migraineurs, as UNITS A1.5 & A1.6.
Number of slices 24
dropping to 5.5% to 29.4% for those under 40
Slice gap 1 mm compared to the range of 4.3% to 11.2% for Anticipated Results
Number of excitations (NEX) 1 controls (Igarashi et al., 1991; Osborn et al., In most cases, the imaging studies in pa-
Number of acquisitions (Nacq ) 1 1991; Robbins and Friedman, 1992; Pavese tients with migraines will be normal or show
et al., 1994; De Benedittis et al., 1995; Cooney nonspecific foci of increased T2 signal that
Read direction Right–left
et al., 1996). Other causes of focal T2 hyper- are typically more prominent in the frontal re-
Slice location Inferior slice to include MCA gions. It is prudent, however, to rule out other
intensities that should be ruled out include hy-
Slice series Interleaved potential causes of multiple foci with increased
pertension, athlerosclerotic heart disease, dia-
Control variables (CV) Ramp sampling = on, betes mellitus, autoimmune disorders, and de- T2 signal such as hypertension, athlerosclerotic
burst sampling = off myelinating diseases. The increased frequency heart disease, diabetes mellitus, autoimmune
Multiphase Yes, 46 phases per location, of T2 hyperintensities in pediatric migraineurs disorders, or demyelinating diseases.
minimum delay between acquisition is controversial, with some stating there is no In patients with basilar artery migraine,
Scan time 1 min, 10 sec increase in incidence (McAbee et al., 1993) hemiplegic migraine, or visually triggered mi-
and others reporting a 50% incidence vs. 17% graines, other imaging findings (see Back-
for age-matched controls (Hamalainen et al., ground Information) may be observed.
4. Program a 5 ml/sec injection rate for both contrast and saline, the contrast dose 1996). Perfusion changes have been detected Literature Cited
required (double dose by weight), the saline flush required (typically 40 ml), and a during visual auras with these changes persist- Arnold, G., Reuter, U., Kinze, S., Wolf, T., and Ein-
10-sec time delay. Start the saline running at TKVO (to keep vein open) and arm the ing for 2.5 hr into the migraine. These perfu- haupl, K.M. 1998. Migraine with aura shows
injection pump. sion studies showed decreased cerebral blood gadolinium enhancement which is reversed
following prophylactic treatment. Cephalalgia
flow (16% to 53%), decreased cerebral blood
Sequences 1 to 5 18:644-646.
volume (6% to 33%), and increased tissue
Chabriat, H., Vahedi, K., Clark, C.A., Poupon,
5. Perform sequences 1 to 4 and sequence 6 as described in the Basic Protocol. mean transit time (10% to 54%) in the gray
C., Ducros, A., Denier, C., Le Bihan, D., and
matter of the occipital cortex (Cutrer et al., Bousser, M.G. 2000. Decreased hemispheric
Sequence 6: Transverse perfusion
1998). Migraines without aura showed no sig- water mobility in hemiplegic migraine related
6. Using the transverse T2 -weighted FSE images, find the slice location where the nificant hemodynamic changes (Sanchez del to mutation of CACNA1A gene. Neurology
middle cerebral arteries are best seen. Using the imaging parameters as shown in Rio et al., 1999). Rarely, as in the case of basi- 54:510-512.
Table A1.7.10, set up the transverse images starting at the same location where the lar artery migraine, reversible cortical swelling Cooney, B.S., Grossman, R.I., Farber, R.E., Goin,
MCA were identified on the T2 FSE. Typically only ten or eleven slices can be may be seen (Maytal et al., 1998). Hemiplegic J.E., and Galetta, S.L. 1996. Frequency of mag-
netic resonance imaging abnormalities in pa-
obtained in the given TR interval. The slice thickness and gap can be modified so that migraine also has uncommon findings with de-
tients with migraine. Headache 36:616-621.
the perfusion study covers any diffusion abnormality. creased diffusion, decreased blood flow, and
Crawford, J.S. and Konkol, R.J., 1997. Familial
leptomeningeal enhancement reported (Craw-
hemiplegic migraine with crossed cerebellar di-
7. Check that the injection pump is ready to inject and that the injection is set to start ford and Konkol, 1997, Arnold et al., 1998, aschisis and unilateral meningeal enhancement.
after a 10-sec delay. Chabriat et al., 2000). In a similar group of pa- Headache 37:590-593.
tients with visually triggered migraines, neu- Cutrer, F.M., Sorensen, A.G., Weisskoff, R.M.,
8. Warn the patient that this sequence is starting and that, partway through the scan, the ronal suppression was accompanied by base- Ostergaard, L., Sanchez del Rio, M., Lee,
i.v. injection will occur. Begin the scan. line contrast intensity increases on functional E.J., Rosen, B.R., and Moskowitz, M.A.
MRI that suggested vasodilatation and tissue 1998. Perfusion-weighted imaging defects dur-
ing spontaneous migrainous aura. Ann. Neurol.
Intracranial Rule Out (R/O) hyperoxygenation.
43:25-31.
Arterial Disease Migraine

A1.7.9 A1.7.10
Current Protocols in Magnetic Resonance Imaging Supplement 13 Supplement 13 Current Protocols in Magnetic Resonance Imaging
De Benedittis, G., Lorenzetti, A., Sina, C., and
Bernasconi, V. 1995. Magnetic resonance imag-
ing in migraine and tension-type headache.
Headache 35:264-268.
Hamalainen, M.L., Autti, T., Salonen, O., and San-
tavuori, P. 1996. Brain MRI in children with
migraine: A controlled morphometric study.
Cephalalgia 16:541-544.
Igarashi, H., Sakai, F., Kan, S., Okada, J., and
Tazaki, Y. 1991. White matter MRI hyperinten-
sities in a hundred-and-twenty-nine consecutive
migraine patients. Cephalalgia 14:318.
Maytal, J., Libman, R.B., and Lustrin, E.S. 1998.
Basilar artery migraine and reversible imag-
ing abnormalities. A.J.N.R. Am. J. Neuroradiol.
19:1116-1119.
McAbee, G.N., Siegel, S.E., Kadakia, S., and Can-
tos, E. 1993. Value of MRI in pediatric migraine.
Headache 33:143-144.
Osborn, R.E., Alder, D.C., and Mitchell, C.S. 1991.
MR imaging of the brain in patients with mi-
graine headaches. A.J.N.R. Am. J. Neuroradiol.
12:521-524.
Current Protocols in Magnetic Resonance Imaging
Pavese, N., Canapicchi, R., Nuti, A., Bibbiani, F.,
Lucetti, C., Collavoli, P., and Bonuccelli, U.
1994. White matter MRI hyperintensities in a
hundred and twenty-nine consecutive migraine
patients. Cephalalgia 14:342-345.
Robbins, L. and Friedman, H. 1992. MRI in mi-
graineurs. Headache 32:507-508.
Sanchez del Rio, M., Bakker, D., Wu, O., Agosti, R.,
Mitsikostas, D.D., Ostergaard, L., Wells, W.A.,
Rosen, B.R., Sorensen, G., Moskowitz, M.A.,
and Cutrer, F.M. 1999. Perfusion weighted
imaging during migraine: Spontaneous visual
aura and headache. Cephalalgia 19:701-707.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Contributed by Steven Thibodeau and
Ellen Grant
Massachusetts General Hospital
Boston, Massachusetts
Intracranial
Arterial Disease
A1.7.11
Supplement 13
CHAPTER A2
Cerebral Venous Lesions
INTRODUCTION
erebral venous imaging includes evaluation of intracerebral venous thrombosis as
C well as imaging cavernous and venous angiomas. This may be accomplished by
conventional MR imaging and magnetic resonance venography (MRV). Cerebral venous
occlusive disease is distinctly different from arterial occlusive disease. With venous
occlusion, the blood-brain barrier often remains intact. Signal abnormalities seen on
neuroimaging studies are often reversible and hemorrhage is more common. Venous
infarcts do not usually affect discrete, well defined territories like arterial infarcts.
However, some patterns do tend to emerge. Transverse sinus thrombosis tends to affect
the temporal lobes. Superior sagittal sinus thrombosis tends to affect the parasagittal
frontal lobes. Vein of Galen/straight sinus thrombosis tends to affect the thalami.
While the cause is not always found, many of these cases fall into certain categories.
Intracerebral venous thrombosis is found to have a number of predisposing factors. These
can be divided into septic and aseptic causes. Aseptic causes include hematologic
disorders, neoplasm, dehydration, trauma, drugs, and inflammatory conditions. Septic
causes include encephalitis, meningitis, mastoiditis, sinusitis, and cellulitis.
Patients with cerebral venous thrombosis may have presentations that range from insidi-
ous onset to rapid loss of altered mental status and focal neurologic deficits. Those
presenting with gradual onset of symptoms have headache, confusion, and papilledema.
Venous occlusive disease often is present with an intact blood-brain barrier. With arterial
occlusive disease, the blood-brain barrier is more frequently broken down. In venous
occlusive disease, initially detected abnormalities are often reversible. Hemorrhage is
much more common with venous than arterial occlusion.
Cerebral venous imaging has previously been performed with invasive angiographic tech-
niques. Many of the patients who need venous circulation evaluation do not need to have an
invasive procedure, a test that involves exposure to ionizing radiation or risk reaction to
iodinated contrast media. This is particularly true in the pregnant patient who is suspected of
having venous sinus thrombosis. Magnetic resonance venography (MRV) offers a welcome
alternative to invasive techniques. It has proven to be a valuable tool in evaluating the
intracranial cerebral venous system. Imaging dural sinus thrombosis is the topic of UNIT A2.1.
Often MRI is used to clarify a confusing finding identified on another imaging study.
Frequently, these lesions are first seen on MRI examinations and may be completely
worked up or followed with MRI. This is related to the frequency with which patients are
being examined by MR for unexplained headaches. Cavernomas may be followed
sequentially with MRI, particularly after a bleed. Most venous angiomas are small and
self-limiting, and no further examination is necessary. The sequences described in UNITS
A2.1, A2.2, and A2.3 are based on experience with the Siemens 1.5 T Vision and Sonata
scanners, but should be adaptable to high-field-strength machines from other manufac-
turers. Cavernous angiomas and venous angiomas may be completely worked up with
MRI without need for additional diagnostic studies. Imaging of cavernous angiomas and
cavernomas are the topics of UNITS A.2.2 and A2.3 respectively.
F. Allan Midyett, Laurie Fisher, and Suresh Mukherji Cerebral Venous
Lesions
Contributed by F. Allan Midyett, Laurie Fisher, and Suresh Mukherji A2.0.1
Current Protocols in Magnetic Resonance Imaging (2003) A2.0.1
Copyright © 2003 by John Wiley & Sons, Inc. Supplement 9
Imaging Dural Sinus Thrombosis
The acute onset of irreversible neurological deficit is referred to as a “stroke” and is the
third leading cause of death and the major cause of adult long-term disability in the United
States. (Fischbein et al., 2000). Obstruction of the venous sinuses by thrombosis accounts
for ∼1% of strokes. Thrombosis of the superior saggital sinus is most frequent, with
involvement of transverse, sigmoid, and cavernous sinuses less often. Venous thrombosis
may involve the dural sinuses, the deep venous system, and the superficial cortical veins
separately or in any combination.
Patients presenting with seizures more often have thrombosis of the superficial cortical
veins. These patients have a worse prognosis than those with isolated dural sinus
thrombosis. Early diagnosis and treatment significantly improve prognosis. Ideally,
therapy can be started before the patient has developed significant edema and/or hemor-
rhage. Mortality figures range from 20% to 75% with a trend of lowered rates with
increased availability of MRI.
In ∼40% of cases of dural sinus thrombosis, there is retrograde clot propagation into the
cortical or deep medullary veins. While the causes of venous thrombosis can be divided
into septic and aseptic etiologies, 25% remain unknown, being classified as idiopathic
(Fischbein et al., 2000).
IMAGING DURAL SINUSES BY TIME-OF-FLIGHT MRV
Magnetic resonance imaging with magnetic resonance venography (MRV) is the tech-
nique of choice in the diagnosis and follow-up of cerebral venous thrombosis. Thrombosis
may be identified on spin echo images as an absence of the normal flow void. On MRV,
thrombosis manifests itself as lack of signal. Thrombus signal intensity differs on T1- and
T2-weighted spin echo images and evolves with the phases of hemoglobin degradation.
Recognition of artifacts and pitfalls related to MRI techniques is crucial to correctly
interpret the presence or absence of venous sinus thrombosis.
The following sequences comprise the authors’ preferred protocol for evaluating dural
sinus thrombosis, (1) three-plane positioning scout; (2) time-of-flight MRV; (3) non-
contrast spin echo sagittal T1-weighted; (4) non-contrast coronal turbo spin echo proton
density/T2-weighted; and (5) spin echo coronal with flow compensation. These imaging
sequences take ∼16 min and 3 sec of imaging time in addition to set up time between
sequences.
Table A2.1.1 lists the equipment needed to perform the following imaging sequences. The
available gradient strength will depend on the scanner, and the echo times given below
will be modified accordingly (i.e., the smaller the gradient strength, the longer the echo
Table A2.1.1 Equipment Specifications Needed to Perform Imaging Sequences
Coil type Circularly polarized head coil
Field strength 1.5 T
Gradient strength Minimum of 15 mT/m
Positioning cushion or head Yes
stabilizers
Knee cushion Yes
Use of contrast agents No
Pulse oximeter If patient requires sedation
Contributed by F. Allan Midyett, Suresh Mukherji, and Laurie Fisher
Current Protocols in Magnetic Resonance Imaging (2001) A2.1.1-A2.1.12
Copyright © 2001 by John Wiley & Sons, Inc.
UNIT A2.1 time for a particular scan). The authors use a Siemens Vision 1.5 T system, but they expect
to run the same sequences in this unit on other machines made by other MR manufacturers.
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.
Set up patient and equipment
1. Interview the patient to assess for contraindications such as cardiac pacemaker,
implanted mechanical devices, and/or ferromagnetic materials. Also, determine if the
patient will need sedation medication, which will necessitate the need for appropriate
monitoring equipment.
Each patient or legal guardian, prior to bringing the patient into the exam area, must sign
a screening form.
The presence of ferromagnetic materials may be a health hazard to the patient while in the
magnetic field and/or adversely affect image quality. To determine the safety of scanning
such ferromagnetic materials, see Shellock (1996).
The presence of ferromagnetic materials in the globe of the eye is contraindicated for MRI.
Patients with prior metal exposure to the eye should have plain X-rays of the orbital area
to ensure that all metal has been removed prior to placing them in the magnetic field
(Shellock, 1996).
BASIC
2. If the procedure is a research protocol, have the patient sign any necessary consent
PROTOCOL form.
3. Request the patient to change into a gown and remove all personal effects such as,
jewelry, hearing aids, glasses, etc., prior to entering the MRI scan room. All personal
belongings should be secured during the examination.
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating
and artifacts.
5. Explain the procedure to the patient and record relevant clinical history. Ensure that
the patient understands what is expected and ask if he or she has questions. Answer
appropriately.
6. Set up the exam room by securing the circularly polarized (CP)-head coil onto the
table and providing a clean exam table.
7. Escort the patient to the MR examination room and position him or her according to
the exam to be performed. Review the following items with the patient:
a. Provide earplugs or headphones to the patient to minimize the gradient noise but
assure the patient that the patient should still be able to hear you.
b. Provide the patient with a safety squeeze-bulb and demonstrate how it works.
Explain to the patient that he or she may use the squeeze-bulb if assistance is
needed during the exam.
c. Explain to the patient that you will be talking to him or her between imaging
sequences that will be when the loud knocking noise stops.
d. Explain to the patient that it is imperative that he or she remains motionless during
the loud knocking noise to ensure good results. Also, explain that they should not
reposition their body between imaging sequences.
e. The patient may call out at any time if he or she feels it necessary.
Cerebral Venous Imaging Dural
Lesions Sinus Thrombosis
A2.1.1 A2.1.2
Current Protocols in Magnetic Resonance Imaging
 f. Provide the patient with an estimate of the examination length. Although the total Table A2.1.2 Imaging Parameters for Sequence 1: Scout Sequence
scan time is 16 min and 3 sec, because of set-up time between sequences, tell the
patient that this exam will take ∼20 min. Patient position Supine
Scan type Gradient echo
8. Position a support under the patient’s knees to enhance comfort. Imaging plane (orientation) Transverse, sagittal, coronal
9. If the patient is unable to hold still, provide appropriate sedation. (various)
Central slice or volume center Nasion
10. Secure the patient’s head with positioning sponges or head stabilizers. Echo time (TE) 6 msec
11. Center the patient’s nasion to the coil using the laser light. Repeat time (TR) 15 msec
Flip angle (FA) 30°
12. Advance the patient couch to isocenter.
Field of view (FOVx, FOVy) 300 mm, 300 mm
Sequence 1: Rapid three-plane positioning scout Resolution (Δx, Δy) 1.17 mm, 2.34 mm
13. To validate the patient’s position and to have a reference to prescribe successive Number of data points collected (Nx, Ny) 256, 128
imaging sequences, acquire a three-plane orthogonal scout sequence. See Table Slice thickness (Δz) 8 mm
A2.1.2 for specific parameters. Run the scan. Number of slices 3 (one per orientation)
Slice gap Not applicable
Most MR scanners can be programmed to acquire the scout automatically after coil tuning Number of acquisitions (Nacq) 1
or after the patient has been placed in isocenter (for systems that do not require tuning). Swap read and phase encoding No
Siemens acquires the scout in three planes by selecting “various.” Slice location Not applicable
Saturation pulses Not applicable
Sequence 2: Time-of-flight MRV–sagittal sinus
Slice series Ascending caudal to cranial
14. Display both the sagittal and transverse scout images in two separate quadrants on
Scan time 7 sec
the scan monitor.
15. Change imaging parameters to those listed in Table A2.1.3. Position slices to midline
of the transverse scout with a sagittal to coronal angle of approximately −12° (this Table A2.1.3 Imaging Parameters for Sequence 2: Time-of-Flight MRV of the
will reduce in-plane saturation effects). Position the presaturation band off the sagittal Sagittal Sinus
scout image angled in a manner to saturate the facial sinuses (be sure not to “clip”
any of the sagittal sinus). Patient position Supine
Scan type 2-D gradient echo
16. Instruct the patient to remain motionless, as the scan will begin and last for ∼9 min. Imaging plane (orientation) Sagittal
Run a time-of-flight magnetic resonance venogram (MRV) of the superior sagittal Central slice or volume center Slices centered to midline on
sinus according to the parameters in Table A2.1.3. transverse scout
This sequence is most important in demonstrating flow within the dural sinuses as well as Echo time (TE) 9 msec
potentially showing the location and cause of obstruction including thrombi. Repeat time (TR) 30 msec
Flip angle (FA) 50°
Sequence 3: Noncontrast spin echo sagittal T1-weighted sequence Field of view (FOVx, FOVy) 210 mm, 210 mm
17. Some patients have this examination because of findings noted on a routine cranial Resolution (Δx, Δy) 0.82 mm, 0.82 mm
MR. If not already provided as a part of a recent study, run a noncontrast spin echo Number of data points collected (Nx, Ny) 256, 256
sagittal T1-weighed scan using the parameters in Table A2.1.4. Slice thickness (Δz) 3 mm
Number of slices 64
Sequence 4: Noncontrast coronal multiecho spin echo Slice gap 1 mm
18. Run a noncontrast coronal multiecho scan according to Table A2.1.5. This is a dual Number of acquisitions (Nacq) 1
echo turbo spin echo sequence that acquires both proton density and T2-weighted Swap read and phase encoding No
images simultaneously. Slice location Angle sagittal to coronal −12° (to
reduce in-plane saturation) off
In this study, look for normal flow voids in the venous sinuses. These flow voids may be transverse scout image
compromised by signal from blood products on T2-weighted images. There tend to be fewer Saturation pulses Yes, one angled through facial
of these artifacts on the proton density weighted images. The proton density weighted sinus area off sagittal scout image,
images come as a bonus with the T2-weighted images without any additional expenditure 60-mm thick
of time and they are useful in this situation.
Slice series Ascending
Scan time 6 min, 43 sec
Sequence 5: Spin echo coronal with flow compensation
19. Run a spin echo coronal scan according to the parameters in Table A2.1.6.
Cerebral Venous Imaging Dural
Lesions Sinus Thrombosis

A2.1.3 A2.1.4
Current Protocols in Magnetic Resonance Imaging Supplement 9 Supplement 9 Current Protocols in Magnetic Resonance Imaging
 Table A2.1.4 Imaging Parameters for Sequence 3: Spin Echo Sagittal Table A2.1.6 Imaging Parameters for Sequence 5: T1-Weighted Spin Echo
T1-Weighted Sequence Coronal with Flow Compensation

Patient position Supine Patient position Supine

Scan type Spin echo Scan type Spin echo
Imaging plane (orientation) Sagittal Imaging plane (orientation) Coronal
Central slice or volume center Slices centered to midline on Central slice or volume center Slices centered to midbrain on
transverse scout sagittal scout
Echo time (TE) 14 msec Echo time (TE) 17 msec
Repeat time (TR) 510 msec Repeat time (TR) 532 msec
Flip angle (FA) 90° Flip angle (FA) 90°
Field of view (FOVx, FOVy) 230 mm, 230 mm Field of view (FOVx, FOVy) 230 mm, 230r mm, with r = 3/4
Resolution (Δx, Δy) 0.90 mm, 0.95 mm (rectangular field of view)
Number of data points collected (Nx, Ny) 256, 243 Resolution (Δx, Δy) 0.90 mm, 0.90 mm
Slice thickness (Δz) 6 mm Number of data points collected (Nx, Ny) 256, 256r, with r = 3/4
Number of slices 17 (rectangular field of view)
Slice gap 1.8 mm Slice thickness (Δz) 5 mm
Number of acquisitions (Nacq) 1 Number of slices 19
Swap read and phase encoding No Slice gap 1.5 mm
Slice location Equidistant between left and right Number of acquisitions (Nacq) 2
brain Swap read and phase encoding No
Saturation pulses No Slice location Center midway anterior to
Slice series Interleaved posterior brain
Scan time 2 min, 7 sec Flow compensation Yes
Saturation pulses No
Slice series Interleave
Table A2.1.5 Imaging Parameters for Sequence 4: Coronal Turbo Spin Echo Scan time 3 min, 26 sec
Proton Density (PD)/T2-Weighted Images

Patient position Supine

Scan type Turbo spin echo (dual echo)
Imaging plane (orientation) Coronal Siemens calls this sequence gradient motion rephasing (GMR). It is very helpful in reducing
Central slice or volume center Slices centered to midbrain on flow related artifact that may hinder evaluation of dural sinuses, particularly in the
sagittal scout posterior fossa. This sequence also accentuates the visibility of vessels, especially rapidly
Echo time (TE) 15 msec and 105 msec flowing blood. This is useful when the vein is narrow or there is a thin rim of flow around
Echo train length (ETL) 7 the periphery of a thrombus.
Repeat time (TR) 4500 msec
Flip angle (FA) 180°a COMMENTARY
Field of view (FOVx, FOVy) 230 mm, 230r mm, with r = 3/4
Background Information infarction, hemorrhage, hydrocephalus, and
(rectangular field of view)
Expectations have risen in the medical com- death. Predisposing factors are listed in the
Resolution (Δx, Δy) 0.90 mm, 0.90 mm munity to the point where clinicians expect to following table (Table A2.1.7; Cure and Van
Number of data points collected (Nx, Ny) 256, 256r, with r = 3/4 rule in or out dural thrombosis using noninva- Tassel, 1994).
(rectangular field of view) sive techniques. MR is the modality of choice In involved veins, flow voids may be re-
Slice thickness (Δz) 5 mm in identifying and following treatment of dural placed by abnormal signal. The specific appear-
Number of slices 24 thrombosis (Bianchi et al., 1998). MR may ance of the abnormality varies according to the
Slice gap 1.5 mm identify both the intravascular thrombus and the age of the thrombus and the imaging parameters
Number of acquisition (Nacq) 1 concomitant parenchymal complications. Pa- used. The parenchymal findings vary with the
Swap read and phase encoding No tients may present with rapid or insidious onset stage of the process from normal to swelling
Slice location Equidistant between anterior and of symptoms. Some patients present with rapid with interstitial edema and subcortical hyper-
posterior brain onset of altered consciousness and focal neuro- intensity on proton density weighted and T2-
Saturation pulses No logical deficits. Others present with gradual weighted images, and in the more severe case
Slice series Interleaved confusion, headache, papilledema, and cranial will include the presence of hematoma (Tsai et
Scan time 2 min, 3 sec nerve palsies from increased intracranial pres- al., 1995).
sure. Complications include cerebral edema,
aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this Cerebral Venous Imaging Dural
sequence is 90°. Lesions Sinus Thrombosis

A2.1.5 A2.1.6
Current Protocols in Magnetic Resonance Imaging Supplement 9 Supplement 9 Current Protocols in Magnetic Resonance Imaging
 Table A2.1.7 Factors Predisposing to Cerebral Venous Thrombosis
Aseptic
Hematologic disorders
Sickle cell disease
Leukemia
Thrombocytopenia
Disseminated intravascular coagulopathy
Anticardiolipin syndrome
Deficiencies: protein S, protein C,
antithrombin III
Neoplasia
Direct compression
Invasion of dural sinus
Hypercoagulability
Dehydration
Infants
Pregnancy & puerperium
Trauma
Drugs
Oral contraceptives
L–asparaginase
Dural AVF
Inflammatory conditions
Ulcerative colitis
Behcet’s disease
Idiopathic
Critical Parameters
The authors routinely use time-of-flight for
MRV. Occasionally, phase-contrast techniques
are used if there is a question about direction
of flow or if there is a problem with shine-
through from thrombus (for more information
about phase contrast sequences, see the exam-
ple given in UNIT A1.1, Table A1.1.8). The diag-
nosis of venous sinus thrombosis is supported
by the absence of flow-related enhancement or
motion-related phase shift. Because of its T1
shortening, methemoglobin may exhibit shine-
through and mimic a patent sinus on time-of-
flight MRV. Findings must be correlated with
those of T1-weighted images. Phase-contrast
MRV is a background subtraction technique
and tends to eliminate signal from thrombus
and hematomas. Phase-contrast MRV could be
Current Protocols in Magnetic Resonance Imaging
a patent venous channel. The background sub- increased T1 signal related to thrombus with
traction technique of phase-contrast MRV may methemoglobin formation (Fig. A2.1.1).
Septic be helpful. This technique eliminates bright A 25-year-old pregnant female who has a
signal from stationary thrombus and hema- hypercoagulable state presented with superior
Encephalitis tomas (Fischbein et al., 2000). The following sagittal sinus thrombosis. There is total occlu-
Meningitis are some examples. sion of most of her superior sagittal sinus with
Mastoiditis An 18-year-old male with sickle cell anemia increased T1 signal (Fig. A2.1.2). This is less
Sinusitis has occlusion of his superior sagittal sinus with intense than on the previous examination sug-
Sepsis
Cellulitis
Empyema subdural or epidural
Figure A2.1.1 Superior sagittal thrombosis with bright T1 signal from
thrombus.
used when unable to resolve this problem with
time-of-flight MRV and conventional T1-
weighted images (Yuh et al., 1994).
Anticipated Results
The goal in studying cerebral dural sinuses
is to identify thrombosis and to allow the choice
of timely treatment options. Interpretation of
cerebral MRI examinations has potential for
both false-positive and false-negative errors.
Slow flow may simulate venous thrombosis on
spin echo images creating a false positive find-
ing. A false negative finding may be present
with time-of-flight MRV in patients who have
actual venous thrombosis. Here methemoglo-
bin is a natural breakdown product present in Figure A2.1.2 Superior sagittal sinus thrombosis with in-
clot. The T1-shortening effects of methemoglo- creased T1 signal. This is less bright than the previous exam-
Cerebral Venous Imaging Dural
bin can “shine-through” giving the illusion of Lesions Sinus Thrombosis ple.
A2.1.7 A2.1.8
Current Protocols in Magnetic Resonance Imaging
gesting that the clot is older than the one seen
on the previous example (Fig. A2.1.1).
A 34-year-old man with leukemia has supe-
rior sagittal sinus thrombosis with clot that is
almost isointense with brain (Fig. A2.1.3). The
clot is older than the two previous examples.
This is considerably less conspicuous than the
two previous examples. With less signal throm-
bus may be overlooked on conventional T1- contraceptives and kept this case from being
weighted images. classified as idiopathic.
A 28-year-old female has occlusion of her A 68-year-old male with thrombocytopenia
transverse sinus with increased T1 signal related has superior sinus thrombosis with partial
to methemoglobin within the thrombus (Fig. blockage of the sinus (Fig. A2.1.5). Some flow
A2.1.4). This much T1 signal has the potential is still demonstrated in the expected lumen of
to cause “shine-through” artifact on MRV. Her the sinus. The presence or absence of flow is
only known risk factor was the use of oral better appreciated on MRV than on other se-
quences.
A 40-year-old female with disseminated in-
travascular coagulopathy has total occlusion of
the superior sagittal sinus, the inferior sagittal
sinus, the straight sinus and parts of the trans-
verse sinuses (Fig. A2.1.6).
A 25-year-old male with mastoiditis had
thrombosis of the right transverse sinus (Fig.
A2.1.7 and Fig. A2.1.8). A conventional cranial
MRI had demonstrated mastoiditis with in-
creased T2 signal in the region of mastoid air
cells on the side of the thrombosis.

Figure A2.1.3 Superior sagittal sinus thrombosis with T1 signal,

which is almost isointense with brain.

Figure A2.1.5 Superior sagittal sinus thrombosis with partial blockage

on MRV.

Figure A2.1.4 Occlusion of transverse sinus with high T1

Figure A2.1.6 Superior sagittal sinus thrombosis with total occlu-
signal in clot. Cerebral Venous Imaging Dural sion on MRV.
Lesions Sinus Thrombosis

A2.1.9 A2.1.10
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Literature Cited
Bianchi, D., Maeder, P., Bogousslavsky, J., Schny-
der, P., and Meuli, R. 1998. Diagnosis of cerebral
venous thrombosis with routine magnetic reso-
nance: An update. Eur. Neurol. 40:179-190.
Cure, J.K. and Van Tassel, P. 1994. Congenital and
acquired abnormalities of the dural venous si-
nuses. Seminars US, CT, Magn. Reson. Imaging
15:520-539.
Fischbein, N.J., Dillon, W.P., and Barkovich, A.J.
2000. Teaching Atlas of Brain Imaging. pp. 248-
249, 329-332. Thieme, New York.
Shellock, F.G. and Kanal, E. 1996. Magnetic Reso-
nance Bioeffects, Safety and Patient Manage-
ment. Lippincott Williams and Wilkins, Phila-
delphia.
Tsai, F.Y., Wang, A., Matovich, V.B., Lavin, M.,
Berberian, B., Simonson, T.M., and Yuh, W.T.
1995. MR staging of acute dural sinus thrombo-
sis: Correlation with venous pressure measure-
ments and implications for treatment and prog-
nosis. Am. J. Neuroradiol. 16:1021-1029.
Yuh, W.T., Simonson, T.M., Wang, A.M., Koci,
T.M., Tali, E.T., Fisher, D.J., Simon, J.H., Jink-
ins, J.R., and Tsai, F. 1994. Venous sinus occlu-
sive disease: MR findings. Am. J. Neuroradiol.
15:309-316.
Figure A2.1.7 Occlusion transverse sinus on transverse
MRV.
Figure A2.1.8 Occlusion transverse sinus on coronal MRV.
Cerebral Venous Imaging Dural
Lesions Sinus Thrombosis
A2.1.11 A2.1.12
Current Protocols in Magnetic Resonance Imaging
Key References
Bianchi et al., 1998. See above.
These authors describe the importance of MRI and
MRV in the diagnosis and follow-up of cerebral
venous thrombosis. They point out how different
signals on spin echo T1- and T2-weighted images
evolve according to hemoglobin degradation caus-
ing pitfalls and artifacts that must be recognized.
Fischbein et al., 2000. See above.
These authors afford a current concise look at dural
sinus thrombosis. In addition to imaging findings,
they discuss, clinical findings, treatment, complica-
tions, pearls, and pitfalls.
Shellock and Kanal, 1996. See above.
Covers a number of important patient management
issues related to MR imaging, including recom-
mended safety procedures, a list of metallic implants
that have been tested for MR compatibility, and a
list of other sources on MR safety.
Contributed by F. Allan Midyett, Suresh
Mukherji, and Laurie Fisher
University of North Carolina at Chapel Hill
Chapel Hill, North Carolina
Current Protocols in Magnetic Resonance Imaging
Imaging Cavernous Malformations UNIT A2.2 Materials
Normal saline (0.9% NaCl)
Cavernous malformations (CMs) represent ∼10% to 15% of vascular malformations. Sterile extravascular contrast agent (e.g., Magnevist, Omniscan, Prohance)
Their incidence is ∼0.5% at autopsy. They have low pressure, slow flow, and tend to be Set up patient and equipment
angiographically occult (Tomlinson et al., 1994). They are commonly referred to as 1. Interview the patient to assess for contraindications such as cardiac pacemaker,
cavernous angiomas or cavernomas. CMs consist of enlarged sinusoidal vascular spaces implanted mechanical devices, and/or ferromagnetic materials. Also, determine if the
that have thin walls devoid of smooth muscle and normal endothelium. These form patient will need sedation medication, which will necessitate the need for appropriate
compact masses within central nervous system (CNS) parenchyma or associated struc- monitoring equipment.
tures without normal interspersed tissue. The thin walls lack normal endothelium and are
Each patient or legal guardian prior to bringing the patient into the exam area signs a
prone to leakage of blood elements. Deposition of blood products in and around CMs screening form.
may be extensive (Glendhill et al., 2000).
The presence of ferromagnetic materials may be a health hazard to the patient while inside
the magnet and/or adversely affect image quality. To determine the safety of scanning such
CAVERNOUS MALFORMATION BASIC ferromagnetic materials see Shellock (1996).
PROTOCOL
The following sequences comprise our basic protocol for evaluating cavernous malfor- The presence of ferromagnetic materials in the globe of the eye is contraindicated for MRI.
mations: Patients with prior metal exposure to the eye should have plain x-rays of the orbital area
to ensure that all metal has been removed prior to placing them in the magnetic field.
1. Three-plane positioning scout. 2. If the procedure is a research protocol, have the patient sign any necessary consent
2. Noncontrast T1-weighted transverse. form.
3. Transverse gradient echo.
4. Noncontrast coronal proton density (PD)/T2-weighted turbo spin echo. 3. Request the patient to change into a gown and remove all personal effects, such as
5. Post contrast T1-weighted transverse. jewelry, hearing aids, glasses, etc., prior to entering the MRI scan room. All personal
belongings should be secured during the examination.
And an optional:
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating
6. Post contrast T1-weighted transverse with flow compensation. and artifacts.
Although the actual scan time is 12 min and 19 sec (or 15 min and 46 sec with the optional 5. Explain the procedure to the patient and record relevant clinical history. Ensure that
sequence), because of setup time between sequences and the time necessary to inject the the patient understands what is expected and ask them if they have had questions
gadolinium, we tell the patient that this exam will take ∼30 min. Except for the three-plane answered appropriately.
positioning scout, if the patient just had a “routine” cranial MR exam, we perform only 6. Set up the exam room by securing the circularly polarized (CP)-head coil onto the
those sequences necessary to complete this protocol. Table A2.2.1 lists the necessary table and providing a clean exam table.
equipment needed to perform the imaging sequences in this unit. 7. Escort the patient to the MR examination room and position them according to the
NOTE: Be sure that technologists and nurses have immediate access to any emergency exam to be performed. Review the following items with the patient:
equipment that may be relevant to a given study, or that may be needed for a particular a. Provide earplugs or headphones to the patient to minimize the gradient noise but
patient, such as a crash cart or oxygen. assure them that they will still be able to hear you.
b. Provide the patient with a safety squeeze-bulb and demonstrate how it works.
Explain to the patient that they may use the squeeze-bulb if they need assistance
during the exam.
Table A2.2.1 Equipment Specifications Needed to Perform the
Following Imaging Sequences c. Explain to the patient that you will be talking to them between imaging sequences
(i.e., when the loud knocking noise stops).
Coil type Circularly polarized head coil d. Explain to the patient that it is imperative that they remain motionless during the
Field strength 1.5 T loud knocking noise to ensure good results. Also, explain that they should not
Gradient strength Minimum of 15 mT/m (or whatever the reposition their body between imaging sequences.
system permits)
Positioning cushion or Yes e The patient may call out at any time if he or she feels it necessary.
head stabilizers f. Provide the patient with an estimate of the examination length. If this entire
Knee cushion Yes protocol will be followed, we suggest that the exam will take approximately half
Use of contrast agents Yes an hour.
Pulse oximeter If patient requires sedation
8. Position a support under the patient’s knees to enhance comfort.
Imaging 9. If the patient is unable to hold still, provide appropriate sedation.
Cerebral Venous Cavernous
Lesions Malformations 10. Secure the patient’s head with positioning sponges or head stabilizers.
Contributed by F. Allan Midyett, Laurie Fisher, and Suresh Mukherji A2.2.1 A2.2.2
Current Protocols in Magnetic Resonance Imaging (2002) A2.2.1-A2.2.14
Copyright © 2002 by John Wiley & Sons, Inc. Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging
11. Use the laser light to center on the patient’s nasion. Table A2.2.3 Imaging Parameters for Sequence 2 (Transverse T1-Weighted Spin
Echo)
12. Advance the patient couch to isocenter of the magnet.
Once this step has been performed, so long as the patient does not move on the table, the Patient position Supine
table itself can be moved and then replaced in the same position as before without Scan type Spin echo
jeopardizing the positioning of one scan relative to another. Imaging plane (orientation) Transverse
Central slice or volume center Slices centered to midbrain on sagittal
Sequence 1: Rapid three-plane positioning scout
scout image
13. To validate the patient’s position and to have a reference to prescribe successive
Echo time (TE) 15 msec
imaging sequences, acquire a three-plane orthogonal scout sequence. See Table
Repeat time (TR) 560 msec
A2.2.2 for specific parameters.
Flip angle (FA) 90°
Most MR scanners can be programmed to acquire the scout automatically after coil tuning Fields of view (FOVx, FOVy) 230 mm, 230r mm, with r = 3/4
or after the patient has been placed in isocenter (for systems that do not require tuning). (rectangular field of view)
Siemens acquires the scout in three planes by selecting “various”. Resolution (Δx, Δy) 0.90 mm, 0.90 mm
Sequence 2: Noncontrast T1-weighted transverse Number of data points collected (Nx, Ny) 256, 256r, with r = 3/4 (rectangular
14. Run a noncontrast T1-weighted transverse spin echo sequence according to Table field of view)
A2.2.3. Slice thickness (Δz) 6 mm
Number of slices 20
On this study we will be looking for multiple round areas of high T1-weighted signal, which Slice gap 1.8 mm (−30%)
are usually seen in the central portion of these lesions. These deposits often demonstrate
Number of acquisition (Nacq) 2
a “popcorn” appearance that is characteristic of many of these lesions.
Read direction Anterior–posterior
Sequence 3: Transverse gradient echo Slice location Vertex to skull base
15. Display both the sagittal and transverse scout images in two separate quadrants on Saturation pulses None
the scan monitor. Change imaging parameters to those listed in Table A2.2.4. Position Slice series Interleaved
slices to the midline of the sagittal scout image. Adjust position of field of view (FOV) Scan time 3 min, 38 sec
off transverse scout image.
16. Instruct the patient to remain motionless as the scan will begin and last for ∼5 min. Table A2.2.4 Imaging Parameters for Sequence 3 (Transverse Gradient Echo)

17. Run a transverse gradient echo according to Table A2.2.4. Patient position Supine
Scan type 2-D Gradient echo
Imaging plane (orientation) Transverse
Table A2.2.2 Imaging Parameters for Sequence 1 (Scout Sequence) Central slice or volume center Slices centered to midline on sagittal
scout image
Patient position Supine Echo time (TE) 26 msec
Scan type Gradient echo Repeat time (TR) 750 msec
Imaging planes (orientation) Transverse, sagittal, and coronal Flip angle (FA) 30°
(various) Fields of view (FOVx, FOVy) 230 mm, 230r mm, with r = 3/4
Central slice or volume center Nasion (rectangular field of view)
Echo time (TE) 6 msec Resolution (Δx, Δy) 0.90 mm, 0.90 mm
Repeat time (TR) 15 msec Number of data points collected (Nx, Ny) 256, 256r, with r = 3/4 (rectangular
Flip angle (FA) 30° field of view)
Fields of view (FOVx, FOVy) 300 mm, 300 mm Slice thickness (Δz) 5 mm
Resolution (Δx, Δy) 1.17 mm, 2.34 mm Number of slices 18
Number of data points collected (Nx, Ny) 256, 128 Slice gap 1.5 mm (30%)
Slice thickness (Δz) 8 mm Number of acquisition (Nacq) 2
Number of slices 3 (one per orientation) Swap read and phase encoding No
Slice gap Not applicable Read direction Posterior–anterior
Number of acquisitions (Nacq) 1 Slice locations Vertex to base of skull
Swap read and phase encoding No Saturation pulses No
Slice location Not applicable Slice series Interleaved
Saturation pulses Not applicable Scan time 4 min, 50 sec
Slice series Ascending (caudal to cranial) Imaging
Cerebral Venous Cavernous
Scan time 7 sec Lesions Malformations

A2.2.3 A2.2.4
Current Protocols in Magnetic Resonance Imaging Supplement 9 Supplement 9 Current Protocols in Magnetic Resonance Imaging
 Table A2.2.5 Imaging Parameters for Sequence 4: Coronal Proton
Density/T2-Weighted Turbo Spin Echo Images

Patient position Supine

Scan type Turbo spin echo (dual)
Imaging plane (orientation) Coronal
Central slice or volume center Slices centered to midbrain on sagittal
scout image
Echo time (TE) 15 msec and 105 msec
Echo train length (ETL) 7
Repeat time (TR) 4500 msec
Flip angle (FA) 180°a
Fields of view (FOVx, FOVy) 230 mm, 230r mm, with r = 3/4
(rectangular field of view)
Resolution (Δx, Δy) 0.90 mm, 0.90 mm
Number of data points collected (Nx, Ny) 256, 256r, with r = 3/4 (rectangular
field of view)
Slice thickness (Δz) 5 mm
Number of slices 24
Slice gap 1.5 mm
Number of acquisition (Nacq) 1
Swap read and phase encoding No
Slice location Equidistant between anterior-posterior
brain
Saturation pulses No
Slice series Interleaved
Scan time 2 min, 3 sec
aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this sequence Figure A2.2.1 Coronal T2-weighted image.
is 90°.
Table A2.2.6 Imaging Parameters for Sequence 5 (Transverse T1-Weighted Spin
Echo)
Gradient echo sequences are the most sensitive sequences for cavernomas. This makes
them the preferred sequence for locating cavernomas in patients who tend to have multiple Patient position Supine
lesions. We will be looking for areas of “blooming” where the lesions appear larger and
Scan type Spin echo
more indistinct on the gradient echo and T2-weighted spin echo images than on the
T1-weighted spin echo images. Imaging plane (orientation) Transverse
Central slice or volume center Slices centered to midbrain on sagittal
Blooming refers to the increase in signal loss centered on the source that occurs as echo scout image
times get longer. For example, a small black dot from a microhemorrhage would grow or
Echo time (TE) 15 msec
“bloom” into a larger black region as the echo time increases.
Repeat time (TR) 560 msec
Sequence 4: Noncontrast coronal multiecho Flip angle (FA) 90°
18. Run a noncontrast coronal multiecho scan according to Table A2.2.5. Fields of view (FOVx, FOVy) 230 mm, 230r mm, with r = 3/4
(rectangular field of view)
This is a dual echo turbo spin echo sequence that acquires both proton density weighted
and T2-weighted images simultaneously. On this study, we will be looking for mixed high Resolution (Δx, Δy) 0.90 mm, 0.90 mm
signal within the center of the lesions, low signal around their periphery, and progressive Number of data points collected (Nx, Ny) 256, 256r, with r = 3/4 (rectangular
“blooming” with increased T2-weighting between the PD and T2-weighted images. The field of view)
proton density weighted images come as a bonus with the T2-weighted images without any Slice thickness (Δz) 6 mm
additional expenditure of time and may be useful in this situation. Number of slices 20
Figure A2.2.1 is the T2-weighted image from the noncontrast coronal multiecho scan and Slice gap 1.8 mm (−30%)
beautifully demonstrates the cavernous angioma in the left temporal region. Number of acquisition (Nacq) 2
Read direction Anterior–posterior
Sequence 5: Post contrast T1-weighted transverse Slice location Vertex to skull base
Although cavernomas may exhibit enhancement, we will be looking mainly for other Saturation pulses None
enhancement related lesions such as venous angiomas or hemorrhagic tumors. Some of Imaging Slice series Interleaved
these other lesions may not be appreciated on noncontrast examinations. Cerebral Venous Cavernous Scan time 3 min, 38 sec
Lesions Malformations

A2.2.5 A2.2.6
Current Protocols in Magnetic Resonance Imaging Supplement 9 Supplement 9 Current Protocols in Magnetic Resonance Imaging
 Table A2.2.7 Imaging Parameters for Sequence 6 (Transverse T1-Weighted Spin tients of Hispanic extraction more commonly the two previous examples. While an increasing
Echo with Flow Compensation) have multiple lesions. Some of these individu- number of these lesions are found incidentally
als have been linked to a genetic mutation on on routine imaging studies, others will still
Patient position Supine chromosome seven (Mason et al., 1988). present with seizures, hemorrhage, focal
Scan type Spin echo While cavernomas may occur anywhere in neurologic deficit, or headache between the
Imaging plane (orientation) Transverse the CNS, ∼75% are located in the cerebral hemi- ages of thirty and sixty.
Central slice or volume center Slices centered to midline on sagittal spheres, with 25% in the cerebellum or brain- Of those presenting with symptoms, ∼50%
scout image stem. When multiple, they may be seen in the have had seizures. Frequent occult hemorrhages
Echo time (TE) 17 msec spinal cord. Rarely they may be dural, subdural, make the frequency of intracranial hemorrhage
Repeat time (TR) 532 msec subarachnoid, or intraventricular. Figures difficult to access. While the incidence of sub-
Flip angle (FA) 90° A2.2.2, A2.2.3, and A2.2.4 demonstrate the clinical hemorrhage is nearly 100%, the rate of
Fields of view (FOVx, FOVy) 230 mm, 230r mm, with r = 3/4 classic appearance of a cavernoma in an unusual overt hemorrhage leading to clinical symptoms
(rectangular field of view) intraventricular location. is ∼0.5% to 2% per year. Approximately 10% to
Resolution (Δx, Δy) 0.90 mm, 0.90 mm A 66-year-old female presented with a his- 15% of patients with the disease will develop
tory of seizures (Fig. A2.2.2). The scan demon- clinically significant hemorrhage (Robinson et
Number of data points collected (Nx, Ny) 256, 256r, with r = 3/4 (rectangular
strates the classic appearance of a cavernoma in al., 1991). The clinical presentation of hemor-
field of view)
an unusual intraventricular location. rhage varies according to location with those in
Slice thickness (Δz) 5 mm
Figure A2.2.3 shows the T2-weighted image more vital areas such as the brainstem more
Number of slices 19
appearance of the same patient as in the previous likely to be symptomatic. Overt hemorrhage is
Slice gap 1.5 mm (−30%) example. This demonstrates the classic appear- more common in females and is prone to occur
Number of acquisitions (Nacq) 2 ance of high central signal surrounded by dark during pregnancy.
Read direction Anterior–posterior hemosiderin rim. Grossly the lesions appear multilobulated,
Slice locations Vertex to skull base Figure A2.2.4 shows the unusual intraven- like blue-black mulberries, and are more fre-
Flow compensation Yes tricular position of the cavernous angiomas on quent in frontal and temporal lobes. Microscopi-
Saturation pulses None T1-weighted imaging of the same patient as in cally, the surrounding brain parenchyma is
Slice series Interleaved
Scan time 3 min, 27 sec

19. Remove the patient from the scanner. Do not allow the patient to move on the table.
Establish an intravenous line from which the contrast agent can be injected, and attach
this line securely to the patient so that movement into or out of the magnet will not
pull at the patient’s arm. Move the patient back into the scanner.
It is preferable to insert the line prior to imaging and to leave the patient in the magnet,
with no intervening motion between the scans run before contrast agent injection and those
run after injection.
20. If the patient has moved, run another rapid three-plane positioning pilot scan (see
Sequence 1).
21. Leaving the patient in the magnet, inject the contrast agent, flush the line with 10 ml
saline, and then immediately run a post contrast T1-weighted transverse sequence
according to Table A2.2.6.
A dose of 0.1 mmol/kg of contrast agent is usually given.
Sequence 6: Post contrast T1-weighted with flow compensation (optional)
22. In atypical cases where the diagnosis is still in doubt, if indicated to accentuate flow
within a lesion, run a post contrast T1-weighted sequence with flow compensation
according to Table A2.2.7

COMMENTARY
Background Information in ∼50% of cases. A familial predilection is seen
Cavernomas are usually isolated findings but where up to 80% of affected family members
they may be multiple and familial. They are have multiple lesions. This is particularly seen Imaging
relatively rare in the general population, but in the Desert Southwest where there is a strong Cerebral Venous Cavernous Figure A2.2.2 Transverse T1-weighted spin echo image. The classic appearance of a cavernoma
when they occur there tend to be multiple lesions geographic preponderance for cavernomas. Pa- Lesions Malformations in an unusual intraventricular location.
A2.2.7 A2.2.8
Current Protocols in Magnetic Resonance Imaging Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging
 Figure A2.2.5 Noncontrast coronal multiecho T2-weighted image.

Figure A2.2.3 T2-weighted transverse image.

Imaging Figure A2.2.6 T2-weighted sagittal image.

Cerebral Venous Cavernous
Figure A2.2.4 T1-weighted sagittal image without gadolinium. Lesions Malformations

A2.2.9 A2.2.10
Current Protocols in Magnetic Resonance Imaging Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging
 Figure A2.2.9 T1-weighted transverse image with gadolinium.

Figure A2.2.7 T2-weighted transverse image.

Figure A2.2.10 T1-weighted sagittal image without gadolinium.

Imaging
Cerebral Venous Cavernous
Lesions Malformations
Figure A2.2.8 T1-weighted transverse image without gadolinium.
A2.2.11 A2.2.12
Current Protocols in Magnetic Resonance Imaging Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging

Figure A2.2.11 T2-weighted transverse image.
gliotic with hemosiderin staining (Fischbein et
al., 2000).
Critical Parameters and
Troubleshooting
Cavernous angiomas often demonstrate sus-
ceptibility artifacts that result from incorrect
frequency encoding. Heterogeneous suscepti-
bility causes a heterogeneous field. This causes
some of the proton’s frequency to be mapped
incorrectly and result in an inappropriate signal
void in the image. A characteristic white band
is often seen at the edge of this void. This is due
to overlapping signals of protons that are inap-
propriately mapped. Imagers are familiar with
this type of artifact around small bits of metal.
Normally we try to minimize susceptibility ar-
tifact by using the highest possible read gradient
(which often corresponds to the shortest possi-
ble TE in some systems) or by using a fast spin
echo technique with the shortest possible spac-
ing (Mitchell, 1999). In this case, we do the
Current Protocols in Magnetic Resonance Imaging
opposite, accentuating the artifact and making
the lesion more conspicuous. Susceptibility ar-
tifact is generally more pronounced with gra-
dient echo techniques. This artifact causes
“blooming” and can be seen to be progressively
more pronounced with gradient echo sequences
and longer TE’s of T2-weighted images.
Figure A2.2.5 is a close up of Figure A2.2.1
and demonstrates indistinctness of the periph-
eral margin of the hemosiderin ring caused by
“blooming.” This tends to accentuate the con-
spicuity of cavernomas with long TE T2-
weighted turbo spin echo and gradient echo
sequences.
This increased sensitivity for cavernomas
makes gradient echo sequence being the se-
quence of choice for finding additional lesions
that may not be appreciated with conventional
imaging techniques. Whereas lesions may be so
characteristic that no other lesions are consid-
ered, occasionally their appearance may be con-
fusing. Extra-axial lesions typically have high
signal on T2-weighted images, more often en-
hance after gadolinium, and may mimic men-
ingiomas. Intra-axially, classically appearing le-
sions may turn out to represent hemorrhagic
metastasis. In symptomatic patients surgical ex-
cision is recommended if feasible with the role
of radiosurgery being evaluated (Wilson, 1992).
Anticipated Results
These lesions usually have a characteristic
imaging appearance that is related to the pres-
ence of blood products at various evolutionary
stages (Rigamonti et al., 1987). The lesion is
usually round or lobulated with a mixed signal
intensity core.
Figure A2.2.6 shows the T2-weighted image
appearance in a 28-year-old male whose lesion
was found at a routine MRI examination. This
again demonstrates the classic appearance of
high central signal surrounded by a dark he-
mosiderin rim.
There are typically multiple round areas that
are bright on both T1-weighted and T2-weighted
images. These rounded deposits cause many of
these lesions to have a characteristic “popcorn”
appearance. A peripheral ferratin or he-
mosiderin rim varies in thickness and contour.
Sometimes the hemosiderin rim is so pro-
nounced that only a “black hole” is appreciated
on some of the image slices, as is seen on Figure
A2.2.7. The adjacent image demonstrated a
more classic appearance. The patient is a 38-
year-old Hispanic female who experienced
CNS symptoms each time when she went
through one of her pregnancies.
After injection of gadolinium, there may be
partial enhancement and an associated venous
malformation may be demonstrated.
The next noncontrast T1-weighted transverse
image (Fig. A2.2.8) demonstrates the cavernous
angiomas in the right temporal fossa.
The associated venous angioma was not ap-
preciated until the gadolinium enhanced T1-
weighted transverse image (Fig. A2.2.9) was
obtained.
Figure A2.2.10 shows the T1-weighted image
appearance of a 25-year-old female who pre-
sented with brainstem symptoms. This again
demonstrates the classic appearance in what is
not an unusual location.
Figure A2.2.11 shows the T2-weighted image
appearance of the previous 25-year-old female
who presented with brainstem symptoms. The
center of the lesion is high signal on both T1-
and T2-weighted images and there is a dark
Imaging
Cerebral Venous Cavernous
Lesions Malformations
A2.2.13 A2.2.14
Supplement 5 Supplement 5
hemosiderin rim around the periphery of the
lesion.
Literature Cited
Fischbein, N.J., Dillon, W.P., Barkovich, A.J. (eds.)
2000. Teaching Atlas of Brain Imaging, pp. 272-
275. Thieme, New York.
Glendhill, K., Moore, K.R., Jacobs, M., and Orrison,
W.W., Jr. (eds.) 2000. Chapter 21. In Neuroimag-
ing, pp. 252-755. W.B. Saunders, Philadelphia.
Mason, I., Aase, J.M., Orrison, W.W., Wicks, J.D.,
Seigel, R.S., and Bicknell, J.M. 1988. Familial
cavernous Angiomas of the brain in an Hispanic
family. Neurology 38:324-326.
Mitchell, D.G. 1999. MRI Principles, pp. 266-267.
W.B. Saunders, Philadelphia.
Rigamonti, D., Drayer, B.P., Johnson, P.C., Hadley,
M.N., Zabramski, J., and Spetzler, R.F. 1987.
The MRI appearance of cavernous malforma-
tions (Angiomas). J. Neurosurgery 67:518-524.
Robinson, J.R., Awad, I.A., Little, J.R. 1991 Natural
history of the cavernous angioma. J. Neurosur-
gery 75:709-714.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Tomlinson, F.H., Houser, O.W., Scheithauer, B.W.,
Sundt, T.M., Jr., Okozaki, H., and Parisa, J.E.
1994. Angiographically occult vascular malfor-
mations: A correlative study of features on mag-
netic resonance imaging and histological exami-
nation. Neurosurgery 34:792-800.
Wilson, C.B. 1992 Cryptic vascular malformations.
Clinics Neurosurg. 38:49-84.
Key References
Fischbein, 2000. See above.
These authors afford a current concise look at in-
tracranial vascular malformations. In addition to
imaging findings, they discuss etiology, clinical find-
ings, treatment, complications, pearls and pitfalls.
Shellock, 1996. See above.
Covers a number of important patient management
issues related to MR imaging, including recom-
mended safety procedures, a list of metallic implants
that have been tested for MR compatibility, and a
list of other sources on MR safety.
Contributed by F. Allan Midyett
UA North Texas Healthcare
Dallas, Texas
Laurie Fisher
Siemens Uptime Service Center
Cary, North Carolina
Suresh Mukherji
University of Michigan
Ann Arbor, Michigan
Current Protocols in Magnetic Resonance Imaging
Imaging Venous Angiomas UNIT A2.3

Venous angiomas or venous malformations are probably not true malformations. They
appear to be variants or developmental anomalies that provide anomalous venous drainage
of normal brain parenchyma. (Glendhill et al., 2000). Their incidence is ∼3% at autopsy.
They are frequent incidental findings on post-contrast studies of the brain.

VENOUS ANGIOMAS BASIC

PROTOCOL
The following sequences are useful for evaluating venous angiomas:
1. Three-plane positioning scout.
2. Transverse T1-weighted post-gadolinium with flow compensation.
3. Coronal T1-weighted post-gadolinium coronal with flow compensation.
4. Sagittal post-T1-weighted gadolinium with flow compensation.
5. MRV (magnetic resonance venography).

Although the actual scan time is 15 min and 46 sec, because of setup and injection time,
we tell the patient that the examination will take approximately one-half hour. Except for
the three-plane positioning scout, if the patient has recently had a routine cranial MR
examination, we perform only these sequences necessary to complete this protocol. (As
an example, a routine cranial MR consists of sequences 1 to 5 in UNIT A5.1, without using
contrast agents.) It is actually more common for us to find venous angiomas incidentally
Figure A2.3.1 A transverse T2–weighted image demonstrates abnormal “flow void” through the
on contrasted MRI examinations than to have found them on CT (computed tomography), level of the pons just posterior to the smaller but normal flow void in the basilar artery in the
angiography, or previous MRI, and to look for them prospectively. In the less common prepontine cistern.
situation where the patient has not had a routine cranial MR, we add T2-weighted
sequences (without contrast agents) to provide a more complete examination. For exam-
ple, one can run a T2-weighted sequence whose imaging parameters are listed in Table
A2.2.5, with imaging plane set to transverse and with a possible change of number of
slices or slice thickness in order to cover the entire brain. Figures A2.3.1 and A2.3.2 are
example images.
Table A2.3.1 lists the hardware necessary to perform the procedure, along with appropri-
ate parameters. The available gradient strength will depend on the scanner, and the echo
times given in other tables below may be varied accordingly (the smaller the gradient
strength, the longer the echo time for a particular scan).
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as a crash cart or oxygen.
Materials
Normal saline (0.9% NaCl)
Sterile extravascular contrast agent (e.g., Magnevist, Ominiscan, Prohance)

Set up patient and equipment

1. Interview (screen) the patient to assess for contraindications such as cardiac pace-
maker, implanted mechanical devices, and/or ferromagnetic materials. Also, deter-
mine if the patient will need sedation, necessitating the need for appropriate
monitoring equipment. Figure A2.3.2 Close-up of the transverse T2–weighted image through the pons at lower level
(caudal) to the image shown in Figure A2.3.1. This shows continuation of the same abnormal “flow
Each patient or legal guardian must sign a screening form (see, e.g., APPENDIX 1) before
void” seen on the image of the same patient as shown Figure A2.3.1.
the patient is brought into the exam area.
Cerebral Venous Imaging Venous
Lesions Angiomas
Contributed by F. Allan Midyett, Suresh Mukherji, and Laurie Fisher A2.3.1 A2.3.2
Current Protocols in Magnetic Resonance Imaging (2003) A2.3.1-A2.3.12
Copyright © 2003 by John Wiley & Sons, Inc. Supplement 9 Supplement 9 Current Protocols in Magnetic Resonance Imaging
 Table A2.3.1 Equipment Specifications Needed to Perform the
Following Imaging Sequences
Coil type Circularly polarized head coil
Field strength 1.5 T
Gradient strength Minimum of 15 mT/m (or whatever the
system permits)
Positioning cushion or Yes
head stabilizers
Knee cushion Yes
Pulse oximeter If patient requires sedation
Use of contrast agents Yes
The presence of ferromagnetic materials may be a health hazard to the patient while in the
magnetic field and/or adversely affect image quality. To determine the safety of scanning
such ferromagnetic materials see Shellock (2001).
The presence of ferromagnetic materials in the globe of the eye is contraindicated for MRI.
Patients with prior metal exposure to the eye should have plain X rays of the orbital area
to ensure that all metal has been removed, prior to placing them in the magnetic field.
2. If the procedure is a research protocol, have the patient sign any necessary consent
form.
3. Request that the patient change into a gown and remove all personal items such as,
jewelry, hearing aids, glasses, any metal, etc., prior to entering the MRI scan room.
All personal belongings should be secured during the examination.
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating
and image artifacts.
5. Explain the procedure to the patient and record relevant clinical history. Ensure that
the patient understands what is expected and ask him or her if he or she has any
questions; answer appropriately.
6. Set up the exam room by providing a clean exam (scanning) table and securing the
circularly polarized (CP) head coil to the table.
7. Escort the patient to the MR examination room and ask him or her to lie down on the
table accordingly with respect to the exam to be performed. Review the following
items with the patient:
a. Provide earplugs or headphones to the patient to minimize the loud knocking noise
that will be produced by the gradients but ensure him or her that he or she will
still be able to hear you.
b. Provide the patient with a safety squeeze-bulb and demonstrate how it works;
explain to the patient when to use the squeeze-bulb (i.e., if they need assistance
during the exam).
c. Explain to the patient that you will be talking to him or her between imaging
sequences when the loud knocking noise stops.
d. Explain to the patient that it is imperative that he or she remain motionless during
the loud knocking noise to ensure good results; also explain that he or she should
not reposition his or her body (especially head) between imaging sequences.
e. Provide the patient with an approximate time that the examination will take. If this
entire protocol will be followed, we suggest that the exam will take approximately
half an hour.
f. Nevertheless, the patient may call out at any time if he or she feels it necessary.
Current Protocols in Magnetic Resonance Imaging
8. Either before or right after the patient lies down, set up any triggering devices or other
monitoring equipment to be used.
9. Establish an intravenous line from which the contrast agent can be injected, and attach
this line securely to the patient so that movement into or out of the magnet will not
pull at the patient’s arm.
It is preferable to insert the line prior to imaging and to leave the patient in the magnet, so
that there is no intervening motion between the scans run before contrast agent injection
and those run after injection.
10. Center the patient in the head coil, at the region where the key information is desired.
Make sure that the head is constrained to prevent motion, especially if high-resolution
scans are to be run. Secure the patient’s head with positioning sponges or head
stabilizers.
Generally, the patient’s head is fixed so that it is horizontal (not tilted) and lies along the
axis of the patient table; other positions may be appropriate depending on the needs at
hand.
11. If needed, place a pillow or other support under the knees to make the patient more
comfortable.
12. Use the centering light to position the patient’s nasion and advance the patient into
the center of the magnet.
Once this step has been performed, so long as the patient does not move on the table, the
table itself can be moved and then returned to the same position as before without
jeopardizing the positioning of one scan relative to another.
13. If the patient is unable to hold still, provide appropriate sedation.
Sequence 1: Three–plane positioning scout
14. To validate the patient’s position and to have a reference to prescribe successive
imaging sequences, acquire a three-plane orthogonal scout sequence. See Table
A2.3.2 for specific parameters.
Most MR scanners can be programmed to acquire the scout automatically after coil tuning
or after the patient has been placed in isocenter (for systems that do not require tuning).
This runs in 7 sec and is used to position the remainder of the sequences. It is particularly
useful to correct off-axis positioning in the coronal plane.
Sequence 2: Transverse post-gadolinium with flow compensation
15. Display both the sagittal and transverse scout images in two separate quadrants on
the scan monitor. Change imaging parameters to those listed in Table A2.3.3. Position
slices to midline of the sagittal scout image. Adjust position of field of view off
transverse scout image.
16. Leaving the patient in the magnet, inject the contrast agent either by hand or using a
mechanical injector. Observe the injection to insure there is no extravasation of the
contrast agent. Flush the line with 10 ml of sterile normal saline. Begin the scan as
soon as the injection is completed.
Alternatively, the scanning table may be advanced out of the magnet for the injection, but
the patient must remain in place.
A dose of 0.1 mmol/kg of contrast agent is usually given.
17. Instruct the patient to remain motionless, as the scan will begin and last for ∼5 min.
Cerebral Venous Imaging Venous
Lesions Angiomas
A2.3.3 A2.3.4
Supplement 9 Supplement 9 Current Protocols in Magnetic Resonance Imaging
 Table A2.3.2 Imaging Parameters for Sequence 1 (Scout Sequence)

Patient position Supine

Scan type Gradient echo
Imaging plane (orientation) Transverse, sagittal, and coronal
Central slice or volume center Nasion
Echo time (TE) 6 msec
Repeat time (TR) 15 msec
Flip angle (FA) 30°
Fields of view (FOVx, FOVy) 300 mm, 300mm
Resolution (Δx, Δy) 1.17 mm, 2.34 mm
Number of data points collected (Nx, Ny) 256, 128
Slice thickness (Δz) 8 mm
Number of slices 3 (one per orientation)
Slice gap Not applicable
Number of acquisition (Nacq) 1
Swap read and phase encoding No
Slice location Not applicable
Saturation pulses Not applicable
Scan time 7 sec

Table A2.3.3 Imaging Parameters for Sequence 2 (Transverse T1-Weighted Spin

Echo with Flow Compensation) Figure A2.3.3 A post-contrast transverse T1–weighted image demonstrates typical venous
angioma of the right cerebellar hemisphere.
Patient position Supine
Scan type Spin echo Table A2.3.4 Imaging Parameters for Sequence 3 (Coronal T1-Weighted Spin
Imaging plane (orientation) Transverse Echo with Flow Compensation)
Central slice or volume center Slices centered to midline on sagittal
scout Patient position Supine
Echo time (TE) 17 msec Scan type Spin echo
Repeat time (TR) 532 msec Imaging plane (orientation) Coronal
Flip angle (FA) 90° Central slice or volume center Slices centered to midline on sagittal
Fields of view (FOVx, FOVy) 230 mm, 230r mm, with r = 3/4 scout
(rectangular field of view) Echo time (TE) 17 msec
Resolution (Δx, Δy) 0.90 mm, 0.90 mm Repeat time (TR) 532 msec
Number of data points collected (Nx, Ny) 256, 256r, with r =3/4 (rectangular Flip angle (FA) 90°
field of view) Fields of view (FOVx, FOVy) 230 mm, 230r mm, with r = 3/4
Slice thickness (Δz) 5 mm (rectangular field of view)
Number of slices 19 Resolution (Δx, Δy) 0.90 mm, 0.90 mm
Slice gap 1.5 mm (–30%) Number of data points collected (Nx, Ny) 256, 256r, with r =3/4 (rectangular
Number of acquisitions (Nacq) 2 field of view)
Read direction Anterior–posterior Slice thickness (Δz) 5 mm
Slice locations Vertex to skull base Number of slices 19
Flow compensation Yes Slice gap 1.5 mm (–30%)
Saturation pulses None Number of acquisitions (Nacq) 2
Slice series Interleaved Read direction Right–left
Scan time 3 min, 26 sec Slice locations Equidistant between anterior–posterior
brain
Flow compensation Yes
Saturation pulses None
Slice series Interleaved
Scan time 3 min, 26 sec
Cerebral Venous Imaging Venous
Lesions Angiomas

A2.3.5 A2.3.6
Current Protocols in Magnetic Resonance Imaging Supplement 9 Supplement 9 Current Protocols in Magnetic Resonance Imaging
 Table A2.3.5 Imaging Parameters for Sequence 4 (Sagittal T1-Weighted Spin
Echo with Flow Compensation)

Patient position Supine

Scan type Spin echo
Imaging plane (orientation) Sagittal
Central slice or volume center Slices centered to midline on
transverse scout
Echo time (TE) 17 msec
Repeat time (TR) 633 msec
Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 210 mm, 210r mm, with r = 3/4
(rectangular field of view)
Resolution (Δx, Δy) 0.82 mm, 0.82 mm
Number of data points collected (Nx, Ny) 256, 256r, with r = 3/4 (rectangular
field of view)
Slice thickness (Δz) 5 mm
Number of slices 25
Slice gap 1.5 mm (–30%)
Number of acquisitions (Nacq) 1
Read direction Anterior–posterior
Slice locations Equidistant between left and right brain
Flow compensation Yes
Saturation pulses None Figure A2.3.5 Close-up lateral view of the venous phase from conventional angiogram demon-
Slice series Interleaved strates a large venous angioma in a transpontine location. This contrast-enhanced vascular
Scan time 2 min, 4 sec structure corresponds to the abnormality seen on the previous three images (Figures A2.3.1,
A2.3.2, and A2.3.4).

Figure A2.3.6 A sagittal T1–weighted image with gadolinium shows venous angioma extending
in periventricular area from the anterior portion of the lateral ventricle posteriorly into the lateral
ventricle. From same patient as image shown in Figure A2.3.7; look back at the previous image to
see some of the flow voids now filled with a contrast agent.

Figure A2.3.4 Close-up of a sagittal T1–weighted image shows large abnormal area of low signal
extending through the pons posteriorly and inferiorly toward the fourth ventricle. This corresponds
to the abnormality (vascular flow void) seen on images in Figure A2.3.1. and Figure A2.3.2. Cerebral Venous Imaging Venous
Lesions Angiomas

A2.3.7 A2.3.8
Current Protocols in Magnetic Resonance Imaging Supplement 9 Supplement 9 Current Protocols in Magnetic Resonance Imaging
 Table A2.3.6 Imaging Parameters for Sequence 5 Time-of-Flight MRV of the
Sagittal Sinus

Patient position Supine

Scan type 2-D gradient echo
Imaging plane Sagittal
Central slice or volume center Slices centered to midline on
transverse scout
Echo time (TE) 9 msec
Repeat time (TR) 30 msec
Flip angle (FA) 50°
Field of view (FOVx, FOVy) 210 mm, 210 mm
Resolution (Δx, Δy) 0.82 mm, 0.82 mm
Number of data collected (Nx, Ny) 256, 256
Slice thickness (Δz) 3 mm
Number of slices 64
Slice gap 0.34 mm
Number of acquisitions (Nacq) 1
Swap read and phase encoding No
Figure A2.3.7 A sagittal T1–weighted image shows abnormal flow voids paralleling superior Slice locations Angle sagittal to coronal, –12° (to
aspect of the ventricle. It may be easy to overlook venous angiomas on T1-weighted images without reduce in-plane saturation) off
gadolinium. transverse scout image
Saturation pulses Yes (one angled through facial sinus
area off sagittal scout image, 60-mm
thick)
Scan time 6 min, 43 seca
aIf the number of slices is set to be 1, the scan time is 7.8 sec.

18. Run sequence 2 according to Table A2.3.3.

An example image is shown in Figure A2.3.3.
This is the basic post-contrast, flow-compensated sequence which will be performed in
transverse, coronal, and sagittal planes. If a venous angioma is present, these three
sequences will probably define it with the least artifact.

Sequence 3: Coronal post-contrast T1-weighted with flow compensation

19. Run sequence 3 according to Table A2.3.4.

Sequence 4: Sagittal post-contrast T1-weighted with flow compensation

20. Position slices to midline of the transverse scout image. Run sequence 4 according
to Table A2.3.5.
Some example images are shown in Figure A2.3.4, Figure A2.3.5, Figure A2.3.6, Figure
A2.3.7, and Figure A2.3.8.
Figure A2.3.8 Coned lateral view of the venous phase from conventional cerebral angiogram.
This demonstrates a typical venous angioma with a “carrot’s top” appearance and mirrors the The term used by Siemens for flow compensation is gradient motion rephasing (GMR).
appearance seen on a sagittal MRI or MRV. Multiple radicles converge on an anomalous dilated This technique is very helpful in reducing flow-related artifact that may hinder evaluation
vein which in turn drains into dilated central venous structures. of venous structures, particularly in the posterior fossa. These sequences also accentuate
the visibility of vessels, especially where there is rapidly flowing blood. This is useful when
the vein is narrow or where there is a thin rim of flow around the periphery of a thrombus.

Sequence 5: Time-of-flight MRV of the sagittal sinus

21. Run sequence 5 according to Table A2.3.6.
This sequence offers the possibility of demonstrating a venous angioma and showing its
Cerebral Venous Imaging Venous relationship to other major venous structures.
Lesions Angiomas

A2.3.9 A2.3.10
Current Protocols in Magnetic Resonance Imaging Supplement 9 Supplement 9 Current Protocols in Magnetic Resonance Imaging
COMMENTARY
Background Information
Venous angiomas are most often located
adjacent to the frontal horns of the lateral ven-
tricles or within the cerebellum. Most com-
monly small, their size ranges from micro-
scopic to gigantic with involvement of an entire
hemisphere. Unless large, they may be subtle
on conventional angiography. These are usually
more conspicuous on contrast-enhanced CT or
MR images.
Venous angiomas are associated with focal
neuronal abnormalities. The association with
cavernous malformation is strong, and ∼7% of
venous angiomas have associated cavernous
malformations. Approximately one-third of
cavernous angiomas are associated with venous
angiomas.
Venous angiomas have no arterial compo-
nents and are theorized to arise from an arrest
of venous development after arterial develop-
ment is complete (Dillon, 1997). This could
cause retention of primitive medullary veins
that are structurally intact but anomalous in
location. Because they are physiologically nor-
mal, they are rarely associated with significant
problems.
Venous angiomas are commonly found on
contrast-enhanced cerebral examinations.
While these patients commonly have headache,
a causal relationship is unclear. Intracranial
hemorrhage has been occasionally associated
with venous angiomas.
Occasionally, the abnormal venous channel
may be compromised by an adjacent structure,
resulting in venous hypertension and parenchy-
mal ischemia. Seizures are probably related to
chronic ischemia. These changes can result in
encephalomalacia and/or calcification. Patients
with cerebellar cavernous angiomas have expe-
rienced ataxia, diplopia, and dizziness without
demonstrable hemorrhage.
Grossly, there is a tuft of enlarged radially
arranged anomalous medullary veins converg-
ing on a central draining vein which empties
into a subependymal vein, cortical vein, or
dural sinus. Microscopically, normal interven-
ing brain parenchyma separates venous radi-
cles, with venous walls being slightly thickened
and hyalinized.
Critical Parameters and
Troubleshooting
Use of intravenous contrast agent (gadolin-
ium-DTPA) is essential when evaluating pa-
Current Protocols in Magnetic Resonance Imaging
Literature Cited Truwit, C.L. 1992. Venous angioma of the brain:
Dillon, W.P. 1997. Cryptic vascular malformations: History, significance, and imaging findings. Am.
tients for venous angiomas. Lesions of signifi- Controversies in terminology, diagnosis, patho- J. Roentgenol. 159:1299-1307.
cant size may be overlooked if contrast-en- physiology, and treatment. Am. J. Neuro-
hanced images are not obtained. Contrast-en- roentgenol. 18:1839-1846. Key References
hanced images should be obtained in patients Fischbein, N.J., Dillon, W.P., and Barkovich, A.J. Dillon, 1997. See above.
who have cavernous angiomas to demonstrate 2000. Teaching Atlas of Brain Imaging, pp. 282- The author presents a broad overview of intracra-
associated venous angiomas. Post-contrast im- 285. Thieme Medical Publishers, New York. nial vascular malformations with regard to termi-
ages should be performed in three planes (sagit- Glendhill, K., Moore, K.R., Jacobs, M., and Orrison, nology, diagnosis, pathophysiology and treatment.
W.W., Jr. 2000. Chapter 21. In Neuroimaging Like opening Pandora’s box we find things are more
tal, coronal, and transverse). complicated than we had first thought.
(W.W. Orrison, Jr., ed.) pp.750-752. W.B. Saun-
ders, Philadelphia.
Imaging with MRV
Rigamonti, D., Spetzler, R.F., Medina, M., et al.
MRV may be helpful in evaluation of venous 1998. Cerebral venous malformations. J. Neuro- Contributed by F. Allan Midyett
malformations. It is particularly important to surg. 73:560-564. VA North Texas Health Care System
identify a venous malformation when associ- Rothfus, W.E., Albright, A.L, Caey, K.F., et al. 1984. Dallas, Texas
ated with a cavernous malformation. The ve- Cerebellar venous angioma: “Benign” entity?
nous malformation drains normal brain paren- Am. J. Neuroroentgenol. 5:61-66. Suresh Mukherji
chyma and needs to be avoided by the neuro- Shellock, F.G. 2001. Pocket Guide to MR Proce- University of Michigan
surgeon resecting the adjacent cavernous lesion dures and Metallic Objects. Lippincott-Raven, Ann Arbor, Michigan
(Rigamonti et al., 1998). Philadelphia.
Laurie Fisher
Siemens Uptime Service Center
Anticipated Results Cary, North Carolina
On pre-contrast images, flow voids and mis-
registration artifacts may be seen. Many venous
angiomas may only be detected after injection
of a contrast agent when the characteristic “Me-
dusa head” appearance of venous radicles con-
verging on a larger draining vein is seen (Tru-
wit, 1992). When venous hypertension causes
ischemia, T2-weighted signal may be increased
in adjacent parenchyma (Fischbein et al.,
2000). The risk of hemorrhage rises with in-
creasing venous pressure secondary to occlu-
sion.
Transitional morphology may be present
with lesions exhibiting arteriovenous shunting
at the precapillary level.
The prognosis is generally excellent with no
treatment required, except when associated
with cavernous angioma or when there is some
type of venous stenosis. In these cases, there is
an increased rate of hemorrhage. Although a
minority of lesions bleed, the rate of hemor-
rhage may be more frequent in the cerebellum
(Rothfus et al., 1984).
Time Considerations
Although the actual scan time is 15 min and
46 sec, because of setup and injection time, we
tell the patient that the examination will take
approximately one-half hour.
Acknowledgements
The authors would like to thank the assis-
tance from Shibu Mathew of Dallas VA Medi-
cal Center.
Cerebral Venous Imaging Venous
Lesions Angiomas
A2.3.11 A2.3.12
Supplement 9 Supplement 9 Current Protocols in Magnetic Resonance Imaging

CHAPTER A3
Cerebral Neoplastic Disease
INTRODUCTION
agnetic resonance imaging (MRI) is virtually always the imaging study of choice
M in the detection and localization of intracerebral tumors. The unique capability of
imaging in multiple planes provides superior localization and definition of intracranial
tumors. Conventional MR sequences coupled with intravenous contrast remains the
optimal technique for the detection and characterization of the majority of intracranial
neoplasms. The role of newer MR techniques (proton MR spectroscopy, diffusion and
perfusion imaging) in the work-up of intracerebral neoplasms is yet to be defined.
This chapter is organized into units based on imaging intracranial tumors by location.
Intra-axial tumors are lesions found within the brain parenchyma and may be metastatic
(UNIT A3.1) or primary (UNIT A3.3). Extra-axial (or extra-cerebral) tumors may be metastatic
to the meninges, calvarium, or skull base (UNIT A3.2), or lesions may primarily arise from
these structures (UNIT A3.4). UNIT A3.5 presents the preferred set of MR sequences for imaging
the brain following therapy for brain tumors. The preferred protocols for evaluating
primary and metastatic brain tumors are discussed with specific modifications where
necessary.
Annette O. Nusbaum and Scott W. Atlas
Cerebral
Neoplastic Disease
Contributed by Annette O. Nusbaum and Scott W. Atlas A3.0.1
Current Protocols in Magnetic Resonance Imaging (2002) A3.0.1
Copyright © 2002 by John Wiley & Sons, Inc. Supplement 5
Metastatic Intra-Axial Neoplasia UNIT A3.1
Magnetic resonance imaging (MRI) is a sensitive noninvasive means by which to evaluate
intracranial pathology. The role of imaging in the metastatic work-up is to detect the
spread of tumor to the brain parenchyma, and define the location. Intravenous contrast
(gadolinium-DTPA) provides the greatest sensitivity for detecting brain lesions (Healy et
al., 1987) and is almost always indicated except when there is no intravenous access. This
unit presents the set of MR sequences used for imaging intra-axial brain metastases (see
Basic Protocol) and specific modifications will be discussed where necessary (see
Alternate Protocol). The sequences described in this unit are based on the authors’
experience with a 1.5 T scanner (Echospeed GE Medical Systems, Milwaukee, Wiscon-
sin), but can be expected to be equally applicable to other field strengths and scanners
from other manufacturers.
RULE OUT (R/O) NON-HEMORRHAGIC METASTATIC DISEASE BASIC
PROTOCOL
The vast majority of metastases are located in the intra-axial supra-tentorial compartment,
are typically multiple and subcortical, and have a considerable amount of surrounding
vasogenic edema (Atlas and Lavi, 1996). In general, T1-weighted and T2-weighted images
provide tissue-specific information by elucidating the water content, hemorrhagic com-
ponents, and cellularity of lesions. Contrast enhancement of intra-axial brain lesions
indicates breakdown of the blood-brain barrier. Imaging following triple-dose MRI
contrast agent administration detects a greater number of small (<10 mm) metastatic
lesions than the standard dose (0.10 mmol/kg) (Yuh et al., 1995). Delayed imaging (10
to 30 min) after the administration of either the standard dose or high dose contrast
improves the detection of small metastatic lesions. In most institutions, patients receive
the standard dose and are scanned without any significant delay in the transverse and
coronal planes. Beyond generating a second plane of imaging, the coronal plane acquisi-
tion following the transverse sequence allows for a small delay (several minutes) follow-
ing intravenous contrast administration. In pediatric patients and other subjects requiring
sedation, the authors feel that a third post-contrast acquisition, in the sagittal plane, is
worthwhile in evaluating brain tumor patients. Sequences 1 to 5 comprise the preferred
protocol.
Table A3.1.1 lists the hardware necessary to perform the procedure, along with appropri-
ate parameters.
NOTE: Be sure that technologists and nurses have immediate access to any equipment
such as crash carts or oxygen that may be necessary in the event of an emergency.
Reactions to contrast agents are rare, but the resources are necessary.
Table A3.1.1 Equipment Parameters for Imaging Brain Tumors
Coil type Quadrature head coil
Gradient coil strength 25 mT/m (or whatever the system permits)
Cardiac gating No
Peripheral gating For safety only
Respiratory gating No
Respirator If required by patient
Oxygen If required by patient
Motion cushions No
Cerebral
Neoplastic Disease
Contributed by Annette O. Nusbaum and Scott W. Atlas A3.1.1
Current Protocols in Magnetic Resonance Imaging (2001) A3.1.1-A3.1.7
Copyright © 2001 by John Wiley & Sons, Inc.
 Materials
Normal saline (0.9% NaCl), sterile
Intravenous contrast agent (e.g., Magnevist, Omniscan, or Prohance)
Set up patient and equipment
1. Interview (screen) the patient to ensure that he or she has no contraindications such
as cardiac pacemakers or other implants containing ferromagnetic materials. Also be
sure to find out if the patient has any health conditions that may require the presence
of special emergency equipment during the scanning procedure, or necessitate any
other precautions.
Generally, standard screening forms are used for all patients scanned in a magnetic
resonance system.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any implants; see Shellock (1996)
for discussion of what implants may be safely scanned using magnetic resonance.
8. If needed, place a pillow or other support under the knees to make the patient more
comfortable.
9. Use the centering light to position the patient (centered on the nasion) and put him
or her into the center of the magnet.
Once this step has been performed, so long as the patient does not move on the table, the
table itself can be moved and then replaced in the same position as before without
jeopardizing the positioning of one scan relative to another.
10. If the patient is unable to hold still, provide an appropriate sedative.
Sequence 1: Localizer
11. Run sequence 1 according to Table A3.1.2. The sagittal scout view is used to prescribe
the transverse or coronal planes.
Sequence 2: T1-weighted scan
12. Run sequence 2 according to Table A3.1.3.

Patients may be accompanied into the magnet room by a friend or family member, who can Sequence 3: T2-weighted scan
sit in the room during the scan and comfort the patient as needed. This companion must 13. Run sequence 3 according to Table A3.1.4.
be screened as well to ensure the absence of loose metal objects on the body or clothing.
2. If the procedure is a research protocol, have the patient sign any necessary consent Sequence 4: Fast FLAIR (fluid-attenuated inversion recovery) scan
form. 14. Run sequence 4 according to Table A3.1.5.

3. Have the patient remove all jewelry and change into a gown to eliminate any metal Sequence 5: Post-contrast imaging
that might be found in clothing. 15. Remove the patient from the scanner. The patient should not move on the table.
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating Establish an intravenous line from which the contrast agent can be injected, and attach
and image artifacts. this line securely to the patient so that movement into or out of the magnet will not
pull at the patient’s arm. Move the patient back into the scanner.
5. Inform the patient about what will occur during the procedure, what he or she will
experience in the magnet, and how to behave, including the following:
Table A3.1.2 Primary Clinical Imaging Parameters for Sequence 1
a. If earphones or headphones are used to protect the ears from the loud sounds (T1-Weighted)
produced by the gradients, the patient will be asked to wear these, but will be able
to communicate with you at any time during the imaging. Patient position Supine
b. The patient will be given a safety squeeze-bulb or similar equipment to request Scan type Spin echo
assistance at any time (demonstrate how this works). Imaging plane (orientation) Sagittal
c. For good results the patient should not talk, and should avoid or minimize Central slice or volume center Laser light centered on nasion
swallowing or other movement, during each scan—i.e., as long as the banging Echo time (TE) 11 msec (or select “minimum
full” echo time)
sounds continue. Between scans, talking and swallowing are allowed in most
Receiver bandwidth (RBW) 10 kHz
cases, but should be avoided when comparative positional studies are being
Repeat time (TR) 500 msec
performed; the patient will be informed when this is the case.
Flip angle (FA) 90°
d. Nevertheless, the patient may call out at any time if he or she feels it necessary. Fields of view (FOVx, FOVy) 240 mm, 240 mm
Resolution (Δx, Δy) 0.94 mm, 1.25 mm
6. Have the patient mount onto the table in the supine position. Either before or right
Number of data points collected (Nx, Ny) 256, 192
after the patient lies down, set up any triggering devices or other monitoring
Slice thickness (Δz) 5 mm
equipment that is to be used.
Number of slices 20 or as many as needed to cover
7. Center the patient in the head coil at the region where the key information is desired. the region of interest
Make sure the head and neck are constrained to prevent motion. Slice gap 2 mm
Number of acquisitions (Nacq) 1
Generally the patient’s head is fixed so that the head is horizontal (not tilted) and the neck
Swap read and phase encoding No
and head lie along the axis of the patient table.
Metastatic Saturation pulses Not applicable
Intra-Axial Scan time 1 min, 28 sec Cerebral
Neoplasia Neoplastic Disease

A3.1.2 A3.1.3
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A3.1.3 Primary Clinical Imaging Parameters for Sequence 2 Table A3.1.5 Primary Clinical Imaging Parameters for Sequence 4 (Fast
(T1-Weighted) FLAIR)

Patient position Supine Patient position Supine

Scan type Spin echo Scan type Inversion recovery fast spin echo
Imaging plane (orientation) Transverse Imaging plane (orientation) Transverse
Central slice or volume center Laser light centered on nasion Central slice or volume center Laser light centered on nasion
Echo time (TE) 11 msec (or select “minimum Echo time (TE) 120 msec (effective)
full” echo time) Receiver bandwidth (RBW) 16 kHz
Receiver bandwidth (RBW) 10 kHz Echo train length (ETL) 8
Repeat time (TR) 500 msec Repeat time (TR) 10000 msec
Flip angle (FA) 90° Inversion time (TI) 2200 msec
Fields of view (FOVx, FOVy) 240 mm, 240 mm Flip angle (FA) 180°
Resolution (Δx, Δy) 0.94 mm, 1.25 mm Fields of view (FOVx, FOVy) 240 mm, 240 mm
Number of data points collected (Nx, Ny) 256, 192 Resolution (Δx, Δy) 0.94 mm, 1.25 mm
Display matrix (Dx, Dy) 256, 192 Number of data points collected (Nx, Ny) 256, 192
Slice thickness(Δz) 5 mm Display matrix (Dx, Dy) 256, 192
Number of slices 20 or as many as needed to cover Slice thickness (Δz) 5 mm
the region of interest Number of slices 20 or as many as needed to cover
Slice gap 2.5 mm the region of interest
Number of acquisitions (Nacq) 1 Slice gap 2.5 mm
Swap read and phase encoding Yes Number of acquisitions (Nacq) 1
Saturation pulses Not applicable Swap read and phase encoding Yes
Scan time 1 min, 28 sec Saturation pulses Not applicable
Scan time ∼5 min

Table A3.1.4 Primary Clinical Imaging Parameters for Sequence 3

(T2-Weighted) It is preferable to insert the line prior to imaging and to leave the patient in the magnet,
with no intervening motion, between the scans run before and after contrast agent injection
Patient position Supine and those run after injection.
Scan type Fast spin echo
16. Leaving the patient in the magnet, inject the contrast agent, flush the line with 10 ml
Imaging plane (orientation) Transverse
saline.
Central slice or volume center Laser light centered on nasion
Echo time (TE) 102 msec (effective) A dose of 0.1 mmol/kg of contrast agent is usually given.
Receiver bandwidth (RBW) 16 kHz
A delay in scanning may actually be beneficial when evaluating for metastases, which is
Echo train length (ETL) 8 one of the reasons we scan in multiple planes after intravenous contrast administration.
Repeat time (TR) 3600 msec
Flip angle (FA) 90° 17. Acquire the post-contrast images using the same parameters as sequence 2 (T1-
Fields of view (FOVx, FOVy) 240 mm, 240 mm weighted).
Resolution (Δx, Δy) 0.94 mm, 1.25 mm In addition to the transverse plane, the coronal plane is routinely obtained, with the
Number of data points collected (Nx, Ny) 256, 192 following changes to the parameters in sequence 2: (a) flow compensation is on; (b) TE is
Display matrix (Dx, Dy) 256, 192 20 msec (prolonged due to flow compensation gradients); and (c) it is not necessary to
Slice thickness (Δz) 5 mm swap read and phase encoding directions. We also routinely obtain a third post-contrast
Number of slices 20 or as many as needed to cover plane of imaging (sagittal) in patients requiring sedation and in all pediatric patients.
the region of interest
Slice gap 2.5 mm R/O HEMORRHAGIC METASTATIC DISEASE/METASTATIC MELANOMA ALTERNATE
Number of acquisitions (Nacq) 1 PROTOCOL
Swap read and phase encoding Yes A small percentage of metastatic brain tumors will demonstrate evidence of hemorrhage.
Saturation pulses Not applicable These most commonly include melanoma, choriocarcinoma, renal cell, bronchogenic,
Scan time ∼2 min and thyroid carcinomas. The appearance will vary according to the stage of the hemor-
rhage (acute-chronic) on the T1-weighted and T2-weighted images (Thulborn et al., 1990;
Thulborn and Atlas, 1996; Atlas and Thulborn, 1998) and gradient-echo imaging demon-
Metastatic
strates evidence of hemorrhage that may not be seen with conventional spin-echo imaging
Intra-Axial (Atlas et al., 1988). Intracranial metastatic melanoma (if melanotic) exhibits the typical Cerebral
Neoplasia Neoplastic Disease

A3.1.4 A3.1.5
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A3.1.6 Primary Clinical Imaging Parameters for Sequence 6 (Gradient
Echo)
Patient position
Scan type
Imaging plane (orientation)
Central slice or volume center
Echo time (TE)
Receiver bandwidth (RBW)
Repeat time (TR)
Flip angle (FA)
Fields of view (FOVx, FOVy)
Resolution (Δx, Δy)
Number of data points collected (Nx, Ny)
Display matrix (Dx, Dy)
Slice thickness (Δz)
Number of slices
Slice gap
Number of acquisitions (Nacq)
Swap read and phase encoding
Saturation pulses
Scan time
imaging appearance of any paramagnetic lesion (Atlas et al., 1987), causing shortening
of T1 and T2 on MR images. However, the imaging appearance may be quite variable
depending on melanin content and hemorrhage.
Setup equipment and patient
1. Use the same equipment and perform patient setup as for the previous method (see
Basic Protocol).
Pre-contrast T1-weighted scan should be acquired in the same plane as the post-contrast
T1-weighted scan to look for intra-lesional enhancement.
Sequence 6: Gradient echo sequence
2. In addition to sequences 1 to 5, obtain gradient echo images (sequence 6, Table
A3.1.6).
COMMENTARY
Background Information
Brain metastases are common intracranial
tumors (Johnson and Young, 1996) affecting
between 80,000 and 170,000 individuals in the
United States per year (Davey, 1999). While
computed tomography (CT) may often be more
readily available and therefore the first exami-
nation performed for suspected intracranial pa-
thology, magnetic resonance imaging (MRI)
has proven to be invaluable in refining the
diagnosis. MRI with intravenous contrast pro-
Metastatic vides the greatest sensitivity and exquisite ana-
Intra-Axial tomical information for the detection of brain
Neoplasia
extent of metastatic disease. A delay in scan-
ning is actually beneficial and at least two
planes of imaging should be obtained following
Supine the administration of intravenous contrast.
2-D gradient recalled echo
Three planes of imaging (transverse, coronal,
Transverse and sagittal) should be acquired post-contrast
Laser light centered on nasion in patients requiring sedation and in all pediat-
30 msec ric patients. Not uncommonly, metastatic le-
4 kHz sions (particularly to the cortex) are detected
500 msec solely on the post-contrast images, which is due
15° to similar signal characteristics with the adja-
cent brain tissue and the lack of surrounding
240 mm, 240 mm
edema. Where small lesions (1 mm) can be
0.94 mm, 1.25 mm
mistaken for vessels, the use of a first and
256, 192 second dose of contrast can be used to distin-
256, 256 guish lesions (more conspicuous after the sec-
5 mm ond dose) from vessels (no change in appear-
20 or as many as needed to cover
ance between doses).
the region of interest
2.5 mm
Anticipated Results
1 The goal in evaluating the brain for metas-
Yes tases is to demonstrate the total number of
Not applicable lesions, and to depict the full extent of disease
∼2 min burden in terms of edema, brain herniation and
hemorrhage. Magnetic resonance imaging of-
fers superb anatomical detail and tissue char-
acterization of brain metastases and provides
the greatest sensitivity for detecting lesions.
Time Considerations
The protocols detailed in this unit should
take ∼30 min to complete.
Literature Cited
Atlas, S.W. and Lavi, E. 1996. Intra-axial brain
tumors. In Magnetic Resonance Imaging of the
Brain and Spine, 2nd ed. (S.W. Atlas, ed.) pp.
315-422. Lippincott-Raven, Philadelphia.
Atlas, S.W. and Thulborn, K.R. 1998. MR detection
of hyperacute parenchymal hemorrhage of the
brain. Am. J. Neuroradiol. 19:1471-1477.
Atlas, S.W., Grossman, R.I., Gomori, J.M., Guerry,
D., Hackney, D.B., Goldberg, H.I., Zimmerman,
metastases. The neuroanatomic location of me-
tastases and the associated edema and mass
effect determine the clinical presentation of the
patient. The most common primary sites of
brain metastases in adults include lung, breast,
skin (melanoma), and the gastrointestinal tract.
In younger patients, brain metastases often
arise from sarcomas and germ cell tumors.
Critical Parameters and
Troubleshooting
Intravenous contrast (gadolinium-DTPA) is
essential when evaluating the brain for the full
R.A., and Bilaniuk, L.T. 1987. MR imaging of
intracranial metastatic melanoma. J. Comput.
Assist. Tomogr. 11:577-582.
Atlas, S.W., Mark, A.S., Grossman, R.I., and Go-
mori, J.M. 1988. Intracranial hemorrhage: Gra-
dient-echo MR imaging at 1.5 T. Radiology
168:803-807.
Davey, P. 1999. Brain metastases. Curr. Probl. Can-
cer 23:59-98.
Healy, M.E., Hesselink, J.R., Press, G.A., and Mid-
dleton, M.S. 1987. Increased detection of intrac-
ranial metastases with intravenous Gd-DTPA.
Radiology 165:619-624.
Johnson, J.D. and Young, B. 1996. Demographics of
brain metastasis. Neurosurg. Clin. N. Am. 7:337-
344.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Thulborn, K.R. and Atlas, S.W. 1996. Intracranial
hemorrhage. In Magnetic Resonance Imaging of
the Brain and Spine, 2nd ed. (S.W. Atlas, ed.) pp.
265-314. Lippincott-Raven, Philadelphia.
Thulborn, K.R., Sorensen, A.G., Kowall, N.W.,
McKee, A., Lai, A., McKinstry, R.C., Moore, J.,
Rosen, B.R., and Brady, T.J. 1990. The role of
ferritin and hemosiderin in the MR appearance
of cerebral hemorrhage: A histopathologic bio-
chemical study in rats. Am. J. Neuroradiol.
2:291-297.
Yuh, W.T., Tali, E.T., Nguyen, H.D., Simonson,
T.M., Mayr, N.A., and Fisher, D.J. 1995. The
effect of contrast dose, imaging time, and lesion
size in the detection of intracerebral metastasis.
Am. J. Neuroradiol. 16:373-380.
Contributed by Annette O. Nusbaum
New York Presbyterian Hospital
New York, New York
Scott W. Atlas
Stanford University Medical Center
Stanford, California
Cerebral
Neoplastic Disease

A3.1.6 A3.1.7
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
Metastatic Extra-Axial Neoplasia
There are a multitude of tumors which may spread to the extra-axial compartment (i.e.,
meninges, skull base, and calvarium). The role of imaging in the search for metastases is
to detect the full extent of and to correctly localize the disease. This unit presents the set
of MR sequences (see Basic Protocols 1, 2, and 3) used for imaging extra-axial metastatic
tumors, and specific modifications will be discussed where necessary. The sequences
described in this unit are based on the authors’ experience with a 1.5 T scanner (Echospeed
GE Medical Systems, Milwaukee, Wisconsin), but can be expected to be equally appli-
cable to other field strengths and scanners from other manufacturers.
RULE OUT (R/O) LEPTOMENINGEAL CARCINOMATOSIS
Spread of tumor to the leptomeninges is frequently seen in association with any adeno-
carcinoma (i.e., breast, lung, gastrointestinal tumors) and lymphoma or leukemia (Gold-
berg et al., 1996). MR imaging in suspected leptomeningeal carcinomatosis is essential
not only to demonstrate the disease, but also to assess for potential complications
including hydrocephalus. The post-gadolinium T1-weighted images and fluid-attenuated
inversion recovery (FLAIR) images are most useful when evaluating spread of tumor,
particularly along the subarachnoid space (leptomeninges; Singer et al., 1998). Sequences
1 to 5 comprise the preferred protocol.
Table A3.2.1 lists the hardware necessary to perform the procedure, along with appropri-
ate parameters.
NOTE: Be sure that technologists and nurses have immediate access to any equipment
such as crash carts or oxygen that may be necessary in the event of an emergency. Contrast
reactions are rare, but the resources are necessary.
Materials
Normal saline (0.9% NaCl), sterile
Intravenous contrast agent (e.g., Magnevist, Omniscan, or Prohance)
Set up equipment and patient
1. Interview (screen) the patient to ensure that he or she has no contraindications to the
MRI examination, such as a cardiac pacemaker or other implants or foreign bodies
containing ferromagnetic materials. Question the patient regarding any health con-
ditions that may require the presence of specific emergency equipment during the
scanning procedure, or necessitate any other precautions.
Generally standard screening forms are used for all patients scanned in a magnetic
resonance system.
Table A3.2.1 Equipment Parameters for Imaging Brain Tumors
Coil type Quadrature head coil
Gradient coil strength 25 mT/m (or whatever the system permits)
Cardiac gating No
Peripheral gating For safety only
Respiratory gating No
Respirator If required by patient
Oxygen If required by patient
Motion cushions No
Contributed by Annette O. Nusbaum and Scott W. Atlas
Current Protocols in Magnetic Resonance Imaging (2001) A3.2.1-A3.2.9
Copyright © 2001 by John Wiley & Sons, Inc.
UNIT A3.2 The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(1996) for discussion of what implants may be safely scanned using magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
2. If the procedure is a research protocol, have the patient sign any necessary consent
form.
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
BASIC
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating
PROTOCOL 1
and image artifacts.
5. Inform the patient about what will occur during the procedure, what he or she will
experience in the magnet, and how to behave, including the following:
a. Ear protection will be provided (earphones, headphones, earplugs) to dampen the
sounds from the gradients, but the patient will be able to communicate with you
at any time during the imaging.
b. The patient will be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this works).
c. For good results the patient should not talk, and should avoid or minimize
swallowing or other movement, during each scan—i.e., as long as the banging
sounds continue. Between scans, talking and swallowing are allowed in most
cases, but should be avoided when comparative positional studies are being
performed; the patient will be informed when this is the case.
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
6. Help the patient mount onto the table and lie in the supine position. Set up all
monitoring equipment or any triggering devices before or right after the patient lies
down.
7. Center the patient in the head coil at the region where the key information is desired.
Make sure the head and neck are constrained to prevent motion.
Generally the patient’s head is fixed so that the head is horizontal (not tilted) and the neck
and head lie along the axis of the patient table.
8. If needed, place a pillow or other support under the knees to make the patient more
comfortable.
9. Use the centering light to position the patient (centered on the nasion) and put him
or her into the center of the magnet.
Once this step has been performed, so long as the patient does not move on the table, the
table itself can be moved and then replaced in the same position as before without
jeopardizing the positioning of one scan relative to another.
10. If the patient is unable to hold still, provide an appropriate sedative.
Sequence 1: Localizer
11. Run sequence 1 according to Table A3.2.2. The sagittal scout view is used to prescribe
Metastatic the transverse and coronal sequences.
Cerebral Extra-Axial
Neoplastic Disease Neoplasia
A3.2.1 A3.2.2
Current Protocols in Magnetic Resonance Imaging
 Table A3.2.2 Primary Clinical Imaging Parameters for T1-Weighted Scan Table A3.2.4 Primary Clinical Imaging Parameters for T2-Weighted Scan
(Sequence 1 Localizer) (Sequence 3)

Patient position Supine Patient position Supine

Scan type Spin echo Scan type Fast spin echo
Imaging plane (orientation) Sagittal Imaging plane (orientation) Transverse
Central slice or volume center Laser light centered on nasion Central slice or volume center Laser light centered on nasion
Echo time (TE) 11 msec (or select “minimum Echo time (TE) 102 msec (effective)
full” echo time) Receiver bandwidth (RBW) 16 kHz
Receiver bandwidth (RBW) 10 kHz Echo train length (ETL) 8
Repeat time (TR) 500 msec Repeat time (TR) 3600 msec
Flip angle (FA) 90° Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 240 mm, 240 mm Fields of view (FOVx, FOVy) 240 mm, 240 mm
Resolution (Δx, Δy) 0.94 mm, 1.25 mm Resolution (Δx, Δy) 0.94 mm, 1.25 mm
Number of data points collected (Nx, Ny) 256, 192 Number of data points collected (Nx, Ny) 256, 192
Slice thickness (Δz) 5 mm Slice thickness (Δz) 5 mm
Number of slices 20 or as many as needed to cover Number of slices 20 or as many as needed to cover
the region of interest the region of interest
Slice gap 2 mm Slice gap 2.5 mm
Number of acquisitions (Nacq) 1 Number of acquisitions (Nacq) 1
Swap read and phase encoding No Swap read and phase encoding Yes
Saturation pulses Not applicable Saturation pulses Not applicable
Scan time 1 min, 28 sec Scan time ∼2 min

Table A3.2.3 Primary Clinical Imaging Parameters for T1-Weighted Scan Table A3.2.5 Primary Clinical Imaging Parameters for Fast FLAIR Scan
(Sequence 2) (Sequence 4)

Patient position Supine Patient position Supine

Scan type Spin echo Scan type Inversion recovery fast spin echo
Imaging plane (orientation) Transverse Imaging plane (orientation) Transverse
Central slice or volume center Laser light centered on nasion Central slice or volume center Laser light centered on nasion
Echo time (TE) 11 msec (or select “minimum Echo time (TE) 120 msec (effective)
full” echo time) Receiver bandwidth (RBW) 16 kHz
Receiver bandwidth (RBW) 10 kHz Echo train length (ETL) 8
Repeat time (TR) 500 msec Repeat time (TR) 10000 msec
Flip angle (FA) 90° Inversion time (TI) 2200 msec
Fields of view (FOVx, FOVy) 240 mm, 240 mm Flip angle (FA) 180°
Resolution (Δx, Δy) 0.94 mm, 1.25 mm Fields of view (FOVx, FOVy) 240 mm, 240 mm
Number of data points collected (Nx, Ny) 256, 192 Resolution (Δx, Δy) 0.94 mm, 1.25 mm
Slice thickness (Δz) 5 mm Number of data points collected (Nx, Ny) 256, 192
Number of slices 20 or as many as needed to cover Slice thickness (Δz) 5 mm
the region of interest Number of slices 20 or as many as needed to cover
Slice gap 2.5 mm the region of interest
Number of acquisitions (Nacq) 1 Slice gap 2.5 mm
Swap read and phase encoding Yes Number of acquisitions (Nacq) 1
Saturation pulses Not applicable Swap read and phase encoding Yes
Scan time 1 min, 28 sec Saturation pulses Not applicable
Scan time ∼5 min

Metastatic
Cerebral Extra-Axial
Neoplastic Disease Neoplasia

A3.2.3 A3.2.4
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
Sequence 2: T1-weighted scan Table A3.2.6 Primary Clinical Imaging Parameters for T2-Weighted Scan
12. Run sequence 2 according to Table A3.2.3. The pre-contrast T1-weighted scan should (Sequence 6)
be acquired in the same plane as the post-contrast T1-weighted scan to look for
Patient position Supine
intra-lesional enhancement.
Scan type Fast spin echo
Sequence 3: T2-weighted scan Imaging plane (orientation) Transverse
13. Run sequence 3 according to Table A3.2.4. Central slice or volume center Laser light centered on nasion
Echo time (TE) 102 msec (effective)
Sequence 4: Fast fluid-attenuated inversion recovery (FLAIR) scan Receiver bandwidth (RBW) 16 kHz
14. Run sequence 4 according to Table A3.2.5. FLAIR images are useful when evaluating Echo train length (ETL) 16
spread of tumor, particularly along the subarachnoid space (leptomeninges). Repeat time (TR) 4000 msec
Flip angle (FA) 90°
Sequence 5: Post-contrast imaging Fields of view (FOVx, FOVy) 180 mm, 180 mm
15. Remove the patient from the scanner. The patient should not move on the table. Resolution (Δx, Δy) 0.70 mm, 0.70 mm
Establish an intravenous line from which the contrast agent can be injected, and attach Number of data points collected (Nx, Ny) 256, 256
this line securely to the patient so that movement into or out of the magnet will not Slice thickness (Δz) 3 mm
pull at the patient’s arm. Move the patient back into the scanner. Number of slices 20 or as many as needed to cover
entire skull base
It is preferable to insert the line prior to imaging and to leave the patient in the magnet, Slice gap 0.5 mm
with no intervening motion, between the scans run before contrast agent injection and those Number of acquisitions (Nacq) 3
run after injection. Swap read and phase encoding Yes
16. Leaving the patient in the magnet, inject the contrast agent, flush the line with 10 ml Saturation pulses Fat
saline. Fat suppression Yes
Scan time ∼5 min
A dose of 0.1 mmol/kg of contrast agent is usually given.
A delay in scanning may actually be beneficial when evaluating for metastases, which is
one of the reasons we scan in multiple planes after intravenous contrast administration. Table A3.2.7 Primary Clinical Imaging Parameters for T1-Weighted Scan
(Sequence 7)
17. Acquire the post-contrast images using the same parameters as in sequence 2
(T1-weighted), Table A3.2.3. In addition to the transverse plane, the coronal plane is Patient position Supine
routinely obtained, with the following changes to the parameters in sequence 2: (a) Scan type Spin echo
flow compensation is on; (b) TE is 15 msec (prolonged due to flow compensation Imaging plane (orientation) Transverse
gradients); and (c) it is not necessary to swap read and phase encoding directions. We Central slice or volume center Laser light centered on nasion
also routinely obtain a third post-contrast plane of imaging (sagittal) in patients Echo time (TE) 11 msec (or select “minimum
full” echo time)
requiring sedation and in all pediatric patients.
Receiver bandwidth (RBW) 10 kHz
Repeat time (TR) 600 msec
R/O BASE OF SKULL METASTASIS BASIC Flip angle (FA) 90°
PROTOCOL 2 Fields of view (FOVx, FOVy) 180 mm, 180 mm
Metastases to the base of the skull frequently arise from the local spread of pharyngeal
tumor, or from hematogenous spread to the bone. The main purpose of imaging is to Resolution (Δx, Δy) 0.70 mm, 0.94 mm
demonstrate disease and determine the full extent of spread of the lesion (i.e., cranial Number of data points collected (Nx, Ny) 256, 192
nerve or meningeal involvement). Fat-suppressed high-resolution imaging (including Slice thickness (Δz) 3 mm
T2-weighted imaging and post-contrast T1-weighted imaging) is essential when evaluating Number of slices 20 or as many as needed to cover
tumor involvement of skull base structures. entire skull base
Slice gap 0.5 mm
Set up patient and equipment Number of acquisitions (Nacq) 3
1. Use the same equipment and perform equipment and patient setup as for the previous Swap read and phase encoding Yes
method (see Basic Protocol 1). Saturation pulses None
Fat suppression No
2. Obtain sequences 1, 2, 3 outlined in Basic Protocol 1 to evaluate the entire brain. Scan time ∼5 min
3. For high-resolution imaging of the skull base, perform transverse pre-contrast se-
quences 6 and 7 outlined in Tables A3.2.6 and A3.2.7. The entire extent of the lesion
must always be included. Metastatic
Cerebral Extra-Axial
Neoplastic Disease Neoplasia

A3.2.5 A3.2.6
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A3.2.8 Primary Clinical Imaging Parameters for Post-Contrast Table A3.2.9 Primary Clinical Imaging Parameters for Gradient Recalled Echo
T1-Weighted Scan (Sequence 8) (GRE) Scan (Sequence 9)

Patient position Supine Patient position Supine

Scan type Spin echo Scan type 2-D gradient recalled echo
Imaging plane (orientation) Transverse and coronal Imaging plane (orientation) Transverse
Central slice or volume center Laser light centered on nasion Central slice or volume center Laser light centered on nasion
Echo time (TE) 20 msec Echo time (TE) 30 msec
Receiver bandwidth (RBW) 10 kHz Receiver bandwidth (RBW) 4 kHz
Repeat time (TR) 500 msec Repeat time (TR) 500 msec
Flip angle (FA) 90° Flip angle (FA) 15°
Fields of view (FOVx, FOVy) 180 mm, 180 mm Fields of view (FOVx, FOVy) 240 mm, 240 mm
Resolution (Δx, Δy) 0.70 mm, 0.94 mm Resolution (Δx, Δy) 0.94 mm, 1.25 mm
Number of slices 20 or as many as needed to cover Number of data points collected (Nx, Ny) 256, 192
the area of interest Slice thickness (Δz) 5 mm
Number of data points collected (Nx, Ny) 256, 192 Number of slices 20 or as many as needed to cover
Slice thickness (Δz) 3 mm the region of interest
Slice gap 0.5 mm Slice gap 2.5 mm
Number of acquisitions (Nacq) 3 Number of acquisitions (Nacq) 1
Swap read and phase encoding Yes if it is a transverse scan Swap read and phase encoding Yes
Saturation pulses Fat Saturation pulses Not applicable
Fat suppression Yes Scan time ∼2 min
Scan time ∼5 min

Set up equipment and patient

Sequence 6: T2-weighted scan 1. Use the same equipment and perform patient setup as in Basic Protocol 1. Obtain
4. Run sequence 6 according to Table A3.2.6. Fat-suppressed high-resolution imaging sequences 1 to 4.
is essential to clearly delineate the full extent of disease in the skull base.
In addition, gradient echo (GRE) images (sequence 9; Table A3.2.9) should be obtained.
Sequence 7: T1-weighted scan
Sequence 9: Gradient echo scan
5. Run sequence 7 according to Table A3.2.7.
2. Run sequence 9 according to Table A3.2.9.
Sequence 8: T1-weighted scan (post-contrast)
6. Repeat steps 15 and 16 in Basic Protocol 1. Sequence 10: Post-contrast imaging
3. Run sequence 5 (see Basic Protocol 1, steps 15 to 17).
7. Post-contrast T1-weighted sequence 8 (Table A3.2.8) are obtained in two planes
(transverse and coronal). The entire extent of the lesion must always be included. COMMENTARY
Fat-suppressed T1-weighted images are critical to evaluate for abnormal enhancement
of cranial nerves, since high signal intensity from fat found along the skull base or Background Information Critical Parameters and
chemical shift artifact from fat-containing structures can obscure enhancement of Metastatic extra-axial disease is common in Troubleshooting
cranial nerves. clinical practice and lesions are frequently nu- The MR evaluation for extra-axial spread of
merous and infiltrative. Obtaining detailed in- tumor to the meninges (leptomeningeal carci-
8. In addition, post-contrast images (sequence 8 outlined in Table A3.2.8) of the entire formation regarding the entire extent of disease nomatosis) should always include FLAIR or
brain should be obtained. is critical to the planning of appropriate treat- post-contrast T1-weighted images to provide
ment. MRI with intravenous contrast is essen- the greatest sensitivity in disease detection.
tial for documenting meningeal disease and When evaluating the skull base, fat-suppressed
R/O HEMORRHAGIC METASTATIC DISEASE/MELANOMA BASIC spread along cranial nerves. high-resolution imaging (including T2-
PROTOCOL 3 High-resolution computed tomography weighted and post-contrast T1-weighted) is es-
The MR appearance of hemorrhage will vary according to the stage of the hemorrhage
(acute-chronic) on the T1-weighted and T2-weighted images. On the other hand, gradient- (CT) of the skull base is often an integral part sential to clearly delineate the full extent of
echo images not uncommonly demonstrate evidence of hemorrhage that may not be seen of the work-up of a skull base lesion. CT pro- disease. With a resolution of less than 1 mm in
with conventional spin-echo images (Atlas et al., 1987, 1988; Thulborn et al., 1990; vides important information regarding the in- the x and y directions, there is superb visuali-
tegrity of bone architecture and depicts bone zation of the fine, intricate structures of the
Thulborn and Atlas, 1996). Intracranial metastatic melanoma (if melanotic) exhibits the
irregularity, erosion, intra-tumoral calcification skull base (foramina and cranial nerves). Fat-
typical imaging appearance of any paramagnetic lesion (Atlas et al., 1987), causing or hyperostosis. suppressed T1-weighted images are critical to
shortening of T1 and T2. However, the imaging appearance may be quite variable Metastatic
Cerebral Extra-Axial
depending on melanin content and hemorrhage. Neoplastic Disease Neoplasia

A3.2.7 A3.2.8
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
evaluate for abnormal enhancement of cranial
nerves, since high signal intensity on the T1-
weighted images from fat found along the skull
base or chemical shift artifact from fat-contain-
ing structures can obscure enhancement of cra-
nial nerves. Inhomogeneities in the magnetic
field can result in incomplete fat suppression
or unintended suppression of water. With expe-
rience this does not usually represent a concern.
Anticipated Results
The goal of imaging when evaluating for
extra-axial metastases is to accurately localize
and detect the full extent of disease. Magnetic
resonance imaging offers the necessary con-
trast and high resolution to fully evaluate the
extra-cranial compartment for tumor spread.
Time Considerations
The majority of these protocols should take
∼30 min to complete.
Literature Cited
Atlas, S.W., Grossman, R.I., Gomori, J.M., Guerry,
D., Hackney, D.B., Goldberg, H.I., Zimmerman,
R.A., and Bilaniuk, L.T. 1987. MR imaging of
intracranial metastatic melanoma. J. Comput.
Assist. Tomogr. 11:577-582.
Atlas, S.W., Mark, A.S., Grossman, R.I., and Go-
mori, J.M. 1988. Intracranial hemorrhage: Gra-
dient-echo MR imaging at 1.5 T. Radiology
168:803-807.
Current Protocols in Magnetic Resonance Imaging
Goldberg, H.I., Lavi, E., and Atlas, S.W. 1996. Ex-
tra-axial brain tumors. In Magnetic Resonance
Imaging of the Brain and Spine (S.W. Atlas, ed.).
Lippincott-Raven, Philadelphia.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Singer, M.B., Atlas, S.W., and Drayer, B.P. 1998.
Subarachnoid space disease: Diagnosis with
fluid-attenuated inversion-recovery MR imag-
ing and comparison with gadolinium-enhanced
spin echo MR imaging: Blinded reader study.
Radiology 208:417-422.
Thulborn, K.R. and Atlas, S.W. 1996. Intracranial
hemorrhage. In Magnetic Resonance Imaging of
the Brain and Spine, 2nd ed. (S.W. Atlas, ed.).
Lippincott-Raven, Philadelphia.
Thulborn, K.R., Sorensen, A.G., Kowall, N.W.,
McKee, A., Lai, A., McKinstry, R.C., Moore, J.,
Rosen, B.R., and Brady, T.J. 1990. The role of
ferritin and hemosiderin in the MR appearance
of cerebral hemorrhage: A histopathologic bio-
chemical study in rats. Am. J. Neuroradiol.
11:291-297.
Contributed by Annette O. Nusbaum
New York Presbyterian Hospital
New York, New York
Scott W. Atlas
Stanford University Medical Center
Stanford, California
Cerebral
Neoplastic Disease
A3.2.9
Intra-Axial Primary Brain Tumors UNIT A3.3
MRI provides the greatest sensitivity for the detection of brain tumors and information
regarding detailed anatomical location and extent, and often elucidates important under-
lying histopathologic features (Atlas and Lavi, 1996). The majority of primary brain
tumors in adults are found in the supratentorial compartment, while tumors in pediatric
patients are infratentorial in location. This unit presents the set of sequences used for
imaging all types of primary intra-axial brain tumors, whether infiltrative (i.e., astrocy-
toma, oligodendroglioma, lymphoma) or circumscribed (i.e., ganglioglioma, cystic astro-
cytoma), and specific modifications will be discussed where necessary. The sequences
described in this unit are based on the authors’ experience with a 1.5 T scanner (Echospeed
GE Medical Systems), but can be expected to be equally applicable to other field strengths
and scanners from other manufacturers.
RULE OUT (R/O) PRIMARY CEREBRAL TUMOR BASIC
PROTOCOL 1
Spin-echo and fast spin-echo imaging essentially provide all of the information needed
to make an accurate diagnosis when evaluating for primary brain tumors. Contrast-en-
hanced imaging is a critical part of the MR evaluation of brain tumors. The non-contrast
spin-echo and fast-spin echo images (sequences 2 and 3) provide information regarding
tissue characteristics and extent of tumor. Fluid attenuated inversion recovery (FLAIR)
images (sequence 4) and the post-contrast T1-weighted images (sequence 5) are most
sensitive for evaluating spread of tumor to the adjacent meninges (Singer, 1998). Enhance-
ment of a brain lesion indicates breakdown of the blood-brain barrier and non-enhancing
portions of a lesion may indicate intra-tumoral cysts or necrosis. Sequences 1 to 5
comprise the preferred protocol.
Gradient echo images depict regions of hypointensity within a tumor that may represent
either calcification, blood products or intra-tumoral vessels (Atlas et al., 1988). Gradient
echo images (sequence 6) are added on to the Basic Protocol described in this unit when
tumors that calcify (i.e., oligodendroglioma, ganglioglioma) or hemorrhage (high grade
gliomas, ependymomas) are suspected.
Table A3.3.1 lists the hardware necessary to perform the procedure, along with appropri-
ate parameters.
NOTE: Be sure that technologists and nurses have immediate access to any equipment
such as crash carts or oxygen that may be necessary in the event of an emergency. Contrast
reactions are rare, but the resources are necessary.
Materials
Intravenous contrast agent (e.g., Magnevist, Omniscan, or Prohance)
Normal saline (0.9% NaCl), sterile
Table A3.3.1 Equipment Parameters for Imaging Brain Tumors
Coil Type Quadrature head coil
Gradient coil strength 25 mT/m (or whatever the system permits)
Cardiac gating No
Peripheral gating For safety only
Respiratory gating No
Respirator If required by patient
Oxygen If required by patient
Motion cushions No Cerebral
Neoplastic Disease
Contributed by Annette O. Nusbaum and Scott W. Atlas A3.3.1
Current Protocols in Magnetic Resonance Imaging (2001) A3.3.1-A3.3.10
Copyright © 2001 by John Wiley & Sons, Inc.
 Set up patient and equipment Once this step has been performed, so long as the patient does not move on the table, the
1. Interview (screen) the patient to ensure that he or she has no contraindications to the table itself can be moved and then replaced in the same position as before without
MRI examination, such as a cardiac pacemaker or other implants containing ferro- jeopardizing the positioning of one scan relative to another.
magnetic materials. Question the patient regarding any health conditions that may 10. If the patient is unable to hold still, provide an appropriate sedative.
require the presence of specific emergency equipment during the scanning procedure,
or necessitate any other precautions. Sequence 1: Localizer
11. Run sequence 1 according to Table A3.3.2. The sagittal scout view is used to prescribe
Generally standard screening forms are used for all patients scanned in a magnetic the transverse and coronal planes.
resonance system.
Sequence 2: T1-weighted scan
The presence of any ferromagnetic metals may be a health hazard to the patient when he 12. Run sequence 2 according to Table A3.3.3.
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock Sequence 3: T2-weighted scan
(1996) for discussion of what implants may be safely scanned using magnetic resonance. 13. Run sequence 3 according to Table A3.3.4.
Patients may be accompanied into the magnet room by a friend or family member, who can Sequence 4: Fast fluid attenuated inversion recovery (FLAIR)
sit in the room during the scan and comfort the patient as needed. This companion must
14. Run sequence 4 according to Table A3.3.5.
be screened as well to ensure the absence of loose metal objects on the body or clothing.
2. If the procedure is a research protocol, have the patient sign any necessary consent Sequence 5: Post-contrast imaging
form. 15. Remove the patient from the scanner. The patient should not move on the table.
Establish an intravenous line from which the contrast agent can be injected, and attach
3. Have the patient remove all jewelry and change into a gown to eliminate any metal this line securely to the patient so that movement into or out of the magnet will not
that might be found in clothing. pull at the patient’s arm. Move the patient back into the scanner.
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating It is preferable to insert the line prior to imaging and to leave the patient in the magnet,
and image artifacts. with no intervening motion, between the scans run before contrast agent injection and those
run after injection.
5. Inform the patient about what will occur during the procedure, what he or she will
16. Leaving the patient in the magnet, inject the contrast agent, flush the line with 10 ml
experience in the magnet, and how to behave, including the following:
saline.
a. If earphones or headphones are used to protect the ears from the loud sounds A dose of 0.1 mmol/kg of contrast agent is usually given.
produced by the gradients, the patient will be asked to wear these, but will be able
to communicate with you at any time during the imaging. A delay in scanning may actually be beneficial when evaluating for metastases, which is
one of the reasons we scan in multiple planes after intravenous contrast administration.
b. The patient will be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this works). Table A3.3.2 Primary Clinical Imaging Parameters for Sequence 1: Localizer
(T1-Weighted)
c. For good results the patient should not talk, and should avoid or minimize
swallowing or other movement, during each scan—i.e., as long as the banging Patient position Supine
sounds continue. Between scans, talking and swallowing are allowed in most Scan type Spin echo
cases, but should be avoided when comparative positional studies are being Imaging plane (orientation) Sagittal
performed; the patient will be informed when this is the case. Central slice or volume center Laser light centered on nasion
d. Nevertheless, the patient may call out at any time if he or she feels it necessary. Echo time (TE) 11 msec (or select “minimum
full” echo time)
6. Have the patient lie in the supine position on the table. Set up all monitoring Receiver bandwidth (RBW) 10 kHz
equipment before or right after the patient lies down. Repeat time (TR) 500 msec
7. Center the patient in the head coil at the region where the key information is desired. Flip angle (FA) 90°
Make sure the head and neck are constrained to prevent motion. Fields of view (FOVx, FOVy) 240 mm, 240 mm
Resolution (Δx, Δy) 0.94 mm, 1.25 mm
Generally the patient’s head is fixed so that the head is horizontal (not tilted) and the neck Number of data points collected, (Nx, Ny) 256, 192
and head lie along the axis of the patient table. Slice thickness (Δz) 5 mm
8. If needed, place a pillow or other support under the knees to make the patient more Number of slices 20 or as many as needed to cover
comfortable. the region of interest
Slice gap 2 mm
9. Use the centering light to position the patient (centered on the nasion) and put him Number of acquisitions (Nacq) 1
or her into the center of the magnet. Swap read and phase encoding No
Intra-Axial
Primary Brain Saturation pulses Not applicable Cerebral
Tumors Scan time ∼1 min, 36 sec Neoplastic Disease

A3.3.2 A3.3.3
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A3.3.3 Primary Clinical Imaging Parameters for Sequence 2 17. Acquire the post-contrast images using the same parameters as in sequence 2
(T1-Weighted) (T1-weighted).
Patient position Supine In addition to the transverse plane, the coronal plane is routinely obtained, with the
Scan type Spin echo following changes to the parameters in sequence 2: (a) flow compensation is on; (b) TE is
Imaging plane (orientation) Transverse 15 msec (prolonged due to flow compensation gradients); and (c) it is not necessary to
Central slice or volume center Laser light centered on nasion
Echo time (TE) 11 msec (or select “minimum
Table A3.3.5 Primary Clinical Imaging Parameters for Sequence 4 (Fast
full” echo time) FLAIR)
Receiver bandwidth (RBW) 10 kHz
Repeat time (TR) 500 msec Patient position Supine
Flip angle (FA) 90° Scan type Inversion recovery fast spin echo
Fields of view (FOVx, FOVy) 240 mm, 240 mm Imaging plane (orientation) Transverse
Resolution (Δx, Δy) 0.94 mm, 1.25 mm Central slice or volume center Laser light centered on nasion
Number of data points collected, (Nx, Ny) 256, 192 Echo time (TE) 120 msec (effective)
Slice thickness (Δz) 5 mm Receiver bandwidth (RBW) 16 kHz
Number of slices 20 or as many as needed to cover Echo train length (ETL) 8
the region of interest Repeat time (TR) 10,000 msec
Slice gap 2.5 mm Inversion time (TI) 2200 msec
Number of acquisitions (Nacq) 1 Flip angle (FA) 180°
Swap read and phase encoding Yes Fields of view (FOVx, FOVy) 240 mm, 240 mm
Saturation pulses Not applicable Resolution (Δx, Δy) 0.94 mm, 1.25 mm
Scan time 1 min, 36 sec Number of data points collected, (Nx, Ny) 256, 192
Slice thickness (Δz) 5 mm
Number of slices 20 or as many as needed to cover
Table A3.3.4 Primary Clinical Imaging Parameters for Sequence 3
(T2-Weighted)
the region of interest
Slice gap 2.5 mm
Patient position Supine Number of acquisitions (Nacq) 1
Scan type Fast spin echo Swap read and phase encoding Yes
Imaging plane (orientation) Transverse Saturation pulses Not applicable
Central slice or volume center Laser light centered on nasion Scan time ∼5 min
Echo time (TE) 102 msec (effective)
Receiver bandwidth (RBW) 16 kHz
Table A3.3.6 Primary Clinical Imaging Parameters for Sequence 6
Echo train length (ETL) 8
Repeat time (TR) 3600 msec Patient position Supine
Flip angle (FA) 90° Scan type 2-D gradient recalled echo
Fields of view (FOVx, FOVy) 240 mm, 240 mm Imaging plane (orientation) Transverse
Resolution (Δx, Δy) 0.94 mm, 1.25 mm Central slice or volume center Laser light centered on nasion
Number of data points collected, (Nx, Ny) 256, 192 Echo time (TE) 30 msec
Slice thickness (Δz) 5 mm Receiver bandwidth (RBW) 4 kHz
Number of slices 20 or as many as needed to cover Repeat time (TR) 500 msec
the region of interest Flip angle (FA) 15°
Slice gap 2.5 mm Fields of view (FOVx, FOVy) 240 mm, 240 mm
Number of acquisitions (Nacq) 1 Resolution (Δx, Δy) 0.94 mm, 1.25 mm
Swap read and phase encoding Yes Number of data points collected, (Nx, Ny) 256, 192
Saturation pulses Not applicable Slice thickness (Δz) 5 mm
Scan time ∼1 min, 30 sec Number of slices 20 or as many as needed to cover
the region of interest
Slice gap 2.5 mm
Number of acquisitions (Nacq) 1
Swap read and phase encoding Yes
Intra-Axial Saturation pulses Not applicable
Primary Brain Cerebral
Tumors Scan time ∼2 min Neoplastic Disease

A3.3.4 A3.3.5
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 swap read and phase encoding directions. In patients requiring sedation and in all pediatric the voxel is positioned over the solid (enhancing) portion of the lesion so that theoretically
patients, we also routinely obtain a third post-contrast plane of imaging (sagittal). signal is acquired from viable tumor tissue.

Sequence 6: Gradient echo imaging to examine calcified or hemorrhagic tumor
18. Run sequence 6 according to Table A3.3.6.
Gradient echo images depict regions of hypointensity within tumors that may represent
either calcification, blood products, or flow voids from intra-tumoral vessels. This sequence
is useful to obtain in lesions that frequently calcify (i.e., oligodendroglioma, ganglio-
glioma) or hemorrhage (i.e., high grade gliomas and ependymomas).
2. Perform automated shimming and water suppression, center the voxel over the
enhancing portion of the lesion, and run sequence 7 according to Table A3.3.7.
3. After completion of the scan, a spectrum will appear on the screen and the values for
the spectrum should be saved.
4. Place the voxel in the contralateral hemisphere and repeat the scan.

ALTERNATE OPTIONAL SEQUENCES

PROTOCOL
The set of imaging sequences described in Basic Protocol 1 often provide virtually all the A
information necessary for the diagnosis of primary intra-axial brain tumors. Studies are
being conducted to investigate the potential value of diffusion-weighted imaging (DWI;
Stejskal and Tanner, 1965; Tien et al., 1994) and perfusion-weighted imaging (Alsop and
Detre, 1998; using either contrast bolus techniques or continuous arterial spin labeling)
in routine tumor imaging. Preliminary studies suggest that DWI may have potential for
detecting intra-tumoral necrosis.

The main indication for obtaining proton MR spectroscopy (MRS) data is to help
distinguish a neoplastic from a non-neoplastic lesion (see Figure A3.3.1); however, proton
MR spectroscopy (sequence 7), does provide biochemical information regarding brain
tumors that may be useful to further categorize a specific lesion (i.e., high or low grade
tumor) (Meyer et al., 1999). Whole brain chemical-shift imaging (CSI) maps may also
be potentially useful in the investigation of brain tumors by providing visual metabolite
maps, rather than spectral data.

Set up patient and equipment

1. Follow the setup for Basic Protocol 1 and obtain the necessary pre- and post-contrast
images (steps 1 to 17). The lesion is localized on a transverse image (typically the
post-contrast transverse T1-weighted image).
Voxel placement requires physician input to ensure the correct region of interest is being
studied. Two separate scans are performed: one with the voxel placed over the area in B C
question (typically the enhancing portion of the lesion), the other scan is performed with
the voxel placed in the contralateral hemisphere (in a “normal appearing” analogous area Cho NAA
of the brain). The screen should always be saved to document placement of the voxel. Cho Cre
MyoI
Sequence 7: Proton (1H) MR spectroscopy
The Proton Brain Exam-Single Voxel (PROBE-SV) on GE systems is a fully automated Lac
software package that is widely available and is performed using the single voxel
technique (8 cm3). A high field system is required to obtain good spectra. There are 2
possible localization methods: STEAM (stimulated echo acquisition mode) or PRESS
(point-resolved spectroscopy). The sequence parameters can be set by the user (TR, TE,
voxel size). Short TE spectroscopy has the advantage of providing more information, since 3.0 2.0 1.0 3.0 2.0 1.0
more peaks are visible for analysis while maximizing the detection of lipid-breakdown
products. Long-TE spectra have less baseline distortion and are easier to quantify. Since Figure A3.3.1 (A) T2-weighted fast spin echo image (TR = 2000 msec, TE = 105 msec, Nacq = 2)
there is a √
⎯⎯2 decrease in the signal-to-noise ratio with STEAM sequences, the PRESS through a large mass in the right thalamus and basal ganglia of a 35 year old female with headaches.
sequence is usually preferred. Spectra obtained following the administration of intrave- Water suppressed protocon spectroscopy was performed using PRESS sequence (point resolved
nous contrast may be altered (Sijens et al., 1997). In particular, studies have shown a mean spectroscopy) in an 8-cm3 voxel with a TE = 144 msec positioned (B) over the tumor mass and (C)
on the contralateral side at the equivalent location. Resonances are labeled as: Lac = lactate; NAA
loss of 15% of the peak area of choline-containing compounds without significant changes
= N-acetyl aspartate; Cre = creatine; Cho = choline; MyoI = myoinositol. The horizontal axis is given
Intra-Axial
in creatine (Cre) or N-acetyl aspartate (NAA) (Sijens et al., 1997). In routine clinical in parts per million (ppm). Automated software used creatine as reference signal intensity quantified
Primary Brain practice, the spectra are obtained following the administration of intravenous contrast and the NAA as significantly decreased in the tumor and NAA/Cre ratio of 1.47 in the normal tissue. The Cerebral
Tumors presence of lactate, markedly reduced NAA and elevated choline, is characteristic of neoplasm. Neoplastic Disease

A3.3.6 A3.3.7
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A3.3.7 Primary Clinical Imaging Parameters for Sequence 7 (1H-MR R/O PRIMARY CEREBELLAR AND BRAINSTEM TUMOR BASIC
Spectroscopy) PROTOCOL 2
In pediatric patients, tumors are most commonly found in the infra-tentorial compartment
Patient position Supine (posterior fossa). In adults, the majority of lesions found in the posterior fossa are
Scan type PROBE-SV metastases; however, primary tumors do occur and the most common primary cerebellar
Voxel center See text (steps 2 and 4) tumor in adults is hemangioblastoma. The sequences described in Basic Protocol 1 should
Echo time (TE) 35 msec and 144 msec be obtained to evaluate posterior fossa tumors.
Repeat time (TR) 2000 msec
Flip angle (FA) 90° Set up patient and equipment
Voxel size (Δx, Δy, Δz) 2 cm, 2 cm, 2 cm 1. Follow the setup for Basic Protocol 1 (steps 1 to 10).
Number of acquisitions (Nacq) 8a and 128b 2. Run sequences 1 to 5.
Saturation pulses Water
Scan time ∼5 min
aWithout water suppression.
COMMENTARY
bWith water suppression.
Background Information Anticipated Results
The incidence of brain tumors varies accord- The goal in evaluating cerebral neoplasms
ing to age and recent literature suggests inci- is to image the entire lesion, localize it to the
Table A3.3.8 Primary Clinical Imaging Parameters for Sequence 8 (3-D Short dence trends have stabilized for all age groups appropriate compartment of the brain, and at-
TR Gradient Echo)
except in individuals greater than 85 years tempt to predict the underlying histology. Con-
Patient position Supine where it is increasing (15.7 cases of cancer per ventional MR sequences offer superb anatomi-
100,000 individuals). Improvements in diagno- cal detail and tissue characterization of brain
Scan type 3-D gradient echo
sis may explain the observed patterns in brain tumors. Contrast-enhanced imaging is critical
Imaging plane (orientation) Transverse
cancer trends (Legler et al., 1991). While com- to evaluate all brain tumors and helps to assess
Central slice or volume center Laser light centered on nasion puted tomography (CT) is often more readily the integrity of the blood-brain barrier and
Echo time (TE) 5 msec (or select the minimum available, and therefore the first examination depict intra-tumoral cystic or necrotic regions.
value when possible) performed for suspected intracranial pathol- Conventional MR techniques cannot reliably
Repeat time (TR) 35 msec ogy, magnetic resonance imaging (MRI) has distinguish high-grade from low-grade tumors.
Flip angle (FA) 45° proven to be invaluable in refining the diagno- Proton MR spectroscopy offers metabolic in-
Fields of view (FOVx, FOVy) 240 mm, 240 mm sis. MRI provides greater sensitivity for the formation and helps predict the relative aggres-
Resolution (Δx, Δy) 0.94 mm, 1.25 mm detection of brain tumors and detailed anatomi- siveness of a lesion by evaluating metabolite
Number of data points collected (Nx, Ny) 256, 192 cal information (Atlas and Lavi, 1996). Con- ratios (NAA/Cre) and identifying metabolites
Slice thickness (Δz) 1.5 mm trast enhanced imaging in more than one plane such as lactate (indicator of anaerobic gly-
Number of slices ∼40 is essential to evaluate intra-axial brain tumors. colysis; Meyer et al., 1999).
Slice gap 0
Number of acquisitions (Nacq) 1 Critical Parameters and Time Considerations
Troubleshooting The Basic Protocol 1 detailed in this unit
Swap read and phase encoding Yes
In pediatric patients and other subjects re- should take ∼30 min to complete. Proton MR
Saturation pulses Not applicable
quiring sedation, the authors feel that 3 planes Spectroscopy may add ∼10 to 15 min to the
Scan time ∼5 min
(transverse, coronal, sagittal) of imaging imaging session.
should be obtained in the post-contrast acqui-
sition when evaluating brain tumor patients. Literature Cited
Proton MR spectroscopy studies have shown Alsop, D.C. and Detre, J.A. 1998. Multisection cere-
5. After completion of the scan, another spectrum will appear on the screen and the that contrast administration will alter spectra bral blood flow MR imaging with continuous
values for the spectrum should be saved. arterial spin labeling. Radiology 208:410-416.
by decreasing the peak area of choline- contain-
ing compounds. Since tumors are frequently Atlas, S.W., Mark, A.S., Grossman, R.I., Gomori,
Sequence 8: 3-D short TR gradient echo J.M. 1988. Intracranial hemorrhage: Gradient-
heterogeneous histologically, placement of the
6. In cases where it is difficult to differentiate an intra-axial peripheral exophytic echo MR imaging at 1.5 T. Radiology 168:803-
voxel in different regions of the tumor may 807.
tumor from an extra-axial lesion impinging on the brain, obtain a thin slice 3-D
yield different MR spectroscopy results. The
gradient echo sequence to better evaluate the brain-tumor interface outlined in Atlas, S.W. and Lavi, E. 1996. Intra-axial brain
relatively large voxel size also results in volume tumors. In Magnetic Resonance Imaging of the
Table A3.3.8. averaging with the adjacent brain. In routine Brain and Spine, 2nd Edition. ( S.W. Atlas, ed.)
clinical use, the post-contrast images are used pp. 315-422. Lippincott-Raven, Philadelphia,
to identify the solid appearing portion of the PA.
lesion and the voxel is placed in the solid Legler, J.M., Gloeckler Ries, L.A., Smith, M.A.,
(enhancing portion) of the lesion. Warren, M.A., Warren, J.L., Heineman, E.F.,
Intra-Axial Kaplan, R.S., and Linet, M.S. 1991. Brain and
Primary Brain other central nervous system cancers: Recent Cerebral
Tumors Neoplastic Disease

A3.3.8 A3.3.9
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 trends in incidence and mortality. J. Natl. Cancer
Inst. 91:1382-1390.
Meyer, M.E., Pipas, J.M., Mamourian, A., Tosteson,
T.D., and Dunn, J.F. 1999. Classification of bi-
opsy-confirmed brain tumors using single-voxel
MR spectroscopy. Am. J. Neuroradiol. 20:117-
123.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Sijens, P.E., van den Bent, M.J., Nowak, P.J., van
Dijk, P., and Ouderk, M. 1997. 1H chemical shift
imaging reveals loss of brain tumor choline sig-
nal after administration of Gd-contrast. Magn.
Reson. Med. 37:222-225.
Singer, M.B., Atlas, S.W., and Drayer, B.P. 1998.
Subarachnoid space disease: Diagnosis with
fluid-attenuated inversion-recovery MR imaging
and comparison with gadolinium-enhanced spin
Intra-Axial
Primary Brain
Tumors
A3.3.10
echo MR imaging-blinded reader study. Radiol-
ogy 208:417-422.
Stejskal, E.O. and Tanner, J.E. 1965. Spin diffusion
measurements: spin echoes in the presence of
time-dependent field gradient. J. Chem. Phys.
42:288-292.
Tien, R.D., Flesberg, G.J., Friedman, H., Brown, H.,
and MacFall, J. 1994. MR imaging of high-grade
cerebral gliomas: value of diffusion-weighted
echoplanar pulse sequences. Am. J. Roentgenol.
162:671-677.
Contributed by Annette O. Nusbaum
New York Presbyterian Hospital
New York, New York
Scott W. Atlas
Stanford University Medical Center
Stanford, California
Current Protocols in Magnetic Resonance Imaging
Extra-Axial Tumors UNIT A3.4
Extra-axial tumors most commonly arise from the meninges, calvarium, or skull base.
Localization of a lesion as extra-axial or extra-cerebral in origin, has significant clinical
importance in terms of treatment planning and predicting prognosis. High-resolution
fat-suppressed imaging is crucial for specific evaluation of the skull base, and correlation
with computed tomography (CT) is often helpful in these cases. This unit presents the set
of magnetic resonance sequences used for imaging extra-axial tumors found in specific
locations; sequence modifications will be discussed where necessary. The sequences
described in this unit are based on the authors’ experience with a 1.5 T scanner (GE
Medical Systems), but can be expected to be equally applicable to other field strengths
and scanners from other manufacturers.
RULE OUT (R/O) EXTRA-AXIAL TUMOR BASIC
PROTOCOL
Spin-echo and fast spin-echo imaging (sequences 1 to 4) essentially provide all of the
information needed to make an accurate diagnosis when evaluating for primary extra-axial
tumors. The fluid-attenuated inversion recovery (FLAIR) images (sequence 4) and
post-contrast T1-weighted images (sequence 5) are most sensitive for evaluating tumor
spread particularly along the subarachnoid space (leptomeninges; Singer et al., 1998).
Sequences 1 to 5 comprise the preferred protocol.
Table A3.4.1 presents equipment parameters for imaging brain tumors.
NOTE: Be sure that technologists and nurses have immediate access to any equipment
such as crash carts or oxygen that may be necessary in the event of an emergency. Contrast
reactions are rare, but the resources are necessary.
Materials
Intravenous contrast agent (e.g., Magnevist, Omniscan, Prohance)
Normal saline (0.9% NaCl), sterile
Set up patient and equipment
1. Interview (screen) the patient to ensure that he or she has no contraindications such
as cardiac pacemaker or other implants containing ferromagnetic materials. Also be
sure to find out if the patient has any health conditions that may require the presence
of specific emergency equipment during the scanning procedure, or necessitate any
other precautions.
Generally standard screening forms are used for all patients scanned in a magnetic
resonance system.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
Table A3.4.1 Equipment Parameters for Imaging Brain Tumors
Coil Type Quadrature head coil
Gradient coil strength 25 mT/m (or whatever the system permits)
Cardiac gating No
Peripheral gating For safety only
Respiratory gating No
Respirator If required by patient
Oxygen If required by patient
Motion cushions No Cerebral
Neoplastic Disease
Contributed by Annette O. Nusbaum and Scott W. Atlas A3.4.1
Current Protocols in Magnetic Resonance Imaging (2001) A3.4.1-A3.4.11
Copyright © 2001 by John Wiley & Sons, Inc.
 composition of the items, it is best to exclude patients with metal implants; see Shellock Table A3.4.2 Primary Clinical Imaging Parameters for Sequence 1 (Localizer;
(1996) for discussion of what implants may be safely scanned using magnetic resonance. T1-Weighted)
Patients may be accompanied into the magnet room by a friend or family member, who can
Patient position Supine
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing. Scan type Spin echo
Imaging plane (orientation) Sagittal
2. If the procedure is a research protocol, have the patient sign any necessary consent Central slice or volume center Laser light centered on nasion
form. Echo time (TE) 11 msec (or select “minimum
full” echo time)
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
Receiver bandwidth (RBW) 10 kHz
that might be found in clothing.
Repeat time (TR) 500 msec
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating Flip angle (FA) 90°
and image artifacts. Fields of view (FOVx, FOVy) 240 mm, 240 mm
Resolution (Δx, Δy) 0.94 mm, 1.25 mm
5. Inform the patient about what will occur during the procedure, what he or she will
Number of data points collected (Nx, Ny) 256, 192
experience in the magnet, and how to behave, including the following:
Slice thickness (Δz) 5 mm
a. If earphones or headphones are used to protect the ears from the loud sounds Number of slices 20 (or as many as needed to cover
produced by the gradients, the patient will be asked to wear these, but will be able the region of interest)
to communicate with you at any time during the imaging. Slice gap 2 mm
Number of acquisitions (Nacq) 1
b. The patient will be given a safety squeeze-bulb or similar equipment to request
Swap read and phase encoding No
assistance at any time (demonstrate how this works).
Saturation pulses Not applicable
c. For good results the patient should not talk, and should avoid or minimize Scan time ∼1 min, 36 sec
swallowing or other movement, during each scan—i.e., as long as the banging
sounds continue. Between scans, talking and swallowing are allowed in most
cases, but should be avoided when comparative positional studies are being Table A3.4.3 Primary Clinical Imaging Parameters for Sequence 2
performed; the patient will be informed when this is the case. (T1-Weighted)

d. Nevertheless, the patient may call out at any time if he or she feels it necessary. Patient position Supine
6. Have the patient lie in the supine position on the table. Setup all monitoring equipment Scan type Spin echo
before or right after the patient lies down. Imaging plane (orientation) Transverse
Central slice or volume center Laser light centered on nasion
7. Center patient in the head coil at the region where key information is desired. Echo time (TE) 11 msec (or select “minimum
Constrain the head and neck to prevent motion. Generally, fix the patient’s head so full” echo time)
that the head is horizontal (not tilted) and the neck and head lie along the axis of the Receiver bandwidth (RBW) 10 kHz
patient table. Repeat time (TR) 500 msec
Flip angle (FA) 90°
8. If needed, place a pillow or other support under the knees to make the patient more Fields of view (FOVx, FOVy) 240 mm, 240 mm
comfortable. Resolution (Δx, Δy) 0.94 mm, 1.25 mm
9. Use the centering light to position the patient (centered on the nasion) and put him Number of data points collected (Nx, Ny) 256, 192
or her into the center of the magnet. Slice thickness (Δz) 5 mm
Number of slices 20 (or as many as needed to cover
Once this step has been performed, so long as the patient does not move on the table, the the region of interest)
table itself can be moved and then replaced in the same position as before without Slice gap 2.5 mm
jeopardizing the positioning of one scan relative to another.
Number of acquisitions (Nacq) 1
10. If the patient is unable to hold still, provide an appropriate sedative. Swap read and phase encoding Yes
Saturation pulses Not applicable
Sequence 1: Localizer Scan time ∼1 min, 36 sec
11. Run sequence 1 according to Table A3.4.2.

Sequence 2: T1-weighted scan

12. Run sequence 2 according to Table A3.4.3.

Extra-Axial Cerebral
Tumors Neoplastic Disease

A3.4.2 A3.4.3
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Sequence 3: T2-weighted scan Table A3.4.5 Primary Clinical Imaging Parameters for Sequence 4 (FLAIR)
13. Run sequence 3 according to Table A3.4.4.
Patient position Supine
Sequence 4: Fluid-attenuated inversion recovery (FLAIR) scan Scan type Inversion-recovery fast-spin echo
14. Run sequence 4 according to Table A3.4.5. Imaging plane (orientation) Transverse
Central slice or volume center Laser light centered on nasion
Sequence 5: Post-contrast imaging Echo time (TE) 120 msec (effective)
15. Remove the patient from the scanner. Do not allow the patient to move on the table. Receiver Bandwidth (RBW) 16 kHz
Establish an intravenous line from for injection of contrast agent. Attach this line Echo train length (ETL) 8
securely to the patient so that movement into or out of the magnet will not pull at the Repeat time (TR) 10,000 msec
patient’s arm. Move the patient back into the scanner. Inversion time (TI) 2200 msec
Flip angle (FA) 180°
It is preferable to insert the line prior to imaging and to leave the patient in the magnet
Fields of view (FOVx, FOVy) 240 mm, 240 mm
with no intervening motion between the scans run before contrast agent injection and those
run after injection. Resolution (Δx, Δy) 0.94 mm, 1.25 mm
Number of data points collected (Nx, Ny) 256, 192
16. Leaving the patient in the magnet, inject the contrast agent, flush the line with 10 ml Slice thickness (Δz) 5 mm
saline. Number of slices 20 (or as many as needed to cover
the region of interest)
A dose of 0.1 mmol/kg of contrast agent is usually given.
Slice gap 2.5 mm
A delay in scanning may actually be beneficial when evaluating for metastases, which is Number of acquisitions (Nacq) 1
one of the reasons we scan in multiple planes after intravenous contrast administration. Swap read and phase encoding Yes
17. Acquire the post-contrast images using the same parameters as in sequence 2 Saturation pulses Not applicable
(T1-weighted scan; Table A3.4.3). Scan time ∼5 min

In addition to the transverse plane, the coronal plane is routinely obtained, with the
following changes to the parameters in sequence 2: (a) flow compensation is on; (b) TE is MENINGIOMA ALTERNATE
20 msec (prolonged due to flow compensation gradients); and (c) it is not necessary to PROTOCOL 1
The majority of meningiomas are found incidentally and are located in the supratentorial
swap read and phase encoding directions. In patients requiring sedation and in all pediatric compartment (along the falx or cerebral convexity). Sequences 1 to 5 (see Basic Protocol)
patients, we also routinely obtain a third post-contrast plane of imaging (sagittal).
comprise the preferred protocol. The location of the lesion determines whether further
sequences should be obtained (i.e., if the tumor is adjacent to the orbit or internal auditory
canals, then additional sequences examining the specific region should be obtained). Not
Table A3.4.4 Primary Clinical Imaging Parameters for Sequence 3 uncommonly, meningiomas may be inseparable from the adjacent dural venous sinus and
(T2-Weighted)
the question arises whether or not the tumor has invaded the venous sinus. In these cases,
Patient position Supine magnetic resonance venography (MRV) is useful to confirm whether or not the venous
Scan type Fast spin echo sinus is patent (see Chapter A2 for evaluating the cerebral venous sinuses).
Imaging plane (orientation) Transverse
For procedure, see Basic Protocol.
Central slice or volume center Laser light centered on nasion
Echo time (TE) 102 msec (effective)
Receiver Bandwidth (RBW) 16 kHz NERVE SHEATH TUMORS—E.G., ACOUSTIC NEUROMA ALTERNATE
Echo train length (ETL) 8 PROTOCOL 2
Patients with nerve sheath tumors typically present with sensorineural hearing loss, arise
Repeat time (TR) 3600 msec within the vestibular or acoustic portions of cranial nerve VIII, and are found within the
Flip angle (FA) 90° internal auditory canals. High-resolution noncontrast (sequence 6) and post-contrast
Fields of view (FOVx, FOVy) 240 mm, 240 mm imaging with fat suppression (sequence 7) is essential to evaluate this region. Imaging of
Resolution (Δx, Δy) 0.94 mm, 1.25 mm the entire brain with pre-contrast (sequences 1, 3, and 4) and post-contrast sequences
Number of data points collected (Nx, Ny) 256, 192 (sequence 5; see Basic Protocol) is a standard component of the complete evaluation of
Slice thickness (Δz) 5 mm these patients.
Number of slices 20 (or as many as needed to cover
the region of interest) Set up patient and equipment
Slice gap 2.5 mm 1. Use the same equipment and perform the same patient setup as for the previous
Number of acquisitions (Nacq) 1 method (see Basic Protocol, steps 1 to 10).
Swap read and phase encoding Yes
Saturation pulses Not applicable 2. Run sequences 1, 3, and 4 in Basic Protocol.
Extra-Axial Scan time ∼1 min, 30 sec Cerebral
Tumors Neoplastic Disease

A3.4.4 A3.4.5
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A3.4.6 Primary Clinical Imaging Parameters for Sequence 6 Sequence 6: High resolution T2-weighted scan
(T2-Weighted) 3. Run sequence 6 (high-resolution transverse T2-weighted images through the internal
auditory canals) according to Table A3.4.6. This sequence nicely depicts the VII and
Patient position Supine
VIII cranial nerve roots.
Scan type Fast spin echo
Imaging plane (orientation) Transverse Sequence 7: High resolution T1-weighted scan (post-contrast)
Central slice or volume center Laser light centered on nasion 4. Repeat steps 15 and 16 in the Basic Protocol.
Echo time (TE) 102 msec (effective)
Receiver Bandwidth (RBW) 16 kHz 5. Run sequence 7 according to Table A3.4.7. Following the administration of contrast,
Echo train length (ETL) 16 obtain sequence 7 in both the transverse and coronal planes through the internal
Repeat time (TR) 4000 msec auditory canals.
Flip angle (FA) 90° 6. In addition, obtain transverse T1-weighted images (sequence 5, step 17 in the Basic
Fields of view (FOVx, FOVy) 180 mm, 180 mm Protocol) through the entire brain.
Resolution (Δx, Δy) 0.70 mm, 0.70 mm
Number of data points collected (Nx, Ny) 256, 256 PITUITARY/SELLA TURCICA MASSES ALTERNATE
Slice thickness (Δz) 3 mm PROTOCOL 3
Pituitary/sella turcica lesions are best evaluated by dedicated high resolution imaging of
Number of slices As many as needed to cover entire
skull base the sella in the sagittal and coronal planes. Non-contrast imaging including high resolution
Slice gap 0.5 mm sagittal T1-weighted localizer (sequence 8), coronal T1-weighted (sequence 9) and coronal
Number of acquisitions (Nacq) 3 T2-weighted images (sequence 6, with orientation changed to coronal and with no swap
Swap read and phase encoding Yes read and phase encoding directions) beautifully depict complicated anatomic regions.
Saturation pulses None except fat Contrast enhanced images (sequence 9) are essential and may demonstrate small pituitary
Fat suppression Yes lesions (microadenomas) not seen otherwise. Post-contrast imaging (sequence 5) of the
Scan time 3 min, 20 sec entire brain is always performed.
Set up patient and equipment
1. Use the same equipment and perform the same patient setup as for the previous
Table A3.4.7 Primary Clinical Imaging Parameters for Sequence 7
(Post-Contrast T1-Weighted) method (see Basic Protocol, steps 1 to 10).

Patient position Supine

Sequence 8: T1-weighted localizer
Scan type Spin echo
2. Run sequence 8 (high resolution sagittal T1-weighted localizer of the sella) according
Imaging plane (orientation) Transverse and coronal
to Table A3.4.8.
Central slice or volume center Laser light centered on nasion
Table A3.4.8 Primary Clinical Imaging Parameters for Sequence 8
Echo time (TE) 20 msec (T1-Weighted Localizer)
Receiver Bandwidth (RBW) 10 kHz
Repeat time (TR) 500 msec Patient position Supine
Flip angle (FA) 90° Scan type Spin echo
Fields of view (FOVx, FOVy) 180 mm, 180 mm Imaging plane (orientation) Sagittal
Resolution (Δx, Δy) 0.70 mm, 0.94 mm Central slice or volume center Laser light centered on nasion
Number of data points collected (Nx, Ny) 256, 192 Echo time (TE) 20 msec
Slice thickness (Δz) 3 mm Receiver Bandwidth (RBW) 10 kHz
Number of slices As many as needed to cover the Repeat time (TR) 500 msec
area of interest Flip angle (FA) 90°
Slice gap 0.5 mm Fields of view (FOVx, FOVy) 160 mm, 160 mm
Number of acquisitions (Nacq) 3 Resolution (Δx, Δy) 0.63 mm, 0.83 mm
Swap read and phase encoding Yes, when it is transverse Number of data points collected (Nx, Ny) 256, 192
Saturation pulses None except fat Slice thickness (Δz) 3 mm
Fat suppression Yes Number of slices 20
Scan time 5 min, 20 sec Slice gap 0
Number of acquisitions (Nacq) 3
Swap read and phase encoding No
Saturation pulses Not applicable
Slice series Interleaved
Extra-Axial Cerebral
Tumors Scan time 5 min Neoplastic Disease

A3.4.6 A3.4.7
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A3.4.9 Primary Clinical Imaging Parameters for Sequence 9 Table A3.4.10 Primary Clinical Imaging Parameters for Sequence 11 (DWI)
(T1-Weighted)
Patient position Supine
Patient position Supine Scan type Diffusion-weighted EPI-spin echo
Scan type Spin echo Imaging plane (orientation) Transverse
Imaging plane (orientation) Coronal Central slice or volume center Laser light centered on nasion
Central slice or volume center Laser light centered on nasion Echo time (TE) 125 msec
Echo time (TE) 20 msec Repeat time (TR) 10,000 msec
Receiver Bandwidth (RBW) 10 kHz Flip angle (FA) 90°
Repeat time (TR) 500 msec Fields of view (FOVx, FOVy) 240 mm, 240 mm
Flip angle (FA) 90° Resolution (Δx, Δy) 1.88 mm, 1.88 mm
Fields of view (FOVx, FOVy) 180 mm, 180 mm Number of data points collected (Nx, Ny) 128, 128
Resolution (Δx, Δy) 0.70 mm, 0.94 mm Slice thickness (Δz) 5 mm
Number of data points collected (Nx, Ny) 256, 192 Number of slices 30 with the diffusion gradients in
Slice thickness (Δz) 3 mm 3 orthogonal directions
Number of slices As many as needed to cover the Slice gap 0
area of interest Number of acquisitions (Nacq) 1
Slice gap 0.5 mm Swap read and phase encoding No (to decrease peripheral nerve
Number of acquisitions (Nacq) 3 stimulation)
Swap read and phase encoding No Read direction Left–right
Saturation pulses No Saturation pulses Not applicable
Fat suppression No b-value 1000 sec/mm2
Scan time 5 min, 20 sec δ 31 msec
Δ 36.6 msec
Scan time ∼45 sec
3. Run sequence 3 (transverse T2-weighted of the entire brain) in Basic Protocol.

Sequence 9: High resolution T1-weighted scan sequence first acquires a set of transverse T2-weighted images (Beneviste et al., 1992)
4. Run sequence 9 (coronal images through the sella) according to Table A3.4.9. without the diffusion gradients (thus b = 0). The diffusion gradients are then applied
sequentially in three orthogonal axes in the directions of slice select, phase encoding, and
Sequence 10: High resolution T2-weighted scan read gradients to generate a set of 3 orthogonal-axis diffusion-weighted echo-planar
5. Run sequence 6 (coronal T2-weighted images of the sella) based on Table A3.4.6. images (EPI). The b-value (diffusion sensitivity) is 1000 sec/mm2. A set of isotropic
images are calculated and displayed.
6. Repeat steps 15 and 16 in the Basic Protocol.
7. Run sequence 9 (post-contrast coronal T1-weighted images through the sella) accord- 2. Obtain the DWI using parameters in Table A3.4.10 (sequence 11) to demonstrate that
ing to Table A3.4.9. the lesion remains iso-intense to CSF.

8. Run sequence 5 (transverse images of the entire brain). 3. Repeat steps 15 to 17 (sequence 5) in the Basic Protocol.

ALTERNATE ARACHNOID CYST EPIDERMOID/DERMOID ALTERNATE

PROTOCOL 4
Arachnoid cysts are benign, developmental anomalies of the subarachnoid space and
typically demonstrate signal characteristics identical to that of cerebrospinal fluid (CSF).
Occasionally, difficulties arise in differentiating an arachnoid cyst from an epidermoid
cyst. FLAIR images or diffusion-weighted images may be useful to identify slight
differences in signal intensity from that of CSF, which would indicate an epidermoid cyst,
rather than an arachnoid cyst.
Set up patient and equipment
1. Follow steps 1 to 10 in the Basic Protocol and run sequences 1 to 4.
Sequence 11: Diffusion-weighted imaging (DWI)
PROTOCOL 5
Epidermoid and dermoid cysts are congenital lesions frequently found along the cerebel-
lopontine angle, supra- and para-sellar regions, and middle cranial fossa. Epidermoid
cysts appear similar, but not identical, in signal intensity to that of CSF and in the majority
of cases epidermoids may be differentiated from arachnoid cysts by MRI. Dermoid cysts
contain fat within them and fat suppressed images may be useful to confirm the contents
of this lesion.
Set up patient and equipment
1. Repeat steps 1 to 10 and run sequences 1 to 5 (Basic Protocol).
2. Obtain fat suppressed T1-weighted images (parameters in sequence 2) in the trans-
verse or sagittal plane.

Extra-Axial Multislice single-shot spin echo echo-planar diffusion-weighted sequences are obtained Cerebral
Tumors with two diffusion gradient pulses applied before and after the 180° pulse. The pulse Neoplastic Disease

A3.4.8 A3.4.9
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
ALTERNATE CHONDROSARCOMA/CHORDOMA Literature Cited spin echo MR imaging-blinded reader study.
PROTOCOL 6 Beneviste, H., Hedlund, L., and Johnson, G. 1992. Radiology 208:417-422.
Primary skull base lesions such as chordoma and chondrosarcoma can be infiltrative and Mechanism of detection of acute cerebral is-
spread to involve skull base foramina. The main purpose of imaging is to determine the chemia in rats by diffusion-weighted magnetic Contributed by Annette O. Nusbaum
full extent of the lesion (i.e., cranial nerve or meningeal involvement). The entire brain resonance microscopy. Stroke 23:746-754.
New York Presbyterian Hospital
should always be imaged (sequences 1 to 5). Fat suppressed high-resolution imaging, Shellock, F.G. 1996. Pocket Guide to MR Proce- New York, NY
dures and Metallic Objects. (F.G. Shellock, ed.).
including T2-weighted images (sequence 6) and post-contrast T1-weighted images (se- Lippincott-Raven, Philadelphia. Scott W. Atlas
quence 9) is essential when evaluating tumor involvement of skull base structures. Stanford University Medical Center
Singer, M.B., Atlas, S.W., and Drayer, B.P. 1998.
Subarachnoid space disease: Diagnosis with Stanford, CA
Set up patient and equipment fluid-attenuated inversion-recovery MR imag-
1. Use the same equipment and perform the same patient setup as for the previous ing and comparison with gadolinium-enhanced
method (see Basic Protocol, steps 1 to 10).
2. Run sequences 1 to 4 (Basic Protocol) to evaluate the entire brain.
3. For high-resolution imaging of the skull base, run sequence 6 according to Table
A3.4.6 (transverse T2-weighted images) and sequence 9 according to Table A3.4.9
(with imaging plane changed to transverse, and with read and phase encoding
directions swapped). The entire extent of the lesion must always be included.
4. Repeat steps 15 and 16 in Basic Protocol.
5. Run post-contrast T1-weighted scans in two planes (transverse and coronal) according
to Table A3.4.9 adding the fat suppression option.
6. Run sequence 5 (post-contrast T1-weighted scan of the entire brain, step 17) in the
transverse plane.

COMMENTARY
Background Information
MRI is the preferred imaging modality for
defining extra-axial tumors. Obtaining detailed
information regarding the entire extent of dis-
ease is critical to the determination of appropri-
ate treatment. High resolution computed to-
mography (CT) of the skull base is often an
integral part of the work-up of a skull-base
lesion. MRI provides superior contrast and is
essential for documenting meningeal disease
and spread along cranial nerves (see UNITS A7.1-
A7.4). Intravenous contrast is critical to help
predict the full extent of disease.
Critical Parameters and
Troubleshooting
Fluid attenuated inversion recovery
(FLAIR) or post-contrast T1-weighted images
provide the greatest sensitivity in detecting ex-
tra-axial spread of tumor to the meninges.
When evaluating the skull base, fat-suppressed
high-resolution imaging (including T2-
weighted and post-contrast T1-weighted im-
ages) is essential to clearly delineate the full
extent of disease. With a resolution of less than
1 mm in the read and phase encoding directions,
Extra-Axial there is superb visualization of the fine intricate
Tumors
structures of the skull base (i.e., foramina and
cranial nerves). Fat-suppressed T1-weighted
images are critical to evaluate for abnormal
enhancement of cranial nerves, since high sig-
nal intensity on the T1-weighted images from
fat found along the skull base or chemical-shift
artifact from fat-containing structures can ob-
scure enhancement of cranial nerves. Inhomo-
geneities in the magnetic field can result in
incomplete fat suppression or unintended sup-
pression of water.
Anticipated Results
The goal of imaging when evaluating for
extra-axial tumors is to accurately localize and
detect the full extent of disease. Magnetic reso-
nance imaging offers superb anatomical detail
and tissue characterization of lesions. Fat sup-
pressed high-resolution MR imaging is a key
component of the skull-base examination.
FLAIR images and post-contrast T1-weighted
images are most sensitive for detecting
subarachnoid space disease.
Time Considerations
The majority of these protocols should take
∼30 min to complete. Cerebral
Neoplastic Disease

A3.4.10 A3.4.11
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
Monitoring Tumor Therapy UNIT A3.5 Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
This unit presents the preferred set of MR sequences for imaging the brain following the
institution of chemotherapy, radiation therapy, or surgery for brain tumors (see Basic 2. If the procedure is a research protocol, have the patient sign any necessary consent
Protocol 1) and specific modifications will be discussed where necessary (see Basic form.
Protocol 2). The sequences described in this unit are based on the authors’ experience 3. Have the patient remove all jewelry and change into a gown to eliminate any metal
with a 1.5 T scanner (Echospeed GE Medical Systems), but can be expected to be equally that might be found in clothing.
applicable to other field strengths and scanners from other manufacturers. Sequences 1 4. Have the patient wash off any mascara and other makeup to avoid local tissue heating
to 5 comprise the preferred Basic Protocol 1. and image artifacts.
5. Inform the patient about what will occur during the procedure, what he or she will
FOLLOW-UP POST SURGERY BASIC experience in the magnet, and how to behave, including the following:
PROTOCOL 1
One of the most important issues for initial post-operative MRI of the brain is that it should a. If earphones or headphones are used to protect the ears from the loud sounds
be performed within 24 to 48 hr of surgery to assess for any enhancing residual tumor. produced by the gradients, the patient will be asked to wear these, but will be able
Noncontrast T1-weighted images are important to obtain in the same imaging plane as the to communicate with you at any time during the imaging.
post-contrast T1-weighted images to assess for residual tumor and not mistake blood
b. The patient will be given a safety squeeze-bulb or similar equipment to request
products for enhancing tumor. Sequences 1 to 5 comprise the preferred protocol.
assistance at any time (demonstrate how this works).
Table A3.5.1 lists the hardware necessary to perform the procedure, along with appropri- c. For good results the patient should not talk, and should avoid or minimize
ate parameters. swallowing or other movement, during each scan—i.e., as long as the banging
sounds continue. Between scans, talking and swallowing are allowed in most
NOTE: Be sure that technologists and nurses have immediate access to any equipment cases, but should be avoided when comparative positional studies are being
such as crash carts or oxygen that may be necessary in the event of an emergency. performed; the patient will be informed when this is the case.
Reactions to contrast agents are rare, but the resources are necessary.
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
Materials
6. Have the patient mount onto the table in the supine position. Either before or right
Intravenous contrast agent (e.g., Magnevist, Omniscan, Prohance) after the patient lies down, set up any triggering devices or other monitoring
Normal saline (0.9% NaCl), sterile equipment that is to be used.
Setup patient and equipment 7. Center the patient in the head coil at the region where the key information is desired.
1. Interview (screen) the patient to ensure that he or she has no contraindications to the Make sure the head and neck are constrained to prevent motion.
MRI examination, such as a cardiac pacemaker or other implants containing ferro- Generally the patient’s head is fixed so that the head is horizontal (not tilted) and the neck
magnetic materials. Question the patient regarding any health conditions that may and head lie along the axis of the patient table.
require the presence of specific emergency equipment during the scanning procedure, 8. If needed, place a pillow or other support under the knees to make the patient more
or necessitate any other precautions. comfortable.
Generally standard screening forms are used for all patients scanned in a magnetic 9. Use the centering light to position the patient (centered on the nasion) and put him
resonance system. or her into the center of the magnet.
The presence of ferromagnetic metals may be a health hazard to the patient when he or she Once this step has been performed, so long as the patient does not move on the table, the
is inside the magnet, and will also affect the imaging. If in doubt as to the exact composition table itself can be moved and then replaced in the same position as before without
of the items, it is best to exclude patients with any metal implants; see Shellock (1996) for jeopardizing the positioning of one scan relative to another.
discussion of what implants may be safely scanned using magnetic resonance.
10. If the patient is unable to hold still, provide an appropriate sedative.
Table A3.5.1 Equipment Parameters for Imaging Brain Tumors Sequence 1: Localizer
11. Run sequence 1 according to Table A3.5.2.
Coil type Quadrature head coil
Gradient coil strength 25 mT/m (or whatever the system permits) Sequence 2: T1-weighted scan
Cardiac gating No 12. Run sequence 2 according to Table A3.5.3.
Peripheral gating For safety only Sequence 3: T2-weighted scan
Respiratory gating No 13. Run sequence 3 according to Table A3.5.4.
Respirator If required by patient
Oxygen If required by patient Sequence 4: Fluid attenuated inversion recovery (FLAIR) scan
Motion cushions No Cerebral 14. Run sequence 4 according to Table A3.5.5.
Monitoring
Neoplastic Disease Tumor Therapy
Contributed by Annette O. Nusbaum and Scott W. Atlas A3.5.1 A3.5.2
Current Protocols in Magnetic Resonance Imaging (2001) A3.5.1-A3.5.7
Copyright © 2001 by John Wiley & Sons, Inc. Supplement 1 Supplement 1 Current Protocols in Magnetic Resonance Imaging
 Table A3.5.2 Primary Clinical Imaging Parameters for Sequence 1 Table A3.5.4 Primary Clinical Imaging Parameters for Sequence 3
(T1-Weighted) (T2-Weighted)

Patient position Supine Patient position Supine

Scan type Spin echo Scan type Fast spin echo
Imaging plane (orientation) Sagittal Imaging plane (orientation) Transverse
Central slice or volume center Laser light centered on nasion Central slice or volume center Laser light centered on nasion
Echo time (TE) 11 msec (or select “minimum Echo time (TE) 102 msec (effective)
full” echo time) Receiver bandwidth (RBW) 16 kHz
Receiver bandwidth (RBW) 10 kHz Echo train length (ETL) 8
Repeat time (TR) 500 msec Repeat time (TR) 3600 msec
Flip angle (FA) 90° Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 240 mm, 240 mm Fields of view (FOVx, FOVy) 240 mm, 240 mm
Resolution (Δx, Δy) 0.94 mm, 1.25 mm Resolution (Δx, Δy) 0.94 mm, 1.25 mm
Number of data points collected (Nx, Ny) 256, 192 Number of data points collected (Nx, Ny) 256, 192
Slice thickness (Δz) 5 mm Slice thickness (Δz) 5 mm
Number of slices 20 or as many as needed to cover Number of slices 20 or as many as needed to cover
the region of interest the region of interest
Slice gap 2 mm Slice gap 2.5 mm
Number of acquisitions (Nacq) 1 Number of acquisitions (Nacq) 1
Swap read and phase encoding No Swap read and phase encoding Yes
Saturation pulses Not applicable Saturation pulses Not applicable
Scan time 1 min, 36 sec Scan time 1 min, 30 sec

Table A3.5.5 Primary Clinical Imaging Parameters for Sequence 4 (Fast

Table A3.5.3 Primary Clinical Imaging Parameters for Sequence 2 FLAIR)
(T1-Weighted)
Patient position Supine
Patient position Supine Scan type Inversion recovery fast spin echo
Scan type Spin echo Imaging plane (orientation) Transverse
Imaging plane (orientation) Transverse Central slice or volume center Laser light centered on nasion
Central slice or volume center Laser light centered on nasion Echo time (TE) 120 msec (effective)
Echo time (TE) 11 msec (or select “minimum Receiver bandwidth (RBW) 16 kHz
full” echo time) Echo train length (ETL) 8
Receiver bandwidth (RBW) 10 kHz Repeat time (TR) 10,000 msec
Repeat time (TR) 500 msec Inversion time (TI) 2200 msec
Flip angle (FA) 90° Flip angle (FA) 180°
Fields of view (FOVx, FOVy) 240 mm, 240 mm Fields of view (FOVx, FOVy) 240 mm, 240 mm
Resolution (Δx, Δy) 0.94 mm, 1.25 mm Resolution (Δx, Δy) 0.94 mm, 1.25 mm
Number of data points collected (Nx, Ny) 256, 192 Number of data points collected (Nx, Ny) 256, 192
Slice thickness (Δz) 5 mm Slice thickness (Δz) 5 mm
Number of slices 20 or as many as needed to cover Number of slices 20 or as many as needed to cover
the region of interest the region of interest
Slice gap 2.5 mm Slice gap 2.5 mm
Number of acquisitions (Nacq) 1 Number of acquisitions (Nacq) 1
Swap read and phase encoding Yes Swap read and phase encoding Yes
Saturation pulses Not applicable Saturation pulses Not applicable
Scan time 1 min, 36 sec Scan time ∼5 min

Cerebral Monitoring
Neoplastic Disease Tumor Therapy

A3.5.3 A3.5.4
Current Protocols in Magnetic Resonance Imaging Supplement 1 Supplement 1 Current Protocols in Magnetic Resonance Imaging
Sequence 5: Post-contrast imaging
15. Remove the patient from the scanner. The patient should not move on the table.
Establish an intravenous line from which the contrast agent can be injected, and attach
this line securely to the patient so that movement into or out of the magnet will not
pull at the patient’s arm. Move the patient back into the scanner.
It is preferable to insert the line prior to imaging and to leave the patient in the magnet,
with no intervening motion, between the scans run before contrast agent injection and those
run after injection.
16. Leaving the patient in the magnet, inject the contrast agent, flush the line with 10 ml
saline.
A dose of 0.1 mmol/kg of contrast agent is usually given.
A delay in scanning may actually be beneficial when evaluating for metastases, which is
one of the reasons we scan in multiple planes after intravenous contrast administration.
17. Acquire the post-contrast images using the same parameters as sequence 2 (T1-
weighted).
In addition to the transverse plane, the coronal plane is routinely obtained, with the
following changes to the parameters in Sequence 2: (a) flow compensation is on; (b) TE is
20 msec (prolonged due to flow compensation gradients); and (c) it is not necessary to
swap read and phase encoding directions. In patients requiring sedation and in all pediatric
patients, we also routinely obtain a third post-contrast plane of imaging (sagittal).
(stimulated-echo acquisition mode) or PRESS (point-resolved spectroscopy). The se-
quence parameters can be set by the user (TR, TE, voxel size). Short TE spectroscopy has
the advantage of providing more information, since more peaks are visible for analysis
and detection of lipid-breakdown products is maximized. Long TE spectra have less
baseline distortion and are easier to quantify. Since there is a √ ⎯⎯2 decrease in the
signal-to-noise ratio with STEAM sequences, the PRESS sequence is usually preferred.
Spectra obtained following the administration of intravenous contrast may be altered
(Sijens et al., 1997). In particular, studies have shown a mean loss of 15% of the peak area
of choline-containing compounds without significant changes in creatine (Cre) or N-ace-
tyl aspartate (NAA) (Sijens et al., 1997). In routine clinical practice, the spectra are
obtained following the administration of intravenous contrast and the voxel is placed over
the solid (enhancing) portion of the lesion so that theoretically signal is acquired from
viable tumor tissue.
4. Perform automated shimming and water suppression, center the voxel over the
enhancing portion of the lesion, and run sequence 6 according to Table A3.5.6.
5. After completion of the scan, a spectrum will appear on the screen and the values for
the spectrum should be saved.
6. Place the voxel in the contralateral hemisphere and repeat the scan.
7. After completion of the scan, another spectrum will appear on the screen and the
values for the spectrum should be saved.

RADIATION NECROSIS VERSUS TUMOR RECURRENCE BASIC Table A3.5.6 Primary Clinical Imaging Parameters for Sequence 6 (1H MR
PROTOCOL 2 Spectroscopy)
Radiation necrosis and recurrent tumor are indistinguishable by conventional MR tech-
niques and on the basis of paramagnetic enhancement, since both may appear mass like Patient position Supine
and enhance in most circumstances (Grossman et al., 1988). The time interval between Scan type PROBE
radiation therapy and the MR scan is an essential part of making an intelligent presumptive Voxel Center See text (steps 4 and 6)
diagnosis of either radiation necrosis or recurrent tumor (Atlas and Lavi, 1996). Several Echo time (TE) 35 msec and 144 msec
studies suggest that proton MR spectroscopy may be a potentially useful method for Repeat time (TR) 2000 msec
differentiating tumor recurrence from radiation necrosis (Taylor et al., 1996; Nelson et Flip angle (FA) 90°
al., 1999; Ricci et al., 2000). Whole brain chemical-shift imaging (CSI) maps may also Voxel size (Δx, Δy, Δz) 2 cm, 2 cm, 2 cm
be potentially useful in the investigation of brain tumors by providing visual metabolite Number of acquisitions (Nacq) 8a and 128b
maps, rather than spectral data. Saturation pulses Water
Scan time ∼5 min
Setup patient and equipment aWithout water suppression.
1. Use the same equipment and perform the same setup as for the previous method (see bWith water suppression.

Basic Protocol 1).

2. Run sequences 1 to 5 (see Basic Protocol 1). COMMENTARY
Background Information
3. Localize the lesion on a transverse image (typically the post-contrast transverse
Magnetic resonance imaging is a well-es-
T1-weighted image). tablished tool in the evaluation of brain tumors
The voxel placement requires physician input to ensure the correct region of interest is by offering superb anatomical detail and tissue
being studied. Two separate scans are performed: one with the voxel placed over the area characterization of brain lesions. Conventional
in question (typically the enhancing portion of the lesion), the other scan is performed with MRI techniques have limitations in the ability
the voxel placed in the contralateral hemisphere (in a “normal appearing” analogous area to discriminate between recurrent tumor and
of the brain). The screen should always be saved to document placement of the voxel. radiation necrosis as well as determining the
exact limits of tumor extension. Postoperative
Sequence 6: Proton (1H) MR Spectroscopy patients being studied by MRI for evaluation of
The proton brain exam (PROBE) on GE systems is an automated acquisition that is widely residual or recurrent tumor do not pose as
available and is performed using the single voxel technique (8 cm3). A high field system difficult a problem as patients who receive
Cerebral Monitoring radiation therapy. Contrast enhanced imaging
is required to obtain good spectra. There are 2 possible localization methods: STEAM Neoplastic Disease Tumor Therapy
is critical to the evaluation of post-therapy brain
tumors. Proton MR spectroscopy may be a
potentially useful tool for differentiating tumor
recurrence from radiation necrosis (Taylor et
al., 1996; Nelson et al., 1999).
Critical Parameters and
Troubleshooting
A critical component of the MR examina-
tion is an adequate history of the therapeutic
intervention instituted. Precontrast T1-
weighted images frequently demonstrate sub-
acute blood products and need to be carefully

A3.5.5 A3.5.6
Current Protocols in Magnetic Resonance Imaging Supplement 1 Supplement 1 Current Protocols in Magnetic Resonance Imaging
compared with post-contrast images to assess
for enhancing portions of the area of interest.
In pediatric patients and other subjects requir-
ing sedation, 3 planes (transverse, coronal,
sagittal) of imaging should be obtained in the
post-contrast acquisition. Post-operative imag-
ing in brain tumors should ideally be performed
within 24 to 48 hr of the operation. Long-term
follow-up imaging relies on the availability of
prior imaging studies for comparison, and his-
tory regarding treatment protocols.
Proton MR spectroscopy studies have
shown that contrast administration will alter
spectra by decreasing the peak area of choline-
containing compounds. Tumors are frequently
heterogeneous histologically and placement of
the voxel in different regions of the tumor may
yield different MR spectroscopy results. The
relatively large voxel size also results in volume
averaging with the adjacent brain. In routine
clinical use, the post-contrast images are used
to identify the solid appearing portion of the
lesion and the voxel is placed in the solid
(enhancing portion) of the lesion (Ricci et al.,
2000).
Anticipated Results
The goal in evaluating post-therapy brain
tumors is to image and localize the site of
documented tumor and assess for residual/re-
current tumor. The dilemma of distinguishing
radiation necrosis from residual/recurrent tu-
mor remains. Correlation with the timing of a
given tumor therapy is helpful, however fre-
quently a tissue biopsy is needed to solve the
issue.
Time Considerations
The Basic Protocol 1 should take ∼30 min
to complete. Proton MR spectroscopy (see Ba-
sic Protocol 2) adds ∼10 to 15 min to the
imaging session.
Current Protocols in Magnetic Resonance Imaging
Literature Cited
Atlas, S.W. and Lavi, E. 1996. Intra-axial brain
tumors. In Magnetic Resonance Imaging of the
Brain and Spine, 2nd Edition (S.W. Atlas, ed.)
pp. 315-422. Lippincott-Raven, Philadelphia.
Grossman, R.I., Hecht-Leavitt, C.M., Evans, S.M.,
Lenkinski, R.E., Holland, G.A., Van Winkle, T.J.,
McGrath, J.T., Curran, W.J., Shetty, A., and
Joseph, P.M. 1988. Experimental radiation in-
jury: Combined MR imaging and spectroscopy.
Radiology 169:305-309.
Nelson, S.J., Vigneron, D.B., and Dillon, W.P. 1999.
Serial evaluation of patients with brain tumors
using volume MRI and 3D 1H MRSI. N.M.R.
Biomed. 12:123-138.
Ricci, P.E., Pitt, A., Keller, P.J., Coons, S.W., and
Heiserman, J.E. 2000. Effect of voxel position
on single-voxel MR spectroscopy findings. Am.
J. Neuroradiol. 21:367-374.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Sijens, P.E., van den Bent, M.J., Nowak, P.J., van
Dijk, P., and Ouderk, M. 1997. 1H chemical shift
imaging reveals loss of brain tumor choline sig-
nal after administration of Gd-contrast. Magn.
Reson. Med. 37:222-225.
Taylor, J.S., Langston, J.W., Reddick, W.E., Wing-
sley, P.B., Ogg, R.J., Pui, M.H., Kun, L.E.,
Jenkins, J.J. 3rd, Chen, G., Ochs, J.J., Sanford,
R.A., and Heideman, R.L. 1996. Clinical value
of proton magnetic resonance spectroscopy for
differentiating recurrent or residual brain tumor
from delayed cerebral necrosis. Int. J. Radiat.
Oncol. Biol. Phys. 36:1251-1261.
Contributed by Annette O. Nusbaum
New York Presbyterian Hospital
New York, New York
Scott W. Atlas
Stanford University Medical Center
Stanford, California
Cerebral
Neoplastic Disease
A3.5.7
Supplement 1
CHAPTER A4
Infectious Diseases of the Brain
INTRODUCTION
n this chapter, magnetic resonance imaging (MRI) for the accurate and safe detection
I and characterization of infectious diseases of the brain is covered, with an additional
section on the imaging of cerebral trauma included. Cerebral infections may take many
forms, including cerebritis, abscess, or meningitis, and can involve any area of the central
nervous system. Though rare today, these infections are of clinical concern and interest
in specific patient populations (e.g., the immunocompromised). The timely and accurate
diagnosis of such infections is also critical, as they may be rapidly progressive, often
culminating in serious or permanent neurologic deficiencies or death. Cerebral infec-
tions are fortunately treatable in the majority of patients if detected and addressed early
enough, thus placing a premium on quality imaging studies (primarily MRI) for the
correct identification and triage of these patients.
In the first unit, UNIT A4.1, the imaging of cerebral abscess is outlined, utilizing a protocol
that is simple, versatile, and largely applicable to the remainder of the chapter. Also
presented are several more complex studies utilizing spectroscopy or diffusion/perfusion
sequences to add specificity in the diagnosis of patients with this condition.
For the imaging of patients suspected of meningitis, the protocol outlined in the first
unit (UNIT A4.1) is largely sufficient; however, in UNIT A4.2, several additional (and often
quite subtle) “tricks” to elucidate the findings of meningeal infection are described. Also,
the optional use of a venous angiographic sequence for the detection of dural sinus
thrombosis, a serious and not uncommon sequela of meningitis, is outlined.
The emergence of the AIDS epidemic has brought with it a marked increase in the need
for imaging studies which are both accurate and reproducible for these patients at high
risk for atypical cerebral infections and tumors. In UNIT A4.3, preferences for optimizing
the MRI examination in this patient population are presented.
Finally, though not an inflammatory condition, cerebral trauma presents a particular set
of challenging clinical problems regarding the patient with neurologic abnormalities that
cannot clearly be explained by the trauma sustained, or that may indeed have preceded and
possibly caused the traumatic event. The rendering of an accurate prognosis for the patient
with cerebral trauma is also an overlooked but important area for optimization by means
of directed and comprehensive imaging. MRI adds valuable diagnostic information in
the evaluation of head-injured patients because of its high sensitivity for traumatic brain
lesions. The optimal MR imaging protocol depends on the timing of the scan (acute
versus chronic), the ability of the patient to cooperate during the scan, and the clinical
question to be answered. In UNIT A4.4 a head-injury protocol is presented. Based upon the
examination outlined in UNIT A4.1, volumetric T1 -weighted gradient echo, diffusion echo
planar, MR angiography (MRA), and spectroscopy sequences found to have utility in
certain cases are also described.
The sequences described in this chapter are optimized for a 1.5 T magnet, but can easily
be modified to perform high-quality imaging on most MR systems.
Andrew E. Auber and Clifford Belden Infectious Diseases
of the Brain
Contributed by Andrew E. Auber and Clifford Belden A4.0.1
Current Protocols in Magnetic Resonance Imaging (2005) A4.0.1
C 2005 by John Wiley & Sons, Inc.
Copyright  Supplement 13
Brain Abscess
Magnetic resonance imaging (MRI) in cerebral abscess, as with most other forms of
intracranial inflammatory or infectious diseases, is a powerful though largely nonspecific
diagnostic tool. This unit presents a variant of a previously published (Castillo, 1998)
standard imaging protocol, to include gadolinium-enhanced sequences, for imaging of
these patients. Several optional sequences, including diffusion (dMRI), perfusion (pMRI),
and spectroscopic (MRS) sequences are outlined that can be employed should patient
tolerance allow and if specific clinical situations exist to be further clarified. The
parameters given here are derived from experience at 1.5 T and may need to be altered
slightly depending on the field strength available and the specific equipment manufac-
turer.
IMAGING OF CEREBRAL ABSCESS
This unit presents what amounts to the authors’ routine MR head examination for patients
with intracranial inflammatory disease, with gadolinium contrast agent sequences. The
total number of sequences presented is seven—scout, T1-weighted sagittal and transverse,
fluid-attenuated inversion recover (FLAIR) and fast spin-echo (FSE) T2-weighted trans-
verse, and post-contrast T1-weighted transverse and coronal. This examination may be
shortened considerably by limiting the use of two planes for post-contrast imaging to a
single transverse plane for exams with a lower clinical suspicion for disease, or if review
of the images obtained through the transverse gadolinium-enhanced images proves
unremarkable. The sequences described herein are based on the authors’ experience with
a Marconi Medical Systems 1.5 T scanner, but are expected to be equally applicable to
machines from other manufacturers.
Table A4.1.1 lists the hardware necessary to perform the examination, along with
appropriate parameters.
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as a crash cart or oxygen. Also ensure only magnetic field compatible oxygen
tanks are utilized if wall-source oxygen is unavailable.
Materials
Intravenous MRI contrast agent: gadolinium chelate (e.g., Magnevist, Omniscan,
or Prohance)
Normal sterile saline (0.9% NaCl)
Table A4.1.1 Equipment Parameters for Contrast-Enhanced Head Imaging
Sequences
Coil type Head
Gradient coil strength 25 mT/m (or whatever the system permits)
Gating (cardiac, respiratory, peripheral) No
Respirator or oxygen If required by patient
Motion cushions Recommended
Contrast agents Yes
Contributed by Andrew E. Auber and Clifford Belden
Current Protocols in Magnetic Resonance Imaging (2001) A4.1.1-A4.1.19
Copyright © 2001 by John Wiley & Sons, Inc.
UNIT A4.1 Set up patient and equipment
1. Interview (screen) the patient to ensure that he/she has no contraindication for MR
scanning such as cardiac pacemaker or other implants containing ferromagnetic
materials. Also be sure to find out if the patient has any health conditions that may
require the presence of special emergency equipment during the scanning procedure,
or that necessitate other special precautions. Screening forms are often employed for
all patients before scanning in an MRI system.
Generally standard screening forms are used for all patients scanned in a magnetic
resonance system.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(1996) for discussion of what implants may be safely scanned using magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member, who can
BASIC
sit in the room during the scan and comfort the patient as needed. This companion must
PROTOCOL
be screened as well to ensure the absence of loose metal objects on the body or clothing.
2. If the procedure is a research protocol, have the patient sign any necessary consent
form.
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
4. Have the patient remove any mascara or other metal-containing makeup to avoid local
tissue heating and image artifacts.
5. Inform the patient about what will occur during the procedure, what he/she will
experience while in the magnet, and how to behave, to include the following:
a. If earphones/headphones are used to protect the ears from loud sounds produced
by the scanner, the patient will be asked to wear these, but will be able to
communicate with the technologist at any/all times during the procedure.
b. The patient will be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this equipment works).
c. For optimum results the patient should not talk, and should avoid/minimize
swallowing or other movement during each scan—i.e., as long as the banging
sounds continue. Between scans, talking and swallowing are allowed in most
cases, but should be avoided when comparative positional studies are being
performed. The patient will be informed in these instances.
d. Nevertheless, the patient may call out at any time if he/she feels it necessary.
6. Have the patient lay supine on the scanner table. Either before or directly after the
patient is positioned on the table, set up any triggering devices or other necessary
monitoring equipment.
7. Establish an i.v. line through which the contrast agent can be injected, and attach this
line securely to the patient so that movement into or out of the magnet will not pull
at the patient’s arm.
It is preferable to insert the line prior to imaging. This allows no intervening motion
between those scans performed before and those run after the contrast agent injection.
8. Center the patient in the head coil. Make sure that the head and neck are constrained
to prevent unnecessary motion, especially if high-resolution scans are to be run.
9. If needed, place a pillow or other support under the knees to make the patient more
Infectious comfortable.
Diseases of the
Brain Brain Abscess
A4.1.1 A4.1.2
Current Protocols in Magnetic Resonance Imaging
10. Use the centering light to position the patient and place them into the center of the Sequence 3: T1-transverse head
magnet. The nasion is the landmark and utilized here. 16. Bring the sequence for a transverse T1-weighted scan up onto the console. Set the
imaging parameters as shown in Table A4.1.4.
Once this step has been performed, so long as the patient does not move on the table, the
table itself can be moved and then replaced in the same position as before without 17. Use the pilot sequence to set up the scan levels.
jeopardizing the positioning of one scan relative to another.
18. Let the patient know you are ready, and begin the scan.
11. If the patient is unable to hold still, either provide an appropriate sedative, or arrange
with anesthesiology for conscious or general anesthesia.
Table A4.1.3 Primary Clinical Imaging Parameters for Sequence 2: T1-Sagittal
Alternatively, a low-field open magnet may be sought for scanning the claustrophobic Head
patient. There are also fast scanning techniques that may be employed for imaging of these
patients (see Chapter A5). Patient position Supine
Scan type Spin echo
This Basic Protocol can be performed in less than 30 min.
Imaging plane (orientation) Sagittal
Sequence 1: Rapid three-plane pilot Central slice or volume center Midline head
12. Run the three-plane pilot (Table A4.1.2) sequence to evaluate the patient positioning Echo time (TE) 12 msec
in the magnet. Repeat time (TR) 300 msec
Flip angle (FA) 90°
This sequence runs in less than 10 sec and is used to position the remainder of the Fields of view (FOVx, FOVy) 240 mm, 240 mm
sequences. It is particularly useful to correct off-axis positioning in the coronal plane. Resolution (Δx, Δy) 1.25 mm, 0.94 mm
Number of data points collected (Nx, Ny) 192, 256
Sequence 2: T1-sagittal head Display matrix (Dx, Dy) 256, 256
13. This sequence serves as a true T1-weighted sagittal study of the head. Bring the Slice thickness (Δz) 4 mm
sequence for a sagittal T1-weighted scan up onto the console and utilize the parame- Number of slices 15
ters in Table A4.1.3. Slice gap 1 mm
14. Use the pilot sequence to set up the scan levels. Number of acquisitions (Nacq) 1
Read direction Superior–inferior
15. Let the patient know you are ready, and begin the scan. Slice location Cover brain parenchyma
Saturation pulses None
Scan time 1 min, 55 sec
Table A4.1.2 Primary Clinical Imaging Parameters for Sequence 1: Rapid
Three-Plane Pilot
Table A4.1.4 Primary Clinical Imaging Parameters for Sequence 3:
Patient position Supine T1-Transverse Head
Scan type Gradient echo
Patient position Supine
Imaging plane (orientation) Sagittal, transverse, and coronal
Scan type Spin echo
Central slice or volume center Midline head
Imaging plane (orientation) Transverse (planum sphenoidale
Echo time (TE) 3.7 msec
line)
Repeat time (TR) 16 msec
Central slice or volume center Mid-cranium
Flip angle (FA) 20°
Echo time (TE) 12.1 msec
Fields of view (FOVx, FOVy) 300 mm, 300 mm
Repeat time (TR) 500 msec
Resolution (Δx, Δy) 2.34 mm, 2.34 mm
Flipe angle (FA) 90°
Number of data points collected (Nx, Ny) 128, 128
Fields of view (FOVx, FOVy) 240 mm, 180 mm
Display matrix (Dx, Dy) 256, 256
Resolution (Δx, Δy) 0.94 mm, 0.94 mm
Slice thickness (Δz) 10 mm
Number of data points collected (Nx, Ny) 256, 192
Number of slices 3
Display matrix (Dx, Dy) 256, 256
Slice gap NA
Slice thickness (Δz) 5 mm
Number of acquisitions (Nacq) 1
Number of slices 24
Read direction Left–right
Slice gap 1 mm
Slice location At isocenter, 3 orthogonal planes
Number of acquisitions (Nacq) 2
Saturation pulses None
Read direction Anterior–posterior
RAMa 2×
Slice location Foramen magnum to vertex
Scan time 6 sec Infectious
Diseases of the Saturation pulses None
aZero padded from 128 by 128 points to 256 by 256 points.
Brain Brain Abscess Scan time 2 min, 36 sec
A4.1.3 A4.1.4
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
Sequence 4: FLAIR transverse head Table A4.1.6 Primary Clinical Imaging Parameters for Sequence 5: FSE
19. Bring the sequence for a transverse FLAIR scan up onto the console. Set the imaging Transverse Head
parameters as shown in Table A4.1.5.
Patient position Supine
20. Use the pilot sequence to set up the scan levels and a caudal saturation pulse. Scan type Fast spin echo (FSE)
Imaging plane (orientation) Transverse (planum sphenoidale
21. Let the patient know you are ready, and begin the scan. line)
Central slice or volume center Mid-cranium
Sequence 5: FSE transverse head
Echo time (TE) 105 msec (effective)
22. Bring the sequence for a transverse FSE scan up onto the console. Set the imaging
Echo train length (ETL) 16
parameters as shown in Table A4.1.6.
Repeat time (TR) 5616 msec
23. Use the pilot sequence to set up the scan levels. Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 240 mm, 180 mm
24. Let the patient know you are ready, and begin the scan. Resolution (Δx, Δy) 0.63 mm, 0.70 mm
Number of data points collected (Nx, Ny) 384, 256
Sequence 6: T1-transverse post-gadolinium head Display matrix (Dx, Dy) 384, 384
25. Bring the sequence for a transverse T1-weighted scan up onto the console. Set the Slice thickness (Δz) 5 mm
imaging parameters as shown in Table A4.1.7. Number of slices 24
26. Use the pilot sequence to set up the scan levels. Slice gap 1 mm
Number of acquisitions (Nacq) 2
27. Leaving the patient in the magnet, inject the contrast agent either by hand, or using Read direction Anterior–posterior
a mechanical injector. Observe the injection to insure there is no extravasation of the Slice location Foramen magnum to vertex
contrast agent. Flush the line with 10 ml of sterile normal saline. Begin the scan as Saturation pulses Caudal to saturate arterial flow
soon as the injection is completed. Scan time 3 min, 22 sec
Alternatively, the scanning table may be advanced out of the magnet for the injection, but
the patient must remain in place.
A dose of 0.1 mmol/kg of contrast agent is usually given. Table A4.1.7 Primary Clinical Imaging Parameters for Sequence 6:
T1-Transverse Post-Gadolinium Head

Patient position Supine

Table A4.1.5 Primary Clinical Imaging Parameters for Sequence 4: FLAIR
Transverse Head Scan type Spin echo
Imaging plane (orientation) Transverse (planum sphenoidale
Patient position Supine line)
Scan type FLAIR-FSE Central slice or volume center Mid-cranium
Imaging plane (orientation) Transverse (parallel to AC-PC line) Echo time (TE) 12.1 msec
Central slice or volume center Mid-cranium Repeat time (TR) 500 msec
Echo time (TE) 125 msec (effective) Flip angle (FA) 90°
Echo train length (ETL) Fixed (generally 12) Fields of view (FOVx, FOVy) 240 mm, 180 mm
Repeat time (TR) 6000 msec Resolution (Δx, Δy) 0.94 mm, 0.94 mm
Inversion time (TI) 1900 msec Number of data points collected (Nx, Ny) 256, 192
Flip angle (FA) 180° Display matrix (Dx, Dy) 256, 256
Fields of view (FOVx, FOVy) 240 mm, 197 mm Slice thickness (Δz) 5 mm
Resolution (Δx, Δy) 0.94 mm, 0.97 mm Number of slices 24
Number of data points collected (Nx, Ny) 256, 204 Slice gap 1 mm
Display matrix (Dx, Dy) 256, 256 Number of acquisitions (Nacq) 2
Slice thickness (Δz) 5 mm Read direction Anterior–posterior
Number of slices 24 Slice location Foramen magnum to vertex
Slice gap 1 mm Saturation pulses None
Number of acquisitions (Nacq) 1 Scan time 2 min
Read direction Anterior–posterior
Slice location Foramen magnum to vertex
Saturation pulses Caudal to saturate arterial flow Infectious
Diseases of the
Scan time 2 min, 48 sec Brain Brain Abscess

A4.1.5 A4.1.6
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A4.1.8 Primary Clinical Imaging Parameters for Sequence 7: T1-Coronal strating the microhemorrhage. This can lead not only to the helpful diagnostic clue of a
Post-Gadolinium Head hemorrhagic lesion but also can bring other lesions to the attention of the radiologist. This
can impact patient prognosis, or guide more effective follow-up imaging. The second
Patient position Supine
optional sequence, the 3-dimensional gradient-echo post-contrast sequence, is usually
Scan type Spin echo
reserved for higher resolution of small lesions or adjacent structures that may be affected
Imaging plane (orientation) Coronal (perpendicular to planum
by the pathology. It can also be an effective aid in operative planning, and may indeed
sphenoidale line)
serve as the base MRI sequence for several of the commercially available image-guided
Central slice or volume center Mid-cranium
surgery tools presently on the market and utilized by many neurosurgeons. This sequence
Echo time (TE) 12.1 msec
Repeat time (TR) 500 msec
can be utilized in lieu of the transverse and coronal post-gadolinium sequences outlined
Flip angle (FA) 90°
in the Basic Protocol, if desired. It is easily manipulated in post-processing to yield a
Fields of view (FOVx, FOVy) 220 mm, 178 mm
desired projection of a lesion, or simply to give a multiplanar assessment of the brain in
Resolution (Δx, Δy) 0.86 mm, 0.93 mm
a given patient.
Number of data points collected (Nx, Ny) 256, 192
Set up patient and equipment
Display matrix (Dx, Dy) 256, 256
1. Use the same equipment and the same patient set-up as for the previous method (see
Slice thickness (Δz) 5 mm
Basic Protocol, steps 1 to step 11).
Number of slices 24
Slice gap 1 mm 2. Run the pilot scan as sequence 1 in the Basic Protocol (see Table A4.1.2).
Number of acquisitions (Nacq) 2
Read direction Left–right Sequence 8: Coronal gradient echo (head)
Slice location Frontal to occipital cerebral poles 3. Bring the sequence for a gradient-echo scan up onto the console. Set the imaging
Saturation pulses None parameters as shown in Table A4.1.9.
Scan time 2 min, 36 sec
4. Use the pilot sequence to set up the scan levels.

This is essentially the same sequence as the pre-gadolinium scan, but can be performed
with the addition of a flow compensation (FC) pulse to better delineate the cerebral vessels,
and a magnetization transfer (MT) pulse to optimize enhancing lesion detection.
Sequence 7: T1-coronal post-gadolinium head
28. Bring the sequence for a T1-weighted scan up onto the console. Set the imaging
parameters as shown in Table A4.1.8.
29. Use the pilot sequence to set up the scan levels.
30. Let the patient know you are ready, and begin the scan.
Small abscesses may be much better delineated with a smaller field of view (FOV; i.e., FOVx
= 200 mm, FOVy = 150 mm) and higherresolution, thinner slice thickness (3 mm). Please
note that this modification should be performed with one acquisition, and limited to the
area of abnormality to avoid prolonged scan times.
When scanning pediatric patients (i.e., under 24 months) modify the transverse and coronal
sequences by placing the FOV at 200 mm, and the slice thickness and gap at 4 mm and 1
mm, respectively. For neonates, premature infants, and other extremely small patients, an
extremity coil may be used.
5. Let the patient know you are ready, and begin the scan.
Sequence 9: 3-Dimensional post-gadolinium transverse gradient echo (head)
6. Repeat Basic Protocol, step 27.
When surgery is contemplated, or in order to better delineate a smaller lesion, a useful
alternative sequence to the post-gadolinium transverse and coronal scans described in the
Basic Protocol is a post-gadolinium 3-D gradient-echo sequence.
7. Perform scan with high resolution parameters as described in Table A4.1.10.
This scan can be reformatted into any desired plane as it is a volume-acquired sequence.
This scan generates 120 images as a whole brain study, but higher resolution image sets
with 32 or 64 images, with a resolution as low as 1 mm can be produced, but may be
restrictive in area of coverage.
8. Perform planar reformations of the 3-D data set on-line or off-line using the refor-
mation software.
One can coordinate with the neurosurgeon to obtain the most useful projections, that can
then be saved and/or filmed.

This sequence may be performed with the addition of a flow compensation (FC) pulse to Note that the cerebral vasculature is quite bright on this sequence as this 3-D gradient echo
better delineate the cerebral vessels, and a magnetization transfer (MT) pulse to optimize sequence is the base-sequence for time-of-flight (TOF) magnetic resonance angiography
enhancing lesion detection. (MRA).

SPECIAL SITUATIONS ALTERNATE ALTERNATE IMAGING OF CEREBRAL ABSCESS BY MAGNETIC RESONANCE

PROTOCOL 1
The two sequences that follow are presented here for consideration by the radiologist in
those patients suspected of significant intracranial inflammatory conditions for whom the
basic protocol has failed to elucidate a clearly defined abnormality. For instance, in
patients with microabscesses related to septic emboli that may be associated with Infectious
Diseases of the
pseudoaneurysm formation, the gradient-echo sequence is extremely useful for demon- Brain
PROTOCOL 2 DIFFUSION, PERFUSION, AND SPECTROSCOPY
The Basic Protocol is quite sufficient in most instances of cerebral abscess imaging and
represents the standard for imaging of the patient suspected for cerebral abscess. However,
in certain clinical instances there may remain confusion as to the diagnosis, or in order
Brain Abscess to gain a higher degree of specificity in labeling a lesion as an abscess, further imaging

A4.1.7 A4.1.8
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A4.1.9 Primary Clinical Imaging Parameters for Sequence 8: Coronal can be performed utilizing dMRI, pMRI, or MRS. Recent literature (Ernst et al., 1998;
Gradient Echo Head Burtscher and Holtas, 1999; Desprechins et al., 1999) has indicated a clinical benefit in
some patients with these new approaches. Higher-field-strength magnets with enhanced
Patient position Supine
gradient packages (e.g., generally >20 mT/m for dMRI and pMRI) are preferred for
Scan type 2-D gradient echo
optimum benefit in the performance of these sequences. Also, magnetic field homogeneity
Imaging plane (orientation) Coronal
is a crucial factor in the successful performance of these sequences, mandating routine
Central slice or volume center Mid-cranium
assessment and optimization of magnet shim prior to each acquisition. Indeed, as the
Echo time (TE) 24.6 msec
hardware and software to obtain these specialized sequences becomes more ubiquitous
Repeat time (TR) 719 msec
and further data is gathered, one or more of these sequences may prove invaluable in the
Flip angle (FA) 25°
work-up of the patient suspected of cerebral abscess. Described here is the authors’
Fields of view (FOVx, FOVy) 220 mm, 220 mm
experience with a 1.5 T whole-body system with an enhanced gradient system. Note that
Resolution (Δx, Δy) 0.86 mm, 1.15 mm
a detailed presentation on the performance and processing of these complex sequences
Number of data points collected (Nx, Ny) 256, 192
goes beyond the scope of this unit. For such a discussion, the reader is referred to their
Display matrix (Dx, Dy) 256, 256
respective scanner and workstation manufacturers, as well as the literature cited in this
Slice thickness (Δz) 5 mm
unit. Additionally, having an MRI-trained physicist available to insure the proper quality
Number of slices 21
management of these sequences is highly recommended.
Slice gap 1 mm
Number of acquisitions (Nacq) 2 Additional Materials
Read direction Left–right
Enhanced gradient equipped MRI scanner
Slice location Frontal to occipital cerebral poles
Specialized post-processing software, especially for pMRI image processing (a
Saturation pulses None separate workstation is preferred for off-line processing of data sets, enabling
Scan time 4 min, 36 sec continuous patient scanning to proceed)
Power injector for dynamic contrast administration (i.e., while scanning)

Table A4.1.10 Primary Clinical Imaging Parameters for Sequence 9: 3-D

Set up patient and equipment
Post-Gadolinium Transverse Gradient Echo Head 1. Perform the patient setup as for the previous method (see Basic Protocol, steps 1 to
11).
Patient position Supine
2. Run sequence 1 as in Basic Protocol (see Table A4.1.2).
Scan type 3-D gradient echo
Imaging plane (orientation) Transverse (option sagittal or Sequence 10: pMRI of the brain
coronal)
3. Perform this scanutilizing the parameters outlined in Table A4.1.11 to assess vascular
Central slice or volume center Mid-cranium for whole brain, or
area of abnormality
perfusion of the brain and pathologic foci.
Echo time (TE) 5 msec pMRI can help in differentiating abscess from necrotic neoplasm.This is a rapid acquisition
Repeat time (TR) 11.2 msec (or minimum) with moderate spatial resolution that can yield information on the entire brain and multiple
lesions, even in agitated patients.
Flip angle (FA) 35°
Fields of view (FOVx, FOVy) 240 mm, 187 mm This sequence must be the first sequence performed after the gadolinium injection, and is
Resolution (Δx, Δy) 0.94 mm, 0.97 mm acquired dynamically. After a baseline brain image acquisition is performed, an injection
Number of data points collected (Nx, Ny) 256, 192 time of <8 sec is used to deliver 0.2 mmol/kg gadolinium chelate with a 10 ml normal saline
bolus flush via the injector. This step may be performed by hand, but timing of the sequence
Display matrix (Dx, Dy) 256, 256
is problematic.
Slice thickness (Δz) 1.2 mm
Number of slices 120 4. Perform image analysis visually or off-line on a workstation using commercially
Slice gap 0 mm available programs.
Number of acquisitions (Nacq) 1 Regional cerebral blood volume (rCBV) maps can be reconstructed using an appropriate
Read direction Anterior–posterior, if imaging analysis of the MR signal intensity as a function of time.
plane is transverse The rCBV maps can be aligned with the anatomic images to obtain region-of-interest (ROI)
Slice location Whole brain data.
No phase wrap (NPW)a Yes
Saturation pulses NA Sequence 11: Echo-planar imaging (EPI) dMRI of the brain
Scan time 4 min, 19 sec 5. Run this sequence utilizing the parameters outlined in Table A4.1.12 to assess water
aPhase oversampling or anti-aliasing. motion within the brain parenchyma.
Infectious Diffusion imaging may help in the differential diagnosis of cerebral abscess versus necrotic
Diseases of the neoplasm.
Brain Brain Abscess

A4.1.9 A4.1.10
Current Protocols in Magnetic Resonance Imaging Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging
 Table A4.1.11 Primary Clinical Imaging Parameters for Sequence 10: pMRI of Table A4.1.13 Five Frame Preset b-Value Results
the Brain
Phase encoding
Image Appearance Read b-value Slice select b-value
Patient position Supine b-value
Scan type Single-Shot Echo Planar Imaging Frame no. 1. No diffusion 0 0 0
(SS-EPI)
Frame no. 2. No diffusion 0 0 0
Imaging plane (orientation) Transverse or coronal
Frame no. 3. Diffusion along read 1000 sec/mm2 0 0
Central slice or volume center Mid-cranium
Frame no. 4. Diffusion along phase encoding 0 1000 sec/mm2 0
Echo time (TE) 60 msec
Frame no. 5. Diffusion along slice select 0 0 1000 sec/mm2
Repeat time (TR) 1400 msec
Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 240 mm, 240 mm
This is a rapid acquisition (scan time 28 sec) with moderate spatial resolution that can
Resolution (Δx, Δy) 1.28 mm, 1.88 mm yield information on the entire brain and multiple lesions, even in agitated patients.
Number of data points collected (Nx, Ny) 188, 128
Display matrix (Dx, Dy) 376, 256 Optimizing echo-planar scan parameters (phase offset and delay) is important to obtain
high-quality diffusion images. For this reason, the scan is often simulated prior to being
Slice thickness (Δz) 5 mm
run to allow optimization of these parameters. This may take 1 to 3 min. The values obtained
Number of slices 10 are then passed to the scanner and the sequence performed. (This is unique to Marconi’s
Slice gap 5 mm system.)
Number of acquisitions (Nacq) 39 (time resolution 1.4 sec)
6. Acquire images at b-values (diffusion gradients; Table A4.1.13) of both 0 and 1000
Read direction Left–right
sec/mm2, each applied in three orthogonal directions.
Slice location Area of interest
Saturation pulses NA Nacq may range from 1 to 40 depending on required SNR.
Scan time 56 sec These data provide raw diffusion-weighted multishot echo-planar images and is used to
process apparent diffusion coefficient (ADC) maps.

Sequence 12: MRS of the brain

Table A4.1.12 Primary Clinical Imaging Parameters for Sequence 11: dMRI of 7. Use previously acquired multiplanar sequences to place the single voxel over a region
the Brain
of interest. The voxel is 2 × 2 × 2 cm3 and should be located in the brain, avoiding
Patient position Supine contamination of signal and magnetic field inhomogeneities from cerebrospinal fluid,
Scan type Diffusion EPI fat, bone, and air (sinuses).
Imaging plane (orientation) Transverse To assume the proper voxel placement, it is recommended that its position be checked in
Central slice or volume center Mid-cranium three planes prior to running the sequence.
Echo time (TE) 100 msec 8. Before the sequence can be run, two important steps must be completed: shimming
Repeat time (TR) 5629 msec over the voxel and water suppression. Many scanners have automated packages to
Flip angle (FA) 90° perform these functions. However, these preliminary steps may take 5 min to perform.
Fields of view (FOVx, FOVy) 240 mm, 240 mm
9. Run this sequence utilizing the parameters outlined in Table A4.1.14 to provide
Resolution (Δx, Δy) 2.96 mm, 3 mm
information as to the specific molecular substrate within a volume of interest that can
Number of data points collected (Nx, Ny) 81, 80
indicate the presence of abnormal tissue or metabolic processes.
Display matrix (Dx, Dy) 81, 80
Slice thickness (Δz) 5 mm MRS can be used in the diagnosis and follow-up of cerebral abscesses.
Number of slices 20 10. Sequence acquisition time is 10 min per TE performed (including the water suppres-
Slice gap 1 mm sion and shimming, with the sequence acquisition), making useful results difficult to
Number of acquisitions (Nacq) 1 (see Table A4.1.13 below)a obtain in agitated patients.
Read direction Left–right
Spectral post-processing is usually automated within the scanner and may include apodi-
Slice location Whole brain or area of interest zation with a gaussian filter, frequency-shift correction, Fourier transformation, and
Saturation pulses Fat adjustment of phase relations.
Scan time 28 sec
A tracing of the magnetic spectra of the molecular contents of the voxel is produced for
aWith each slice, the MR machine scans five times (five frames), and each frame has a different b-value
analysis.
associated with it (see Table A4.1.13).
In general, longer TE sequences (e.g., 270 msec) isolate spectra that are more specific for
the most commonly assessed peaks of N-acetyl aspartate (NAA), choline (Cho), and
creatine (Cr). Lower TE sequences provide more complex spectra with greater sensitivity
Infectious
Diseases of the for many other substances of potential interest, such as myoinositol.
Brain Brain Abscess

A4.1.11 A4.1.12
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A4.1.14 Primary Clinical Imaging Parameters for Sequence 12: MRS of
the Brain
Patient position
Scan type
Voxel center
Echo time (TE)
Repeat time (TR)
Flip angle (FA)
Voxel size (Δx, Δy, Δz)
Number of acquisitions (Nacq)
Saturation pulses
aThese 8 acquisitions are done without water suppression and are used for phase correction of the free
induction decay scans.
bThese 192 acquisitions are done with water suppression.
COMMENTARY
Background Information
Cerebritis and abscess, though not common
diagnoses, remain disturbingly common today
despite the advent of, and significant advances
in, the field of microbiology. This is primarily
related to the increased prevalence of immuno-
compromised patient populations and resistant
strains of microorganisms. The patient with
cerebral abscess may present in a nonspecific
fashion. The most commonly noted symptom
is headache, which may be the only complaint.
More disturbing signs and symptoms of fever
(i.e., <50% in some series), altered sensorium,
focal neurologic deficit, and seizure may also
manifest. The primary source of the intracranial
infection may be an important clue and should
be sought.
MR has emerged as the study of choice in
patients suspected of harboring a cerebral ab-
scess(es), primarily as it is significantly more
sensitive for the detection of inflammatory dis-
ease in the central nervous system (CNS) than
computed tomography (CT) and the other im-
aging and laboratory studies (Mathisen and
Johnson, 1997). The results of the Basic Proto-
col outlined above are often nonspecific
though, yielding a differential diagnosis that
also includes primary and metastatic brain tu-
mors, resolving hematoma, infarction, and de-
myelinating disease. However, MR is crucial
for defining not only the presence of disease,
but also the extent, and is used in surgical
planning (and timing) with these patients. With
the recent advancements of diffusion, perfu-
sion, and spectroscopic sequences, and newly
available research data, the lack of specificity
of the technique is rapidly diminishing. pMRI
Current Protocols in Magnetic Resonance Imaging
Supine
Single voxel Point-Resolved Spectroscopic
(PRESS)
Area of abnormality
35 msec
1500 msec
90°
2 cm, 2 cm, 2 cm
8 (reference)a, 192 (signal)b
Fat and water
parallels the published results of nuclear medi-
cine studies differentiating abscess from neo-
plasm using single-photon emission CT
(SPECT) and positron emission tomography
(PET; Ernst et al., 1998). However, pMRI is a
more convenient and rapid exam (when per-
formed in conjunction with the MR study). In
addition, the anatomic information and higher
spatial resolution provided by MRI make the
technique a powerful tool in these patients.
Additionally, MR can be used to follow cere-
bral abscess patients through their therapy. Not
only does MR demonstrate reduced size and
number of lesions and mass effect with effec-
tive therapy, but also the status of organism
activity within abscesses and their response to
antibiotics may be assessed (Burtscher and
Holtas, 1999). This in turn may lead to earlier
intervention with percutaneous or surgical
drainage of abscess cavities or more rapid as-
signment of the proper differential diagnosis of
a noninfectious lesion, such as tumor, allowing
for directed therapies to be instituted. The sen-
sitive detection and proper classification of
infectious lesions and abscess by MR can also
provide more timely and directed determina-
tion of the source of an infection, with institu-
tion of the necessary treatments, as is seen with
cardiac valvular disease causing mycotic aneu-
rysm and abscess. Laboratory analysis is still
required to determine the specific infectious
organism.
Critical Parameters and
Troubleshooting
There are several parameters that enhance
the sensitivity and utility of the Basic Protocol
Figure A4.1.1 A T2-weighted spin-echo transverse (TR = 2500 msec, TE = 80 msec, Nacq = 0.75) image of the brain (A)
at the level of the midbrain shows several subtle abnormal foci of low signal intensity in the posterior brainstem and cerebellar
peduncles (arrows) that on a gradient-echo coronal (TR = 500 msec, TE = 25 msec, Nacq = 1) image through the posterior
midbrain (B) demonstrate marked magnetic susceptibility effect with “blooming” of the lesions (arrows). In addition there are
numerous other foci demonstrated at multiple levels consistent with diffuse axonal injury in this 12-year-old boy who had
sustained a severe closed head injury 2 months previously.
(Orrison et al., 1995). These include the proper improves the sensitivity for early detection of
selection of the TI (inversion time) for the subtle foci of abnormal enhancement. The MT
FLAIR images to effect nulling of hyperintense pulse works by suppressing overall background
cerebrospinal fluid (CSF) signal, producing signal intensity, thus leaving the high T1-signal
dark CSF. This enables the detection of abnor- of enhancing tissue more apparent.
mal periventricular high T2-signal intensity for When considering the optional sequences,
the important determination of ventriculi- the optimal choice of TE on the gradient-echo
tis/ependymitis, indicating spread of infection coronal sequence is important to allow for de-
to the ventricles or possibly rupture of the tection of foci of microhemorrhage. These foci
abscess into the CSF spaces. Interestingly, most may indicate the presence of satellite lesions of
patients have a thin rim of this hyperintensity infection, and their detection may narrow the
in the normal state. Additionally, the use of the differential diagnostic possibilities for infec-
MT pulse for the post-gadolinium sequences tious organism (such as Herpes simplex virus,
Figure A4.1.2 (on right) FLAIR (TR = 6000 msec, TE = 96 msec, Nacq = 1;TI = 1800 msec) and
FSE T2-weighted (TR = 5616 msec, TE = 105 msec, Nacq = 1; echo train length = ETL = 12)
transverse images (A,B) of the brain at the level of the corona radiata show an abnormal rounded
low T2 signal intensity lesion (arrow) adjacent the left frontal horn of the lateral ventricle with
surrounding edema. Precontrast (C) and post-contrast (D) transverse images at the same level (TR
= 500 msec, TE = 12 msec, Nacq = 1) show intense enhancement of the lesion. The patient is a
41-year-old male with chronic sarcoidosis being treated with high-dose corticosteroids. The
resulting chronic immunosuppression enabled infection and cerebral abscess formation by Bipo-
laris fungus. This patient previously underwent a left frontal lobe biopsy and resulting encepha-
lomalacia is noted (asterisk). Magnetic resonance spectroscopy (PRESS, TR = 1600 msec, TE =
135 msec) of one of the abscesses (E) demonstrates a decrease in the NAA peak with elevation
Infectious of the creatine and choline peaks consistent with a loss of neuronal tissue within the lesion.
Diseases of the
Brain Brain Abscess
A4.1.13 A4.1.14
Current Protocols in Magnetic Resonance Imaging
 Figure A4.1.3 FSE (TR = 3380 msec, TE = 90 msec, Nacq = 2; ETL = 8) T2-weighted (A) image through the occipital lobe
shows a large heterogeneous lesion with prominent surrounding edema. There are layers of low T2 signal within its wall
(arrows) suggesting not only abscess but possible microhemorrhage. The corresponding post-contrast T1-weighted (TR =
600 msec, TE = 14 msec, Nacq = 2) transverse (B) image reveals a relatively defined deep wall (arrow) compared to the
more superficial wall (white arrow) and a target configuration (arrowhead), as well. This 40-year-old male with AIDS was
treated for presumed toxoplasmosis with lesion resolution over the ensuing 8 weeks (not shown).

which is angioinvasive and thus predisposed to
causing hemorrhage). A longer TE will demon-
strate increased T2-based contrast and magnetic
susceptibility with more pronounced low signal
“blooming” from foci of hemorrhage contain-
ing hemosiderin (Fig. A4.1.1), though at the
expense of overall tissue signal-to-noise ratio
(SNR). Generally, TE values of <25 msec (the
authors use 24.6 msec) are less sensitive for this
purpose. For similar reasons, this “blooming”
is prominent from densely calcified cortex,
such as in the calvarium, and makes for poor
evaluation of signal at the skull-brain interface,
particularly at the skull base.
An option with speed and utility that can be
performed in place of both the gradient-echo
coronal and FLAIR/FSE T2 transverse se-
quences is a standard proton-density/T2 dual-
echo spin-echo sequence (see UNIT A5.1). For a
time expenditure similar to that of the Basic
Protocol, the sensitivity for magnetic suscepti-
bility (better on T2 spin-echo than FSE, but best
on the gradient-echo coronal sequence) is im-
Infectious proved, but at the expense of some sensitivity
Diseases of the in the detection of periventricular disease (bet-
Brain Brain Abscess
Figure A4.1.2 (See legend on facing page.)
A4.1.15 A4.1.16
Current Protocols in Magnetic Resonance Imaging
ter with the FLAIR). Also, if detection of brain
calcifications is important, this is still best ac-
complished with CT, as the gradient-echo coro-
nal sequence, though quite sensitive for the
magnetic susceptibility of calcified foci, is non-
specific in differentiating calcium from blood
products. Additional time savings for imaging
critically ill or noncompliant patients can be
found by limiting pre-contrast and post-con-
trast T1-weighted sequences to one plane each,
or performing only one acquisition (Nacq) on
each sequence, or alternatively by utilizing FSE
T1-weighted sequences rather than routine T1-
weighted sequences. These maneuvers will all
limit SNR and sensitivity for detection of ab-
normal contrast enhancement, but will afford
faster protocol scan times (for a more detailed
discussion of fast-scan techniques, see chapter
A5 of this series).
Anticipated Results
Brain abscesses have been shown to pro-
gress through four evolutionary stages on im-
aging studies, with the first being the acute or
cerebritis stage (Taveras and Pile-Spellman,
Current Protocols in Magnetic Resonance Imaging
 1996; Wong and Quint, 1999). This occurs
during the first 5 days of the infection and is
characterized by a loss of the blood-brain-bar-
rier integrity with edema formation and (rarely)
microhemorrhage. Imaging is charac-
teristically normal on CT, though either CT or
MR may demonstrate edema without promi-
nent mass effect or contrast enhancement. The
late cerebritis stage begins at the end of the first
week and is characterized by progression to
necrosis with a thin rim of granulation tissue.
Imaging will show increased edema and mass
effect and minimal, diffuse enhancement. Sat-
ellite lesions may emerge (Fig. A4.1.2). Any
extension of the process is usually noted within
the white matter, and it is in this second stage
that rupture into the ventricular system is most
likely. By the end of the second week (capsular
stage) a capsule can be defined on imaging
studies that thickens over time, especially along
the more superficial (i.e., gray matter) border.
The capsule is now considered complete, with
three definable layers: inner granulation layer,
middle collagenous layer, and outer gliotic
layer. Imaging will show this well defined and
prominently enhancing capsule encompassing
the liquefied abscess with decreased surround-
ing edema. Surgery is often delayed until this
stage to allow for greater success in lesion
resection and to minimize hemorrhage in adja-
cent structures. Finally, in the chronic stage,
abscesses most often continue to grow, eventu-
ally resulting in a significant morbid event such
as rupture into the ventricular or subarachnoid
CSF. However, growth arrest and mineraliza-
tion of the capsule may occur, as can resolution
and resorption, especially with successful anti-
biotic therapy. Imaging will parallel these
changes with continued enhancement of the
capsule, while resorbing lesions may lose their
necrotic center, eventually becoming small foci
of enhancement or mineralization.
At initial presentation, the typical cerebral
abscess appears on MR imaging as a well-de-
fined rounded lesion with fluid-signal material
centrally and a rim of thin, uniform low signal
intensity on T2-weighted images that demon-
strates smooth, homogeneous enhancement
(Fig. A4.1.3). This is usually surrounded by a
variable amount of parenchymal edema. The
lesion is most commonly located in the frontal
Figure A4.1.4 This 23-year-old female presented with headache and ataxia. (A) The FSE-T2 (TR = 5616 msec, TE = 105
msec, Nacq = 1; ETL = 8) and (B) post-contrast T1-weighted (TR = 500 msec, TE = 12 msec, Nacq = 1) transverse images lobe (adjacent an infected sinus, usually the
show a heterogeneous rounded lesion of the right posterior midbrain (arrows) without definite enhancement. However, the frontal), the parietal lobe (reflecting hemato-
genous spread from the infection source, and
transverse dMRI sequence (TR = 3000 msec, TE = 100 msec; b = 1000 sec/mm2) discloses corresponding high signal on
the ADC map (C) and low signal on the diffusion-weighted image (D) compatible with a tumoral cyst. A low-grade astrocytoma centering at the gray-white junction), or less
was subsequently proven at surgical biopsy. This lesion presents a dMRI signal pattern opposite to that expected with commonly in the temporal lobe adjacent an
cerebral abscess.
Brain Abscess infected temporal bone. On pMRI, a cerebral
A4.1.17 A4.1.18
Current Protocols in Magnetic Resonance Imaging
abscess is expected to demonstrate decreased
perfusion relative to normal brain, and mark-
edly decreased perfusion relative to hypervas-
cular tumors that generally show increased per-
fusion (Ernst et al., 1998). On dMRI, cerebral
abscess will typically demonstrate high signal
on diffusion-weighted images, with markedly
decreased abscess cavity signal on ADC images
(Desprechins et al., 1999). This is the opposite
of that found in necrotic tumors such as
glioblastoma multiforme (Fig. A4.1.4). And
finally, MRS tracings may show specific prod-
ucts of infectious organism metabolism (i.e.,
acetate, lactate, pyruvate and succinate—meta-
bolites not reported in human brain tissue other
than in infectious states) in the acute/pretreat-
ment phase, with regression over the course of
therapy to a single lactate peak of a “dead”
abscess, indicating the cessation of viable in-
fectious organism activity (Burtscher and Hol-
tas, 1999). Also, the presence of a group of
resonances at 0.9 ppm on the TE = 270 msec
sequence demonstrating inversion on a 135
msec sequence is diagnostic (95% specific) of
bacterial abscess. Several amino acids, includ-
ing valine, leucine and isoleucine, produce this
peak and are found in pus within bacterial
abscesses at concentrations 20 to 80 times
greater than that found in necrotic tumor cavi-
ties. Unfortunately, this relationship is not
found in abscesses caused by non-bacterial or-
ganisms (Grand et al., 1999).
Literature Cited
Burtscher, I.M. and Holtas, S. 1999. In vivo proton
MR spectroscopy of untreated and treated brain
abscesses. Am. J. Neuroradiol. 20:1049-1053.
Castillo, M. 1998. Neuroradiology Companion:
Methods, Guidelines, and Imaging Fundamen-
tals, 2nd ed. Lippincott, New York.
Desprechins, B., Stadnik, T., Koerts, G., Shabana,
W., Breucq, C., and Osteaux, M. 1999. Use of
diffusion-weighted MR imaging in differential
diagnosis between intracerebral necrotic tumors
and cerebral abscesses. Am. J. Neuroradiol.
20:1252-1257.
Ernst, T.M., Chang, L., Witt, M.D., Aronow, H.A.,
Cornford, M.E., Walot, I., and Goldberg, M.A.
1998. Cerebral toxoplasmosis and lymphoma in
AIDS: Perfusion MR experience in 13 patients.
Radiology 208:663-669.
Grand, S., Passaro, G., Ziegler, A., Esteve, F., Bou-
jet, C., Hoffmann, D., Rubin, C., Segebarth, C.,
Decorps, M., Le Bas, J., and Remy, C. 1999.
Necrotic tumor versus brain abscess: Importance
of amino acids detected at 1H MR spectros-
copy—initial results. Radiology 213:785-793.
Mathisen, G.E. and Johnson, J.P. 1997. Brain ab-
scess. Clin. Infect. Dis. 25:763-781.
Current Protocols in Magnetic Resonance Imaging
Orrison, W.W., Lewine, J.D., Sanders, J.A., and
Hartshorne, M.F. 1995. Functional Brain Imag-
ing. Mosby, St. Louis.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Taveras, J.M. and Pile-Spellman, J. 1996. Inflamma-
tory diseases. In Neuroradiology, 3rd ed. pp.
259-326. Williams and Wilkins, Baltimore,
Maryland.
Wong, J. and Quint, D.J. 1999. Imaging of central
nervous system infections. Semin. Roentgenol.
34:123-143.
Current Protocols in Magnetic Resonance Imaging
Key References Meningitis UNIT A4.2
Orrison et al., 1995. See above
Contains lucid explanations for the physics and
basic scan parameters of standard and advanced Meningitis is an inflammation of the pia and arachnoid maters (leptomeningitis) or the
magnetic resonance imaging studies. dura mater (pachymeningitis). Magnetic resonance imaging (MRI) in meningitis, as with
most other forms of intracranial inflammatory or infectious diseases, is a powerful though
largely non-specific diagnostic tool. The technique is used to detect the presence of
Contributed by Andrew E. Auber and disease, and also complications related to the primary process, as well as assess for
Clifford Belden alternative diagnoses. The actual diagnosis of meningitis remains largely a clinical and
Brooke Army Medical Center
San Antonio, Texas
laboratory one. For imaging of these patients the Basic Protocol presented previously in
UNIT A4.1 is utilized, to include gadolinium-enhanced sequences. An optional sequence,
diffusion MRI (dMRI), is outlined, which can be employed if specific clinical situations
exist to be further clarified. The parameters given here are derived from experience at
1.5T and may need to be altered slightly depending on the field strength available and the
specific equipment manufacturer.
IMAGING OF MENINGITIS BASIC
PROTOCOL
Presented here is the authors’ routine MR head examination for inflammatory disease,
with gadolinium contrast agent sequences. The total number of sequences presented is
seven (scout, T1-weighted sagittal and transverse, fluid-attenuated inversion recover
(FLAIR) and fast spin-echo (FSE) T2-weighted transverse, and post-contrast T1-weighted
transverse and coronal). This examination may be shortened considerably by limiting the
use of two planes for post-contrast imaging to a single transverse plane for exams with a
lower clinical suspicion for disease, or if review of the images obtained through the
transverse gadolinium-enhanced images proves unremarkable. The sequences described
herein are based on the authors’ experience with a Marconi Medical Systems 1.5T scanner,
but are expected to be equally applicable to machines from other manufacturers. The same
basic scanning protocol to image for meningitis as that for cerebral abscess (UNIT A4.1) is
utilized, which follows below.
Table A4.2.1 lists the hardware necessary to perform the examination, along with
appropriate parameters.
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as a crash cart or oxygen. Also ensure only magnetic field compatible oxygen
tanks are utilized if wall-source oxygen is unavailable.
Materials
Extravascular contrast agent (e.g., 0.1 mmol/kg patient body weight of gadolinium
chelate from Mangevist, Omniscan, or Prohance)
Normal saline (0.9% NaCl) sterile
Table A4.2.1 Equipment Parameters for Contrast-Enhanced Head Imaging Sequences
Coil type Head
Gradient coil strength 25 mT/m (or whatever the system permits)
Gating (cardiac, respiratory, peripheral) No
Respirator or oxygen If required by patient
Motion cushions Recommended
Infectious Contrast agents Yes Infectious
Diseases of the Diseases of the
Brain Brain
A4.1.19 Contributed by Andrew E. Auber and Clifford Belden A4.2.1
Current Protocols in Magnetic Resonance Imaging (2001) A4.2.1-A4.2.13
Copyright © 2001 by John Wiley & Sons, Inc. Supplement 1
 Set up patient and equipment
1. Interview (screen) the patient to ensure that he/she has no contraindication for MR
scanning such as cardiac pacemaker or other implants containing ferromagnetic
materials. Also be sure to find out if the patient has any health conditions that may
require the presence of special emergency equipment during the scanning procedure,
or that necessitate other special precautions. Screening forms are often employed for
all patients before scanning in an MRI system.
Generally standard screening forms are used for all patients scanned in a magnetic
resonance system.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(1996) for discussion of what implants may be safely scanned using magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
2. If the procedure is a research protocol, have the patient sign any necessary consent
form.
9. If needed, place a pillow or other support under the knees to make the patient more
comfortable.
10. Use the centering light to center on the nasion of the patient and place them into the
center of the magnet.
Once this step has been performed, so long as the patient does not move on the table, the
table itself can be moved and then replaced in the same position as before without
jeopardizing the positioning of one scan relative to another.
11. If the patient is unable to hold still, either provide an appropriate sedative, or arrange
with anesthesiology for conscious or general anesthesia.
Alternatively, a low-field open magnet may be sought for scanning the claustrophobic
patient. There are also fast scanning techniques that may be employed for imaging of these
patients (see Chapter A5).
12. This Basic Protocol can be performed in <30 min.
Sequence 1: Rapid three-plane pilot
13. Run the three-plane pilot (Table A4.2.2) sequence to evaluate the patient positioning
in the magnet.

3. Have the patient remove all jewelry and change into a gown to eliminate any metal 14. This sequence runs in <10 sec and is used to position the remainder of the sequences.
that might be found in clothing. It is particularly useful to correct off-axis positioning in the coronal plane.

4. Have the patient remove any mascara or other metal-containing makeup to avoid local Sequence 2: T1-sagittal head
tissue heating and image artifacts. 15. This sequence serves as a true T1-weighted sagittal study of the head. Bring the
5. Inform the patient about what will occur during the procedure, what he/she will sequence for a sagittal T1-weighted scan up onto the console and utilize the parame-
experience while in the magnet, and how to behave, to include the following: ters in Table A4.2.3.

a. If earphones/headphones are used to protect the ears from loud sounds produced 16. Use the pilot sequence to set up the scan levels.
by the scanner, the patient will be asked to wear these, but will be able to
communicate with the technologist at any/all times during the procedure. Table A4.2.2 Primary Clinical Imaging Parameters for Sequence 1: Rapid
Three Plane Pilot
b. The patient will be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this equipment works). Patient position Supine
c. For optimum results the patient should not talk, and should avoid/minimize Scan type Gradient echo
swallowing or other movement during each scan, i.e., as long as the banging Imaging plane (orientation) Sagittal, transverse, and coronal
sounds continue. Between scans, talking and swallowing are allowed in most Central slice or volume center Midline head
cases, but should be avoided when comparative positional studies are being Echo time (TE) 3.7 msec
performed. The patient will be informed in these instances. Repeat time (TR) 16 msec
Flip angle (FA) 20°
d. Nevertheless, the patient may call out at any time if he/she feels it necessary.
Fields of view (FOVx, FOVy) 300 mm, 300 mm
6. Have the patient lay supine on the scanner table. Either before or directly after the Resolution (Δx, Δy) 2.34 mm, 2.34 mm
patient is positioned on the table, set up any triggering devices or other monitoring Number of data points collected (Nx, Ny) 128, 128
equipment necessary. Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 10 mm
7. Establish an i.v. line through which the contrast agent can be injected, and attach this Number of slices 3
line securely to the patient so that movement into or out of the magnet will not pull Slice gap NA
at the patient’s arm. It is preferable to insert the line prior to imaging. This allows no Number of acquisitions (Nacq) 1
intervening motion between those scans performed before and those run after the Read direction Left–right
contrast agent injection. Slice locations At isocenter, 3 orthogonal planes
8. Center the patient in the head coil. Make sure that the head and neck are constrained Saturation pulses None
to prevent unnecessary motion, especially if high-resolution scans are to be run. RAMa 2X
Scan time 6 sec Infectious
aZero padded from 128 by 128 points to 256 by 256 points.
Diseases of the
Meningitis Brain

A4.2.2 A4.2.3
Supplement 1 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 1
 17. Let the patient know you are ready, and begin the scan. 20. Let the patient know you are ready, and begin the scan.
Sequence 3: T1-transverse head Sequence 4: FLAIR transverse head
18. Bring the sequence for a transverse T1-weighted scan up onto the console. Set the 21. Bring the sequence for a transverse FLAIR scan up onto the console. Set the imaging
imaging parameters as shown in Table A4.2.4. parameters as shown in Table A4.2.5.
19. Use the pilot sequence to set up the scan levels.
22. Use the pilot sequence to set up the scan levels and a caudal saturation pulse.
Table A4.2.3 Primary Clinical Imaging Parameters for Sequence 2: T1-Sagittal 23. Let the patient know you are ready, and begin the scan.
Head
Sequence 5: FSE transverse head
Patient position Supine
24. Bring the sequence for a transverse FSE scan up onto the console. Set the imaging
Scan type Spin echo
parameters as shown in Table A4.2.6.
Imaging plane (orientation) Sagittal
Central slice or volume center Midline head 25. Use the pilot sequence to set up the scan levels.
Echo time (TE) 12 msec
26. Let the patient know you are ready, and begin the scan.
Repeat time (TR) 300 msec
Flip angle (FA) 90°
Sequence 6: T1-transverse post-gadolinium head
Fields of view (FOVx, FOVy) 240 mm, 240 mm
27. Bring the sequence for a transverse T1-weighted scan up onto the console. Set the
Resolution (Δx, Δy) 1.25 mm, 0.94 mm
imaging parameters as shown in Table A4.2.7.
Number of data points collected (Nx, Ny) 192, 256
Display matrix (Dx, Dy) 256, 256 28. Use the pilot sequence to set up the scan levels.
Slice thickness (Δz) 4 mm
29. Leaving the patient in the magnet, inject the contrast agent. This may be by either
Number of slices 15
Slice gap 1 mm
hand or mechanical injector. Observe the injection to insure there is no extravasation
Number of acquisitions (Nacq) 1
of the contrast agent. Flush the line with 10 ml of sterile normal saline. The scan may
Read direction Superior–inferior
begin as soon as the injection is completed. Alternatively, the scanning table may be
Slice locations Cover brain parenchyma
advanced out of the magnet for the injection, but the patient must remain in place.
Saturation pulses None A dose of 0.1 mmol/kg of contrast agent is usually given.
Scan time 1 min, 55 sec

Table A4.2.5 Primary Clinical Imaging Parameters for Sequence 4: FLAIR

Table A4.2.4 Primary Clinical Imaging Parameters for Sequence 3: Transverse Head
T1-Transverse Head
Patient position Supine
Patient position Supine Scan type FLAIR-FSE
Scan type Spin echo Imaging plane (orientation) Transverse (parallel to AC-PC line)
Imaging plane (orientation) Transverse (planum sphenoidale Central slice or volume center Mid-cranium
line)
Echo time (TE) 125 msec (effective)
Central slice or volume center Mid-cranium
Echo train length (ETL) Fixed (generally 12)
Echo time (TE) 12.1 msec
Repeat time (TR) 6000 msec
Repeat time (TR) 500 msec
Inversion time (TI) 1900 msec
Flip angle (FA) 90°
Flip angle (FA) 180°
Fields of view (FOVx, FOVy) 240 mm, 180 mm
Fields of view (FOVx, FOVy) 240 mm, 197 mm
Resolution (Δx, Δy) 0.94 mm, 0.94 mm
Resolution (Δx, Δy) 0.94 mm, 0.97 mm
Number of data points collected (Nx, Ny) 256, 192
Number of data points collected (Nx, Ny) 256, 204
Display matrix (Dx, Dy) 256, 256
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 5 mm
Slice thickness (Δz) 5 mm
Number of slices 24
Number of slices 24
Slice gap 1 mm
Slice gap 1 mm
Number of acquisitions (Nacq) 2
Number of acquisitions (Nacq) 1
Read direction Anterior–posterior
Read direction Anterior–posterior
Slice locations Foramen magnum to vertex
Slice locations Foramen magnum to vertex
Saturation pulses None
Saturation pulses Caudal to saturate arterial flow Infectious
Scan time 2 min, 36 sec Diseases of the
Meningitis Scan time 2 min, 48 sec Brain

A4.2.4 A4.2.5
Supplement 1 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 1
 Table A4.2.6 Primary Clinical Imaging Parameters for Sequence 5: FSE Table A4.2.8 Primary Clinical Imaging Parameters for Sequence 7: T1-Coronal
Transverse Head Post-Gadolinium Head

Patient position Supine Patient position Supine

Scan type Fast spin echo (FSE) Scan type Spin echo
Imaging plane (orientation) Transverse (planum sphenoidale Imaging plane (orientation) Coronal (perpendicular to planum
line) sphenoidale line)
Central slice or volume center Mid-cranium Central slice or volume center Mid-cranium
Echo time (TE) 105 msec (effective) Echo time (TE) 12.1 msec
Echo train length (ETL) 16 Repeat time (TR) 500 msec
Repeat time (TR) 5616 msec Flip angle (FA) 90°
Flip angle (FA) 90° Fields of view (FOVx, FOVy) 220 mm, 178 mm
Fields of view (FOVx, FOVy) 240 mm, 180 mm Resolution (Δx, Δy) 0.86 mm, 0.93 mm
Resolution (Δx, Δy) 0.63 mm, 0.70 mm Number of data points collected (Nx, Ny) 256, 192
Number of data points collected (Nx, Ny) 384, 256 Display matrix (Dx, Dy) 256, 256
Display matrix (Dx, Dy) 384, 384 Slice thickness (Δz) 5 mm
Slice thickness (Δz) 5 mm Number of slices 24
Number of slices 24 Slice gap 1 mm
Slice gap 1 mm Number of acquisitions (Nacq) 2
Number of acquisitions (Nacq) 2 Read direction Superior–inferior
Read direction Anterior–posterior Slice locations Frontal to occipital cerebral poles
Slice locations Foramen magnum to vertex Saturation pulses None
Saturation pulses Caudal to saturate arterial flow Scan time 2 min, 36 sec
Scan time 3 min, 22 sec

This is essentially the same sequence as the pre-gadolinium scan, but can be performed
Table A4.2.7 Primary Clinical Imaging Parameters for Sequence 6: with the addition of a flow compensation (FC) pulse to better delineate the cerebral vessels,
T1-Transverse Post-Gadolinium Head and a magnetization transfer (MT) pulse to optimize enhancing lesion detection.

Patient position Supine Sequence 7: T1-coronal post-gadolinium head

Scan type Spin echo 30. Bring the sequence for a T1-weighted scan up onto the console. Set the imaging
Imaging plane (orientation) Transverse (planum sphenoidale parameters as shown in Table A4.2.8.
line)
Central slice or volume center Mid-cranium 31. Use the pilot sequence to set up the scan levels.
Echo time (TE) 12.1 msec 32. Let the patient know you are ready, and begin the scan.
Repeat time (TR) 500 msec
Small abscesses may be much better delineated with a smaller field of view (FOV) (e.g.,
Flip angle (FA) 90°
FOVx = 200 mm, FOVy = 150 mm) and higher resolution thinner slice thickness (e.g., 3
Fields of view (FOVx, FOVy) 240 mm, 180 mm mm). Please note that this modification should be performed with one acquisition, and
Resolution (Δx, Δy) 0.94 mm, 0.94 mm limited to the area of abnormality to avoid prolonged scan times.
Number of data points collected (Nx, Ny) 256, 192
When scanning pediatric patients (under 24 months) modify the transverse and coronal
Display matrix (Dx, Dy) 256, 256 sequences by placing the FOV at 200 mm, and the slice thickness and slice gap at 4 mm
Slice thickness (Δz) 5 mm and 1 mm, respectively. For neonates, premature infants, and other extremely small patients
Number of slices 24 an extremity coil may be used.
Slice gap 1 mm
This sequence may be performed with the addition of a flow compensation (FC) pulse to
Number of acquisitions (Nacq) 2 better delineate the cerebral vessels, and a magnetization transfer (MT) pulse to optimize
Read direction Anterior–posterior enhancing lesion detection.
Slice locations Foramen magnum to vertex
Saturation pulses None
SPECIAL SITUATIONS ALTERNATE
Scan time 2 min, 36 sec
The two sequences that follow (gradient echo and magnetic resonance angiogram or PROTOCOL
venogram) are presented here for consideration by the radiologist in those patients
suspected of meningitis for whom the Basic Protocol has failed to elucidate a clearly
defined abnormality. Patients with meningitis can suffer venous hemorrhage, which can Infectious
be obvious and massive but is often subtle with microhemorrhage. The gradient echo Diseases of the
Meningitis Brain

A4.2.6 A4.2.7
Supplement 1 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 1
 sequence is the most sensitive MR sequence for detecting the magnetic susceptibility COMMENTARY
effect of the products of hemorrhage, particularly hemosiderin, and can display not only ment, the ependyma is at risk and ependymitis
Background Information
the helpful diagnostic clue of a hemorrhagic lesion but may well bring other lesions to Meningitis may affect the dura mater may ensue, particularly in neonatal disease.
the attention of the radiologist. This may in turn impact upon patient prognosis, or guide (pachymeningitis), or more commonly the pia Obstruction to normal CSF drainage pathways
more effective follow-up imaging. Magnetic resonance angiography or venography can and arachnoid maters (leptomeningitis). The is the norm because of loculations that may
aid in detection of an associated (inflammatory) angiitis, most often manifest in the most common presentation is headaches, form in the subarachnoid or ventricular spaces,
meningitis patient as venous thrombosis. Gadolinium administration may disclose such which may be accompanied by a stiff neck or sludging of inflammatory debris within these
abnormalities more effectively, as it provides greater signal-to-noise within the vascular (meningismus), confusion, and disorientation. pathways or the arachnoid granulations. This
tree. This improves conspicuity of subtle findings in the larger vessels and increases Fever is often present and cranial neuropathies may result in hydrocephalus and increased in-
visibility of the smaller vascular structures. may develop (Hansman-Whiteman et al., 1996; tracranial pressure, as well as sterile sympa-
Taveras and Pile-Spellman, 1996; Wong and thetic subdural effusions. These latter fluid col-
Set up patient and equipment Quint, 1999). The disease may be viral, bacte- lections are most often noted in children with
1. Use the same equipment and the same patient setup as for the previous method (see rial, or fungal in origin. The most common H. influenza infection along the frontal and
Basic Protocol, steps 1 to 11). etiology, and the most difficult form to diagnose parietal convexities, and may turn into subdural
is viral meningitis as imaging studies are usu- empyema.
2. Run the pilot scan as sequence 1 in the Basic Protocol. ally negative and cerebralspinal fluid (CSF) While leptomeningitis is a medical emer-
Sequence 8: Coronal gradient echo head abnormalities are not specific. Viral cultures gency, a subdural empyema is a neurosurgical
3. Bring the sequence for a gradient echo scan up onto the console. Set the imaging may reveal the pathogen as mumps or en- emergency as these infections can progress
terovirus (most commonly), and occasionally quite rapidly to vascular thrombosis of many
parameters as shown in Table A4.2.9.
the herpes or human immunodeficiency vi- cortical veins or larger dural venous sinuses.
4. Use the pilot sequence to set up the scan levels. ruses. Bacterial disease is most often caused by This can produce large cerebral venous in-
5. Let the patient know you are ready, and begin the scan. E. coli in infants, H. influenza in children and farctions and hemorrhage. These empyemas
in adults is predominantly secondary to menin- are usually the result of sinusitis (predomi-
Sequence 9: Magnetic resonance angiography (MRA) and venography (MRV) gococcal or pneumococcal species. Tubercu- nantly frontal), though they can occur secon-
6. Repeat step 29 in the Basic Protocol. lous and parasitic meningitides occur in en- dary to otitis media, and trauma.
demic populations. Fungal meningitis is mostly The only significant differential diagnostic
7. These sequences may be employed when the specific history provided for a patient
found in the immunocompromised population considerations in patients with meningitis are
suggests possible arterial or venous pathology, or if review of the spin-echo images with many pathogens described, to include meningeal carcinomatosis, noninfectious
indicates this possibility. In particular, the authors often find utility with MRV in cryptococcus, coccidioidomycosis, and asper- granulomatous meningitis (i.e., sarcoid), post-
evaluation of the meningitis patient. The reader is referred to the scanning parameters gillus (Harris and Enterline, 1997). The basilar traumatic meningeal effusions such as may be
outlined for these sequences in Chapter A1. meninges are most commonly affected by the seen in child abuse, and benign subdural hygro-
tuberculosis and fungal pathogens, a condition mas of infancy. The clinical scenario is usually
referred to as basilar meningitis. Alternate eti- all that is necessary to differentiate among these
Table A4.2.9 Primary Clinical Imaging Parameters for Sequence 8: Coronal ologies for basilar meningitis are other granu- entities. More important than differentiation
Gradient Echo Head lomatous conditions, including sarcoid. Basilar from these other conditions, is the initial detec-
meningitis is the most common form for tion of the often subtle findings of the disease
Patient position Supine chronic meningitis, though it may also present and then proper ascription of an infectious
Scan type 2D Gradient echo acutely. agent, and all of this in a timely manner in these
Imaging plane (orientation) Coronal The route of infection is usually hemato- often gravely ill patients.
Central slice or volume center Mid-cranium genous for leptomeningitis, primarily to the
Echo time (TE) 24.6 msec choroid plexus and then the CSF. But contigu- Critical Parameters and
ous spread of disease may be seen, particularly Troubleshooting
Repeat time (TR) 719 msec
in patients with sinus or middle ear disease, or There are several parameters that enhance
Flip angle (FA) 25°
post-operative or traumatized patients. How- the sensitivity and utility of the Basic Protocol.
Fields of view (FOVx, FOVy) 220 mm, 220 mm ever, all patients with meningitis should be The reader is referred to the prior unit on cere-
Resolution (Δx, Δy) 0.86 mm, 1.15 mm carefully evaluated for the primary source of bral abscess (UNIT A4.1) for further discussion of
Number of data points collected (Nx, Ny) 256, 192 the infection. several key parameters to optimize when imag-
Display matrix (Dx, Dy) 256, 256 Leptomeningitis begins in the pia mater, a ing the patient suspected of meningitis. Specifi-
Slice thickness (Δz) 5 mm vascular structure that allows for the passage of cally, as the early findings of meningitis may
Number of slices 21 leukocytes into the subarachnoid space where be quite subtle (or not yet present) proper at-
Slice gap 1 mm the relatively avascular arachnoid membrane tention to these parameters may make the dif-
Number of acquisitions (Nacq) 2 traps the cells, and pus accumulates. A phlebitis ference between a positive and negative study.
Read direction Superior–inferior eventually ensues in the traversing cortical These include the proper selection of the TI
Slice locations Frontal to occipital cerebral poles veins that leads to thrombosis, cortical venous (inversion time) for the FLAIR images to effect
congestion, and abscess and/or small in- nulling of hyperintense CSF signal and produce Infectious
Saturation pulses None Diseases of the
farctions. With subarachnoid space involve- dark CSF. This enables enhanced detection of
Meningitis Scan time 4 min, 36 sec Brain

A4.2.8 A4.2.9
Supplement 1 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 1
 abnormal periventricular high T2-signal inten-
sity for the important determination of ven-
triculitis/ependymitis, hydrocephalus with
transependymal flow of CSF, as well as better
evaluation of the subarachnoid space. Addi-
tionally, the optional use of the MT pulse for
the post-gadolinium sequences enhances the
sensitivity for early detection of subtle foci of
abnormal enhancement. However, it is the
authors’ experience that obtaining post-gad-
olinium sequences in multiple planes is also
essential to not only aid in the detection of
meningitis, but to confirm subtle findings seen
in another plane. The authors recommend both
transverse and coronal planes as these offer the
best evaluation of the subdural and subarach-
noid spaces, as well as the ependymal surfaces
of the ventricles for the ependymal and
subependymal enhancement of ventriculitis.
This may be difficult or impossible to demon-
strate in a single (transverse) plane. A double
dose of gadolinium contrast to better demon-
strate the abnormal meningeal enhancement is
rarely used.
When considering the optional sequences,
the optimal choice of TE on the gradient echo
sequence (sequence 8) is important to allow for
detection of foci of microhemorrhage. A longer
TE will demonstrate increased T2-contrast, and
magnetic susceptibility with more pronounced
low signal “blooming” from foci of hemor-
rhage containing hemosiderin (though at the
expense of overall tissue signal to noise ratio).
Generally, TE values of <25 msec are less sen-
sitive for this purpose (NOTE: the authors use
a TE of 24.6 msec). For similar reasons, this
“blooming” makes for poor evaluation of signal
at the skull-brain interface, particularly at the
skull base. Thus, the meninges are not well
demonstrated on the gradient echo sequence.
The MR angiographic sequences (MRA and
MRV) are utilized for large vessel status rather
than subtle small vessel findings. The latter are
much better evaluated on catheter angiograms,
and are generally not present unless there are
marked abnormalities noted on spin echo se-
quences (Pomper et al., 1999). Thus, MRA is
used to evaluate the circle-of-Willis, while
MRV is used for the larger dural venous sinuses.
The authors routinely employ MRV in patients
suspected of meningitis as these sinuses are
commonly affected by the disease (Fig.
A4.2.1), but will only obtain an MRA if there
is supportive evidence of arterial vascular pa-
thology on spin-echo sequences, such as suspi-
cious hemorrhage or abnormal vascular con-
Meningitis tours or luminal signal.
A4.2.10
Supplement 1
Imaging of meningitis by magnetic
resonance diffusion
The Basic Protocol outlined above is quite A B
sufficient in most instances of meningitis im-
aging and represents the standard for imaging
of the patient suspected for meningitis. dMRI
has no defined role in the diagnosis of menin-
gitis but can be utilized in the patient suspected
of meningitis to assess for one of the more
common complications of the disease, specifi-
cally cerebral infarction. These are usually ve-
nous infarctions and may often be accompanied
by hemorrhage, especially when larger. Please
refer to the scan parameters and set-up outlined
in detail in UNIT A4.1 of this chapter for perform-
ance of this sequence.
Anticipated Results
The role of MR imaging in the evaluation of
the patient suspected to have meningitis is pre-
dominantly to assess for the presence of ab-
scess, localized empyema and ventriculitis. The
paranasal sinuses and temporal bones may also
be scrutinized, but are mostly better seen with
computed tomography (CT). The other main C
benefit of MR is to assess for the complications
of the disease such as hydrocephalus, cerebral
infarction, and to evaluate treatment efficacy
and disease progression. However, it must al-
ways be kept in mind and readily communi-
cated to referring clinicians, that a negative
imaging result in a patient suspected of menin-
gitis must always be followed by a diagnostic
CSF tap. This test remains the “gold standard”
and has higher sensitivity and specificity for
meningitis than does MRI, as well as the added
necessary benefit of allowing for specific
pathogen identification and culture, and deter-
mination of antibiotic sensitivities.
At initial presentation, meningitis typically
appears on MR imaging as either a thickening
of the meningeal layer(s), or as a collection of
fluid signal material in either the subarachnoid
or subdural compartments. The disease may be
localized or diffuse, and may only involve the
basilar meninges. Depending on the pathogen
and the burden of disease these findings may Figure A4.2.1 This 20-year-old female with known pulmonary tuberculosis presented with new-onset seizures.
be either quite subtle or striking. If the disease MRI shows parenchymal hemorrhage of a venous infarction pattern on this T2-weighted (TR = 2500 msec, TE =
r esid es in the subdural compartment 80 msec, Nacq = 0.75) transverse (A) vertex image (arrows). Also noted is abnormal increased T2 signal intensity
(pachymeningitis), then the dura may appear suggesting dural thickening in bifrontal distribution (arrowheads). Enhancement of the dura confirming pachymen-
abnormally thick and low in signal intensity on ingitis is seen on the T1-weighted (TR = 600 msec, TE = 23 msec, Nacq = 2) post-contrast coronal (B) image
T1-weighted images, while T2-weighting will (arrowheads). The MRV (C) discloses abnormal attenuation of the superior sagittal sinus (arrow) in parietal location,
confirming a stenosis likely causing the venous hypertension and infarction in this patient.
disclose high fluid-like signal. The signal of an
empyema is between that of CSF and brain on
T1-weighted imaging and occasionally, a fluid-
debris level may be demonstrated in the sub-
dural space. Gadolinium sequences will dis-
A4.2.11
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 1
 close abnormal thickening and prominent en-
A hancement of the involved dura. A contiguous
infectious source for the collection should be
sought, with prime candidates being the middle
ear spaces and the frontal sinuses. The gradient
echo and MRV sequences may have utility in
these patients, disclosing hemorrhagic foci
(these may be numerous) and dural sinus
thrombosis (especially the superior sagittal,
straight, and transverse sinuses), respectively.
With leptomeningitis, abnormal fluid-like
signal may be found within the subarachnoid
spaces with abnormally prominent or asym-
metric sulci seen. Similar findings may be
found in the basilar compartment and the
spaces of Virchow-Robin that may be subtly
involved with an unusual prominence and num-
ber to the spaces noted. Gelatinous pseudocysts
have been described with a “soap-bubble” ap-
pearance within the basilar Virchow-Robin
spaces in patients with cryptococcal meningi-
tis. These lesions appear similar in signal to
CSF, but are delineated by a barely discernible
membrane. Gadolinium images will show en-
hancement of the basilar cisterns, the perivas-
B cular Virchow-Robin spaces, or a “sugar-coat-
ing” of the brain and brainstem in the involved
regions (Fig. A4.2.2). The enhancing lep-
tomeninges can appear thickened and even
nodular, though this is less common. This is the
inflamed pia mater with abnormal thickening
and enhancement. Inflammatory septa may be
found within the subarachnoid space or within
the ventricles, and ependymal enhancement
may be present, possibly with periventricular
edema. Cortical hyperintensity on T2-weighted
images may be noted from edema related to
adjacent meningeal inflammation and associ-
ated cerebritis or abscess may be seen. Hydro-
cephalus is a common occurrence eventually in
the meningitis patient and should be sought.
Additionally, leptomeningitis may rarely lead
to vasculitic involvement of the major arterial
tributaries of the brain as they traverse the
affected subarachnoid space. This may mani-
fest as subarachnoid or intraventricular hemor-
rhage due to rupture of the involved vessel,
which can become aneurysmal as a result of the
inflammation. These may be true or false
Figure A4.2.2 This 16-year-old boy presented to the emergency room with subtle mental status
changes and a fever of one week duration. A proton-density weighted (TR = 2500 msec, TE = 30 (pseudo-) aneurysms. Alternatively, infarction
msec, Nacq = 0.75) transverse image (A) through the lateral ventricles shows increased ventricular of perfused areas of the brain may result.
volume relative to the extra-transverse CSF spaces. There is also subtle abnormal increased signal
Chronically, bulky subarachnoid exudates
seen rimming the ventricles (arrowheads) compatible with acute hydrocephalus and transependy- may result, especially with basilar meningitis.
mal flow of CSF. The transverse pre- and post-contrast (TR = 500 msec, TE = 11 msec, Nacq = 2) Abnormal enhancement may be seen with MRI
images (B) at the level of the basilar cisterns demonstrates diffuse “sugar-coating” enhancement even years after disease onset, and atrophy may
of the basilar meninges (arrows on the post-contrast image) diagnostic of basilar meningitis. A
result. Meningeal calcifications can occur, es-
positive tuberculin skin test was obtained and CSF analysis was confirmatory for tuberculosis.
Meningitis pecially with tuberculosis.
A4.2.12
Supplement 1 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
An unusual form of meningitis (though ex-
ceedingly common in endemic regions) is cys-
ticercal meningitis. This disease also com-
monly causes parenchymal infection and has
different appearances depending on when in the
infectious cycle it is imaged (Hansman-White-
man et al., 1996; Taveras and Pile-Spellman,
1996; Wong and Quint, 1999). This cycle runs
between 2 and 10 years for the pork tapeworm
larva (T. solium), averaging 5 years. It usually
demonstrates one or several gray-white matter
junction or subarachnoid/ventricular rounded
lesions measuring 1 to 2 mm in size and har-
boring a small mural nodule, or scolex. This is
the dead parasite, as there is generally no in-
flammatory involvement of the central nervous
system (CNS) until the parasite dies. These
lesions eventually calcify and may reveal inert
imaging findings of isointense signal to normal
brain on T1-weighted images with isointense to
hypointense T2 signal. Variable edema and wall
enhancement are noted until the chronic stage,
however.
Literature Cited
Hansman-Whiteman, M., Bowen, B.C., Donovan-
Post, M.J., and Bell, M.D. 1996. Intracranial
infection. In Magnetic Resonance Imaging of the
Brain and Spine, 2nd edition (Atlas, S.W., ed.)
pp. 707-772. Lippincott-Raven, Philadelphia.
Harris, D.E. and Enterline, D.S. 1997. Neuroimag-
ing of AIDS. I. Fungal infections of the central
nervous system. Neuroimaging Clin. N. Am.
7:187-198.
Pomper, M.G., Miller, T.J., Stone, J.H., Tidmore,
W.C., and Hellmann, D.B. 1999. CNS vasculitis
in autoimmune disease: MR imaging findings
and correlation with angiography. A.J.N.R.
20:75-85.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Taveras, J.M. and Pile-Spellman, J. 1996. Inflamma-
tory diseases. In Neuroradiology, 3rd ed. pp.
259-326. Williams and Wilkins, Baltimore.
Wong, J. and Quint, D.J. 1999. Imaging of central
nervous system infections. Semin. Roentgenol.
34:123-143.
Key Reference
Harris, D.E. and Enterline, D.S. 1997. See above.
Contains lucid explanations for the physics and
basic scan parameters of standard and advanced
magnetic resonance imaging studies.
Contributed by Andrew E. Auber
and Clifford Belden
Brooke Army Medical Center Infectious
Diseases of the
San Antonio, Texas
Brain
A4.2.13
Supplement 1
Human Immunodeficiency Virus (HIV)
Magnetic resonance imaging (MRI) for the evaluation of patients infected with human
immunodeficiency virus (HIV), as with most other forms of intracranial inflammatory or
infectious diseases, is a powerful though largely nonspecific diagnostic tool. For imaging
of these complex patients with the varied and numerous pathologies they may harbor, the
Basic Protocol presented in UNIT A4.1 is utilized to include gadolinium-enhanced se-
quences. Several optional imaging sequences, to include magnetic resonance diffusion
(dMRI), magnetic resonance perfusion (pMRI), and magnetic resonance spectroscopy
(MRS) are outlined, which can be employed should patient tolerance allow and if specific
clinical situations exist to be further clarified. The parameters given here are derived from
experience at 1.5 T and may need to be altered slightly depending on the field strength
available and the specific equipment manufacturer.
IMAGING OF HIV
This protocol presents the authors routine MR head examination for patients with
intracranial inflammatory disease, with gadolinium contrast agent sequences. The total
number of sequences presented is seven (scout, T1-weighted sagittal and transverse,
fluid-attenuated inversion recover (FLAIR) and fast spin-echo (FSE) T2-weighted trans-
verse, and post-contrast T1-weighted transverse and coronal). This examination may be
shortened considerably by limiting the use of two planes for post-contrast imaging to a
single transverse plane for exams with a lower clinical suspicion for disease, or if review
of the images obtained through the transverse gadolinium-enhanced images proves
unremarkable. The sequences described herein are based on the authors’ experience with
a Marconi Medical Systems 1.5-T scanner, but are expected to be equally applicable to
machines from other manufacturers. The authors utilize the same basic scanning protocol
to image HIV patients as that for other cerebral inflammatory conditions (UNIT A4.1), which
follows below.
Table A4.3.1 lists the hardware necessary to perform the examination, along with
appropriate parameters.
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as a crash cart or oxygen. Also ensure only magnetic field compatible oxygen
tanks are utilized if wall-source oxygen is unavailable.
Materials
Extravascular contrast agent (e.g., 0.1 mmol/kg patient body weight of gadolinium
chelate from Mangevist, Omniscan, or Prohance)
Normal saline (0.9% NaCl) sterile
Table A4.3.1 Equipment Parameters for Contrast-Enhanced Head
Imaging Sequences
Coil type Head
Gradient coil strength 25 mT/m (or whatever the
system permits)
Gating (cardiac, respiratory, No
peripheral)
Respirator or oxygen If required by patient
Motion cushions Recommended
Contrast agents Yes
Contributed by Andrew E. Auber and Clifford Belden
Current Protocols in Magnetic Resonance Imaging (2001) A4.3.1-A4.3.16
Copyright © 2001 by John Wiley & Sons, Inc.
UNIT A4.3 Set up patient and equipment
1. Interview (screen) the patient to ensure that he/she has no contraindication for MR
scanning such as cardiac pacemaker or other implants containing ferromagnetic
materials. Also be sure to find out if the patient has any health conditions that may
require the presence of special emergency equipment during the scanning procedure,
or that necessitate other special precautions.
Generally standard screening forms are used for all patients scanned in a magnetic
resonance system.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(1996) for discussion of what implants may be safely scanned using magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
BASIC be screened as well to ensure the absence of loose metal objects on the body or clothing.
PROTOCOL
2. If the procedure is a research protocol, have the patient sign any necessary consent
form.
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
4. Have the patient remove any mascara or other metal-containing makeup to avoid local
tissue heating and image artifacts.
5. Inform the patient about what will occur during the procedure, what he/she will
experience while in the magnet, and how to behave, to include the following:
a. If earphones/headphones are used to protect the ears from loud sounds produced
by the scanner, the patient will be asked to wear these, but will be able to
communicate with the technologist at any/all times during the procedure.
b. The patient will be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this equipment works).
c. For optimum results the patient should not talk, and should avoid/minimize
swallowing or other movement during each scan, i.e., as long as the banging
sounds continue. Between scans, talking and swallowing are allowed in most
cases, but should be avoided when comparative positional studies are being
performed. The patient will be informed in these instances.
d. Nevertheless, the patient may call out at any time if he/she feels it necessary.
6. Have the patient lay supine on the scanner table. Either before or directly after the
patient is positioned on the table, set up any triggering devices or other monitoring
equipment necessary.
7. Establish an intravenous (i.v.) line through which the contrast agent can be injected,
and attach this line securely to the patient so that movement into or out of the magnet
will not pull at the patient’s arm. It is preferable to insert the line prior to imaging.
This allows no intervening motion between those scans performed before and those
run after the contrast agent injection.
8. Center the patient in the head coil. Make sure that the head and neck are constrained
to prevent unnecessary motion, especially if high-resolution scans are to be run.
Infectious Human 9. If needed, place a pillow or other support under the knees to make the patient more
Diseases of the Immunodeficiency
Brain
comfortable.
Virus (HIV)
A4.3.1 A4.3.2
Supplement 2 Supplement 2 Current Protocols in Magnetic Resonance Imaging
10. Use the centering light to center on the nasion of the patient and place them into the Sequence 3: T1-transverse head
center of the magnet. 18. Bring the sequence for a transverse T1-weighted scan up onto the console. Set the
imaging parameters as shown in Table A4.3.4.
Once this step has been performed, so long as the patient does not move on the table, the
table itself can be moved and then replaced in the same position as before without 19. Use the pilot sequence to set up the scan levels.
jeopardizing the positioning of one scan relative to another.
20. Let the patient know you are ready, and begin the scan.
11. If the patient is unable to hold still, either provide an appropriate sedative, or arrange
with anesthesiology for conscious or general anesthesia. Table A4.3.3 Primary Clinical Imaging Parameters for Sequence 2: T1-Sagittal
Head
Alternatively, a low-field open magnet may be sought for scanning the claustrophobic
patient. There are also fast scanning techniques that may be employed for imaging of these
Patient position Supine
patients (see Chapter A5).
Scan type Spin echo
12. This Basic Protocol can be performed in <30 min. Imaging plane (orientation) Sagittal
Central slice or volume center Midline head
Sequence 1: Rapid three-plane pilot Echo time (TE) 12 msec
13. Run the three-plane pilot (Table A4.3.2) sequence to evaluate the patient positioning Repeat time (TR) 300 msec
in the magnet. Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 240 mm, 240 mm
14. This sequence runs in <10 sec and is used to position the remainder of the sequences.
Resolution (Δx, Δy) 1.25 mm, 0.94 mm
It is particularly useful to correct off-axis positioning in the coronal plane.
Number of data points collected (Nx, Ny) 192, 256
Sequence 2: T1-sagittal head Display matrix (Dx, Dy) 256, 256
15. This sequence serves as a true T1-weighted sagittal study of the head. Bring the Slice thickness (Δz) 4 mm
sequence for a sagittal T1-weighted scan up onto the console and utilize the parame- Number of slices 15
ters in Table A4.3.3. Slice gap 1 mm
Number of acquisitions (Nacq) 1
16. Use the pilot sequence to set up the scan levels. Read direction Superior–inferior
Slice locations Cover brain parenchyma
17. Let the patient know you are ready, and begin the scan.
Saturation pulses None
Scan time 1 min, 55 sec
Table A4.3.2 Primary Clinical Imaging Parameters for Sequence 1: Rapid
Three-Plane Pilot
Table A4.3.4 Primary Clinical Imaging Parameters for Sequence 3:
T1-Transverse Head
Patient position Supine
Scan type Gradient echo Patient position Supine
Imaging plane (orientation) Sagittal, transverse, and coronal Scan type Spin echo
Central slice or volume center Midline head Imaging plane (orientation) Transverse (planum sphenoidale
Echo time (TE) 3.7 msec line)
Repeat time (TR) 16 msec Central slice or volume center Mid-cranium
Flip angle (FA) 20° Echo time (TE) 12.1 msec
Fields of view (FOVx, FOVy) 300 mm, 300 mm Repeat time (TR) 500 msec
Resolution (Δx, Δy) 2.34 mm, 2.34 mm Flip angle (FA) 90°
Number of data points collected (Nx, Ny) 128, 128 Fields of view (FOVx, FOVy) 240 mm, 180 mm
Display matrix (Dx, Dy) 256, 256 Resolution (Δx, Δy) 0.94 mm, 0.94 mm
Slice thickness (Δz) 10 mm Number of data points collected (Nx, Ny) 256, 192
Number of slices 3 Display matrix (Dx, Dy) 256, 256
Slice gap NA Slice thickness (Δz) 5 mm
Number of acquisitions (Nacq) 1 Number of slices 24
Read direction Left–right Slice gap 1 mm
Slice locations At isocenter, 3 orthogonal planes Number of acquisitions (Nacq) 2
Saturation pulses None Read direction Anterior–posterior
RAMa 2X Slice locations Foramen magnum to vertex
Scan time 6 sec Infectious Human Saturation pulses None
aZero padded from 128 by 128 points to 256 by 256 points. Diseases of the Immunodeficiency Scan time 2 min, 36 sec
Brain Virus (HIV)

A4.3.3 A4.3.4
Current Protocols in Magnetic Resonance Imaging Supplement 2 Supplement 2 Current Protocols in Magnetic Resonance Imaging
Sequence 4: FLAIR transverse head Table A4.3.6 Primary Clinical Imaging Parameters for Sequence 5: FSE
21. Bring the sequence for a transverse FLAIR scan up onto the console. Set the imaging Transverse Head
parameters as shown in Table A4.3.5.
Patient position Supine
22. Use the pilot sequence to set up the scan levels and a caudal saturation pulse. Scan type Fast spin echo (FSE)
23. Let the patient know you are ready, and begin the scan. Imaging plane (orientation) Transverse (planum sphenoidale
line)
Sequence 5: FSE transverse head Central slice or volume center Mid-cranium
24. Bring the sequence for a transverse FSE scan up onto the console. Set the imaging Echo time (TE) 105 msec (effective)
parameters as shown in Table A4.3.6. Echo train length (ETL) 16
25. Use the pilot sequence to set up the scan levels. Repeat time (TR) 5616 msec
Flip angle (FA) 90°
26. Let the patient know you are ready, and begin the scan. Fields of view (FOVx, FOVy) 240 mm, 180 mm
Sequence 6: T1-transverse post-gadolinium head Resolution (Δx, Δy) 0.63 mm, 0.70 mm
27. Bring the sequence for a transverse T1-weighted scan up onto the console. Set the Number of data points collected (Nx, Ny) 384, 256
imaging parameters as shown in Table A4.3.7. Display matrix (Dx, Dy) 384, 384
Slice thickness (Δz) 5 mm
28. Use the pilot sequence to set up the scan levels. Number of slices 24
29. Leaving the patient in the magnet, inject the contrast agent. This may be by either Slice gap 1 mm
hand or by a mechanical injector. Observe the injection to insure there is no Number of acquisitions (Nacq) 2
extravasation of the contrast agent. Flush the line with 10 ml of sterile normal saline. Read direction Anterior–posterior
The scan may begin as soon as the injection is completed. Slice locations Foramen magnum to vertex
Alternatively, the scanning table may be advanced out of the magnet for the injection, but Saturation pulses Caudal to saturate arterial flow
the patient must remain in place. Scan time 3 min, 22 sec
A dose of 0.1 mmol/kg of contrast agent is usually given.
This is essentially the same sequence as the pre-gadolinium scan, but can be performed Table A4.3.7 Primary Clinical Imaging Parameters for Sequence 6:
with the addition of a flow compensation (FC) pulse to better delineate the cerebral vessels, T1-Transverse Post-Gadolinium Head
and a magnetization transfer (MT) pulse to optimize enhancing lesion detection.
Patient position Supine
Scan type Spin echo
Table A4.3.5 Primary Clinical Imaging Parameters for Sequence 4: FLAIR
Imaging plane (orientation) Transverse (planum sphenoidale
Transverse Head
line)
Patient position Supine Central slice or volume center Mid-cranium
Scan type FLAIR-FSE Echo time (TE) 12.1 msec
Imaging plane (orientation) Transverse (parallel to AC-PC line) Repeat time (TR) 500 msec
Central slice or volume center Mid-cranium Flip angle (FA) 90°
Echo time (TE) 125 msec (effective) Fields of view (FOVx, FOVy) 240 mm, 180 mm
Echo train length (ETL) Fixed (generally 12) Resolution (Δx, Δy) 0.94 mm, 0.94 mm
Repeat time (TR) 6000 msec Number of data points collected (Nx, Ny) 256, 192
Inversion time (TI) 1900 msec Display matrix (Dx, Dy) 256, 256
Flip angle (FA) 180° Slice thickness (Δz) 5 mm
Fields of view (FOVx, FOVy) 240 mm, 197 mm Number of slices 24
Resolution (Δx, Δy) 0.94 mm, 0.87 mm Slice gap 1 mm
Number of data points collected (Nx, Ny) 256, 204 Number of acquisitions (Nacq) 2
Display matrix (Dx, Dy) 256, 256 Read direction Anterior–posterior
Slice thickness (Δz) 5 mm Slice locations Foramen magnum to vertex
Number of slices 24 Saturation pulses None
Slice gap 1 mm Scan time 2 min, 36 sec
Number of acquisitions (Nacq) 1
Read direction Anterior–posterior
Slice locations Foramen magnum to vertex
Saturation pulses Caudal to saturate arterial flow Infectious Human
Diseases of the Immunodeficiency
Scan time 2 min, 48 sec Brain Virus (HIV)

A4.3.5 A4.3.6
Current Protocols in Magnetic Resonance Imaging Supplement 2 Supplement 2 Current Protocols in Magnetic Resonance Imaging
 Table A4.3.8 Primary Clinical Imaging Parameters for Sequence 7: T1-Coronal prognosis, or guide more effective follow-up imaging. The second optional sequence
Post-Gadolinium Head (3-dimensional gradient echo post-contrast) is usually reserved for higher resolution of
small lesions or adjacent structures that may be affected by the pathology. It can also be
Patient position Supine
an effective aid in operative planning, and may indeed serve as the base MRI sequence
Scan type Spin echo
for several of the commercially available image-guided surgery tools presently on the
Imaging plane (orientation) Coronal (perpendicular to planum
market and utilized by many neurosurgeons. This sequence can be utilized in lieu of the
sphenoidale line)
transverse and coronal post-gadolinium sequences outlined in the Basic Protocol, if
Central slice or volume center Mid-cranium
desired. It is easily manipulated in post-processing to yield a desired projection of a lesion,
Echo time (TE) 12.1 msec
Repeat time (TR) 500 msec
or simply to give a multiplanar assessment of the brain in a given patient. Magnetic
Flip angle (FA) 90°
resonance angiography or venography can aid in detection of an associated (inflamma-
Fields of view (FOVx, FOVy) 220 mm, 178 mm
tory) angiitis, most often manifest in the HIV patient as venous thrombosis. Gadolinium
Resolution (Δx, Δy) 0.86 mm, 0.93 mm
administration may disclose such abnormalities more effectively, as it provides greater
Number of data points collected (Nx, Ny) 256, 192
signal-to-noise within the vascular tree. This improves conspicuity of subtle findings in
Display matrix (Dx, Dy) 256, 256
the larger vessels and increases visibility of the smaller vascular structures.
Slice thickness (Δz) 5 mm
Set up patient and equipment
Number of slices 24
1. Use the same equipment and the same patient set-up as for the previous method (see
Slice gap 1 mm
Basic Protocol, steps 1 to 11).
Number of acquisitions (Nacq) 2
Read direction Superior–inferior 2. Run the pilot scan as sequence 1 in the Basic Protocol.
Slice locations Frontal to occipital cerebral poles
Saturation pulses None Sequence 8: Coronal gradient echo (head)
Scan time 2 min, 36 sec 3. Bring the sequence for a gradient echo scan up onto the console. Set the imaging
parameters as shown in Table A4.3.9.

Sequence 7: T1-coronal post-gadolinium head
30. Bring the sequence for a T1-weighted scan up onto the console. Set the imaging
parameters as shown in Table A4.3.8.
31. Use the pilot sequence to set up the scan levels.
32. Let the patient know you are ready, and begin the scan.
Small abscesses may be much better delineated with a smaller field of view (FOV; e.g.,
FOVx = 200 mm, FOVy = 150 mm) and higher resolution thinner slice thickness (e.g., 3
mm). Please note that this modification should be performed with one acquisition, and
limited to the area of abnormality to avoid prolonged scan times.
When scanning pediatric patients (under 24 months) modify the transverse and coronal
sequences by placing the FOV at 200 mm, and the slice thickness and gap at 4 mm and 1
mm, respectively. For neonates, premature infants, and other extremely small patients an
extremity coil may be used.
This sequence may be performed with the addition of a flow compensation (FC) pulse to
better delineate the cerebral vessels, and a magnetization transfer (MT) pulse to optimize
enhancing lesion detection.
SPECIAL SITUATIONS
The three sequences that follow (gradient echo; 3-dimensional gradient echo post-gad-
olinium; magnetic resonance angiogram or venogram) are presented here for considera-
tion by the radiologist in those patients suspected of significant intracranial HIV-related
inflammatory conditions for whom the Basic Protocol has failed to elucidate a clearly
defined abnormality. These patients can suffer intracranial lesional hemorrhage, for which
the gradient echo sequence is extremely useful for demonstrating microhemorrhage. This
can produce not only the helpful diagnostic clue of a hemorrhagic lesion but may well
bring other lesions to the attention of the radiologist. This may in turn impact upon patient
4. Use the pilot sequence to set up the scan levels.
5. Let the patient know you are ready, and begin the scan.
Sequence 9: 3-Dimensional post-gadolinium transverse gradient echo (head)
6. Repeat step 29 in the Basic Protocol.
7. This sequence is often employed when scanning patients with HIV and may do so in
lieu of the routine post-contrast 2-D transverse and coronal spin echo sequences
described in the Basic Protocol, or when surgery is contemplated. It can be used to
better delineate smaller lesions and can be reformatted into any desired plane as it is
a volume-acquired sequence performed with high resolution parameters as described
in Table A.4.3.10.
8. This scan generates 120 images as a whole brain study, but higher resolution image
sets with 32 or 64 images, with a resolution as low as 1 mm can be produced, but
may be restrictive in area of coverage.
9. Planar reformations of the 3-D data set can be performed on- or off-line using the
reformation software. One can coordinate with the neurosurgeon to obtain the most
useful projections, which can then be saved and/or filmed.
ALTERNATE
PROTOCOL 10. Note that the cerebral vasculature is quite bright on this sequence as this 3-D gradient
echo sequence is the base-sequence for time-of-flight (TOF) magnetic resonance
angiography (MRA).
Sequence 10: Magnetic resonance angiography (MRA) and venography (MRV)
11. These sequences may be employed when the specific history provided for a patient
suggests possible arterial or venous pathology, or if review of the spin echo images
Infectious Human indicates this possibility. MRV is routinely employed when evaluating patients with
Diseases of the Immunodeficiency
Brain Virus (HIV) suspected meningitis, but will usually reserve the use of MRA to those patients

A4.3.7 A4.3.8
Current Protocols in Magnetic Resonance Imaging Supplement 2 Supplement 2 Current Protocols in Magnetic Resonance Imaging
 Table A4.3.9 Primary Clinical Imaging Parameters for Sequence 8: Coronal demonstrating concerning arterial findings on spin echo images, or with suspicious
Gradient Echo (Head) hemorrhage noted. The reader is referred to the scanning parameters outlined for
these sequences in Chapter A1 of this series.
Patient position Supine
Scan type 2-D gradient echo
Imaging plane (orientation) Coronal COMMENTARY
Central slice or volume center Mid-cranium Background Information imaging diagnosis of these many varied pa-
Echo time (TE) 24.6 msec Human immunodeficiency virus (HIV) is thologies found in HIV patients is most devel-
Repeat time (TR) 719 msec the virus responsible for the AIDS epidemic. In oped presently with MRI, but may require the
Flip angle (FA) 25° the United States, an estimated 650,000 to addition of other imaging techniques such as
Fields of view (FOVx, FOVy) 220 mm, 220 mm 900,000 people are infected with HIV, and computed tomography (CT), single-photon
Resolution (Δx, Δy) 0.86 mm, 1.15 mm >200,000 of these individuals are unaware of emission computed tomography (SPECT), and
Number of data points collected (Nx, Ny) 256, 192 their infection (Fauci, 1999; Wong and Quint, positron emission tomography (PET). A simple
Display matrix (Dx, Dy) 256, 256 1999). There are 40,000 new infections per year method to add a degree of specificity to the
Slice thickness (Δz) 5 mm with the infection rate in heterosexuals, and diagnostic workup is to search for primary
especially women, accelerating greatly in the infection sources in the patient, such as the
Number of slices 21
1990s. Worldwide in 1998 there were >33 mil- lungs for many of the infectious agents. Differ-
Slice gap 1 mm
lion people worldwide with HIV infection and ent pathologies may occur synchronously or
Number of acquisitions (Nacq) 2 AIDS was listed as the fourth leading cause of sequentially, confusing attempts at simplified
Read direction Superior–inferior death, with an estimated number of 2.3 million. diagnosis in these patients. The most common
Slice locations Frontal to occipital cerebral poles However, the age-adjusted death rate from example of this being the problematic differen-
Saturation pulses None AIDS in the United States declined 48% from tiation of the lesions of toxoplasmosis from
Scan time 4 min, 36 sec 1996 to 1997. This improvement in the death lymphoma. Clinical presentation depends more
rate is multifactorial in origin, but largely the on anatomic location within the CNS than on
result of new treatment advances, in particular lesion etiology. A categorization scheme is dis-
the development of highly active antiretroviral cussed below (see Anticipated Results), con-
Table A4.3.10 Primary Clinical Imaging Parameters for Sequence 9: 3-D therapy (HAART). These developments under- sisting of four lesion classes, (1) focal lesions
Post-Gadolinium Transverse Gradient Echo (Head)
score the need for early and accurate diagnosis with mass effect; (2) focal lesions without sig-
and follow-up of these challenging patients. nificant mass effect; (3) diffuse global CNS
Patient position Supine
There remain subsets of pathologies in the abnormalities; and (4) ventriculitis, meningitis,
Scan type 3-D gradient echo AIDS population that are refractory to treat- and infarcts (Walot et al., 1996).
Imaging plane (orientation) Transverse (option sagittal, or ment at this time. Chief among these is progres-
coronal) sive multifocal leukoencephalopathy (PML). Critical Parameters and
Central slice or volume center Mid-cranium for whole brain, or This fulminant opportunistic central nervous Troubleshooting
area of abnormality system (CNS) infection occurs in 4% of AIDS Please see the discussions presented in UNITS
Echo time (TE) 5 msec patients and is caused by the Jakob-Creutzfeldt A4.1 and A4.2 for useful remarks in optimizing
Repeat time (TR) 11.2 msec (or minimum) (JC) papovavirus (JCV). It carries a mean sur- the Basic Protocol. Some additional discussion
Flip angle (FA) 35° vival of 2.5 to 4 months from the time of related to the imaging of HIV infected patients
Fields of view (FOVx, FOVy) 240 mm, 187 mm diagnosis. Whereas brain biopsy was pre- is warranted. Considering the utility and bene-
Resolution (Δx, Δy) 0.94 mm, 0.97 mm viously the mainstay for the diagnosis of PML, fits of the contrast-enhanced 3-D gradient echo
Number of data points collected (Nx, Ny) 256, 192 recent advances in cerebrospinal fluid (CSF) sequence if used as a whole brain sequence, it
Display matrix (Dx, Dy) 256, 256 analysis (>92% specific for JCV) when coupled can either supplement or replace the post-con-
Slice thickness (Δz) 1.2 mm with typical MR changes of the disease and a trast transverse and coronal spin echo se-
Number of slices 120 clinical presentation consistent with PML, have quences. It affords superior spatial resolution
produced an accepted and much less invasive and some time savings, but at some mild loss
Slice gap 0 mm
diagnostic regimen (Donovan-Post et al., of signal-to-noise contrast. The sequence is
Number of acquisitions (Nacq) 1
1999). routinely performed as a problem-solver in that
Read direction Anterior–posterior, if imaging Approximately 40% to 90% of patients with the high resolution images (especially when
plane is transverse
AIDS will develop CNS manifestations during limited to area of interest with thinner slice
Slice locations Whole brain the course of the illness (Walot et al., 1996). thickness) can give the exact location of subtle
No phase wrap (NPW)a Yes The most common of these opportunistic infec- lesions (i.e., in the subarachnoid space of a
Saturation pulses NA tions are caused by toxoplasma, cytomega- cerebral sulcus or within adjacent cerebral cor-
Scan time 4 min, 19 sec lovirus (CMV), cryptococcus, and the JCV tex). This determination is readily aided by the
aPhase oversampling or anti-alrasing. causing PML. Less commonly observed agents ability to reformat the sequence into any desired
Infectious Human include herpes simplex virus (HSV), varicella- plane.
Diseases of the Immunodeficiency zoster, histoplasmosis and tuberculosis. The
Brain Virus (HIV)

A4.3.9 A4.3.10
Current Protocols in Magnetic Resonance Imaging Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging
 There is a distinct though rare association of
HIV and AIDS with CNS vasculopathy, par-
ticularly in the pediatric AIDS population
(Shah et al., 1996). The prudent MR imager
will remain vigilant for this occurrence and
employ either 3-D gradient echo or directed
MRA studies to further evaluate the patient.
The findings can be subtle when early in their
development, but can rapidly advance to a
markedly abnormal vessel contour, rupture and
associated hemorrhage, or thrombosis and in-
farction. Alternatively, when infarction or hem-
orrhage are noted, clues to the presence, loca-
tion, and severity of vasculopathy should be
assessed. The authors have found it occasion-
ally useful to perform MRA/MRV after the
administration of gadolinium in order to obtain
superior signal-to-noise from the vessels. This
can be especially important when examining
the small and hyperdynamic vessels found in
the pediatric age group. Also, contrast admini-
stration may allow for the evaluation of more
distal arterial vessels of the brain (i.e., M2
segment of the middle cerebral artery and be-
yond) that would not otherwise afford sufficient
signal for scrutiny. Such is the case in coopera-
tive patients, but in uncooperative patients,
there is often a delay obtaining motion sensitive
studies such as MRA until adequate sedation
can be obtained. Though specifically in diffi-
cult patients, gadolinium enhanced MRA may
give sufficient signal in the larger arterial struc-
tures of the circle-of-Willis for sufficient diag-
nostic confidence.
The occurrence of the AIDS dementia com-
plex (ADC) in AIDS patients, as well as altered
sensorium related to more focal pathologies in
this population make fast scanning technique
familiarization important. For a more detailed
discussion of fast-scan techniques, please refer
to Chapter A5 of this series. Additional time
savings for imaging these patients can be found
by limiting pre-contrast and post-contrast T1-
weighted sequences to one plane each, per-
forming only one acquisition (Nacq) on each
sequence, or alternatively by utilizing FSE T1-
weighted sequences rather than routine T1-
weighted sequences. These maneuvers will
limit all SNR and sensitivity for detection of
abnormal contrast enhancement and subtle le-
sions but will afford faster exam times. The
marked speed of acquisition for dMRI make it
a desirable sequence in these patients, as well.
It can provide both T2-weighted and diffusion
data.
For the always difficult differentiation of
toxoplasmosis from lymphoma, several helpful
Current Protocols in Magnetic Resonance Imaging
clues should be sought. The presence of micro-
hemorrhage within lesions is a good indicator
of toxoplasmosis. This determination is aided
by the judicious use of the gradient echo se-
quence (sequence 8), which can readily display
hemorrhagic blood products that might other-
wise be extremely subtle, or even invisible on
routine spin echo sequences (Figure A4.1.3).
MRS and pMRI can also be helpful, if the
reader’s scanner is equipped for these studies.
Finally, the utility of MR for follow-up im-
aging cannot be overstressed for HIV infected
patients. The pathologies found in this group
demand repeat imaging, often over extended
periods of months and years and after multiple
associated interventions such as surgery, bi-
opsy, and of course, anti-infectious therapies.
It is important to attempt standardization of
protocols from one imaging session to the next
to avoid the “apples-to-oranges” effect that can
handicap the radiologist’s ability to assess for
lesion change over time.
Imaging of HIV by magnetic resonance
diffusion, perfusion, and spectroscopy
The Basic Protocol outlined above is quite
sufficient in most instances and represents the
standard for imaging of the patient with HIV.
However, in certain clinical instances there may
remain confusion as to the diagnosis, or in order
to gain a higher degree of specificity in labeling
a lesion, further imaging can be performed
utilizing dMRI, pMRI, or MRS. Recent litera-
ture (Chang et al., 1995; Ernst et al., 1998) has
indicated diagnostic benefit in some clinical
and imaging differential situations with these
functions, with the greatest added utility in HIV
patients presently residing with MRS. Please
refer to the scan parameters and set-up outlined
in detail in UNIT A4.1 of this chapter for perform-
ance of these sequences.
pMRI of the brain. This sequence is per-
formed to assess vascular perfusion of the brain
and pathologic foci. pMRI can help in differ-
entiating abscess from necrotic neoplasm. It is
a rapid acquisition with moderate spatial reso-
lution which can yield information on the entire
brain and multiple lesions, even in agitated
patients.
Echo-planar imaging (EPI) dMRI of the
brain. This sequence is performed to assess
water motion (proton diffusion) within the
brain parenchyma. Diffusion imaging may help
in the differential diagnosis of cerebral abscess
vs. necrotic neoplasm. It is a rapid acquisition
(scan time 28 sec) with moderate spatial reso-
lution that can yield information on the entire
brain and multiple lesions, even in agitated
patients.
MRS of the brain. This sequence will pro-
vide information as to the specific molecular
substrate(s) within an area of interest, which
may indicate the presence of abnormal tissue
or metabolic processes. MRS can be used in the
diagnosis and follow-up of cerebral abscess,
encephalitis, regions of demyelination, or tu-
mors. The sequence acquisition time is ∼10 min
per TE performed, making useful results diffi-
cult to obtain in agitated patients. A tracing of
the magnetic spectra of the molecular contents
of the voxel is produced for analysis.
Anticipated Results
Focal brain lesions with mass effect are
found in 15% to 20% of AIDS patients and, as
a class, represent the most important patholo-
gies for accurate identification as generally
effective therapies are available (Walot et al.,
1996). As AIDS is fundamentally an immune
system compromising illness, there can be dis-
tinct and confusing differences in the appear-
ance of pathologic entities in the AIDS patient
when compared to similar pathologies in the
normal population. Specifically, the inflamma-
tory response can be weak in the immunocom-
promised patient, which may lead to a decrease
in reactive edema found around inflammatory
or neoplastic lesions, and poor contrast en-
hancement. However, tumors that would nor-
mally enhance will usually continue to do so,
as neoplasia involves the formation of abnor-
mal tumor vessels that leak contrast agent and
are not subject to an inflammatory factor.
Toxoplasma gondii is the protozoan responsi-
ble for the most common opportunistic CNS
infection in AIDS patients (13% to 33% of
AIDS patients with CNS disease). Encephalitis
and abscess can result from reactivation of the
organism in the immunocompromised host.
The primary means of differentiating toxoplas-
mosis from lymphoma in the CNS remains a
satisfactory response to antitoxoplasma thera-
pies, with noticeable lesion regression seen on
imaging studies by 2 weeks. Toxoplasmosis
tends to present with multiple lesions and the
presence of small lesional hemorrhages is sup-
portive of the diagnosis (Taveras and Pile-
Spellman, 1996). However, a solitary lesion at
presentation is not uncommon (∼30%). Pri-
mary lymphoma is the most common neoplasm
in AIDS patients, and is almost always of the
high-grade B-cell type. It is also most often
Infectious Human
Diseases of the Immunodeficiency multifocal at presentation (80% to 100%) and
Brain Virus (HIV) hemorrhage in these lesions is distinctly un-
A4.3.11 A4.3.12
Supplement 2 Supplement 2
usual. Though typically solid tumors, lym-
phoma may display necrosis (an unusual find-
ing in non-AIDS patients) and ring enhance-
ment, especially when larger. The ring enhance-
ment is often less distinct and thicker (>3 mm)
than that seen with toxoplasmosis. Lymphoma
is isointense to brain on both T1-weighted and
T2-weighted MR images, with occasional hy-
pointensity seen on T2-weighted images due to
the high cellularity of the process and the high
nuclear/cytoplasmic ratio. It occurs centrally in
the basal ganglia, corpus callosum, and
periventricular white matter. Subependymal
spread and ventricular encasement are also
common.
Despite these appearances and their seeming
differences, the imaging characteristics of
toxoplasmosis and lymphoma overlap to a sig-
nificant degree, and in fact have been reported
to occur together in the same patient (Walot et
al., 1996; Wong and Quint, 1999). With therapy
all lesions must thus be closely monitored.
pMRI and MRS have been used to further
differentiate these pathologies (see discussion
in UNIT A4.1; Chang et al., 1995; Ernst et al.,
1998). Toxoplasmosis tends to demonstrate
marked elevation of the lactate and lipid peaks
on MRS and normal metabolite depletion,
while lymphoma produces an elevated choline
peak with moderate elevation of the lactate and
lipid peaks. These spectra are thought secon-
dary to the more infiltrating and solid nature of
the neoplasm compared to the more typically
necrotic and abscessed infectious lesion.
Cryptococcosis, an infection from the yeast
Cryptococcus neoformans, occurs in 5% of
AIDS patients and can be varied in presentation
and structural involvement of the CNS (Harris
and Enterline, 1997). However, meningitis is
certainly most common at presentation. Inflam-
matory response by the CNS to invasion by
cryptococcus is muted as the organism elabo-
rates a mucinous capsule. Thus, meningeal en-
hancement is rarely noted. However, the organ-
ism proliferates in the subarachnoid space and
characteristically settles in the perivascular
spaces of Virchow-Robin, where it fills these
spaces with organisms and mucoid exudate
leading to lesions termed “gelatinous pseudo-
cysts” (Figure A4.3.1). These pseudocysts have
signal characteristics similar to CSF, but may
be detected on FLAIR or proton-density images
with abnormal high T2 signal in dilated and
unusually numerous Virchow-Robin spaces,
which are seen mostly at the base of the brain
and in the brainstem (predominantly the mid-
brain).
Current Protocols in Magnetic Resonance Imaging
 AIDS patients do not have a particular sus-
ceptibility to pyogenic abscess formation or
tuberculosis, but these infections can still be
seen in the i.v. drug abusing subset of the AIDS
population (Walot et al., 1996). Neurotubercu-
losis presentations are varied with cerebritis, or
focal parenchymal or meningeal lesions seen
typically with prominent and diffuse enhance-
ment, though true abscess formation may occur.
The simultaneous occurrence of basilar menin-
gitis and associated hydrocephalus is a strong
indicator of tuberculous infection, or another
granulomatous process such as fungal infection
by coccidioidomycosis.
The second broad category of CNS abnor-
malities in AIDS patients is focal lesions with-
out significant mass effect. These processes
generally involve only white matter without
producing significant edema, and usually with-
out associated enhancement after administra-
tion of gadolinium chelates. The lesions are of
low signal on T1-weighted MR images, and
demonstrate high T2 signal intensity. This ap-
pearance is nonspecific, being produced by
lesions of multiple sclerosis, post-infectious
demyelinating conditions such as acute dis-
seminated encephalomyelitis (ADEM) and is- A B
chemic processes in the elderly population. Of
paramount concern among these is PML that
represents a reactivation of the dormant JCV
found in 80% of the normal adult population.
The virus infects and destroys oligodendro-
cytes causing the demyelination. The lesions
are usually multiple, confluent and bilateral
(though asymmetric) at presentation, and found
in the white matter, primarily in subcortical,
periventricular, and centrum semiovale loca-
tions. The process rarely enhances (2%) and
may have mild associated cortical atrophy
(Donovan-Post et al., 1999). The main differ-
ential consideration with PML is that of pri-
mary HIV or CMV cerebritis. These conditions
characteristically have advanced associated at-
rophy of the cerebral structures on presentation.
Additionally PML, as the name implies, can
progress. This progression may be startlingly
rapid with serial MR studies depicting marked
progression in many patients within 9 weeks.

Figure A4.3.2 A coronal (A) FSE T2 image through the third ventricle (TR = 5600 msec, TE = 105 msec,
Nacq = 2; echo train length = 8) discloses abnormal increased T2 signal in the subependymal and
immediate periventricular white matter (arrowheads). The corresponding post-contrast T1-weighted (B)
coronal image (TR = 500 msec, TE = 12 msec, Nacq = 1) shows linear ependymal enhancement
(arrowheads) compatible with presumed CMV ventriculitis and ependymitis. This diagnosis was supported
in this 33-year-old female with AIDS by a concurrent CMV retinitis. The authors utilized the additional
coronal FSE sequence (A) in this patient primarily as a means of evaluating the optic pathways in this
patient with CMV retinitis.

MRS has been described for AIDS patients and
PML (Chang et al., 1995) and suggests a sig-
nificant contribution to specificity in distin-
guishing amongst the many differential diag-
nostic possibilities when coupled with the re-
sults of spin-echo examination.
In considering the third broad category of
CNS lesions found in AIDS patients, diffuse
global CNS abnormalities, the most common
presentation is that of iso-intense to hypo-in-
tense T1 signal relative to brain, with subtle and
diffuse abnormal increased T2 signal located
predominantly in the cerebral white matter in a
symmetric manner. This is seen with associated
diffuse and often marked cerebral atrophy
(which in itself may be the only finding). The
Figure A4.3.1 This T2-weighted (TR = 2500 msec, TE = 80 most common cause of this constellation of
msec, Nacq = 0.75) transverse image at the basal ganglia level findings is direct CNS infection by HIV, a
demonstrates a cluster of rounded lesions (arrowheads) in the neurotropic virus. No enhancement is present
left globus pallidus and posterior limb of the internal capsule with in the regions of abnormal signal and MRS
mild increased T2 signal in adjacent brain. Minimal mass effect shows neuronal loss with a decrease of the
is present. The CSF disclosed cryptococcal infection, and “ge- Infectious Human
Diseases of the Immunodeficiency N-acetylaspartate (NAA) peak (a neuronal
latinous pseudocysts” within the perivascular spaces of Virchow-
Brain Virus (HIV) m ar ker). Clinically, an associated
Robin were diagnosed.
A4.3.13 A4.3.14
Current Protocols in Magnetic Resonance Imaging Supplement 2 Supplement 2
encephalopathy may be noted (ADC). Autopsy
studies have shown a high prevalence of a
concurrent CMV infection, though this patho-
gen is believed to invade the brain later in the
course of the illness.
Finally, ventriculitis, meningitis, and in-
farcts comprise the fourth category of CNS
abnormality. CMV may cause an aggressive
ventriculoencephalitis, the diagnostic clue to
which can often be found with CMV infection
elsewhere in the body (most notably CMV
retinitis). The ventricles are enlarged, generally
from associated atrophy, with increased T2 sig-
nal seen in subependymal location along with
poorly defined periventricular enhancement
(Figure A4.3.2). This viral agent also has a
tendency to spread to the meninges and adja-
cent cranial nerves. CNS vasculopathy, which
can lead to ischemic or hemorrhagic infarction,
will rarely affect the larger arterial structures
(circle-of-Willis) at the base of the brain and
can lead to fusiform aneurysm formation, as
well as sclerosis and thrombosis (Figure
Current Protocols in Magnetic Resonance Imaging
 Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
A B Philadelphia.
Taveras, J.M. and Pile-Spellman, J. 1996. Inflamma-
tory diseases. In Neuroradiology, 3rd edition pp.
259-326. Williams and Wilkins, Baltimore.
Walot, I., Miller, B.L., Chang, L., and Mehringer,
C.M. 1996. Neuroimaging findings in patients
with AIDS. Clin. Infect. Dis. 22:906-919.
Wong, J. and Quint, D.J. 1999. Imaging of central
nervous system infections. Semin. Roentgenol.
34:123-143.
Key Reference
Walot et al., 1996. See above.
Presents a directed and concise summary of the
myriad imaging appearances encountered in this
complex patient population.
Contributed by Andrew E. Auber
and Clifford Belden
Brooke Army Medical Center
San Antonio, Texas

Figure A4.3.3 This 9-year-old boy with congenitally acquired HIV infection began complaining of
headaches. The proton-density T2-weighted (A) transverse image (TR = 2500 msec, TE = 30 msec, Nacq
= 0.75) shows prominent fusiform aneurysmal dilatation of the anterior cerebral artery (arrow). A
subsequent cerebral angiogram (B) reveals diffuse aneurysmal involvement of the supraclinoid internal
carotid artery (ICA), as well as the A1 (asterisk) and M1 (arrowhead) portions of the ICA branch vessels.
The child was treated conservatively, but several months later experienced severe subarachnoid hemor-
rhage and died.

A4.3.3). This may be the result of spread to the
vessel wall (presumably the adventitia primar-
ily) via meningeal infection and inflammation,
or possibly a direct vascular infection via the
hematogenous route. This vasculopathy has a
particular incidence in the pediatric AIDS
population and is a late manifestation of the
disease (Shah et al., 1996). It is rare to see
vasculopathy without associated abnormalities
present on spin echo images of the brain (Pom-
per et al., 1999).
Literature Cited
Chang, L., Miller, B.L., McBride, D., Cornford, M.,
Oropilla, G., Buchtal, S., Chiang, F., Aronow, H.,
and Ernst, T. 1995. Brain lesions in patients with
AIDS: H-1 MR spectroscopy. Radiology
197:525-531.
Donovan-Post, M.J., Yiannoutsos, C., Simpson, D.,
Booss, J., Clifford, D.B., Cohen, B., McArthur,
J.C., Hall, C.D., and the AIDS Clinical Trials
Group, 243 Team. 1999. Progressive multifocal
leukoencephalopathy in AIDS: Are there any
MR findings useful to patient management and
predictive of patient survival. A.J.N.R. 20:1896-
1906.
Ernst, T.M., Chang, L., Witt, M.D., Aronow, H.A.,
Cornford, M.E., Walot, I., and Goldberg, M.A.
1998. Cerebral toxoplasmosis and lymphoma in
AIDS: Perfusion MR experience in 13 patients.
Radiology 208:663-669.
Fauci, A.S. 1999. The AIDS epidemic: Considera-
tions for the 21st century. N. Engl. J. Med.
341:1046-1050.
Harris, D.E. and Enterline, D.S. 1997. Fungal infec-
tions of the central nervous system. Neuroimag-
ing Clin. N. Am. 7:187-198.
Pomper, M.G., Miller, T.J., Stone, J.H., Tidmore,
W.C., and Hellmann, D.B. 1999. CNS vasculitis
in autoimmune disease: MR imaging findings
and correlation with angiography. A.J.N.R.
20:75-85.
Shah. S.S., Zimmerman, R.A., Rorke, L.B., and
Vezina, L.G. 1996. Cerebrovascular complica- Infectious Human
tions of HIV in children. A.J.N.R. 17:1913-1917. Diseases of the Immunodeficiency
Brain Virus (HIV)

A4.3.15 A4.3.16
Current Protocols in Magnetic Resonance Imaging Supplement 2 Supplement 2 Current Protocols in Magnetic Resonance Imaging
Traumatic Brain Injury
Magnetic resonance (MR) imaging of the brain following head injury is used in two
distinct clinical contexts, (1) acutely, within days of the injury, to evaluate an unexplained
neurologic deficit or to obtain prognostic information, and (2) chronically, to assess the
degree of brain injury and explain neurologic or neuropsychologic findings. In this unit,
two basic protocols are presented, one for acute imaging (see Basic Protocol 1) and the
other for chronic imaging (see Basic Protocol 2). Advanced MR imaging sequences, such
as MR spectroscopy (MRS) and diffusion-weighted (DW) imaging can add additional
prognostic information in the acute setting and are described. MR angiography (MRA)
and direct vessel wall imaging can be important when the question of traumatic vessel
dissection is raised. These techniques will be mentioned but are covered fully in other
units. The parameters given here are derived from experience at 1.5T and may need to be
altered slightly depending on the field strength available and the specific equipment
manufacturer.
ACUTE AND SUBACUTE INJURY
MR imaging following head injury can be a challenge. Patients are often unable to
cooperate with scan instructions or hold still. Many will need monitoring of vital signs
during the imaging study and may be intubated. External hardware, such as intracranial
pressure monitors, can cause significant artifacts. The goal of this protocol is to expedi-
tiously evaluate the brain for the presence of traumatic injury. Basic Protocol 1 consists
of 5 sequences: a fast 3-plane localizer, sagittal T1-weighted images, transverse fluid-at-
tenuated inversion recovery (FLAIR), fast spin echo (FSE) T2-weighted images, and
coronal gradient echo images. These sequences can be performed on the majority of MR
scanners in use today, including most 1.0T and 0.5T systems. For systems without FLAIR
capabilities, a conventional spin-echo dual-echo sequence can be substituted for the
T2-weighted FSE and FLAIR sequences. Additional scans, such as diffusion-weighted
(DW) and MR spectroscopy (MRS) sequences, can be used if available (see Alternate
Protocol). The sequences described herein are based on the authors’ experience with a
Marconi Medical Systems 1.5T scanner, but are expected to be equally applicable to
machines from other manufacturers.
Table A4.4.1 lists the hardware necessary to perform the examination. Gradient coil
strength is only a factor for the performance of diffusion imaging and for the determination
of the allowable echo train length for FSE sequences.
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as a crash cart or oxygen. Also ensure only magnetic field compatible oxygen
tanks are utilized if wall-source oxygen is unavailable.
Table A4.4.1 Equipment Parameters for Head Imaging Sequences
Coil type Head
Gradient coil strength 25 mT/m (or whatever the system permits)
Gating (cardiac, respiratory, peripheral) No
Respirator or oxygen If required by patient
Motion cushions Recommended
Contrast agents No
Contributed by Andrew E. Auber and Clifford Belden
Current Protocols in Magnetic Resonance Imaging (2002) A4.4.1-A4.4.13
Copyright © 2002 by John Wiley & Sons, Inc.
UNIT A4.4 Set up patient and equipment
1. Interview (screen) the patient to ensure that he/she has no contraindication such as
cardiac pacemaker or other implants containing ferromagnetic materials. Also be sure
to ascertain whether the patient has any health conditions requiring the presence of
either special emergency equipment during the scanning procedure, or necessitating
special precautions.
Generally standard screening forms are used for all patients scanned in a magnetic
resonance system.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(1996) for discussion of what implants may be safely scanned using magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
2. If the procedure is a research protocol, have the patient sign any necessary consent
BASIC form.
PROTOCOL 1
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
4. Have the patient remove any mascara or other metal-containing makeup to avoid local
tissue heating and image artifacts.
5. Inform the patient about what will occur during the procedure, what he/she will
experience while in the magnet, and how to behave, including the following:
a. If earphones/headphones are used to protect the ears from loud sounds produced
by the scanner, the patient will be asked to wear these, but will be able to
communicate with the technologist at any/all times during the procedure.
b. The patient will be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this equipment works).
c. For optimum results the patient should not talk, and should avoid/minimize
swallowing or other movement during each scan–i.e., as long as the banging
sounds continue. Between scans, talking and swallowing are allowed in most
cases, but should be avoided when comparative positional studies are being
performed. The patient will be informed in these instances.
d. Nevertheless, the patient may call out at any time if he/she feels it necessary.
6. Have the patient lay supine on the scanner table. Either before or directly after the
patient is positioned on the table, set up any triggering devices or other monitoring
equipment necessary.
7. Center the patient in the head coil. Make sure that the head and neck are constrained
to prevent unnecessary motion, especially if high-resolution scans are to be run.
8. If needed, place a pillow or other support under the knees to make the patient more
comfortable.
9. Use the centering light to position the patient and place him or her into the magnet
centering on the nasion.
Once this step has been performed, as long as the patient does not move on the table, the
Infectious table itself can be moved and then replaced in the same position as before without
Diseases of the Traumatic Brain
Brain Injury jeopardizing the positioning of one scan relative to another.
A4.4.1 A4.4.2
Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging
 Table A4.4.2 Primary Clinical Imaging Parameters for Sequence 1: Rapid Table A4.4.3 Primary Clinical Imaging Parameters for Sequence 2: T1-Sagittal
Three-Plane Pilot Head

Patient position Supine Patient position Supine

Scan type Gradient echo Scan type Spin echo
Imaging plane (orientation) Sagittal, transverse, and coronal Imaging plane (orientation) Sagittal
Central slice or volume center Midline head Central slice or volume center Midline head
Echo time (TE) 3.7 msec Echo time (TE) 12 msec
Repeat time (TR) 16 msec Repeat time (TR) 300 msec
Flip angle (FA) 20° Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 300 mm, 300 mm Fields of view (FOVx, FOVy) 240 mm, 240 mm
Resolution (Δx, Δy) 2.34 mm, 2.34 mm Resolution (Δx, Δy) 1.25 mm, 0.94 mm
Number of data points collected (Nx, Ny) 128, 128 Number of data points collected (Nx, Ny) 192, 256
Display matrix (Dx, Dy) 256, 256 Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 10 mm Slice thickness (Δz) 4 mm
Number of slices 3 Number of slices 15
Slice gap NA Slice gap 1 mm
Number of acquisitions (Nacq) 1 Number of acquisitions (Nacq) 1
Read direction Left–right Read direction Superior–inferior
Slice locations At isocenter, 3 orthogonal planes Slice location Cover brain parenchyma
Saturation pulses None Saturation pulses None
RAMa 2X Scan time 1 min, 55 sec
Scan time 6 sec
aZero padded from 128 by 128 points to 256 by 256 points.
Table A4.4.4 Primary Clinical Imaging Parameters for Sequence 3: FLAIR
Transverse Head
10. If the patient is unable to hold still, either provide an appropriate sedative, or arrange
with anesthesiology for conscious or general anesthesia. Patient position Supine
Scan type FLAIR-FSE
Alternatively, a low-field open magnet may be sought for scanning the claustrophobic Imaging plane (orientation) Transverse (parallel to AC-PC line)
patient. Central slice or volume center Mid-cranium
11. This protocol can be performed in <20 min. Echo time (TE) 125 msec (effective)
Echo train length (ETL) 12
Sequence 1: Rapid three-plane pilot Repeat time (TR) 6000 msec
12. Run the three-plane pilot (Table A4.4.2) sequence to evaluate the patient positioning Inversion time (TI) 1900 msec
in the magnet. Flip angle (FA) 180°
Fields of view (FOVx, FOVy) 240 mm, 197 mm
This sequence runs in <10 sec and is used to position the remainder of the sequences. It is
Resolution (Δx, Δy) 0.94 mm, 0.97 mm
particularly useful to correct off-axis positioning in the coronal plane.
Number of data points collected (Nx, Ny) 256, 204
Sequence 2: T1-sagittal head Display matrix (Dx, Dy) 256, 256
13. Bring the sequence for a sagittal T1-weighted spin echo scan up on the console. Using Slice thickness (Δz) 5 mm
the 3-plane pilot, plan the location of the slices, correcting for off-axis positioning of Number of slices 24
the head in the coronal and transverse planes. Use parameters as specified in Table Slice gap 1 mm
A4.4.3. Number of acquisitions (Nacq) 1
Read direction Anterior–posterior
14. Let the patient know you are about to begin the study and start the scan. Slice location Planum sphenoidale-foramen
magnum line to vertex
Sequence 3: FLAIR transverse head Saturation pulses Caudal to saturate arterial flow
15. Bring the sequence for a transverse FLAIR scan up onto the console. Set the imaging Scan time 2 min, 48 sec
parameters as shown in Table A4.4.4.
16. Use the sagittal T1-weighted images and 3-plane pilot to set up the scan levels and a
caudal saturation pulse. Infectious
Diseases of the Traumatic Brain
17. Let the patient know you are ready, and begin the scan. Brain Injury

A4.4.3 A4.4.4
Current Protocols in Magnetic Resonance Imaging Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging
 Table A4.4.5 Primary Clinical Imaging Parameters for Sequence 4: FSE Sequence 4: T2-weighted FSE transverse head
Transverse Head 18. Bring the sequence for a transverse FSE scan up onto the console. Set the imaging
parameters as shown in Table A4.4.5.
Patient position Supine
Scan type FSE 19. Copy the slice positions from the transverse FLAIR sequence. Place the inferior
Imaging plane (orientation) Transverse saturation pulse to saturate arterial inflow.
Central slice or volume center Mid-cranium
20. Let the patient know you are ready, and begin the scan.
Echo time (TE) 105 msec (effective)
Echo train length (ETL) 16
Sequence 5: Coronal gradient echo (head)
Repeat time (TR) 5616 msec
21. Bring the sequence for a gradient echo scan up onto the console. Set the imaging
Flip angle (FA) 90°
parameters as shown in Table A4.4.6.
Fields of view (FOVx, FOVy) 240 mm, 180 mm
Resolution (Δx, Δy) 0.63 mm, 0.70 mm This sequence is optimized to detect the presence of blood products that usually accompany
Number of data points collected (Nx, Ny) 384, 256 significant head injury.
Display matrix (Dx, Dy) 384, 384 22. Use the T1-weighted sagittal midline image to set up the scan levels.
Slice thickness (Δz) 5 mm
Number of slices 24 23. Let the patient know you are ready, and begin the scan.
Slice gap 1 mm
Number of acquisitions (Nacq) 3 ALTERNATE OPTIONAL SEQUENCES
Read direction Anterior–posterior PROTOCOL
Slice location Planum sphenoidale-foramen
Several sequences can add additional information to the MR evaluation of a head-injured
magnum line to vertex patient. Both DW imaging and MRS have shown promise in providing information about
Saturation pulses Caudal to saturate arterial flow the degree of cerebral injury and/or prognosis. These techniques usually require additional
Scan time 3 min, 22 sec software and high-performance gradients in the case of DW imaging. Both of these
sequences are very sensitive to magnetic field inhomogeneity, so careful shimming of the
magnet is necessary prior to performing the sequences. In the case of a critically ill patient,
Table A4.4.6 Primary Clinical Imaging Parameters for Sequence 5: Coronal the additional time required to perform these sequences may outweigh the benefits.
Gradient Echo Head
MRA plays an important role in the diagnosis and follow-up of carotid and vertebral artery
Patient position Supine dissections. In practice, a 3-D time-of-flight MRA with selected transverse T1-weighted
Scan type 2-D Gradient echo images are used to diagnose dissections. Contrast-enhanced MRA may be an even more
Imaging plane (orientation) Coronal sensitive method for detecting and following dissections. These MRA techniques are
Central slice or volume center Mid-cranium detailed in other units (see Chapter A1) and will not be discussed here.
Echo time (TE) 24.6 msec
Repeat time (TR) 719 msec Additional materials
Flip angle (FA) 25° The additional materials needed for the performance of MRS and DW MR are detailed
Fields of view (FOVx, FOVy) 220 mm, 220 mm in Table A4.4.7. Many MR manufacturers have integrated semi-automated or completely
Resolution (Δx, Δy) 0.86 mm, 1.15 mm automated packages for the processing of both diffusion and spectroscopic data. If an
Number of data points collected (Nx, Ny) 256, 192 analysis package is not available, the data must be analyzed on a separate workstation
Display matrix (Dx, Dy) 256, 256 using outside software.
Slice thickness (Δz) 5 mm
Number of slices 21 Sequence 6: DW MR of the brain (optional)
Slice gap 1 mm 1. Bring the sequence for an echo-planar diffusion scan up onto the console. Set the
Number of acquisitions (Nacq) 2 imaging parameters as shown in Table A4.4.8.
Read direction Superior–inferior
Slice location Frontal to occipital cerebral poles 2. Use the scout sequence, midline image, to set up the scan levels.
Saturation pulses No 3. Let the patient know you are ready, and begin the scan. Simulate the echo-planar
Scan time 4 min, 30 sec sequence and adjust the phase offset and delay as needed to optimize the image.
Optimizing echo-planar scan parameters (phase offset and delay) is important to obtain
high quality diffusion images. For this reason, the scan is often simulated prior to being
run to allow optimization of these parameters. This may take 1 to 3 min to optimize, even
though the actual scan is only 28 sec in length (this is unique to Marconi’s system).
Infectious
Diseases of the Traumatic Brain 4. Pass the changed values to the scanner and run the sequence.
Brain Injury

A4.4.5 A4.4.6
Current Protocols in Magnetic Resonance Imaging Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging
 Table A4.4.7 Additional Equipment Needs for Advanced Head Table A4.4.9 Primary Clinical Imaging Parameters for Sequence 7: MRS of
Imaging Sequences the Brain

Gradient coil strength High-performance needed for diffusion Patient position Supine
imaging Scan type PRESS single voxel (STEAM
Software For processing diffusion data and MR may also be used)
spectroscopy Voxel center Area of interest
Workstation If analysis package not included in scan Echo time (TE) 35 msec (PRESS) or 20 msec
software (STEAM)
Repeat time (TR) 1500 msec
Flip angle (FA) 90°
The sequence obtains echo-planar images with b-values of 0 and 1000 sec/mm2. The Voxel size (Δx, Δy, Δz) 2 cm, 2 cm, 2 cm
diffusion gradients are applied in 3 directions and from these images, both a diffusion- Number of acquisitions (Nacq) 192 (signala), 8 (referenceb)
weighted “trace” image is obtained as well as a map of the Apparent Diffusion Coefficients
Saturation pulses Water
(ADC) for each pixel. Most manufacturers have built-in automated software for calculating
these images. Scan time 5 min, 50 sec
aThese 192 acquisitions are done with water suppression.
bThese 8 acquisitions are done without water suppression and are used for phase correction of the free
Sequence 7: MRS of the brain (optional)
induction decay scans.
5. Bring the MRS scan up onto the console. Set the imaging parameters as shown in
Table A4.4.9.
placed in a region that avoids contamination from fat and cerebrospinal fluid and
There are 2 basic types of spectroscopic sequences in clinical practice—single voxel and
2-D multi-voxel. Of these, single voxel techniques are the most widely used in non-research
gross magnetic field inhomogeneities, such as above the sinuses.
environments. PRESS and STEAM (stimulated echo acquisition mode) are the two widely 7. Let the patient know you are ready, and begin the scan. Before the sequence can be
available single voxel localization techniques. PRESS (point-resolved spectroscopic) spec- run two important steps need to be completed: shimming over the voxel and water
tra have a better signal-to-noise ratio but localization is often not as precise as with the
STEAM technique. STEAM also allows somewhat shorter TE values. The authors use a
suppression.
PRESS sequence with a TE of 35 msec as a compromise. This allows detection of lactate, Many scanners have somewhat automated packages to perform these functions. These
N-acetyl aspartate (NAA), choline and creatine (Cr) levels. preliminary steps may take 5 to 10 min to perform.
6. Use previously obtained transverse, sagittal, and coronal images and place the voxel 8. Pass the scan values to the scanner and run the spectroscopy sequence.
over the region of interest. The voxel is usually a 2 × 2 × 2–cm cube that should be Often you will want to do more than one voxel—one over a lesion and another in a
“normal” area of brain. This will necessitate repeating steps 5 to 8 again. Overall time for
Table A4.4.8 Primary Clinical Imaging Parameters for Sequence 6: DW MR of the scan (including the water suppression and shimming) is 10 to 15 min per voxel.
the Brain (Optional)
9. The spectroscopy data will need to be analyzed to obtain useful information. Again,
Patient position Supine many manufacturers have automated this process to get reliable spectroscopic data.
Scan type Diffusion EPI
Imaging plane (orientation) Transverse BASIC CHRONIC INJURY
Central slice or volume center Mid-cranium PROTOCOL 2
Months to years after a traumatic injury to the brain the primary focus of imaging is not
Echo time (TE) 100 msec
the acute diagnosis of cerebral injury but rather the assessment of the degree and location
Repeat time (TR) 5629 msec
of brain injury. The protocol used for chronic injury is similar to Basic Protocol 1 for
Flip angle (FA) 90°
acute injury with the exception of the addition of a T1-weighted 3-D gradient echo
Fields of view (FOVx, FOVy) 240 mm, 240 mm
sequence. This sequence provides an excellent map of brain architecture and enables
Resolution (Δx, Δy) 2.96 mm, 3 mm
detailed localization of brain injury. In the chronic phase, diffusion imaging and MR
Number of data points collected (Nx, Ny) 81, 80
spectroscopy play less of a role and at this time are not routinely used.
Display matrix (Dx, Dy) 81, 80
Slice thickness (Δz) 5 mm This protocol consists of 6 sequences: a fast 3-plane localizer, sagittal T1-weighted images,
Number of slices 20 transverse FLAIR and FSE T2-weighted images, coronal gradient echo images, and
Slice gap 1 mm transverse T1-weighted 3-D gradient echo sequence. The protocol takes 20 to 22 min to
Number of acquisitions (Nacq) 1a run.
Read direction Left–right Set up patient and equipment
Slice locations Whole brain or area of interest
1. The equipment needed is identical to that identified for Basic Protocol 1 in Table
Saturation pulses Fat
A4.4.1. Set-up and run the first 5 sequences of Basic Protocol 1, steps 1 to 23.
b value 0 and 1000 sec/mm2 Infectious
Scan time 28 sec Diseases of the Traumatic Brain
aSee also Table A4.1.12. Brain Injury

A4.4.7 A4.4.8
Current Protocols in Magnetic Resonance Imaging Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging
 Table A4.4.10 Primary Clinical Imaging Parameters for Sequence 8: 3-D
Gradient Echo A B
Patient position Supine
Scan type 3-D gradient echo
Imaging plane (orientation) Transverse
Central slice or volume center Mid-cranium for whole brain, or
area of abnormality
Echo time (TE) 5 msec
Repeat time (TR) 11.2 msec (or minimum)
Flip angle (FA) 35°
Fields of view (FOVx, FOVy) 240 mm, 187 mm
Resolution (Δx, Δy) 0.94 mm, 0.97 mm
Number of data points collected (Nx, Ny) 256, 192
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 1–1.5 mm
Number of slices 120
Slab thickness 120–180 mm, depending on
patient size
Slice gap 0 mm
Number of acquisitions (Nacq) 1
Read direction Anterior–posterior
Slice location Whole brain
C D
No phase wrap (NPW)a Yes
Saturation pulses NA
Scan time 4 min, 19 sec
aPhase oversampling or anti-aliasing.

Sequence 8: 3-Dimensional transverse gradient echo head

2. Bring the sequence for a 3-D volumetric gradient echo scan on to the console. Set
the imaging parameters as shown in Table A4.4.10.
This sequence allows thin slice (i.e., 1 to 1.5 mm) images through the entire brain. The
dataset can be re-sliced in any plane desired. The images are T1-weighted and show areas
of encephalomalacia, and in addition, the fact that the images are a gradient echo sequence
provides some blooming artifact around areas of axonal injury, allowing precise localiza-
tion of damage.
To obtain T1-weighting, a relatively large flip angle is needed. This becomes a trade-off
between time and T1-weighting as the TR increases. With the parameters given, decent
T1-weighting is obtained with an imaging time of 7 to 8 min.
3. Use the scout sequence, midline image, to set up the scan levels.
4. Let the patient know you are ready, and begin the scan.

COMMENTARY
Background Information
Traumatic brain injury (TBI) is a major
health problem throughout the world. In the
United States there are over 50,000 deaths per
year related to TBI (Thurman and Guerrero,
1999). The number of patients left with signifi-
cant mobidity following head injury is many
times greater than those who die of their injuries.
Imaging plays an important role in these pa-
tients. Computed tomography (CT) is the main-
stay of imaging in cerebral trauma. CT is able
to quickly assess which patients need emergent
treatment for intracranial injuries (usually ex-
tra-axial hematomas). It is well documented
that CT misses many of the traumatic lesions
in the brain, particularly small subdural collec-
tions and cortical contusions. MR imaging in
patients with normal head CTs show abnor-
Figure A4.4.1 21 year-old male one year following closed head injury. (A-D) and (E-H) demonstrate the
appearance of shear injury using hemosiderin sensitive gradient echo, FSE T2-weighted, FLAIR, and 3-D
gradient echo sequences, respectively, at two imaging levels. (A-D). The large region of shear injury near
the vertex of the left parietal lobe is best seen on the 2-D gradient echo sequence (A, arrows), with several
smaller foci of signal dephasing (arrowheads), likely representing smaller areas of shear injury. The parietal
region is essentially normal on T2-weighted FSE (B) and FLAIR (C) images and barely discernible on a short
echo 3-D gradient echo sequence (D, arrow). (E-H). At the level of the third ventricle, shear injury to both
basal ganglia (E, arrows), an old parenchymal hematoma in the right frontal lobe (E, open arrow) and the
old hemorrhagic contusion in the left frontal lobe (E, arrowhead) are best seen on the 2-D gradient echo
Infectious sequences. The old right frontal lobe hematoma is easily seen on T2-weighted FSE, FLAIR and 3-D gradient
Diseases of the (continued on next page)
Brain

A4.4.9 A4.4.10
Current Protocols in Magnetic Resonance Imaging Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging
 malities 30% to 60% of the time (Levin et al.,
1992; Mittl et al., 1994). In one study, small
E F subdural collections were detected by MR
alone in 58% of cases and CT missed the
majority of cases of shear injury and non-hem-
orrhagic contusions (Kelly et al., 1988).
Currently, there is no specific treatment for
the majority of the additional lesions that MR is
able to detect. This, along with the difficulties
in scanning acutely head-injured patients, has
been the major reason MR has not been adopted
for routine use in acute trauma settings. It is
becoming increasingly clear, however, that sig-
nificant brain damage can occur in patients with
even mild head injuries. Presently, MR is the
only tool we have to document the location and
extent of these injuries. In the future, neuropro-
tective agents may play an important role in
limiting the long-term brain damage in these
patients.
In the chronic state, head-injured patients are
often scanned to depict the location and extent
of brain injury. The impetus for the scan is
usually a persistent impairment in some aspect
G H of cognitive function or abnormal performance
on neuro-psychologic testing. The treatment for
these patients is usually recognition of their
limitation and introduction of skills to help them
cope with their disabilities.
Critical Parameters and
Troubleshooting
The major technical challenge in imaging
patients with head injury is patient cooperation.
Methods to decrease overall scanning time and
ensure relative immobilization are essential to
high-quality imaging. The two most critical se-
quences in Basic Protocols 1 and 2 are the
transverse FLAIR sequence and the coronal
gradient echo sequence. These have the highest
sensitivity for non-hemorrhagic contusions and
diffuse axonal injury (shear damage), respec-
tively (Figure A4.4.1; Ashikaga et al., 1997;
Parizel et al., 1998).
FLAIR sequence
The FLAIR sequence can be one of the more
Figure A4.4.1 (continued) echo sequences as signal dephasing (open arrow F, G, H). The smaller area difficult sequences to optimize. The majority of
of contusion in the left frontal lobe is present on all sequences, but appears bright on T2-weighted FSE (F, FLAIR is done as an FSE acquisition, giving 4
arrowhead), and dark on both FLAIR and 3-D gradient echo sequences (G, H arrowhead). The presence of parameters to adjust: echo train length (ETL),
shear injury in the basal ganglia is only confidently made on the gradient echo sequence, though there is a TR, TE, and TI. The goal of the optimization is
small focus of increased signal on T2-weighted FSE images (F, arrow) and decreased signal on FLAIR (G, to limit the signal in the CSF while maintaining
arrow) and 3-D gradient echo (H, arrow) sequences. (NOTE: The described protocol uses coronal gradient reasonable gray-white matter differentiation
echo images. Transverse images are presented to optimally demonstrate the relative lesion conspicuity and minimizing time. The TI should be selected
between basic sequences.)
to null CSF (cerebrospinal fluid) signal. TR and
Traumatic Brain TE can be adjusted to enhance the gray-white
Injury matter differentiation. The echo train length can
A4.4.11 A4.4.12
Current Protocols in Magnetic Resonance Imaging Supplement 5 Supplement 5
be increased to decrease the time of the scan,
but often at the expense of increased CSF flow
artifact.
Hemosiderin sensitive gradient echo
sequence
The gradient echo sequence has parameters
selected to optimize the blooming effect seen
around blood products. The sequence described
is a T2*-weighted sequence with a relatively long
TE to enhance the blooming artifact.
Anticipated Results
MRI is used in the acute and subacute trauma
setting to identify a cause for neurologic deficits
that are not explained by findings present on CT.
Basic Protocol 1 is designed to give a fairly
quick assessment of the degree of cerebral dam-
age following head injury. The sagittal T1-
weighted images are useful for the detection of
subacute blood products (methemaglobin). The
transverse FSE T2-weighted sequence is the
primary sequence for assessing overall brain
structure and anatomy. The FLAIR sequence
has been shown to be the most sensitive to
cortical contusions as well as deep contusions
of the fornix and corpus callosum. Finally, the
coronal gradient echo scan is optimized to detect
blood products. Blooming around areas of shear
injury are fairly characteristic, and this is often
the only sequence on which they are seen.
The optional sequences for Alternate Proto-
col are designed to give additional prognostic
information. DW images can depict areas of
cellular injury. DW MR has been used in some
patients and animal models following TBI
(Hanstock et al., 1994; Smith et al., 1995; Alsop
et al., 1996; Assaf et al., 1997; Barzo et al., 1997;
Stroop et al., 1998; Werring et al., 1998). It is
known in animal models that following head
injury, the ADC is significantly decreased in
areas of trauma, and that by 7 days following
injury the ADC is larger than that in normal brain
parenchyma (Assaf et al., 1997). At the 1 week
time there was good correlation between regions
of diffusion abnormality and pathologic evi-
dence of brain injury (Assaf et al., 1997).
Investigators have had some success at-
tempting to correlate MRS data with outcome
after TBI (Ross et al., 1998; Friedman et al.,
1999). Friedman showed early gray matter
NAA predicted overall neuropsychologic per-
formance in 14 patients imaged 1 to 2 months
post injury (Friedman et al., 1999). In 25 pa-
tients with TBI, Ross et al. (1998) showed
decreased NAA correlated with outcome fol-
lowing TBI, and that in children, decreased
Current Protocols in Magnetic Resonance Imaging
NAA/Cr and lactate portended a poor outcome
(Ross et al., 1998).
Basic Protocol 2 for use in chronic head injury
is similar to Basic Protocol 1 with the exception
of the addition of a volumetric T1-weighted gra-
dient echo sequence. This sequence demonstrates
the areas of brain injury (encephalomalacia) bet-
ter than most and allows for reslicing in multiple
planes as well as for volumetric measurements to
be made. In at least one study, a sequence similar
to this one was the optimal sequence for defining
the location and extent of parenchymal damage
(Herskovits et al., 1999).
Literature Cited
Alsop, D.C., Murai, H., Detre, J.A., McIntosh, T.K.,
and Smith, D.H. 1996. Detection of acute
pathologic changes following experimental trau-
matic brain injury using diffusion-weighted
magnetic resonance imaging. J. Neurotrauma
13:515-521.
Ashikaga, R., Araki, Y., and Ishida, O. 1997. MRI
of head injury using FLAIR. Neuroradiology
39:239-242.
Assaf, Y., Beit-Yannai, E., Shohami, E., Berman, E.,
and Cohen, Y. 1997. Diffusion- and T2-weighted
MRI of closed-head injury in rats: a time course
study and correlation with histology. Magn.
Reson. Imaging 15:77-85.
Barzo, P., Marmarou, A., Fatouros, P., Ito, J., and
Corwin, F. 1997. MRI diffusion-weighted spec-
troscopy of reversible and irreversible ischemic
injury following closed head injury. Acta Neuro-
chir. Suppl. (Wien) 70:115-118.
Friedman, S.D., Brooks, W.M., Jung, R.E., Chiulli,
S.J., Sloan, J.H., Montoya, B.T., Hart, B.L., and
Yeo, R.A. 1999. Quantitative proton MRS pre-
dicts outcome after traumatic brain injury. Neu-
rology 52:1384-1391.
Hanstock, C.C., Faden, A.I., Bendall, M.R., and
Vink, R. 1994. Diffusion-weighted imaging dif-
ferentiates ischemic tissue from traumatized tis-
sue. Stroke 25:843-848.
Herskovits, E.H., Megalooikonomou, V., Davatz-
ikos, C., Chen, A., Bryan, R.N., and Gerring, J.P.
1999. Is the spatial distribution of brain lesions
associated with closed-head injury predictive of
subsequent development of attention-deficit/hy-
peractivity disorder? Analysis with brain-image
database. Radiology 213:389-394.
Kelly, A.B., Zimmerman, R.D., Snow, R.B., Gandy,
S.E., Heier, L.A., and Deck, M.D. 1988. Head
trauma: Comparison of MR and CT-experience
in 100 patients. A.J.N.R. 9:699-708.
Levin, H.S., Williams, D.H., Eisenberg, H.M., High,
W.M. Jr., and Guinto, F.C. Jr. 1992. Serial MRI
and neurobehavioural findings after mild to
moderate closed head injury. J. Neurol. Neuro-
surg. Psychiatry 55:255-262.
Mittl, R.L., Grossman, R.I., Hiehle, J.F., Hurst,
R.W., Kauder, D.R., Gennarelli, T.A., and Albur-
Current Protocols in Magnetic Resonance Imaging
ger, G.W. 1994. Prevalence of MR evidence of Multiple Sclerosis UNIT A5.1
diffuse axonal injury in patients with mild head
injury and normal head CT findings. A.J.N.R.
15:1583-1589. MRI is unquestionably the imaging modality of choice for evaluation of white matter
Parizel, P.M., Ozsarlak, Van Goethem, J.W., van den diseases. Although a routine brain MR exam would suffice in many applications, specific
Hauwe, L., Dillen, C., Verlooy, J., Cosyns, P., and
De Schepper, A.M. 1998. Imaging findings in protocols for multiple sclerosis have been developed to maximize its contribution. This
diffuse axonal injury after closed head trauma. unit presents one such protocol while recognizing the wide opportunity for variation
Eur. Radiol. 8:960-965. depending on local clinical demand and personal preference.
Ross, B.D., Ernst, T., Kreis, R., Haseler, L.J., Bayer,
S., Danielsen, E., Bluml, S., Shonk, T., Mandigo,
J.C., Caton, W., Clark, C., Jensen, S.W., Lehman, MULTIPLE SCLEROSIS BASIC
N.L., Arcinue, E., Pudenz, R., and Shelden, C.H. PROTOCOL
The basic goal in the evaluation of the suspected multiple sclerosis is one of plaque
1998. 1H MRS in acute traumatic brain injury. J.
Magn. Reson. Imaging 8:829-840. identification followed by characterization. All demyelinating lesions exhibit T2 prolon-
Shellock, F.G. 1996. Pocket Guide to MR Proce- gation, and so the exam begins with some form of T2 weighting. The choice of such
dures and Metallic Objects. Lippincott-Raven, weighting, along with imaging plane, will be discussed. Although many approaches exist
Philadelphia.
for plaque characterization, they generally include a search for enhancement as a measure
Smith, D.H., Meaney, D.F., Lenkinski, R.E., Alsop, of activity. Rather than being an all-or-none phenomenon, enhancement appears to be a
D.C., Grossman, R., Kimura, H., McIntosh,
T.K., and Gennarelli, T.A. 1995. New magnetic matter of degree, and so several tricks have evolved to maximize detection of enhance-
resonance imaging techniques for the evaluation ment. These tricks, and their advantages and limitations, will be discussed. The parameters
of traumatic brain injury. J. Neurotrauma given here are derived from experience at 1.5 T, and may need to be altered depending on
12:573-577.
the field strength and the equipment manufacturer. The entire procedure generally takes
Stroop, R., Thomale, U.W., Pauser, S., Bernarding, ∼45 min.
J., Vollmann, W., Wolf, K.J., Lanksch, W.R., and
Unterberg, A.W. 1998. Magnetic resonance im-
aging studies with cluster algorithm for charac- Table A5.1.1 summarizes the six sequences which comprise the basic protocol along with
terization of brain edema after controlled cortical two optional sequences. Table A5.1.2 lists the hardware necessary to perform the proce-
impact injury (CCII). Acta Neurochir. Suppl. dure. Next, stepwise instructions for performing the protocol are provided. The protocol
(Wien) 71:303-305.
is not particularly demanding on scanner hardware and is of the sort which may be well
Thurman, D. and Guerrero, J. 1999. Trends in hos-
performed on most clinical systems.
pitalization associated with traumatic brain in-
jury [see comments]. J. Am. Med. Assoc.
282:954-957. NOTE: Be sure that technologists and nurses have immediate access to any equipment
such as crash carts or oxygen that may be necessary in the event of an emergency.
Werring, D.J., Clark, C.A., Barker, G.J., Miller,
D.H., Parker, G.J., Brammer, M.J., Bullmore,
E.T., Giampietro, V.P., and Thompson, A.J. 1998. Materials
The structural and functional mechanisms of
Normal saline (0.9% NaCl), sterile
motor recovery: Complementary use of diffu-
sion tensor and functional magnetic resonance Extravascular contrast agent (e.g., Magnevist, Omniscan, or Prohance)
imaging in a traumatic injury of the internal
capsule. J. Neurol. Neurosurg. Psychiatry
65:863-869.
Table A5.1.1 Basic Multiple Sclerosis Protocola
Key References
Mittl et al., 1994. See above. Imaging
Type of weighting and sequence
plane
MRI demonstrates abnormalities in mild head in-
jury patients with normal CT. Axonal injury was 1. Pilot scan (scout) —
found in 30% of patients.
2. T1-weighted spin echo Sagittal
Kelly et al., 1988. See above. 3. Fast FLAIR Sagittal
This article demonstrates the ability of MR to depict 4. Dual echo PD/T2-weighted TSE/FSE Transverse
many more lesions in the traumatized patient com- 5. T1-weighted spin echo Transverse
pared with CT. Over half of contusions and sub-
durals were seen by MR.
6. Post-gadolinium T1-weighted spin echo Transverse
Optional sequences
Contributed by Andrew E. Auber 7. Dual echo PD/T2-weighted CSE Transverse
and Clifford Belden 8. T1-weighted MT spin echo Transverse
Brooke Army Medical Center aAbbreviations: TSE, turbo spin echo; FSE, fast spin echo; CSE, con-
San Antonio, Texas ventional spin echo; MT, magnetization transfer; FLAIR, fluid attenu-
Infectious
Diseases of the ated inversion recovery; PD, proton density. Miscellaneous
Brain Brain Pathology
A4.4.13 Contributed by Danial K. Hallam A5.1.1
Current Protocols in Magnetic Resonance Imaging (2001) A5.1.1-A5.1.15
Supplement 5 Copyright © 2001 by John Wiley & Sons, Inc.
 Table A5.1.2 Equipment Parameters for MRI In Multiple 7. Center the patient in the head coil. Make sure that the head and neck are constrained
Sclerosis to limit motion, especially if high-resolution scans are to be run.
Coil type Quadrature head coil Generally the patient’s head is fixed so that the head is horizontal (not tilted) and the neck
and head lie along the axis of the patient table; other positions may be appropriate
Gradient coil strength Whatever the system permits
depending on the needs at hand.
Cardiac gating No
Peripheral monitoring For safety only 8. If needed, place a pillow or other support under the knees to make the patient more
Respiratory gating No comfortable.
Respirator If required by patient 9. Establish an intravenous (i.v.) line from which the contrast agent can be injected, and
Oxygen If required by patient attach this line securely to the patient so that movement into or out of the magnet will
Motion cushions Useful not pull at the patient’s arm.
This step may be performed prior to entering the magnet room if necessary to save scanner
Setup equipment and patient time.
1. Interview (screen) the patient to ensure that he or she has no contraindications such It is preferable to insert the line prior to imaging and to leave the patient in the magnet,
as cardiac pacemakers or other implants containing ferromagnetic materials. Also be with no intervening motion between the scans run before contrast agent injection and those
sure to find out if the patient has any health conditions that may require the presence run after injection.
of special emergency equipment or other precautions during the scanning procedure. 10. Use the centering light to position the patient’s nasion and put him or her into the
Generally standard screening forms are used for all patients scanned in a magnetic center of the magnet.
resonance system. Once this step has been performed, so long as the patient does not move on the table, the
The presence of any ferromagnetic metals may be a health hazard to the patient when he table itself can be moved and then replaced in the same position as before without
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact jeopardizing the positioning of one scan relative to another.
composition of the items, it is best to exclude patients with any metal implants; see Shellock 11. If the patient is unable to hold still, provide an appropriate sedative.
(1996) for discussion of what implants may be safely scanned using magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member, who can Sequence 1: Rapid three-plane positioning pilot
sit in the room during the scan and comfort the patient as needed. This companion must 12. To validate the patient’s position, run the system’s pilot (or scout) scan to ensure
be screened as well to ensure the absence of loose metal objects on the body or clothing. correct location of the head in three dimensions, using the imaging sequence given
2. If the procedure is a research protocol, have the patient sign any necessary consent in Table A5.1.3 or similar parameters.
form.
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing. Table A5.1.3 Parameters for Pilot Scan (Scout; Sequence 1)

4. Have the patient wash off mascara or other makeup to avoid local tissue heating and Patient position Supine
image artifacts. Scan type Gradient echo
5. Inform the patient about what will occur during the procedure, what he or she will Imaging plane (orientation) 3 planes
experience while in the magnet, and how to behave, including the following: Central slice or volume center Run initially at magnet isocentera
Echo time (TE) Minimum
a. If earphones or headphones are used to protect the ears from the loud sounds
Repeat time (TR) Minimum
produced by the gradients, the patient will be asked to wear these, but will be able
Flip angle (FA) 20°
to communicate with you at any time during the imaging.
Fields of view (FOVx, FOVy) 300 mm, 300 mm
b. The patient will be given a safety squeeze-bulb or similar equipment to request Resolution (Δx, Δy) 1.17 mm, 1.56 mm
assistance at any time (demonstrate how this works). Number of data points collected (Nx, Ny) 256, 192
c. For good results the patient should be instructed not to talk unless absolutely Display matrix (Dx, Dy) 256, 256
necessary and to avoid or minimize swallowing and other movements during each Slice thickness (Δz) 5 mm
scan—i.e., as long as the banging sounds continue. Between scans, talking and Number of slices 3, one in each of 3 cardinal planes
swallowing are acceptable in most cases, but should be avoided when comparative Slice gap Not applicable
positional studies are being performed; the patient will be informed when this is Number of acquisitions (Nacq) 1
the case. Swap read and phase encoding No
d. Nevertheless, the patient may call out at any time if he or she feels it necessary. Slice locations Not applicable
Saturation pulses Not applicable
6. Have the patient mount onto the table. Set up any triggering devices or other Scan time ∼10 sec
monitoring equipment that is to be used either before or just after the patient lies aIf the pilot scan shows the patient’s head to be significantly off-center, then it may be helpful to
Multiple Miscellaneous
Sclerosis down. reposition the patient and repeat the pilot scan. Brain Pathology

A5.1.2 A5.1.3
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 This sequence usually consists of three orthogonal planes to allow localization. The images Sequence 6: T1-weighted spin echo post-contrast transverse
are often also used later to determine where to place the saturation pulses and to set up 18. Simply repeat sequence 5 after injecting contrast using the parameters as given in
total coverage of the volume of interest. Table A5.1.7.
Sequence 2: T1-weighted spin echo sagittal Gadolinium administration is not generally necessary for the diagnosis of multiple sclero-
13. Using images generated in sequence 1, sagittal images are planned to provide whole sis but it is very helpful for assessment of disease activity, as only acute lesions enhance.
brain coverage. Set the parameters as indicated in Table A5.1.4. Several techniques exist which are meant to provide more sensitive evaluation of the
breakdown of blood-brain barrier indicated by enhancement. Use of triple the usual dose
The scan is best positioned first on the transverse pilot scan. Some angulation may be
needed to provide for true sagittal anatomic imaging. The coverage provided should then
be checked on the sagittal scout. This sequence is used to accurately position subsequent Table A5.1.4 Parameters for T1-Weighted Spin Echo Sagittal (Sequence 2)
scans. It provides excellent visualization of midline structures.
Patient position Supine
Sequence 3: Fast FLAIR sagittal
Scan type Spin echo
14. Set parameters for fast FLAIR sagittal as indicated in Table A5.1.5.
Imaging plane (orientation) Sagittal
The scan is positioned just as in sequence 2. The sequence provides superior demonstration Central slice or volume center Midline
of T2-weighted bright lesions, especially in the supratentorial compartment. The suppres- Echo time (TE) 14 msec
sion of bright CSF (cerebrospinal fluid) afforded by the FLAIR technique demonstrates Repeat time (TR) 600 msec
lesions adjacent to CSF that would otherwise be poorly visualized. The use of the sagittal
Flip angle (FA) 90°
plane for the study provides detailed visualization of lesions of the corpus callosum (Figure
A5.1.1). Such lesions are not only common in multiple sclerosis but are judged to be fairly Fields of view (FOVx, FOVy) 230 mm, 230 mm
specific for this diagnosis (Gean-Marton et al., 1991). Resolution (Δx, Δy) 1.2 mm, 0.9 mm
Number of data points collected (Nx, Ny) 192, 256
Sequence 4: Dual echo PD/T2-weighted transverse Slice thickness (Δz) 5 mm
15. The T1-weighted sagittal image is used to position transverse scans. Set parameters Number of slices 19–21
for the PD/T2-weighted dual echo transverse sequence as given in Table A5.1.6. Slice gap (distance factor) 1.5 mm (0.30)
Number of acquisitions (Nacq) 1
The scan is positioned graphically to provide whole brain coverage. The slices should be
Read direction Cranio-caudal
angled according to an institutional standard for transverse images. The author recom-
mends using the “AC-PC” line as such a standard. Here, the anterior and posterior Saturation pulses Inferior may be used
commissures are identified on the T1-weighted sagittal midline image (Fig. A5.1.2). The Scan time ∼2 min
slices of all transverse scans are positioned parallel to a line drawn between the two
landmarks. The sequence provides excellent visualization of the whole brain. The trans-
verse plane is the most useful for evaluating demyelinating lesions involving the brainstem Table A5.1.5 Parameters for Fast FLAIR Sagittal (Sequence 3)
(Figure A5.1.3). Proton-density weighting optimally visualizes lesions on the surface of the
brainstem, whereas the combination of T2-weighted and proton-density-weighted images Patient position Supine
is best for deeper brainstem lesions. (FLAIR, although superior for evaluation of much Scan type Fast FLAIR
intracranial pathology, especially that adjacent to CSF, has proven less sensitive to Imaging plane (orientation) Sagittal
posterior fossa lesions; Bastianello et al., 1997; Tubridy et al., 1998). Central slice or volume center Midline
Echo time (TE)a 100–140 msec
Sequence 5: T1-weighted spin echo transverse Echo train length (ETL) (turbo factor) 7–11
16. Set parameters for the T1-weighted spin echo transverse sequence as indicated in Repeat time (TR)a 8,000–10,000 msec
Table A5.1.7 and position identically to sequence 4, dual PD/T2-weighted transverse Inversion time (TI)a 2300 msec
sequence. Flip angle (FA) 180°
The scan provides visualization of T1-weighted hypointense demyelinating lesions. Such Fields of view (FOVx, FOVy) 230 mm, 230 mm
lesions have been shown to correlate better with the degree of disability than T2-weighted Resolution (Δx, Δy) 0.9 mm, 0.9 mm
lesions (van Walderveen et al., 1995; Truyen et al., 1996; Giugni et al., 1997). The Number of data points collected (Nx, Ny) 256, 256
T1-weighted spin echo transverse image provides the appropriate comparison for evalu- Slice thickness (Δz) 4–5 mm
ation of contrast enhancement. Some demyelinating plaques, especially those that are Number of slices 21
chronic, will exhibit borders of intrinsic T1-weighted hyperintensity (generally thought to Slice gap 1.5 mm (or run twice and
be evidence of myelin breakdown products). Without the benefit of a pre-contrast scan interleave with zero gap)
performed identically to that post-contrast, differentiation of such areas of intrinsic Number of acquisitions (Nacq) 1
T1-weighted hyperintensity from true enhancement may not be possible.
Read direction Cranio-caudal
17. Inject the contrast agent, then flush the intravenous line with 10 ml saline. It is Saturation pulses Inferior may be used
preferable to perform the contrast injection while leaving the patient in the magnet. Scan time ∼4 min
aOptimum choice of parameters for fast FLAIR varies significantly with manufacturer and scanner
Multiple A dose of 0.1 mmol/kg of contrast agent is usually given. Triple dose (0.3 mmol/kg) may Miscellaneous
limitations. For a review of parameter optimization in fast FLAIR (see Rydberg et al., 1995).
Sclerosis be used to improve demonstration of enhancement. Brain Pathology

A5.1.4 A5.1.5
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 of gadolinium is likely the most powerful technique (Bastianello et al., 1998). Please see
discussion of augmenting enhancement under Commentary.
Sequence 7: Dual echo PD/T2-weighted conventional spin echo (optional)
19. Sequence 7 may be used in lieu of sequence 4 using parameters in Table A5.1.8.
The sequence will require more scanning time than the FSE/TSE and a larger slice gap is
required to provide whole brain imaging. It is comparable in lesion detection to sequence
4 (see discussion in the Commentary).

Figure A5.1.2 Midline T1-weighted sagittal image demonstrates placement of AC-PC line
to be used to position transverse images.

Figure A5.1.1 Fast FLAIR sagittal provides optimal A B

visualization of small callosal multiple sclerosis lesions
sometimes referred to as subependymal striations.

Table A5.1.6 Parameters for Dual Echo PD/T2-Weighted TSE/FSE Transverse

(Sequence 4)

Patient position Supine

Scan type TSE/FSE
Imaging plane (orientation) Transverse
Central slice or volume center Position for whole brain coverage
Echo time (TE) 14–17 msec for PD-weighted
image and 85–102 msec for
T2-weighted image
Echo train length (ETL) (turbo factor) 5–8
Repeat time (TR) 2500–3000 msec
Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 180 mm, 230 mm
Resolution (Δx, Δy) 0.94 mm, 0.90 mm
Number of data points collected (Nx, Ny) 192, 256 Figure A5.1.3 Combination of PD (A) and T2-weighted (B) transverse images demonstrates even
Slice thickness (Δz) 5 mm very small brainstem lesions as shown here in the ventral medulla. Such lesions are important for
Number of slices 23 several reasons. Brainstem lesions are frequently clinically apparent and imaging confirmation is
Slice gap (distance factor) 1.5 mm (0.30) generally preferred by the clinician. Also, multiple sclerosis brainstem lesions are often located along
the cisternal or ventricular surface—a feature atypical for senescent/ischemic lesions.
Number of acquisitions (Nacq) 1
Read direction Anterior to posterior
Multiple Saturation pulses Inferior may be used Miscellaneous
Sclerosis Scan time ∼3 min Brain Pathology

A5.1.6 A5.1.7
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Sequence 8: T1-weighted magnetization transfer spin echo transverse (optional) Magnetization transfer imaging has multiple potential uses in everyday clinical imaging
20. Run sequence 8 using parameters in Table A5.1.9 and position similarly to other of multiple sclerosis and still more in the investigational arena. In routine use, MT may be
transverse sequences. helpful in improving specificity in the diagnosis (Mehta et al., 1996) of periventricular
demyelinating lesions (Fig. A5.1.4). Also, it is one of the tricks used to augment enhance-
The increased time required for the magnetization transfer pulse increases TR and reduces ment. Both of these issues are discussed in further detail under Commentary.
the maximum number of slices. An increase in number of acquisitions is generally needed
for satisfactory image quality.
Table A5.1.9 Parameters for T1-Weighted Magnetization Transfer Spin Echo
Table A5.1.7 Parameters for T1-Weighted Spin Echo Transverse (Sequences 5 Transverse (Sequence 8)
and 6)
Patient position Supine
Patient position Supine Scan type Spin echo with MT pulse
Scan type Spin echo Imaging plane (orientation) Transverse
Imaging plane (orientation) Transverse Central slice or volume center Position for whole brain coverage
Central slice or volume center Position for whole brain coverage Echo time (TE) 15 msec
Echo time (TE) 14 msec Repeat time (TR) 715 msec
Repeat time (TR) 600 msec Flip angle (FA) 90°
Flip angle (FA) 90° Fields of view (FOVx, FOVy) 180 mm, 230 mm
Fields of view (FOVx, FOVy) 180 mm, 230 mm Resolution (Δx, Δy) 0.94 mm, 0.90 mm
Resolution (Δx, Δy) 0.94 mm, 0.90 mm Number of data points collected (Nx, Ny) 192, 256
Number of data points collected (Nx, Ny) 192, 256 Slice thickness (Δz) 5 mm
Slice thickness (Δz) 5 mm Number of slices 19
Number of slices 23 Slice gap (distance factor) 1.5 mm (0.30)
Slice gap (distance factor) 1.5 mm (0.30) Number of acquisitions (Nacq) 2
Number of acquistitions (Nacq) 1 Read direction Anterior to posterior
Read direction Anterior to posterior Saturation pulses Inferior may be used
Saturation pulses Inferior may be used Scan time ∼5 min
Scan time ∼3 min

Table A5.1.8 Parameters for Dual Echo PD/T2-Weighted Conventional Spin

Echo (Sequence 7) A B

Patient position Supine

Scan type Conventional spin echo
Imaging plane (orientation) Transverse
Central slice or volume center Position for whole brain coverage
Echo time (TE) 30 msec for PD-weighted image
and 80 msec for T2-weighted
image
Repeat time (TR) 2500 msec
Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 180 mm, 230 mm
Resolution (Δx, Δy) 0.94 mm, 0.90 mm
Number of data points collected (Nx, Ny) 192, 256
Slice thickness (Δz) 5 mm
Number of slices 20
Slice gap (distance factor) 2.5 mm (0.50)
Number of acquisitions (Nacq) 0.75a
Read direction Anterior to posterior
Saturation pulses Inferior may be used Figure A5.1.4 T2-weighted TSE transverse (A) and T1-weighted MT spin echo transverse (B)
images demonstrate multiple callosal and periventricular lesions in this patient with multiple
Scan time ∼5 min
sclerosis. The increased signal on the T1-weighted MT sequence is seen in some but not all
aRefers to 3/4 NEX imaging, also known as partial Fourier imaging—an option available on many
demyelinating lesions. It is not generally seen with senescent/ischemic lesions and so is somewhat
clinical scanners. Partial Fourier imaging reduces imaging time by exploiting the inherent symmetry specific for demyelinating disease.
Multiple of k-space; essentially, fewer phase encoding steps are used to provide the same spatial resolution. The Miscellaneous
Sclerosis time saved is at the expense of signal to noise. Brain Pathology

A5.1.8 A5.1.9
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 COMMENTARY
Background Information
Owing to its high incidence, chronicity, and
tendency to attack young adults, multiple scle-
rosis is one of the most important neurologic
diseases. Although the exact cause is yet to be
elucidated, autoimmune mechanisms and ge-
netic susceptibility are generally regarded to be
central. Prevalence increases with latitude, and
clear racial differences exist with Caucasians at
greatest risk. In the United States, individuals
have a one in one thousand risk of developing
multiple sclerosis during their lifetime. Annual
costs related to the disease, including loss of
productivity, rehabilitation, and medical care,
have been estimated at over 2.5 billion dollars.
MRI has fundamentally altered the evalu-
ation of this patient population. Whereas in the
past, multiple sclerosis diagnosis rested solely
upon clinical criteria, current evaluation and
treatment monitoring always include MRI.
Sensitivity to changes in water content, the
hallmark of MR, allows for detection of lesions
missed by other imaging techniques. MRI also
demonstrates lesion activity through the use of
contrast administration and allows for more
specific diagnosis of demyelinating lesions.
MRI has altered the understanding of multiple
sclerosis by revealing its dynamic nature, in
which lesions progess and resolve undetected
clinically.
The importance of multiple sclerosis and the
central role played by MRI have conspired to
generate myriad investigations directed at im-
proving the contribution of MRI to evaluation
and management. MR techniques applied to
multiple sclerosis include magnetization trans-
fer, proton spectroscopy, phosphorus spectros-
copy, lesion volume quantitation, and diffusion
anisotropy. Some of these techniques have al-
ready contributed to our understanding of mul-
tiple sclerosis and hold promise for more in the
future. Despite these advances, protocols for
basic clinical evaluation consist primarily of
some form of T2-weighted imaging, and, when
appropriate, gadolinium enhancement.
Use of fast FLAIR
T2-weighted imaging provides superior
parenchymal contrast in brain imaging. How-
ever, small T2-weighted hyperintense lesions
may be obscured by adjacent, normally bright
structures, most notably CSF. FLAIR (Fluid
Attenuated Inversion Recovery) was developed
to overcome this limitation through use of a
Multiple
magnetization preparation pulse applied prior
Sclerosis
A5.1.10
rosis is clinically suspected. Considering the disruption in the blood-brain barrier. Gadolin-
advantage of fast FLAIR in evaluating periven- ium enhancement on MRI provides a conven-
to initiation of a T2-weighted imaging se-
tricular disease, as well as its limitations in the ient marker for this disruption and correlates
quence. Here, an inversion recovery pulse is
brainstem, the sequence is applied to best ad- well with the presence of acute inflammation
applied and the spins are then allowed to relax
vantage sagittally (Figure A5.1.1). (Fig. A5.1.5; Nesbit et al., 1991; Katz et al.,
back partially to the equilibrium state as defined
The choice of slice thickness for fast FLAIR 1993). MRI thereby detects early events in the
by their T1 characteristics. The imaging se-
sagittal evokes another discussion. With some- inflammatory process leading eventually to
quence is begun after a delay, TI for inversion
what thicker slices (4 to 5 mm), whole brain myelin loss and other tissue destruction (Tro-
time, when the tissue to be suppressed is at the
coverage is conveniently achieved. Although jano et al., 1996; Giovannoni et al., 1997).
null point (effectively, with no net longitudinal
the focus of the scan is to evaluate the corpus Reflecting the importance of demonstrating
magnetization) on its course back to equilib-
callosum, complete brain coverage may un- lesion activity in multiple sclerosis, the efficacy
rium. In FLAIR the TI is chosen so that CSF is
cover small subcortical lesions not readily ap- of several techniques in improving detection of
at the null point. The end result is a predomi-
parent on the PD/T2-weighted transverse im- enhancement has been investigated. Tech-
nantly T2-weighted image with CSF sup-
ages. Use of thinner slices (3 mm or even 2 mm) niques so investigated include triple-dose gad-
pressed through exploitation of its T1 relaxation
brings out subtle subependymal striations ad- olinium, delayed imaging, and magnetization
characteristics.
vocated by some as representing early demyeli- transfer (MT; see below) (Silver et al., 1997;
The T1 of CSF is long, so a fairly long time
nation which might not otherwise be detected Bastianello et al., 1998; Filippi et al., 1998a,b;
(2000 to 2500 msec) must elapse between the
(Hashemi et al., 1995). Use of such thin slices Rovaris et al., 1998; Rovaris et al., 1999a,b).
inversion recovery pulse and the initiation of
requires increasing the number of acquisitions Of these techniques, triple-dose gadolinium is
the imaging sequence. In early implementation,
to achieve reasonable image quality. In the most effective, with several studies document-
FLAIR was used with a spin echo sequence.
author’s experience, the potential benefit does ing significant increases ranging from 50% to
Although the advantage of the technique was
not warrant the additional imaging time re- 100% in the number of enhancing lesions de-
demonstrated, the imaging time (greater than
quired. tected. Delayed imaging and MT have some
10 min) was too long to gain widespread ac-
effect and may be particularly useful in combi-
ceptance (Hajnal et al., 1992). More recently, RARE (rapid acquisition and resolution nation. Each technique, of course, carries some
FLAIR has been implemented with FSE/TSE enhancement) versus conventional spin echo penalty. For triple-dose gadolinium, the down-
sequence (fast FLAIR) yielding acceptable im-
Hennig et al. (1986) demonstrated a tech- side is simply one of increased cost. Delayed
aging times.
nique for diminishing scan time in which mul- imaging alone is only mildly effective and of
Considering the propensity of multiple scle-
tiple spin echos were measured for a given course requires additional time (20 min or
rosis lesions to occur near ventricles, one might
excitation (that is, multiple lines of k-space) more) for the delay (Silver et al., 1997). For
predict a sequence which suppresses the bright
rather than one spin echo, as is the case with MT, proper interpretation of the presence of
signal of CSF to uncover subtle periventricular
conventional spin echo. The train of spin echos enhancement necessitates performance of a
lesions. Numerous studies have supported this
was created through the use of a series of 180° pre-contrast MT scan. Some of these tech-
expectation, and fast FLAIR is now widely
inversion pulses which were applied after the niques are currently in use in clinical trials but
regarded as providing superior sensitivity to
creation of each echo. Originally labeled are not, so far, part of standard clinical practice.
supratentorial demyelinating lesions. Some-
RARE for Rapid Acquisition and Resolution They are therefore left as options.
what surprisingly, the same cannot be stated for
Enhancement, the approach is now more com-
infratentorial lesions. Several studies have Contribution of magnetization transfer (MT)
monly known by manufacturer labels—fast
demonstrated relative insensitivity of fast
spin echo (FSE) and turbo spin echo (TSE). Protons in water may be detected using
FLAIR in the posterior fossa (Bastianello et al.,
Although clearly faster, early implementations conventional MRI because they spin or precess
1997; Tubridy et al., 1998). In light of this
proved less sensitive in detection of demyeli- at nearly the same frequency. These detectable
limitation, and recognizing the clinical impor-
nating plaques. Following further evaluation, it protons are relatively “free” at the molecular
tance of identifying posterior fossa lesions, a
became apparent that the problem resided not level. A second population of protons exists,
dual-echo PD/T2-weighted transverse se-
with the concept but rather with the long TR but for our purposes may only be detected
quence is included as a key component of the
values (4000 msec or greater) initially em- indirectly. This second population consists of
recommended protocol.
ployed. With proper choice of parameters, protons “bound” to complex macromolecules.
TSE/FSE is very similar to the conventional The association with these macromolecules
Use of sagittal plane
spin echo method in lesion detection, while modulates the magnetic microenvironment of
Evaluation of the corpus callosum in the
maintaining the advantage of reduced imaging the “bound” protons, causing them to spin at
sagittal plane enhances detection of subtle le-
time (Thorpe et al., 1994). Conventional spin varying frequencies. With this variation in fre-
sions and provides some specificity to the di-
echo imaging is comparable in lesion detection quency the protons cannot be directly meas-
agnosis of demyelinating disease relative to
and so is left as an alternative to TSE/FSE in ured. However, the magnetizations of the
other periventricular white matter diseases
the recommended protocol. “bound” and “free” populations are not sepa-
(Gean-Marton et al., 1991; Jackson et al., 1993;
rate, and will exchange to some extent (Wolff
Hashemi et al., 1995; Palmer et al., 1999). A Augmenting enhancement and Balaban, 1989). Because of this exchange,
scan in the sagittal plane is now expected by
Acute multiple sclerosis lesions are primar- manipulations of the signal of one group will Miscellaneous
our referring clinicians whenever multiple scle-
ily inflammatory and carry with them a local be transferred to the other. Brain Pathology
A5.1.11
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging

A B
Figure A5.1.5 T2-weighted TSE transverse (A) and post-contrast T1-weighted spin echo trans-
verse (B) images demonstrate typical enhancing demyelinating plaque. Contrast enhancement can
be subtle in multiple sclerosis. Performance of T2-weighted transverse sequence in identical location
as post-contrast T1-weighted transverse sequence aids recognition of enhancing lesions.
Several types of MT pulses exist. In the most
conceptually straightforward approach, an off-
resonance saturation pulse is applied to the
“bound” population. This saturation will be
exchanged with the measurable “free” popula-
tion, thereby partially decreasing their signal.
The extent of the exchanged saturation depends
on factors relating both to the saturation pulse
and to the tissue to which it is applied. Tissues
with higher number of “bound” protons (i.e.,
white matter) will be more effected by the
magnetization pulse. Pathology, including de-
myelination, decreases the number of bound
protons to be saturated and thereby diminishes
the magnetization transfer effect. These effects
provide several opportunities for application of
MT to improve evaluation of white matter le-
sions. Such applications include but are not
limited to: (1) improved sensitivity to enhance-
ment, (2) improved specificity in diagnosis of
multiple sclerosis plaques versus other periven-
tricular white matter disease, and (3) further
characterization of underlying pathology in
multiple sclerosis plaques. Of these, in the
author’s opinion, the first two items are closest
to becoming part of usual clinical practice and
will be discussed below.
MT may be used to augment enhancement,
predominantly by selectively suppressing
Multiple
nonenhancing structures. The application of
Sclerosis
A5.1.12
MT has been shown to approximately double
the contrast-to-noise ratio of enhancing lesions
compared with non-MT spin echo imaging
(Finelli et al., 1994). The MT effect is not
limited to enhancement, and nonenhancing le-
sions may exhibit increased signal on MT im-
ages. For this reason, precontrast MT imaging
is needed to validly diagnose enhancement
(Mehta et al., 1995).
MT may be helpful in distinguishing de-
myelinating lesions from periventricular white
matter ischemic lesions. This follows from the
increased MT effect in white matter ischemic
lesions compared with demyelinating lesions
of multiple sclerosis (Mehta et al., 1996). With
the MT effect higher in ischemic white matter
lesions, such lesions are suppressed along with
normal brain, more so than demyelinating le-
sions. When inspecting MT T1-weighted spin
echo images, the difference translates into high
signal in some demyelinating lesions, but the
contrast between white matter and gray matter
is suppressed (Figure A5.1.4). In general, mul-
tiple sclerosis is usually recognized by consid-
ering lesion distribution and morphology in
conjunction with the clinical setting. In those
cases in which such considerations are insuffi-
cient for confident diagnosis, MT may be a
helpful option.
Current Protocols in Magnetic Resonance Imaging
Critical Parameters and
Troubleshooting
Motion artifacts may be a problem for nearly
any MRI examination. Some motion may be
present without significantly compromising
the diagnostic utility of the examination. More
frequent motion will be a problem. Proper care
in placing and securing the patient in the head
coil is critical, as is the interaction of the MRI
technologist with the patient. In cases in which
such measures are insufficient, prudent con-
scious sedation should be employed if at all
possible. If sedation fails or is not possible, then
faster imaging may be considered. As a first
step, the provided sequences may be shortened
simply by reducing the acquisition matrix (Nx,
Ny). Fast imaging techniques do exist and will
be described in a later unit. However, such
techniques are particularly limited in matters
of contrast and are clearly inferior in evaluating
white matter disease.
Proper contrast in fast FLAIR imaging is
central to the recommended protocol. As noted
previously, manufacturers vary in their imple-
mentation of this sequence, including the pa-
rameters employed. Utilization of FLAIR
should begin with the manufacturer’s standard
parameters. Beyond this, some modifications
may yield more appealing images with im-
proved T2-weighted lesion contrast. Use of a TR
that is too low (i.e., <7000 msec) compromises
both T2-weighted contrast and CSF suppres-
sion. The TE should be fairly long, at least 100
msec, again for reasons related to T2-weighted
contrast. The inversion time (TI) is dictated by
the T1 of CSF and for a 1.5 Tesla magnet should
be between 2100 msec and 2400 msec. See
Rydberg et al. (1995) for a thorough discussion
of the considerations involved.
As already indicated, in early implementa-
tions of TSE/FSE, the TR and TE values chosen
were too long to provide good T2-weighted
lesion contrast. TR and TE values should be
chosen that are not longer than those given in
Table A5.1.6. Use of longer relaxation parame-
ters yields images which are visually accept-
able; however fairly extensive experience
proves that sensitivity to T2-weighted hyperin-
tense lesions is compromised.
Phase-encoding artifacts may occasionally
be problematic in the posterior fossa on post-
contrast images. Use of flow compensation
usually solves such problems. Flow compensa-
tion requires a small increase in TE (up to ∼17
msec). Such a parameter change does not di-
minish image quality. Alternatively, the addi-
Current Protocols in Magnetic Resonance Imaging
tion of a post-contrast T1-weighted scan in a
different plane, usually coronal, generally al-
lows definitive interpretation of the presence of
enhancement.
Anticipated Results
The exam should provide full imaging char-
acterization of brain-parenchymal demyelinat-
ing lesions. To accomplish this end, effectively
three types of sequences with various T2
weighting are employed: fast FLAIR, proton
density, and T2. Each of these is performed in
the plane in which it is most useful. Lesions of
the corpus callosum, especially at the callosal-
septal interface, are optimally visualized by
FLAIR performed in the sagittal plane. Other
periventricular lesions are well visualized on
both the sagittal FLAIR and the PD/T2-
weighted transverse images. Lesion activity, as
manifested by blood-brain barrier compromise,
will be demonstrated by the presence of en-
hancement.
The exam is not intended to optimally detect
optic neuritis. Generally, this is not an impor-
tant imaging indication as its diagnosis is clini-
cally readily apparent and the MRI is usually
employed to evaluate for the presence of other
clinically occult lesions. If evaluation for optic
neuritis is the clinical indication, then a dedi-
cated orbit exam should be performed (see UNIT
A7.5).
Literature Cited
Bastianello, S., Bozzao, A., Paolillo, A., Giugni, E.,
Gasperini, C., Koudriavtseva, T., Millefiorini, E.,
Horsfield, M.A., Colonnese, C., Toni, D.,
Fiorelli, M., Pozzilli, C., and Bozzao L. 1997.
Fast spin-echo and fast fluid-attenuated inver-
sion-recovery versus conventional spin-echo se-
quences for MR quantification of multiple scle-
rosis lesions. Am. J. Neuroradiol. 18:699-704.
Bastianello, S., Gasperini, C., Paolillo, A., Giugni,
E., Ciccarelli, O., Sormani, M.P., Horsfield,
M.A., Rovaris, M., Pozzilli, C., and Filippi, M.
1998. Sensitivity of enhanced MR in multiple
sclerosis: Effects of contrast dose and magneti-
zation transfer contrast. Am. J. Neuroradiol.
19:1863-1867.
Filippi, M., Rovaris, M., Capra, R., Gasperini, C.,
Yousry, T.A., Sormani, M.P., Prandini, F., Hors-
field, M.A., Martinelli, V., Bastianello, S., Kuhne
I, Pozzilli C, and Comi G. 1998a. A multi-centre
longitudinal study comparing the sensitivity of
monthly MRI after standard and triple dose gad-
olinium-DTPA for monitoring disease activity in
multiple sclerosis. Implications for phase II
clinical trials. Brain 121:2011-2020.
Filippi, M., Rovaris, M., Gasperini, C., Capra, R.,
Bastianello, S., Kuhne, I., and Yousry, T.A.
Miscellaneous
Brain Pathology
A5.1.13
 1998b. A preliminary study comparing the sen-
sitivity of serial monthly enhanced MRI after
standard and triple dose gadolinium-DTPA for
monitoring disease activity in primary progres-
sive multiple sclerosis. Journal of Neuroimaging
8:88-93.
Finelli, D.A., Hurst, G.C., Gullapali, R.P., and Bel-
lon, E.M. 1994. Improved contrast of enhancing
brain lesions on postgadolinium, T1-weighted
spin-echo images with use of magnetization
transfer. Radiology 190:553-559.
Gean-Marton, A.D., Vezina, L.G., Marton, K.I., Sti-
mac, G.K., Peyster, R.G., Taveras, J.M., and
Davis, K.R. 1991. Abnormal corpus callosum: A
sensitive and specific indicator of multiple scle-
rosis. Radiology 180:215-221.
Giovannoni, G., Lai, M., Thorpe, J., Kidd, D.,
Chamoun, V., Thompson, A J., Miller, D.H.,
Feldmann, M., and Thompson, E.J. 1997. Lon-
gitudinal study of soluble adhesion molecules in
multiple sclerosis: correlation with gadolinium
enhanced magnetic resonance imaging. Neurol-
ogy 48:1557-1565.
Giugni, E., Pozzilli, C., Bastianello, S., Gasperini,
C., Paolillo, A., Koudriavtseva, T., Frontoni, M.,
Farina, D., and Bozzao, L. 1997. MRI measures
and their relations with clinical disability in re-
lapsing-remitting and secondary progressive
multiple sclerosis. Multiple Sclerosis 3:221-225.
Hajnal, J.V., Bryant, D.J., Kasuboski, L., Pattany,
P.M., De Coene, B., Lewis, P.D., Pennock, J.M.,
Oatridge, A., Young, I.R., and Bydder, G. M.
1992. Use of fluid attenuated inversion recovery
(FLAIR) pulse sequences in MRI of the brain. J.
Comput. Assisted Tomogr. 16:841-844.
Hashemi, R.H., Bradley, W.G., Jr., Chen, D.Y., Jor-
dan, J.E., Queralt, J.A., Cheng, A.E., and Henrie,
J.N. 1995. Suspected multiple sclerosis: MR im-
aging with a thin-section fast FLAIR pulse se-
quence. Radiology 196:505-510.
Hennig, J., Nauerth, A., and Friedburg, H. 1986.
RARE imaging: A fast imaging method for clini-
cal MR. Magn. Reson. Med. 3:823-833.
Jackson, A., Fitzgerald, J.B., and Gillespie, J.E.
1993. The callosal-septal interface lesion in mul-
tiple sclerosis: Effect of sequence and imaging
plane. Neuroradiology 35:573-577.
Katz, D., Taubenberger, J.K., Cannella, B., McFar-
lin, D.E., Raine, C.S., and McFarland, H.F. 1993.
Correlation between magnetic resonance imag-
ing findings and lesion development in chronic,
active multiple sclerosis. Ann. Neurol. 34:661-
669.
Mehta, R.C., Pike, G.B., and Enzmann, D.R. 1995.
Improved detection of enhancing and nonen-
hancing lesions of multiple sclerosis with mag-
netization transfer. Am. J. Neuroradiol. 16:1771-
1778.
Mehta, R.C., Pike, G.B., and Enzmann, D.R. 1996.
Measure of magnetization transfer in multiple
sclerosis demyelinating plaques, white matter
ischemic lesions, and edema. Am. J. Neurora-
Multiple diol. 17:1051-1055.
Sclerosis
A5.1.14
Nesbit, G.M., Forbes, G.S., Scheithauer, B.W., quences with variable slice thickness. Neurora- Key References
Okazaki, H., and Rodriguez, M. 1991. Multiple diology 40:293-297. Shellock, 1996. See above.
sclerosis: Histopathologic and MR and/or CT van Walderveen, M.A., Barkhof, F., Hommes, O.R.,
correlation in 37 cases at biopsy and three cases Covers a number of important patient management
Polman, C.H., Tobi, H., Frequin, S.T., and Valk, issues related to MR imaging, including recom-
at autopsy. Radiology 180:467-474. J. 1995. Correlating MRI and clinical disease
mended safety procedures, a list of metallic implants
Palmer, S., Bradley, W.G., Chen, D.Y., and Patel, S. activity in multiple sclerosis: Relevance of hy- that have been tested for MR compatibility, and a
1999. Subcallosal striations: Early findings of pointense lesions on short-TR/short-TE (T1- list of other sources on MR safety.
multiple sclerosis on sagittal, thin-section, fast w eig h ted ) spin-ech o images. Neurology
FLAIR MR images. Radiology 210:149-153. 45:1684-1690.
Rovaris, M., Mastronardo, G., Gasperini, C., Prand- Wolff, S.D., and Balaban, R.S. 1989. Magnetization
ini, F., Yousry, T.A., and Filippi, M. 1998. MRI transfer contrast (MTC) and tissue water proton
Contributed by Danial K. Hallam
evolution of new MS lesions enhancing after relaxation in vivo. Magn. Reson. Med. 10:135- University of Washington
different doses of gadolinium. Acta Neurol. 144. Seattle, Washington
Scand. 98:90-93.
Rovaris, M., Mastronardo, G., Prandini, F., Bastian-
ello, S., Comi, G., and Filippi, M. 1999a. Short-
term evolution of new multiple sclerosis lesions
enhancing on standard and triple dose gadolin-
ium-enhanced brain MRI scans. J. Neurol. Sci.
164:148-152.
Rovaris, M., Rodegher, M., Comi, G., and Filippi,
M. 1999b. Correlation between MRI and short-
term clinical activity in multiple sclerosis: Com-
parison between standard- and triple-dose Gd-
enhanced MRI. Eur. Neurol. 41:123-127.
Rydberg, J.N., Riederer, S.J., Rydberg, C.H., and
Jack, C.R. 1995. Contrast optimization of fluid-
attenuated inversion recovery (FLAIR) imaging.
Magn. Reson. Med. 34:868-877.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Silver, N.C., Good, C.D., Barker, G.J., MacManus,
D.G., Thompson, A.J., Moseley, I.F., McDonald,
W.I., and Miller, D.H. 1997. Sensitivity of con-
trast enhanced MRI in multiple sclerosis. Effects
of gadolinium dose, magnetization transfer con-
trast and delayed imaging. Brain 120:1149-
1161.
Thorpe, J.W., Halpin, S.F., MacManus, D.G.,
Barker, G.J., Kendall, B.E., and Miller, D.H.
1994. A comparison between fast and conven-
tional spin-echo in the detection of multiple scle-
rosis lesions. Neuroradiology 36:388-392.
Trojano, M., Avolio, C., Simone, I.L., Defazio, G.,
Manzari, C., De Robertis, F., Calo, A., and
Livrea, P. 1996. Soluble intercellular adhesion
molecule-1 in serum and cerebrospinal fluid of
clinically active relapsing-remitting multiple
sclerosis: correlation with Gd-DTPA magnetic
resonance imaging-enhancement and cerebro-
spinal fluid findings. Neurology 47:1535-1541.
Truyen, L., Van Waesberghe, J.H., Van Walderveen,
M.A., Van Oosten, B.W., Polman, C.H.,
Hommes, O.R., Ader, H.J., and Barkhof, F. 1996.
Accumulation of hypointense lesions (“black
holes”) on T1 spin-echo MRI correlates with
disease progression in multiple sclerosis. Neu-
rology 47:1469-1476.
Tubridy, N., Barker, G.J., MacManus, D.G., Mose-
ley, I.F., and Miller, D.H. 1998. Optimisation of
unenhanced MRI for detection of lesions in mul-
tiple sclerosis: A comparison of five pulse se- Miscellaneous
Brain Pathology
A5.1.15
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
Pituitary UNIT A5.2 Materials
Normal saline (0.9% NaCl), sterile
Benefiting from superior tissue contrast, multi-planar capability and lack of bone artifact, Extravascular contrast agent (e.g., Magnevist, Omniscan, or Prohance)
MRI readily depicts complex anatomy in and about the pituitary gland. This unit presents
three basic protocols for common indications relating to pathology of the sella and Set up equipment and patient
parasellar region. The protocols differ in emphasis, more than in concept, and share a 1. Interview the patient to ensure that he or she has no contraindications such as cardiac
basic theme of thin slice high-resolution imaging including the use of gadolinium. With pacemakers or other implants containing ferromagnetic materials. Also be sure to
the possible exception of dynamic imaging, all protocols may be readily performed on find out if the patient has any health conditions that may require the presence of
any MR scanner. special emergency equipment or special precautions during the scanning procedure.
Generally standard screening forms are used for all patients scanned in a magnetic
MACROADENOMA BASIC resonance system.
PROTOCOL 1 The presence of any ferromagnetic metals may be a health hazard to the patient when he
Imaging of suspected or known pituitary macroadenoma focuses on evaluation of tumor
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
extension into adjacent structures, specifically the suprasellar space and cavernous sinus.
composition of the items, it is best to exclude patients with any metal implants; see Shellock
Identification of the lesion itself, essentially by definition, is not a taxing problem. The (1996) for discussion of what implants may be safely scanned using magnetic resonance.
protocol is quite simple and focuses on demonstrating anatomy in the most useful planes.
The T1-weighted spin echo coronal sequence is the cornerstone of the exam, providing Patients may be accompanied into the magnet room by a friend or family member, who can
the majority of useful information. sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
Table A5.2.1 lists the four sequences which comprise the Basic Protocol. Two optional 2. If the procedure is a research protocol, have the patient sign any necessary consent
sequences are listed. A T2-weighted TSE (turbo spin echo)/FSE (fast spin echo) coronal form.
sequence is suggested as an option, to provide better characterization of intrasellar
pathology, such as differentiating cyst versus tumor. An optional sixth sequence consisting 3. Have the patient remove all jewelry and change into a gown to eliminate any metal
of a PD (proton density)/T2-weighted whole brain transverse sequence should be used in that might be found in clothing.
those cases in which sellar pathology is not necessarily the only consideration. Table
4. Have the patient wash off mascara or other makeup to avoid local tissue heating and
A5.2.2 lists the hardware necessary to perform the procedure. Next, stepwise instructions
image artifacts.
for performing the protocol are provided. The protocol is not terribly demanding of
scanner hardware and may be performed well on most clinical systems. Basic Protocol 1 5. Inform the patient about what will occur during the procedure, what he or she will
generally requires 30 to 45 min to complete. experience while in the magnet, and how to behave, including the following:
a. If earphones or headphones are used to protect the ears from the loud sounds
Table A5.2.1 Basic Pituitary Protocol: Macroadenoma produced by the gradients, the patient will be asked to wear these, but will be able
to communicate with you at any time during the imaging.
Sequence and type of weighting Imaging plane
b. The patient will be given a safety squeeze-bulb or similar equipment to request
1. Pilot scan (scout) assistance at any time (demonstrate how this works).
2. T1-weighted spin echo Sagittal
c. For good results the patient should be instructed not to talk unless absolutely
3. T1-weighted spin echo Coronal
necessary and to avoid or minimize swallowing and other movements during each
4. Post-gadolinium T1-weighted spin echo Coronal
scan—i.e., as long as the banging sounds continue. Between scans, talking and
Optional sequences swallowing are acceptable in most cases, but should be avoided when comparative
5. T2-weighted TSE/FSE Coronal positional studies are being performed; the patient will be informed when this is
6. Dual echo PD/T2-weighted TSE/FSE whole brain Transverse the case.
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
Table A5.2.2 Equipment Parameters
6. Have the patient mount onto the table. Set up any triggering devices or other
Coil type Quadrature head coil monitoring equipment that is to be used either before or just after the patient lies
Gradient coil strength Whatever the system permits down.
Cardiac gating No
Peripheral gating For safety only
7. Center the patient in the head coil. Make sure that the head and neck are constrained
Respiratory gating No
to limit motion, especially if high-resolution scans are to be run.
Respirator If required by patient Generally the patient’s head is fixed so that it is horizontal (not tilted) and the neck and
Oxygen If required by patient head lie along the axis of the patient table; other positions may be appropriate depending
Motion cushions Useful on the needs at hand.
Miscellaneous
Brain Pathology Pituitary
Contributed by Danial K. Hallam A5.2.1 A5.2.2
Current Protocols in Magnetic Resonance Imaging (2001) A5.2.1-A5.2.20
Copyright © 2001 by John Wiley & Sons, Inc. Supplement 3 Supplement 3 Current Protocols in Magnetic Resonance Imaging
 8. If needed, place a pillow or other support under the knees to make the patient more be checked on the sagittal scout. A small field of view (FOV) is used to improve in-plane
comfortable. resolution. In order to avoid aliasing, the (read) oversampling and the no-phase-wrap or
the phase oversampling option should be chosen. Occasionally, large sellar masses may
9. Establish an intravenous line from which the contrast agent can be injected, and attach necessitate increasing the FOV—this should be apparent from the pilot scan. Appropriate
this line securely to the patient so that movement into or out of the magnet will not FOV modifications can then be made for all scans focused on the sella. This sequence is
pull at the patient’s arm. used to accurately position subsequent scans. It provides excellent visualization of midline
structures.
This step may be performed prior to entering the magnet room if necessary to save scanner
time. It is preferable to insert the line prior to imaging and to leave the patient in the magnet, Sequence 3: T1-weighted spin echo coronal
with no intervening motion between the scans run before and after contrast agent injection. 14. Set parameters as indicated in Table A.5.2.5. Using sequence 2, position coronal
10. Use the centering light to position the patient’s nasion and put him or her into the images orthogonal to floor of sella or planum sphenoidale.
center of the magnet. The center slice should be placed in the middle of the sella. Sequence 3 is the most valuable
Once this step has been performed, so long as the patient does not move on the table, the sequence of the protocol. It provides excellent visualization of sellar contents and their
table itself can be moved and then replaced in the same position as before without
jeopardizing the positioning of one scan relative to another.
Table A5.2.4 Parameters for T1-Weighted Spin Echo Sagittal (Sequence 2)
11. If the patient is unable to hold still, provide an appropriate sedative.
Patient position Supine
Sequence 1: Rapid three-plane positioning pilot Scan type Spin echo
12. To validate the patient’s position, run the system’s pilot (or scout) scan to ensure Imaging plane (orientation) Sagittal
correct location of the head in three dimensions, using the imaging sequence given Central slice or volume center Midline
in Table A5.2.3 or similar parameters. Echo time (TE) 14 msec
Repeat time (TR) 500 to 600 msec
This sequence usually consists of three orthogonal planes to allow localization. The images Flip angle (FA) 90°
are often also used later to determine where to place the saturation pulses and to set up
Fields of view (FOVx, FOVy) 220 mm, 220 mm
total coverage of the volume of interest.
Resolution (Δx, Δy) 0.98 mm, 0.86 mm
Sequence 2: T1-weighted spin echo sagittal Number of data points collected (Nx, Ny) 224, 256
13. Using images generated in sequence 1, sagittal images are planned to provide thin Slice thickness (Δz) 3 mm
slice midline coverage. Set the parameters as indicated in Table A5.2.4. Number of slices 11
Slice gap (distance factor) 0.5 mm (0.17)
The scan is best positioned first on the transverse pilot scan. Some angulation may be Number of acquisitions (Nacq) 2
needed to provide for true sagittal anatomic imaging. The coverage provided should then Read direction Cranio-caudal
Saturation pulses Inferior may be used
Table A5.2.3 Parameters for Pilot Scan (Scout; Sequence 1) Scan time 4 to 5 min

Patient position Supine Table A5.2.5 Parameters for T1-Weighted Spin Echo Coronal (Sequence 3)
Scan type Gradient echo
Imaging plane (orientation) 3 planes Patient position Supine
Central slice or volume center Run initially at magnet isocenter Scan type Spin echo
Echo time (TE) Minimum Imaging plane (orientation) Coronal
Repeat time (TR) Minimum Central slice or volume center Mid-sella
Flip angle (FA) 20° Echo time (TE) 14 msec
Fields of view (FOVx, FOVy) 300 mm, 300 mm Repeat time (TR) 500 to 600 msec
Resolution (Δx, Δy) 1.17 mm, 1.56 mm Flip angle (FA) 90°
Number of data points collected (Nx, Ny) 256, 192 Fields of view (FOVx, FOVy) 220 mm, 220 mm
Display matrix (Dx, Dy) 256, 256 Resolution (Δx, Δy) 0.98 mm, 0.86 mm
Slice thickness (Δz) 5 mm Number of data points collected (Nx, Ny) 224, 256
Number of slices 3, one in each of 3 cardinal planes Slice thickness (Δz) 3 mm
Slice gap Not applicable Number of slices 11
Number of acquisitions (Nacq) 1 Slice gap (distance factor) 0.5 mm (0.17)
Swap read and phase encoding No Number of acquisitions (Nacq) 2
Slice locations Not applicable Read direction Cranio-caudal
Saturation pulses Not applicable Saturation pulses Inferior may be used
Miscellaneous
Scan time A few seconds Brain Pathology Pituitary Scan time 4 to 5 min

A5.2.3 A5.2.4
Current Protocols in Magnetic Resonance Imaging Supplement 3 Supplement 3 Current Protocols in Magnetic Resonance Imaging
 relationship to the cavernous sinus. It also provides visualization of the optic apparatus in
the neighborhood of the pituitary (Figure A5.2.1). Invasion by pituitary adenoma of the
clivus is demonstrated due to the intrinsic contrast afforded by fatty clival marrow (Figure
A5.2.2).
15. Inject the contrast agent, then flush the intravenous line with 10 ml saline. It is
preferable to perform the contrast injection while leaving the patient in the magnet.
A dose of 0.1 mmol/kg is usually given.
Sequence 4: Post-contrast T1-weighted spin echo coronal
16. Repeat sequence 3 with nearly identical parameters and positioning.
Flow compensation is recommended to minimize phase encoding artifact from the cavern-
ous internal carotids. This will increase the TE to ∼17 msec. If a mass is present, this
sequence further characterizes the lesion by demonstrating its enhancement pattern
(Figure A5.2.3).
Sequence 5: T2-weighted TSE/FSE coronal (optional)
17. Set parameters as indicated in Table A5.2.6.
The scan is positioned identically to the T1-weighted coronal studies. This sequence allows
further characterization of sellar lesions and is particularly helpful in the setting of
hemorrhage and in differentiating cyst from neoplasm (Figure A5.2.4). Rarely, it will
demonstrate a lesion not otherwise detected. As implemented, the scan will result in some
right-to-left aliasing. This may be solved by employing the no-phase-wrap or the phase-
oversampling option available on many scanners. Such measures will increase the imaging
A B

Figure A5.2.3 (A) The presence of an ademona is suggested by possible mass effect and hypointensity
on non-contrast T1-weighted coronal image. (B) Following contrast, adenoma is clearly demarcated by
decreased enhancement relative to normal pituitary gland.

Table A5.2.6 Parameters for T2-Weighted TSE/FSE Coronal (Sequence 5)

Figure A5.2.1 Displacement of optic apparatus by su-
prasellar extension of a Rathke’s cleft cyst is demonstrated Patient position Supine
on T1-weighted coronal image. Scan type TSE/FSE
Imaging plane (orientation) Coronal
Central slice or volume center Mid-sella
Echo time (TE) 102 msec
Echo train length (ETL; turbo factor) 8 to 11
Repeat time (TR) 3000 msec
Flip angle (FA) 180°a
Fields of view (FOVx, FOVy) 180 mm, 180 mm
Resolution (Δx, Δy) 0.94 mm, 0.70 mm
Number of data points collected (Nx, Ny) 192, 256
Slice thickness (Δz) 3 mm
Number of slices 11
Slice gap (distance factor) 1 mm (0.33)
Number of acquisitions (Nacq) 3 to 4
Read direction Cranio-caudal
Saturation pulses Inferior may be used
Figure A5.2.2 Invasive pituitary adenoma. Involvement of Scan time 3.5 to 5 min
clivus and clear invasion of left cavernous sinus demon- Miscellaneous aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this
strated by pre-contrast T1-weighted coronal image. Brain Pathology Pituitary
sequence is 90°.
A5.2.5 A5.2.6
Current Protocols in Magnetic Resonance Imaging Supplement 4 Supplement 4 Current Protocols in Magnetic Resonance Imaging
 time of an already lengthy scan. Alternatively, one may simply accept the aliasing artifact
as it does not overlap the sella.
Sequence 6: Dual echo PD/T2-weighted TSE/FSE transverse (optional)
18. The T1-weighted sagittal image (sequence 2) is used to position these transverse
scans.
Set parameters for the PD/T2-weighted dual echo transverse sequence as given in Table
A5.2.7. The scan is positioned graphically to provide whole brain coverage. The slices
ALTERNATE
PROTOCOL
should be angled according to an institutional standard for transverse images. The author
recommends using “AC-PC” line as such a standard (see UNIT A5.1). The anterior and
posterior commissures are identified on the T1-weighted sagittal midline image. The slices
of all transverse scans are positioned parallel to a line drawn between the two landmarks.
The sequence provides excellent visualization of the whole brain. It should be used as a
quick whole brain screen when the clinical issue being addressed is not necessarily limited
to the sella.
STATUS POST TRANSPHENOIDAL SURGERY
This alternate protocol as summarized in Table A5.2.8 should be used in patients who
have undergone transphenoidal surgery for pituitary lesions. The protocol differs from
the Basic Protocol predominantly by the use of fat saturation on post-contrast images.
Also, a second plane post-contrast is helpful in the evaluation of often quite distorted
post-surgical anatomy. This Alternate Protocol should include sequences 1 through 3 as
discussed in Basic Protocol 1. Then perform post-gadolinium imaging in coronal and
sagittal planes with fat saturation. For patient set up and contrast agent injection, see Basic
Protocol 1, steps 1 to 15.
Sequences 7 and 8: Post-contrast T1-weighted spin echo with fat saturation (FS)
1. Set parameters for sequences 7 and 8, as given in Tables A5.2.9 and A5.2.10
respectively.
The sequences are positioned identically to their pre-contrast counterparts. A reduction in
total number of slices (thus TR) may help to limit a somewhat lengthy scan time. Evaluation
of pre-contrast images will determine whether or not this can be done without compromis-
ing the diagnostic utility of the scan.

Table A5.2.8 Alternate Protocol (Status Post Transphenoidal Surgery)

Figure A5.2.4 T2-weighted TSE image can be helpful to Sequence and type of weighting Imaging plane
better characterize some lesions, as in this case of an arach-
noid cyst. 7. Post-gadolinium T1-weighted spin echo FS Coronal
8. Post-gadolinium T1-weighted spin echo FS Sagittal

Table A5.2.7 Parameters for Dual Echo PD/T2-Weighted TSE/FSE Transverse

(Sequences 6 and 18) Table A5.2.9 Parameters for FS Post-Contrast T1-Weighted Spin Echo
Coronal (Sequence 7)
Patient position Supine
Scan type TSE/FSE Patient position Supine
Imaging plane (orientation) Transverse Scan type Spin echo
Central slice or volume center Position for whole brain coverage Imaging plane (orientation) Coronal
Echo time (TE) 14–17 msec for PD weighted Central slice or volume center Mid-sella
image and 85–102 msec for Echo time (TE) 17 msec
T2-weighted image Repeat time (TR) 600 msec
Echo train length (ETL) (turbo factor) 5 to 8 Flip angle (FA) 90°
Repeat time (TR) 2500–3000 msec Fields of view (FOVx, FOVy) 220 mm, 220 mm
Flip angle (FA) 90° Resolution (Δx, Δy) 0.98 mm, 0.86 mm
Fields of view (FOVx, FOVy) 180 mm, 230 mm Number of data points collected (Nx, Ny) 224, 256
Resolution (Δx, Δy) 0.94 mm, 0.9 mm Slice thickness (Δz) 3 mm
Number of data points collected (Nx, Ny) 192, 256 Number of slices 11
Slice thickness (Δz) 5 mm Slice gap (distance factor) 0.5 mm (0.17)
Number of slices 23 Number of acquisitions (Nacq) 3
Slice gap (distance factor) 1.5 mm (0.30) Read direction Cranio-caudal
Number of acquisitions (Nacq) 2 Flow compensation Yes (if available)
Read direction Anterior to posterior Saturation pulses Inferior may be used
Saturation pulses Inferior may be used Miscellaneous Fat suppression Yes
Scan time 2 to 4 min Brain Pathology Pituitary Scan time 5 to 6 min
A5.2.7 A5.2.8
Current Protocols in Magnetic Resonance Imaging Supplement 3 Supplement 3 Current Protocols in Magnetic Resonance Imaging
 Table A5.2.10 Parameters for FS Post-Contrast T1-Weighted Spin Echo Set up equipment and patient
Sagittal (Sequence 8) 1. Use the same equipment and perform patient setup as for the previous method (see
Basic Protocol 1).
Patient position Supine
Scan type Spin echo 2. As before (see Basic Protocol 1, step 9), establish an intravenous line.
Imaging plane (orientation) Sagittal
For the current protocol, the patient will be scanned while contrast is injected. As such, the
Central slice or volume center Midline i.v. line must be sufficiently long to extend outside the magnet while the patient is being
Echo time (TE) 17 msec imaged.
Repeat time (TR) 500–600 msec
Flip angle (FA) 90° Sequence 9: Pilot scan
Fields of view (FOVx, FOVy) 220 mm, 220 mm 3. Run a rapid three-plane positioning pilot scan (see Basic Protocol 1, sequence 1).
Resolution (Δx, Δy) 0.98 mm, 0.86 mm
Number of data points collected (Nx, Ny) 224, 256 Sequence 10: T1-weighted spin echo sagittal
Slice thickness (Δz) 3 mm 4. Set parameters as indicated in Table A5.2.12.
Number of slices 11
Using images generated in the pilot scan, sagittal images are planned to provide high-reso-
Slice gap (distance factor) 0.5 mm (0.17)
lution thin slice midline coverage. The scan is best positioned first on the transverse pilot
Number of acquisitions (Nacq) 3 scan. Some angulation may be needed to provide for true sagittal anatomic imaging. A
Read direction Cranio-caudal small FOV is used to improve in-plane resolution. In order to avoid aliasing, the (read)
Flow compensation Yes (if available) oversampling and the no-phase-wrap or the phase oversampling option should be chosen.
Saturation pulses Inferior may be used
Fat suppression Yes Sequence 11: High-resolution T1-weighted spin echo coronal
Scan time 5 to 6 min 5. Set parameters as indicated in Table A5.2.13.

MICROADENOMA BASIC Table A5.2.11 Basic Pituitary Microadenoma Protocol

PROTOCOL 2
In the MR examination for microadenomas, emphasis shifts from evaluation of the extent Type of weighting and sequence Imaging plane
of the lesion, primarily to lesion detection. Several approaches exist for meeting this goal.
9. Pilot scan (scout)
First, the resolution of the T1 sequences may be increased, with the recognition that the
10. T1-weighted spin echo Sagittal
technique used for macroadenomas already represents a reasonably high-resolution exam.
11. T1-weighted spin echo Coronal
Further increases in acquisition matrix and decreases in slice thickness come at predict-
12. Dynamic T1-weighted FSE/TSE Coronal
able costs both in imaging time and in motion-artifact vulnerability. Second, dynamic
13. Post-gadolinium T1-weighted spin echo Coronal
imaging during contrast injection may be applied. In microadenoma detection, dynamic
imaging has received much attention and provides some, albeit limited, increase in
sensitivity (Nagele et al., 1993; Davis et al., 1994; Kucharczyk et al., 1994). Addition of
Table A5.2.12 Parameters for T1-Weighted Spin Echo Sagittal (Sequence 10)
a T2-weighted sequence usually does not improve sensitivity (Peck et al., 1989; Elster,
1993) and so is not included as part of the standard protocol. Patient position Supine
Scan type Spin echo
The type of thin-slice, high-resolution exams utilized here push the limits of signal-to-
Imaging plane (orientation) Sagittal
noise. Some modification of noise-relevant parameters including slice thickness, Nacq, and
Central slice or volume center Midline
acquisition matrix may be needed to yield visually satisfactory images, depending on
Echo time (TE) 14 msec
individual scanner capabilities. Many centers do not employ high-resolution T1-weighted
Repeat time (TR) 500-600 msec
sequences as described here. As an alternative approach, the reader may wish to simply
Flip angle (FA) 90°
add the dynamic imaging scan to the macroadenoma protocol as their version of a
Fields of view (FOVx, FOVy) 180 mma, 180 mm
microadenoma exam (see Background Information, Use of Microadenoma Protocol).
Resolution (Δx, Δy) 0.80 mm, 0.70 mm
Due to higher Nacq’s utilized on several sequences, as well as setup for the dynamic exam, Number of data points collected (Nx, Ny) 224, 256
total scanning time will be increased 10 to 15 min beyond that required for the macroade- Slice thickness (Δz) 2 to 3 mm
noma protocol. Number of slices 11
Slice gap 0.5 mm
Optional sequences for the microadenoma protocol are the same as that for the macroade- Number of acquisitions (Nacq) 2 to 3
noma protocol and will not be repeated. Table A5.2.11 lists the five sequences that will Read direction Cranio-caudal
be run in Basic Protocol 2. Saturation pulses Inferior may be used
Miscellaneous
Scan time 3.5 to 7 min
Brain Pathology Pituitary aNo-phase-wrap or phase-oversampling.

A5.2.9 A5.2.10
Current Protocols in Magnetic Resonance Imaging Supplement 3 Supplement 3 Current Protocols in Magnetic Resonance Imaging
 The scan is positioned on the T1-weighted sagittal images orthogonal to sellar floor or Table A5.2.14 Parameters for Dynamic T1-Weighted TSE/FSE Coronal
planum sphenoidale. This is the most important sequence of a protocol geared towards (Sequence 12)
detecting subtle lesions within the pituitary (Figure A5.2.5).
Patient position Supine
Sequence 12: Dynamic T1-weighted TSE/FSE coronal Scan type TSE/FSE
6. Set parameters as indicated in Table A5.2.14. Imaging plane (orientation) Coronal
Central slice or volume center Mid-sella
Microadenomas usually enhance somewhat later than adjacent normal anterior pituitary
gland. Fast T1-weighted scanning performed during contrast injection may allow detection Echo time (TE) 12 msec
of these subtle abnormalities. The scan is run once before, once during and multiple times Echo train length (ETL) (turbo factor) 8
after the administration of contrast agent (for 2 to 3 min after injection). Contrast agent Repeat time (TR) 500 msec
Flip angle (FA) 180°a
Fields of view (FOVx, FOVy) 180 mm, 180 mm
Table A5.2.13 Parameters for High Resolution T1-Weighted Spin Echo
Coronal (Sequences 11 and 13) Resolution (Δx, Δy) 0.94 mm, 0.70 mm
Number of data points collected (Nx, Ny) 192, 256
Patient position Supine Slice thickness (Δz) 3 mm
Scan type Spin echo Number of slices 7
Imaging plane (orientation) Coronal Slice gap 0 to 0.3 mm
Central slice or volume center Mid-sella Number of acquisitions (Nacq) 1 to 2
Echo time (TE) 14 msec Read direction Cranio-caudal
Repeat time (TR) 500–600 msec Saturation pulses Not applicable
Flip angle (FA) 90° Scan time 14 to 28 sec per sequential
Fields of view (FOVx, FOVy) 180 mma, 180 mm acquisition
Resolution (Δx, Δy) 0.80 mm, 0.70 mm aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this

Number of data points collected (Nx, Ny) 224, 256 sequence is 90°.
Slice thickness (Δz) 2 to 3 mm
Number of slices 11 may be administered manually or through the use of an automated injector. Rate of
Slice gap 0.5 mm administration is not critical but should be brisk, with a target of at least 1 ml/sec and
Number of acquisitions (Nacq) 2 to 3 preferably 2 ml/sec.
Read direction Cranio-caudal
A rough survey of techniques used by a few different centers for this particular application
Saturation pulses Inferior may be used
disclosed wide variation in parameters. Institutions vary in chosen echo train length (3 to
Scan time 3.5 to 7 min 8), TR (380 msec to 800 msec), number of slices (3 to 7), slice thickness (2 to 3 mm), and
aNo phase-wrap or phase-oversampling. acquisition matrix. The general idea is simply to use a TSE sequence with parameters
providing T1 weighting and run it several times sequentially (Davis et al., 1994). As such,
likely many parameter variations will be acceptable.

A B Sequence 13: Post-gadolinium T1-weighted spin echo coronal

7. Use parameters as indicated in Table 5.2.13 with identical positioning in sequence
11 and one parameter modification.
Flow compensation option should be used to minimized phase encoding artifact from
cavernous carotids. This option will increase TE to ∼17 msec.

BASIC CAVERNOUS SINUS

PROTOCOL 3
Figure A5.2.5 (A) Non-contrast T1-weighted coronal image demonstrates a 2.5 mm microadenoma in the
right side of the sella. (B) Following contrast, the lesion is no longer apparent.
As with the macroadenoma protocol, evaluation of the cavernous sinus focuses on
T1-weighted imaging in the most useful planes. In this case, transverse imaging is more
helpful than sagittal. Also, T2-weighted imaging is frequently helpful and is included as
part of the standard protocol rather than as an option. Because the anatomy to be evaluated
is part of the skull base, fat suppression is helpful both for T1-weighted post-contrast
imaging and for T2-weighted TSE imaging. Finally, if either a cavernous carotid fistula
or cavernous carotid aneurism is clinically questioned, then 3-D time-of-flight (TOF) MR
angiography is suggested as an option. Table A5.2.15 lists 7 sequences and one optional
sequence for Basic Protocol 3. Basic Protocol 1 generally requires 40 to 60 min to
complete.
Pituitary

A5.2.11 A5.2.12
Current Protocols in Magnetic Resonance Imaging Supplement 3 Supplement 3 Current Protocols in Magnetic Resonance Imaging
 Table A5.2.15 Basic Cavernous Sinus Protocol Table A5.2.17 Parameters for T1-Weighted Spin Echo Coronal (Sequence 16)

Sequence and type of weighting Imaging plane Patient position Supine

14. Pilot scan (scout) Scan type Spin echo
15. T1-weighted spin echo Transverse Imaging plane (orientation) Coronal
16. T1-weighted spin echo Coronal Central slice or volume center Mid-sella
17. T2-weighted TSE/FSE FS Coronal Echo time (TE) 14 msec
18. Dual echo PD/T2-weighted TSE/FSE whole brain Transverse Repeat time (TR) 500–600 msec
19. Post-gadolinium T1-weighted spin echo FS Coronal Flip angle (FA) 90°
20. Post-gadolinium T1-weighted spin echo FS Transverse Fields of view (FOVx, FOVy) 180 mma, 180 mm
Resolution (Δx, Δy) 0.80 mm, 0.70 mm
Optional sequence
Number of data points collected (Nx, Ny) 224, 256
21. 3-D TOF MR angiography (source images)a Transverse
Slice thickness (Δz) 4 mm
aIf CCF is a specific issue, then 3-D short T spoiled gradient echo imaging is recommended
R Number of slices 15
(should be performed prior to gadolinium administration).
Slice gap (distance factor) 1 mm (0.25)
Number of acquisitions (Nacq) 2
Table A5.2.16 Parameters for T1-Weighted Spin Echo Transverse (Sequence Read direction Cranio-caudal
15) Saturation pulses Inferior
Scan time 4 min
Patient position Supine aNo-phase-wrap or phase-oversampling may be used.
Scan type Spin echo
Imaging plane (orientation) Transverse
Central slice or volume center Sella Table A5.2.18 Parameters for T2-Weighted TSE/FSE Coronal (Sequence 17)
Echo time (TE) 14 msec
Patient position Supine
Repeat time (TR) 500–600 msec
Scan type TSE/FSE
Flip angle (FA) 90°
Imaging plane (orientation) Coronal
Fields of view (FOVx, FOVy) 180 mma, 180 mm
Central slice or volume center Mid-sella
Resolution (Δx, Δy) 0.80 mm, 0.70 mm
Echo time (TE) 102 msec
Number of data points collected (Nx, Ny) 224, 256
Echo train length (ETL) (turbo factor) 8 to 11
Slice thickness (Δz) 3 mm
Repeat time (TR) 3000 msec
Number of slices 15
Flip angle (FA) 180°a
Slice gap (distance factor) 0.5 mm (0.17)
Fields of view (FOVx, FOVy) 180 mmb, 180 mm
Number of acquisitions (Nacq) 2
Resolution (Δx, Δy) 0.94 mm, 0.70 mm
Read direction Anterior to posterior
Number of data points collected (Nx, Ny) 192, 256
Saturation pulses Inferior
Slice thickness (Δz) 4 mm
Scan time 4 min
Number of slices 15
aNo-phase-wrap or phase-oversampling.
Slice gap (distance factor) 1 mm (0.25)
Number of acquisitions (Nacq) 3 to 4
Set up equipment and patient Read direction Cranio-caudal
1. Use the same equipment and perform patient setup as for the previous method (see Saturation pulses Inferior
Basic Protocol 1). Fat suppression Yes
Scan time 4 min
Sequence 14: Pilot scan aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this
2. Run a rapid three-plane positioning pilot scan (see Basic Protocol 1, sequence 1). sequence is 90°.
bNo-phase-wrap or phase-oversampling may be used.
Sequence 15: T1-weighted spin echo transverse
3. Set parameters as indicated in Table A5.2.16.
Transverse images are planned off the sagittal pilot scan. The images are positioned
parallel to the planum sphenoidale and should cover from above the sella to the pons.

Sequence 16: T1-weighted spin echo coronal

4. Set parameters as indicated in Table A5.2.17.
Miscellaneous
Brain Pathology Pituitary

A5.2.13 A5.2.14
Current Protocols in Magnetic Resonance Imaging Supplement 3 Supplement 3 Current Protocols in Magnetic Resonance Imaging
 The scan is positioned on the pilot sagittal with images orthogonal to sellar floor or planum Table A5.2.20 Parameters for FS Post-Contrast T1-Weighted Spin Echo
sphenoidale. The scan should cover from mid-pons to mid-orbit. Transverse (Sequence 20)

Sequence 17: T2-weighted TSE/FSE coronal Patient position Supine

5. Set parameters as indicated in Table A5.2.18. Scan type Spin echo
Imaging plane (orientation) Transverse
The scan is positioned identically to the T1-weighted coronal study. This scan provides
Central slice or volume center Sella
T2-weighted information about the contents of the cavernous sinus. Fat saturation allows
for evaluation of pathology in or adjacent to the skull base. Echo time (TE) 17 msec
Repeat time (TR) 650 msec
Sequence 18: Dual echo PD/T2-weighted TSE/FSE whole brain transverse Flip angle (FA) 90°
6. The scan is identical to that of Sequence 6 of the Basic Protocol 1. Set parameters as Fields of view (FOVx, FOVy) 180 mma, 180 mm
indicated in Table A5.2.7. Resolution (Δx, Δy) 0.94 mm, 0.70 mm
Number of data points collected (Nx, Ny) 192, 256
A quick whole brain screening sequence is often helpful in patients with symptoms related Slice thickness (Δz) 3 mm
to cavernous sinus pathology. Number of slices 15 or as allowed
Slice gap (distance factor) 0.5 mm (0.17)
Sequence 19: FS post-contrast T1-weighted spin echo coronal
Number of acquisitions (Nacq) 3
7. Set parameters as indicated in Table A5.2.19 and run with identical positioning to the
Read direction Anterior to posterior
pre-contrast T1-weighted coronal sequence. Repeat step 15 in Basic Protocol 1.
Flow compensation Yes (if available)
The time required for the FS pulse will increase TE and likely require some increase in TR Saturation pulses Inferior may be used
to provide adequate coverage. The TR should be kept below ∼700 msec to maintain proper Fat suppression Yes
T1 weighting. A decrease in total number of slices is often required and usually acceptable Scan time 6 min
depending upon the anatomy of the patient and any pathology demonstrated on the aNo-phase-wrap or phase-oversampling may be used.
pre-contrast scan. In exceptional circumstances, it may be necessary to run two separate
measurements to provide adequate coverage. An increase in Nacq is often needed to
maintain high image quality. Table A5.2.21 Parameters for 3-D Time-of-Flight MR Angiogram (Sequence 21)

Sequence 20: FS post-contrast T1-weighted transverse Patient position Supine

8. Set parameters as indicated in Table A5.2.20 and run with identical positioning to the Scan type 3-D short TR spoiled gradient echo
pre-contrast T1-weighted transverse sequence. Imaging plane (orientation) Transverse
Central slice or volume center Sella
Table A5.2.19 Parameters for FS Post-Contrast T1-Weighted Spin Echo Echo time (TE) Minimum (or 6.9 msec)
Coronal (Sequence 19) Repeat time (TR) Minimum (or 36 msec)
Flip angle (FA) 25°
Patient position Supine Fields of view (FOVx, FOVy) 160 mm, 220 mm
Scan type Spin echo Resolution (Δx, Δy) 0.83 mm, 0.86 mm
Imaging plane (orientation) Coronal Number of data points collected (Nx, Ny) 192, 256
Central slice or volume center Mid-sella Slice thickness (Δz) 1 mm
Echo time (TE) 17 msec Number of slices 50–64
Repeat time (TR) 650 msec Slice gap 0
Flip angle (FA) 90° Number of acquisitions (Nacq) 1
Fields of view (FOVx, FOVy) 180 mma, 180 mm Read direction Anterior to posterior
Resolution (Δx, Δy) 0.94 mm, 0.70 mm Flow compensationa Yes
Number of data points collected (Nx, Ny) 192, 256 Saturation pulses None
Slice thickness (Δz) 4 mm Scan time ∼6 to 7 min
Number of slices 15 or as allowed aVelocity (flow) compensation in read and slice-select directions.
Slice gap (distance factor) 1 mm (0.25)
Number of acquisitions (Nacq) 3
Read direction Cranio-caudal
Flow compensation Yes (if available)
Saturation pulses Inferior may be used
Fat suppression Yes
Scan time 6 min
aNo-phase-wrap or phase-oversampling may be used.
Miscellaneous
Brain Pathology Pituitary

A5.2.15 A5.2.16
Current Protocols in Magnetic Resonance Imaging Supplement 3 Supplement 3 Current Protocols in Magnetic Resonance Imaging
Sequence 21: 3-D TOF MR angiogram (optional)
9. Set parameters as indicated in Table A5.2.21.
The scan is positioned just as prior transverse acquisitions. If this option is chosen, it must
be run prior to gadolinium enhancement (before sequence 19). When the presence of a
cavernous-carotid fistula (CCF) is questioned, a 3-D TOF angiogram is a helpful step. The
exam is interpreted directly from source images where the observation of flow-related
enhancement in the cavernous sinus indicates the presence of a CCF. Processing the image
data to obtain maximum intensity projections can of course be used to aid evaluation of
arterial anatomy in the usual fashion (see Chapter A1).
COMMENTARY
Background Information
Within the modest limits of the sella turcica
and environs, an impressive array of anatomy
merges, reflecting elements of neural, endo-
crine, vascular, meningeal and bony tissues.
The multi-planar capability and tissue contrast
afforded by MRI provides detailed information
about the pathology and anatomy of this region.
With tumors of the anterior pituitary compris-
ing 10% to 15% of all intracranial neoplasms,
careful attention to the MRI evaluation of the
sella is paramount (Kovacs and Horvath, 1986).
As a testimony to the tremendous intrinsic
soft tissue contrast demonstrated by MRI, the
simplest of sequences—a T1-weighted spin
echo sequence performed in the coronal
plane—provides much of the desired informa-
tion. The T1-weighted coronal sequence iden-
tifies most, but not all, microadenomas. The
sequence also evaluates suprasellar extension
of sellar lesions and the presence of optic ap-
paratus compression. Despite the use of T2
weighting and various contrast enhancement
techniques, the two limitations of this se-
quence, incomplete sensitivity to microade-
nomas and evaluation of cavernous sinus inva-
sion, remain problematic.
T1-weighted non-contrast imaging detects
most microadenomas, with the actual sensitiv-
ity varying in different series between 60% and
80% (Kucharczyk et al., 1986; Nichols et al.,
1988; Lundin et al., 1991; Elster, 1993). Post-
gadolinium imaging improves sensitivity
mildly, detecting ∼5% to 10% more lesions
(Macpherson et al., 1989; Newton et al., 1989;
Steiner et al., 1989; Stadnik et al., 1990). Fi-
nally, use of dynamic enhanced techniques pro-
vides visualization of 5% to 20% more lesions,
increasing the overall sensitivity of the com-
plete MR exam to 80% to 90%. The improve-
ment in sensitivity provides the motivation for
the dedicated microadenoma protocol. How-
ever, with each additional step, the MRI be-
comes lengthier, more expensive, and, in the
Current Protocols in Magnetic Resonance Imaging
the so-called pituitary incidentalomas (Chong sion beyond the lateral wall of the cavernous
et al., 1994; Teramoto et al., 1994). The exist- sinus. For less obvious cases, several classifi-
ence of fairly frequent incidental abnormalities cation schemes have been devised for predict-
should be kept in mind in proper interpretation ing the likelihood of invasion, generally using
of high-resolution pituitary exams. the intracavernous carotid artery as a landmark
Differential temporal enhancement of pitui- (Scotti et al., 1988; Knosp et al., 1993).
tary neoplasms provides another dimension
Pathology primary to the cavernous sinus
along which they may be separated from the
normally enhancing gland. The general ap- The cavernous sinus protocol, Basic Proto-
proach is to perform some form of fast T1- col 3, focuses on pathology primary to the
weighted imaging in the coronal plane using cavernous sinus. Reflecting the wide variety of
just a few slices located in the mid-sella. The potential pathology, the protocol is somewhat
sequence is repeated serially just before, at the longer and uses several types of scan weighting
case of the dynamic study, logistically more
beginning of, and several times after a bolus along with enhancement. The protocol may not
complex. Hence, some consideration should be
injection of contrast agent, up to ∼2 min. In this be ideal for each patient and appropriate modi-
given to the need for such an exhaustive search.
way, microadenomas, which usually enhance fications will be necessary from time to time.
in delayed fashion compared to the normal Symptoms related to the cavernous sinus derive
Use of microadenoma protocol
adenohypophysis, may be identified. Fast T1- from dysfunction of its cranial nerve and ve-
Not all microadenomas must be detected.
weighted imaging may be achieved in a number nous contents. If the symptomatology is limited
The need for definitive microadenoma detec-
of ways, and several different approaches have to just one or a few of the cranial nerves present,
tion is determined by many factors relating to
been applied successfully, including FSE/TSE then one of several other similar protocols may
results of endocrine tests, patient factors, and
sequences, gradient-recalled sequences, and be more appropriate. Symptoms related to cra-
institutional habits. For a recent review of the
keyhole imaging (Nagele et al., 1993; Davis et nial nerves III, IV and VI may be better evalu-
issue, the reader is referred to Elster (1994). In
al., 1994; Kucharczyk et al., 1994). In the ated by an orbit protocol (see UNIT A7.5). If only
many cases, reliable diagnosis of microade-
author’s experience, the application of a T1- cranial nerve V symptoms are present, then a
noma can be made by clinical history and re-
weighted TSE/FSE sequence, as indicated in dedicated protocol for imaging the full extent
sults of serum hormone assays. If an institution
Basic Protocol 2, provides reliable and useful of this important nerve is likely more suitable
treats all such neoplasms medically, then detec-
information, and is available on many current (see UNIT A7.2).
tion of the lesion by imaging is not necessary
scanners. In the past, approaches employing Symptoms and signs of a cavernous-carotid
and the MRI is used only to exclude other
gradient-recalled sequences with fairly long fistula are often dramatic and the diagnosis
pathology and perhaps as a baseline. In this
TE’s suffered from skull base and sphenoid clinically apparent. If MRI evaluation of the
event, an all-out MR search for the lesion is
sinus related susceptibility artifacts. With condition is desired, then 3-D TOF MRA
unnecessary. However, in some institutions,
shorter TE’s (down to 2 msec) afforded by newer should be used (sequence 21 of Basic Protocol
surgery is applied as a front-line therapy. Also,
scanners, such artifacts are minimized and gra- 3). Inspection of source images for flow related
medical therapy may be ineffective or not tol-
dient-echo sequences in this application now enhancement in the cavernous sinus has been
erated by the patient. In such situations, detec-
represent a viable option. Keyhole dynamic shown to be more sensitive to the presence of
tion of the lesion preoperatively becomes im-
imaging likely has advantages, but is not com- a CCF (Hirai et al., 1998). It should be noted
portant.
monly available. that not all CCFs will be positive on an MRA
As a first step in increasing sensitivity to
and catheter angiography may be necessary. In
microadenoma detection, the resolution of the Cavernous sinus invasion by pituitary this fairly specific clinical context, not all of the
T1-weighted coronal sequence may be further adenoma sequences in Basic Protocol 3 will be necessary.
increased. To the author’s knowledge, the util-
Identification of cavernous sinus invasion
ity of this rather straightforward move has not
by pituitary adenomas carries important clini- Critical Parameters and
been subjected to a careful, controlled prospec-
cal implications. For the neurosurgeon, an ade- Troubleshooting
tive study. However, given the small size of
noma invading the cavernous sinus cannot usu- Fat saturation techniques will exaggerate
lesions to be detected and the proven value of
ally be successfully removed, so once invasion susceptibility artifacts related to dental hard-
a T1-weighted coronal sequence, it makes in-
is diagnosed, the goal of transphenoidal surgery ware. Such artifacts are not as common in the
herent sense. Of course, several problems may
shifts from tumor removal to tumor debulking. sella as in the orbits and paranasal sinuses but
arise. First, the quality of implementation on
MRI does not often allow for definitive evalu- nonetheless can occasionally be problematic.
individual scanners will vary. The sequence
ation of invasion as the medial wall of the As a first step to deal with such a problem,
pushes the limits of image signal-to-noise and
cavernous sinus is generally not visualized even simply repeat the sequence without the fat satu-
some experimentation with acquisition matrix
with high-resolution imaging. As with most ration pulse. In the case of post-contrast imag-
(number of data points collected), slice thick-
pituitary pathology, the T1-weighted spin echo ing, if a pre-contrast scan in the same plane
ness, and number of acquisitions will likely be
coronal sequence provides the most useful in- exists, then interpretation is still certainly pos-
necessary to achieve a visually pleasing and
formation. Clear signs of invasion include en- sible, albeit perhaps a bit more demanding. If
hence useful exam. Second, higher resolution
Miscellaneous casement of the intracavernous carotid artery T2-weighted FS imaging is desired, as in the
will doubtless uncover an increasing number of
Brain Pathology Pituitary by tumor, cavernous sinus expansion, and inva- cavernous sinus protocol, then a T2-weighted
A5.2.17 A5.2.18
Supplement 3 Supplement 3 Current Protocols in Magnetic Resonance Imaging
STIR (short tau inversion recovery) scan pro-
vides a comparable substitute.
Anticipated Results
The pituitary protocols will provide excel-
lent visualization of the sella and its key neigh-
boring relationships. The pituitary gland, in-
fundibulum, and sella itself will be demon-
strated. Pathology in the adjacent sphenoid
sinus and clivus will be apparent. The optic
chiasm, as well as adjacent optic nerves and
tracts, will be well identified and their displace-
ment by pituitary pathology characterized. All
macroadenomas will be detected along with
most microadenomas. Some information re-
garding the possibility of cavernous sinus inva-
sion will be present.
The cavernous sinus protocol will identify
and characterize contents of the cavernous si-
nus. Multiple sequences are employed in two
advantageous planes in an effort to detail the
intricate anatomy of this region. Reflecting the
variety of pathology which may be present, T1
weighting, T2 weighting, and contrast agent
enhancement sequences are all utilized. Using
this approach, small tumors and inflammatory
processes of the nerves of the sinus will be
demonstrated. The morphology of the cavern-
ous carotids will be demonstrated, although not
quite as well as by catheter angiography. The
presence of abnormalities of cavernous sinus
venous flow will be visualized.
Literature Cited
Chong, B.W., Kucharczyk, W., Singer, W., and
George, S. 1994. Pituitary gland MR: a compara-
tive study of healthy volunteers and patients with
microadenomas. Am. J. Neuroradiol. 15:675-
679.
Davis, W.L., Lee, J.N., King, B.D., and Harnsberger,
H.R. 1994. Dynamic contrast-enhanced MR im-
aging of the pituitary gland with fast spin-echo
technique. J. Magn. Reson. Imaging 4:509-511.
Elster, A.D. 1993. Modern imaging of the pituitary.
Radiology 187:1-14.
Elster, A.D. 1994. High-resolution, dynamic pitui-
tary MR imaging: Standard of care or academic
pastime? [comment]. Am. J. Roentgenol.
163:680-682.
Hirai, T., Korogi, Y., Hamatake, S., Ikushima, I.,
Sugahara, T., Sigematsu, Y., Higashida, Y., and
Takahashi, M. 1998. Three-dimensional FISP
imaging in the evaluation of carotid cavernous
fistula: comparison with contrast-enhanced CT
and spin-echo MR. Am. J. Neuroradiol. 19:253-
259.
Knosp, E., Steiner, E., Kitz, K., and Matula, C. 1993.
Pituitary adenomas with invasion of the cavern-
Current Protocols in Magnetic Resonance Imaging
ous sinus space: A magnetic resonance imaging Teramoto, A., Hirakawa, K., Sanno, N., and Osa- mended safety procedures, a list of metallic implants
classification compared with surgical findings mura, Y. 1994. Incidental pituitary lesions in that have been tested for MR compatibility, and a
[see comments]. Neurosurgery 33:610-617; dis- 1,000 unselected autopsy specimens. Radiology list of other sources on MR safety.
cussion 617-618. 193:161-164.
Kovacs, K., and Horvath, E. 1986. Tumors of the
Key References
Pituitary Gland. Armed Forces Institute of Pa- Contributed by Danial K. Hallam
thology, Washington, D.C. Shellock, 1996. See above.
University of Washington
Kucharczyk, W., Davis, D.O., Kelly, W.M., Sze, G., Covers a number of important patient management Seattle, Washington
Norman, D., and Newton, T.H. 1986. Pituitary issues related to MR imaging, including recom-
adenomas: high-resolution MR imaging at 1.5 T.
Radiology 161:761-765.
Kucharczyk, W., Bishop, J.E., Plewes, D.B., Keller,
M.A., and George, S. 1994. Detection of pitui-
tary microadenomas: comparison of dynamic
keyhole fast spin-echo, unenhanced, and con-
ventional contrast-enhanced MR imaging [see
comments]. Am. J. Roentgenol. 163:671-679.
Lundin, P., Bergstrom, K., Thuomas, K.A., Lund-
berg, P.O., and Muhr, C. 1991. Comparison of
MR imaging and CT in pituitary macroade-
nomas. Acta Radiol. 32:189-196.
Macpherson, P., Hadley, D.M., Teasdale, E., and
Teasdale, G. 1989. Pituitary microadenomas.
Does Gadolinium enhance their demonstration?
Neuroradiology 31:293-298.
Nagele, T., Petersen, D., Klose, U., Grodd, W.,
Opitz, H., Gut, E., Martos, J., and Voigt, K. 1993.
Dynamic contrast enhancement of intracranial
tumors with snapshot-FLASH MR imaging. Am.
J. Neuroradiol. 14:89-98.
Newton, D.R., Dillon, W.P., Norman, D., Newton,
T.H., and Wilson, C.B. 1989. Gd-DTPA-en-
hanced MR imaging of pituitary adenomas. Am.
J. Neuroradiol. 10:949-954.
Nichols, D.A., Laws, E.R., Jr., Houser, O.W., and
Abboud, C.F. 1988. Comparison of magnetic
resonance imaging and computed tomography in
the preoperative evaluation of pituitary ade-
nomas. Neurosurgery 22:380-385.
Peck, W.W., Dillon, W.P., Norman, D., Newton,
T.H., and Wilson, C.B. 1989. High-resolution
MR imaging of pituitary microadenomas at 1.5
T: experience with Cushing disease. Am. J.
Roentgenol. 152:145-151.
Scotti, G., Yu, C.Y., Dillon, W.P., Norman, D.,
Colombo, N., Newton, T.H., De Groot, J., and
Wilson, C.B. 1988. MR imaging of cavernous
sinus involvement by pituitary adenomas. Am. J.
Roentgenol. 151:799-806.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Stadnik, T., Stevenaert, A., Beckers, A., Luypaert,
R., Buisseret, T., and Osteaux, M. 1990. Pituitary
microadenomas: Diagnosis with two-and three-
dimensional MR imaging at 1.5 T before and
after injection of gadolinium. Radiology
176:419-428.
Steiner, E., Imhof, H., and Knosp, E. 1989. Gd-
DTPA enhanced high resolution MR imaging of
pituitary adenomas. Radiographics 9:587-598.
Miscellaneous
Brain Pathology Pituitary
A5.2.19 A5.2.20
Supplement 3 Supplement 3 Current Protocols in Magnetic Resonance Imaging
Dementia UNIT A5.3 Table A5.3.2 Equipment Parameters

Coil type Quadrature head coil

The main clinical purpose of imaging in the setting of dementia has been and continues Gradient coil strength 25 mT/m or whatever the system permits
to be the identification of treatable structural disease. However, as the population ages, Cardiac gating No
growing interest centers on understanding, characterizing, and potentially treating de- Peripheral gating For safety only
menting illnesses. Although several types of imaging are applied, MRI is the most useful Respiratory gating No
structural imaging modality. In this unit, specific protocols intended to maximize the Respirator If required by patient
contribution of MRI to the clinical evaluation of dementia are presented. Each protocol Oxygen If required by patient
has, as its foundation, a whole brain screen designed to evaluate for treatable causes of Motion cushions Useful
dementia. Reflecting the frequent overlap of vascular disease with dementing illness, the
screening protocol also serves as an optimal evaluation of the patient with suspected
the protocol are provided. The protocol is not terribly demanding of scanner hardware
vascular dementia. Additional sequences or modifications are then used to answer specific
and may be performed on any clinical system. Basic Protocol 1 generally requires 20 to
clinical questions. The evaluation of normal pressure hydrocephalus is reserved for a later
25 min.
unit devoted to cerebrospinal fluid (CSF)-flow related diagnoses. The parameters given
here are derived from experience at 1.5 vT and may need to be altered slightly depending
Set up patient and equipment
on the field strength and the equipment manufacturer.
1. Interview the patient to ensure that he or she has no contraindications such as cardiac
pacemakers or other implants containing ferromagnetic materials. Also be sure to
VASCULAR DEMENTIA BASIC find out if the patient has any health conditions that may require the presence of
PROTOCOL 1 special emergency equipment or other precautions during the scanning procedure.
Evaluation for usual structural treatable causes of dementia is not a demanding imaging
problem. Indeed, one might argue that computed tomography (CT) would suffice for most In general, standard screening forms are used for all patients scanned in a magnetic
applications. However, even a limited and fairly brief MR study will better characterize resonance system (APPENDIX 1).
potential abnormalities and allow more sensitive detection of subtle lesions, such as The presence of any ferromagnetic metals may be a health hazard to the patient when he
subcortical ischemic injury, which is common in this patient group and may be of clinical or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
consequence. Multiplanar imaging and improved sensitivity to blood products are addi- composition of the items, it is best to exclude patients with any metal implants; see Shellock
tional advantages provided by MRI. and Kanal (1996) for a discussion of what implants may be safely scanned using magnetic
resonance.
Table A5.3.1 lists the five sequences that comprise the basic screening protocol along with Patients may be accompanied into the magnet room by a friend or family member, who can
five additional sequences used to address specific imaging issues. Table A5.3.2 lists the sit in the room during the scan and comfort the patient as needed. This companion must
hardware necessary to perform the procedure. Next, stepwise instructions for performing be screened as well to ensure the absence of loose metal objects on the body or clothing.

Table A5.3.1 Screening Protocol/Vascular Dementiaa 2. If the procedure is a research protocol, have the patient sign any necessary consent
forms.
Type of weighting and sequence Imaging plane
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
1. Pilot scan (scout) — that might be found in clothing.
2. T1-weighted spin echo Sagittal
4. Have the patient wash off mascara or other makeup to avoid local tissue heating and
3. T2-weighted TSE/FSE Transverse
4. Fast FLAIR Transverse
image artifacts.
5. T2*-weighted gradient echo Coronal 5. Inform the patient about what will occur during the procedure, what he or she will
Additional sequences for specific dementia diagnoses: experience while in the magnet, and how to behave, including the following:
Alzheimer’s evaluation a. If earphones or headphones are used to protect the ears from the loud sounds
6. 3-D short TR T1-weighted gradient echo Oblique coronal produced by the gradients, the patient will be asked to wear these, but will be able
or to communicate with you at any time during the imaging.
7. T1-weighted spin echo Oblique coronal b. The patient will be given a safety squeeze-bulb or similar equipment to request
Huntington’s evaluation assistance at any time (demonstrate how this works).
8. 3-D short TR T1-weighted gradient echo Coronal c. For good results, the patient should be instructed not to talk unless absolutely necessary
or and to avoid or minimize swallowing and other movements during each scan—i.e.,
9. T1-weighted spin echo Coronal as long as the banging sounds continue. Between scans, talking and swallowing are
Parkinson’s evaluation acceptable in most cases, but should be avoided when comparative positional studies
10. Dual echo PD-weighted/T2-weighted CSE Transverse are being performed; the patient will be informed when this is the case.
aAbbreviations: TSE, turbo spin echo; FSE, fast-spin echo; FLAIR, fluid attenuated d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
inversion recovery; PD, proton density; CSE, conventional spin echo. Miscellaneous
Brain Pathology Dementia
Contributed by Danial K. Hallam and Noriko Salamon-Muramaya A5.3.1 A5.3.2
Current Protocols in Magnetic Resonance Imaging (2002) A5.3.1-A5.3.18
Copyright © 2002 by John Wiley & Sons, Inc. Supplement 6 Supplement 6 Current Protocols in Magnetic Resonance Imaging
 6. Have the patient lie onto the table. Set up any triggering devices or other monitoring Table A5.3.3 Parameters for Pilot Scan (Scout)
equipment that is to be used either before or just after the patient lies down.
Patient position Supine
7. Center the patient in the head coil. Make sure that the head and neck are constrained
Scan type Gradient echo
to limit motion.
Imaging plane (orientation) 3 planes
Generally, the patient’s head is fixed so that the head is horizontal (not tilted) and the neck Central slice or volume center Run initially at magnet isocenter
and head lie along the axis of the patient table; other positions may be appropriate Echo time (TE) Minimum
depending on the needs at hand.
Repeat time (TR) Minimum
8. If needed, place a pillow or other support under the knees to make the patient more Flip angle (FA) 20°
comfortable. Fields of view (FOVx, FOVy) 300 mm, 300 mm
9. Use the centering light to position the patient’s nasion and put him or her into the Resolution (Δx, Δy) 1.17 mm, 1.56 mm
center of the magnet. Number of data points collected (Nx, Ny) 256, 192
Display matrix (Dx, Dy) 256, 256
Once this step has been performed, so long as the patient does not move on the table, the
table itself can be moved and then replaced in the same position as before without Slice thickness (Δz) 5 mm
jeopardizing the positioning of one scan relative to another. Number of slices 3, one in each of 3 cardinal planes
Slice gap Not applicable
10. If the patient is unable to hold still, provide an appropriate sedative. Number of acquisitions (Nacq) 1
Sequence 1: Rapid three-plane positioning pilot Swap read and phase encoding No
11. To validate the patient’s position, run the system’s pilot (or scout) scan to ensure Scan time ∼10 sec
correct location of the head in three dimensions, using the imaging sequence given
in Table A5.3.3 or similar parameters.
Table A5.3.4 Parameters for Sagittal T1-Weighted Spin Echo
This sequence usually consists of three orthogonal planes to allow localization. The images
are often also used later to determine where to place the saturation pulses and to set up Patient position Supine
total coverage of the volume of interest. Scan type Spin echo
If the pilot scan shows the patient’s head to be significantly off-center then it may be helpful Imaging plane (orientation) Sagittal
to reposition the patient and repeat the pilot scan. Central slice or volume center Midline
Echo time (TE) 14 msec
Sequence 2: Sagittal T1-weighted spin echo
Repeat time (TR) 600 msec
12. Using images generated in sequence 1, plan sagittal images to provide whole brain
Flip angle (FA) 90°
coverage. Set the parameters as indicated in Table A5.3.4. Run sequence 2.
Fields of view (FOVx, FOVy) 230 mm, 230 mm
The scan is best positioned first on the transverse pilot scan. Some angulation may be Resolution (Δx, Δy) 1.20 mm, 0.90 mm
needed to provide for true sagittal anatomic imaging. The coverage provided should then Number of data points collected (Nx, Ny) 192, 256
be checked on the sagittal scout. This sequence is used to accurately position subsequent
Slice thickness (Δz) 5 mm
scans. It provides excellent visualization of midline structures.
Number of slices 19–21
Sequence 3: Transverse T2-weighted TSE/FSE Slice gap (distance factor) 1.5 mm (0.30)
13. Use the sagittal T1-weighted image to position transverse scans. Set parameters as Number of acquisitions (Nacq) 1
given in Table A5.3.5. Run sequence 3. Read direction Cranio-caudal
The scan is positioned graphically to provide whole brain coverage. The slices should be Saturation pulses Inferior may be used
angled according to an institutional standard for transverse images. The authors recom- Scan time ∼2 min
mend using the “AC-PC” line as such a standard. See Figure A5.1.2. Here, the anterior
and posterior commissures are identified on the sagittal T1-weighted midline image. The
slices of all transverse scans are positioned parallel to a line drawn between the two Sequence 5: Coronal T2*-weighted gradient echo
landmarks. The sequence provides excellent visualization of the whole brain. 15. Set parameters as indicated in Table A5.3.7. Using sagittal T1-weighted images,
Sequence 4: Transverse fast FLAIR position coronal images orthogonal to transverse images using graphic prescription
14. Set parameters for transverse fast FLAIR as indicated in Table A5.3.6. Run sequence 4. for whole brain coverage. Run sequence 5.

The scan is positioned identically to sequence 3. The sequence provides superior demon-
stration of T2-weighted bright lesions, especially in the supratentorial compartment. The
suppression of bright CSF afforded by the FLAIR technique demonstrates lesions adjacent
to CSF that would otherwise be poorly visualized. Subtle subcortical ischemic injury is
optimally visualized. FLAIR also nicely demonstrates extra-axial fluid collections, such as
subdural hematomas, occasionally seen in this patient population.
Miscellaneous
Brain Pathology
This sequence is designed to provide sensitivity to blood products through exploitation of
their susceptibility effects. Punctate foci of hemorrhage as seen in amyloid angiopathy or
coating of brain surfaces with hemosiderin as seen in superficial siderosis may not be
detected without susceptibility weighting. Use of the coronal plane simplifies interpretation
of skull base susceptibility artifacts inherent to this sequence.
Dementia

A5.3.3 A5.3.4
Current Protocols in Magnetic Resonance Imaging Supplement 6 Supplement 6 Current Protocols in Magnetic Resonance Imaging
 Table A5.3.5 Parameters for Transverse T2-Weighted TSE/FSE Table A5.3.7 Parameters for Coronal T2*-Weighted Gradient Echo

Patient position Supine Patient position Supine

Scan type TSE/FSE Scan type 2-D gradient echo
Imaging plane (orientation) Transverse Imaging plane (orientation) Coronal
Central slice or volume center Position for whole brain coverage Central slice or volume center Position for whole brain coverage
Echo time (TE) 85 msec Echo time (TE) 20–25 msec
Echo train length (ETL) 5 to 8 Repeat time (TR) 600–900 mseca
Repeat time (TR) 2500 to 3000 msec Flip angle (FA) 20°
Flip angle (FA) 90° Fields of view (FOVx, FOVy) 180 mm, 230 mm
Fields of view(FOVx, FOVy) 180 mm, 230 mm Resolution (Δx, Δy) 0.94 mm, 0.90 mm
Resolution (Δx, Δy) 0.94 mm, 0.90 mm Number of data points collected (Nx, Ny) 192, 256
Number of data points collected (Nx, Ny) 192, 256 Slice thickness (Δz) 4–6 mm
Slice thickness (Δz) 5 mm Number of slices 19–21
Number of slices 23 Slice gap 1–2 mm
Slice gap (distance factor) 1.5 mm (0.30) Number of acquisitions (Nacq) 1–2
Number of acquisitions (Nacq) 1–2 Read direction Cranio-caudal
Read direction Anterior to posterior Saturation pulses Inferior may be used
Saturation pulses Inferior may be used Scan time ∼2–6 min
Scan time ∼1–4 min aManufacturers vary in the efficiency of multiplanar slice acquisition in this sequence. Choice of T and
R
slice thickness amounts to a trade-off between imaging time and resolution.

Table A5.3.6 Parameters for Transverse Fast FLAIR

BASIC ALZHEIMER’S DISEASE
Patient position Supine PROTOCOL 2
Scan type Fast FLAIR Routine MRI scans nicely depict atrophy of mesial temporal lobe structures that suffer
Imaging plane (orientation) Transverse the dominant impact of Alzheimer-type degenerative changes. In this protocol, an angled
Central slice or volume center Position for whole brain coverage T1-weighted sequence to provide a cross-section of hippocampal structures is chosen to
Echo time (TE) 105–140 mseca demonstrate these changes (Fig. A5.3.1). Multiple investigators employ a 3-D, short TR,
Echo train length (ETL) 7–11 T1-weighted gradient echo sequence for sophisticated volumetric studies on Alzheimer
Repeat time (TR) 8000–10,000 mseca disease (AD) patients. Such a sequence is recommended in addition to the screening
Inversion time (TI) 2000–2400 mseca protocol (see Basic Protocol 1) as an AD focused exam. Alternatively, a coronal T1-
Flip angle (FA) 180° weighted spin echo sequence will suffice when visual inspection only is planned.
Fields of view (FOVx, FOVy) 180 mm, 230 mm Compared to the 3-D volume sequence, the 2-D spin echo sequence has the advantages
Resolution (Δx, Δy) 0.94 mm, 0.90 mm of being somewhat shorter in duration and overall less vulnerable to motion degradation.
Number of data points collected (Nx, Ny) 192, 256
Slice thickness (Δz) 5 mm Set up patient and equipment
Number of slices 23 1. Use the same equipment and perform patient setup as in Basic Protocol 1, steps 1 to 10.
Slice gap (distance factor) 1.5 mm (0.3) 2. Perform sequences 1 to 5 as described in Basic Protocol 1.
Number of acquisitions (Nacq) 1
Read direction Anterior to posterior Sequence 6: Additional sequence for AD: 3-D oblique coronal short TR T1-weighted
Saturation pulses Inferior may be used gradient echo
Scan time ∼2.5–5 min 3. Set parameters as indicated in Table A5.3.8. The scan is positioned graphically by
aOptimum choice of parameters for fast FLAIR varies significantly with manufacturer and scanner first locating the hippocampus on an off-midline slice from the sagittal T1-weighted
limitations. For a review of parameter optimization in fast FLAIR, please refer to Rydberg et al. (1995). sequence (Fig. A5.3.2). Then position the volume for the current sequence so that the
plane of acquisition is orthogonal to the axis of the hippocampus. If the hippocampus
proves difficult to identify, then as a fall back approach, one may position the scan
to be parallel to the axis of the brainstem—this is usually a very similar plane as that
chosen directly orthogonal to the axis of the hippocampus. Run sequence 6.
This sequence demonstrates anatomy of mesial temporal lobe structures allowing for
optimal identification of the disproportionate atrophy seen in AD. The 3-D technique
Miscellaneous employed provides thin slices and allows for reformations in additional planes.
Brain Pathology Dementia

A5.3.5 A5.3.6
Current Protocols in Magnetic Resonance Imaging Supplement 6 Supplement 6 Current Protocols in Magnetic Resonance Imaging
Figure A5.3.1 A coronal T1-weighted image demonstrates disproportionate atrophy of mesial
temporal lobe structures, most notably the hippocampus (hi). Parahippocampal gyrus is also
marked (PHG). Figure A5.3.2 An off-midline sagittal T1-weighted image allows visualization of long axis of
hippocampus (along arrowheads). Oblique coronal T1-weighted images are positioned perpendicu-
lar to the hippocampal long axis.
Table A5.3.8 Parameters for Sequence 6: 3-D Oblique Coronal T1-Weighted

Patient position Supine Table A5.3.9 Parameters for Oblique Coronal T1-Weighted Spin Echo
Scan type 3-D gradient echo
Imaging plane (orientation) Oblique coronal Patient position Supine
Pulse sequence database SPGR Scan type Spin echo
Central slice or volume center Temporal lobe Imaging plane (orientation) Oblique coronal
Echo time (TE) “Minimum full” (∼6 msec) Central slice or volume center Position for whole brain coverage
Repeat time (TR) 33 msec Echo time (TE) 14 msec
Flip angle (FA) 35°–45° Repeat time (TR) 600 msec
Fields of view (FOVx, FOVy) 165 mm, 220 mm Flip angle (FA) 90°
Resolution (Δx, Δy) 1.15 mm–0.86 mm, 0.86 mm Fields of view (FOVx, FOVy) 180 mm, 230 mm
Number of data points collected (Nx, Ny) 144–192, 256 Resolution (Δx, Δy) 0.94 mm, 0.90 mm
Slice thickness (Δz) 1.61 mm Number of data points collected (Nx, Ny) 192, 256
Number of slices 124 Slice thickness (Δz) 5 mm
Slab thickness 200 mm Number of slices 23
Slice gap 0 mm Slice gap (distance factor) 1.5 mm (0.30)
Number of acquisitions (Nacq) 0.75a Number of acquisitions (Nacq) 1–2
Read direction Cranio-caudal (superior to inferior) Read direction Superior to inferior
Saturation pulses Inferior may be used Saturation pulses Inferior may be used
Scan time 7.5–10 min Scan time ∼2–4 min
4
aRefers to 3⁄ -NEX imaging, also known as partial Fourier imaging—an option available on many clinical

scanners. Partial Fourier imaging reduces imaging time by exploiting the inherent symmetry of k-space;
essentially, fewer phase encoding steps are used to provide the same spatial resolution. The time saved is
at the expense of the signal-to-noise ratio.

Miscellaneous
Brain Pathology Dementia

A5.3.7 A5.3.8
Current Protocols in Magnetic Resonance Imaging Supplement 6 Supplement 6 Current Protocols in Magnetic Resonance Imaging
Sequence 7: Alternate additional sequence for AD: 2-D oblique coronal T1-weighted
spin echo
4. Set parameters as indicated in Table A5.3.9. The scan is positioned in a manner
identical to that described for sequence 6. Run sequence 7.
This sequence demonstrates anatomy of mesial temporal lobe structures allowing for
optimal identification of the disproportionate atrophy seen in AD.
HUNTINGTON’S DISEASE BASIC
PROTOCOL 3
Characteristic pathologic abnormalities for Huntington’s disease consist of bilateral
atrophy of the caudate nucleus and putamen (together denoted striatum), often accompa-
nied by a moderate degree of gyral atrophy in the frontal and temporal regions. The striatal
atrophy is best demonstrated in coronal plane (Fig. A5.3.3). As with Alzheimer’s disease,
a coronal T1-weighted sequence is chosen for its superior demonstration of anatomy.
Unlike the AD evaluation, the sequence is positioned in routine coronal and not oblique
coronal orientation. Again, an option of 3-D, short TR, T1-weighted gradient echo or 2-D
T1-weighted spin echo is provided. Considerations regarding sequence choice are identi-
cal to those noted in Basic Protocol 2.
Set up patient and equipment
1. Use the same equipment and perform patient setup as in Basic Protocol 1, steps 1 to 10.
2. Perform sequences 1 to 5 as described previously.
Sequence 8: Additional sequence for Huntington’s disease: 3-D coronal short TR
T1-weighted gradient echo
3. Set parameters as indicated in Table A5.3.8 with orientation to be normal coronal.
The scan is positioned perpendicular to prior transverse sequences (sequences 3 and
4). Run sequence 8.
A B
Figure A5.3.3 Coronal T1-weighted images of two Huntington’s patients, (A) and (B), demonstrat-
ing caudatal atrophy seen with advanced disease. In the normal, the caudate head bulges into the
lateral aspect of the anterior horn of the lateral ventricles. Atrophy of the caudate head results in
loss of the usual medially oriented convex margin, which may become flattened or with increasing Miscellaneous
atrophy, concave. Reproduced from D.H. Yock (1995) with permission from Mosby. Brain Pathology
A5.3.9
Current Protocols in Magnetic Resonance Imaging Supplement 6
Sequence 9: Alternate additional sequence for Huntington’s disease: 2-D coronal
T1-weighted spin echo
4. Set parameters as indicated in Table A5.3.9 with orientation to be normal coronal.
The scan is positioned perpendicularly to prior transverse sequences (sequences 3
and 4). Run sequence 9.
BASIC PARKINSON’S DISEASE
PROTOCOL 4
Little in the way of structural abnormality is present in Parkinson’s disease. However,
subtle findings in the midbrain reflecting the loss of dopaminergic nigrostriatal neurons
from the pars compacta of the substantia nigra have been identified (Duguid et al., 1986;
Rutledge et al., 1987; Braffman et al., 1989; Olanow, 1992). These findings consist of
narrowing or smudging of the high signal regions separating the substantia nigra from the
red nucleus and are best demonstrated by a sequence providing some susceptibility
weighting (Fig. A5.3.4). A conventional dual echo PD-weighted/T2-weighted spin echo
sequence is chosen for this application.
Set up patient and equipment
1. Use the same equipment and perform equipment and patient setup as in Basic Protocol
1, steps 1 to 10.
Figure A5.3.4 Conventional T2-weighted spin echo demonstrates narrowing and smudging of
pars compacta (arrow) in Parkinson’s Disease. The pars reticulata (black arrowhead) of the
substantia nigra exhibits low signal secondary to mineralization. The red nucleus (white arrowhead)
Dementia is similarly recognized.
A5.3.10
Supplement 6 Current Protocols in Magnetic Resonance Imaging
 Table A5.3.10 Parameters for Dual Echo PD-Weighted/T2-Weighted Routine clinical imaging of demented pa- reasonable specificity in diagnosing AD. Fri-
Conventional Spin Echo tients may not be necessary depending on the soni applied a series of linear measurements
clinician’s level of confidence in the diagnosis. relating to MTL structures in a group of patients
Patient position Supine In general practice, any uncertainty regarding with mild to moderate AD (Frisoni et al., 1996).
Scan type Conventional spin echo the diagnosis usually prompts performance of A compound measure including temporal horn
Imaging plane (orientation) Transverse either CT or MRI. Imaging is specifically rec- and choroid fissure widths along with hippo-
Central slice or volume center Midbrain ommended when focal signs or symptoms sug- campal height was 86% sensitive in discrimi-
Echo time (TE) 30 msec and 80 msec gest either underlying treatable disorder (e.g., nating patients with mild AD from controls. As
Repeat time (TR) 2500 msec subdural hematoma, hydrocephalus, tumor) or a logical next step, Jack et al. applied volumet-
Flip angle (FA) 90° the presence of co-existent pathology (e.g., ric measurement of MTL structures in an effort
Fields of view (FOVx, FOVy) 180 mm, 230 mm stroke, hemorrhage; Larson et al., 1986; Hol- to diagnose patients with the mildest form of
lister and Boutros, 1991). Imaging is also rec- AD as indicated by Clinical Dementia Rating
Resolution (Δx, Δy) 0.94 mm, 0.90 mm
ommended in the setting of acute onset or rapid scores (Jack et al., 1997). Volume measure-
Number of data points collected (Nx, Ny) 192, 256
progression of dementia symptoms. ments of hippocampi were 72% sensitive (with
Slice thickness (Δz) 5 mm In the following, an attempt to present a specificity 90%) in diagnosing patients with
Number of slices 20 complete discussion of the imaging findings in very mild AD compared with controls.
Slice gap (distance factor) 1–1.5 mm (0.20–0.30) dementia will not be made. Rather, specific More recent studies have attempted to iden-
Number of acquisitions (Nacq) 0.75a imaging findings relevant to protocol issues tify individuals at risk for developing AD using
Read direction Anterior to posterior will be emphasized. For the interested reader, MTL atrophy measurements. Convit and col-
Saturation pulses Inferior may be used several recent excellent reviews of neuroimag- leagues studied patients with minimal cognitive
Scan time ∼6 min ing in dementia are available, including discus- impairment (MCI), a clinically defined group
sions representing radiologic (Braffman, 2000) with mild memory and/or word finding defi-
4
aRefers to 3⁄ NEX imaging, also known as partial Fourier imaging—an option available on many clinical

scanners. Partial Fourier imaging reduces imaging time by exploiting the inherent symmetry of k-space; and neurologic (Chawluk, 1999) perspectives. ciency, 75% of whom progress eventually to
essentially, fewer phase encoding steps are used to provide the same spatial resolution. The time saved is AD (Flicker et al., 1991; Convit et al., 1997).
at the expense of the signal-to-noise ratio. Alzheimer’s disease Hippocampal volume measurements correctly
The gross pathologic appearance in AD con- classified MCI subjects compared with nor-
sists of variable atrophy of temporal, frontal, mals 74% of the time. Similar results support-
2. Perform sequences 1 to 5 as described previously. The transverse T2-weighted and parietal lobes with some sparing of periro- ing the ability of hippocampal volume esti-
sequence TSE may be omitted to save time. landic regions. In the past decade, much interest mates to identify individuals at risk for AD have
has centered on use of quantitative MRI tech- been published by other laboratories (Kaye et
Sequence 10: Additional sequence for Parkinson’s disease: Transverse dual echo niques to measure volume of hippocampi and al., 1997; Jack et al., 1999).
PD-weighted/T2-weighted spin echo other mesial temporal lobe (MTL) structures.
The interest in this specific region derives from Huntington’s disease
3. Set parameters as indicated in Table A5.3.10. The scan is positioned in the routine
several observations. First, atrophy of MTL Huntington’s disease (HD) is readily recog-
transverse plane—as done for sequence 3. Run sequence 10.
structures has been observed early and more nized by the clinical triad of chorea, dementia,
The sequence will require more scanning time than the FSE/TSE sequence. With the extensively in AD patients (de Leon et al., and personality disorder along with family his-
T2-weighted coverage provided, the T2-weighted TSE becomes somewhat redundant and 1989). Second, recognizing the central role of tory reflecting its dominant inheritance pattern.
may be omitted. A conventional spin echo sequence is chosen because it provides modest MTL structures in memory (Squire, 1992), In questionable cases, genetic analysis is used
susceptibility weighting allowing visualization of mineralizated areas of the substantia memory impairment is an early and severe to identify the defective gene. Such evaluation
nigra. A more heavily gradient echo sequence may result in some “blooming” of the iron, manifestation of AD. Finally, the anatomy of may even be applied in asymptomatic patients.
thereby obscuring the adjacent high signal region whose narrowing is detected. Sequences this particular region lends itself to volumetric For this reason, imaging evaluation is now
minimizing susceptibility effect (FSE/TSE) might suffice, however, their utility for this analysis more readily than that of other AD rarely utilized in most centers.
application has not been documented in the literature. involved structures. Characteristic imaging findings have been
Techniques applied to evaluation of MTL described that are quite apparent in advanced
COMMENTARY atrophy in AD range from visual inspection to cases (Fig. A5.3.3). Directly paralleling the
sophisticated volumetric quantification. Some- pathologic changes, imaging demonstrates pro-
Background Information years (Evans et al., 1989). Alzheimer’s disease what surprisingly, even simple visual inspec- nounced bilateral atrophy of the caudate and
Expected marked increases in the number of is the most common cause of dementia account- tion has some utility and has been shown to putamen (Simmons et al., 1986). The caudatal
elderly magnify the need for improved evalu- ing for over one half of all cases. Its financial support the diagnosis of AD (sensitivity 81%; atrophy results in a characteristic loss of the
ation, understanding, and, if possible, treatment cost in the U.S. has been estimated at 60 billion specificity 67%) versus control population, es- usual bulge of caudate head into the anterior
of dementia. The world population >65 years dollars annually (Adams et al., 1997). In one pecially in groups with lower baseline atrophy horn of the lateral ventricle. On MRI, signal
of age is anticipated to grow at an annual rate series, the Alzheimer’s incidence rate was 123 (subjects <75 years of age; Scheltens et al., abnormalities have been described in the
of 2.4%, leading to estimates of 37 million cases per 100,000 yearly (Schoenberg et al., 1992). Barclay and colleagues (1992) applied striatum and may consist variably of either
demented people worldwide by the year 2020 1987). Prevalence increases drastically with subjective evaluation of MRI to distinguish decrease or increase in signal intensity (Sa-
(Cooper, 1994). In the U.S., ∼5% of the popu- age. Per 100,000 population, prevalence is patients with AD from those with multi-infarct voiardo et al., 1991). Diffuse cerebral atrophy,
lation >65 years of age has severe dementia 3200 in ages 70 to 79 years and 10,800 in dementia or mixed diagnostic features yielding especially of the frontal lobes, occurs in later
Miscellaneous
with estimates of ≥15% for the population >85 individuals >80 years (Adams et al., 1997). Brain Pathology Dementia

A5.3.11 A5.3.12
Current Protocols in Magnetic Resonance Imaging Supplement 6 Supplement 6 Current Protocols in Magnetic Resonance Imaging
stages of HD and correlates with cognitive
decline (Starkstein et al., 1992; Aylward et al.,
1998).
Volumetric techniques are useful in mild
cases and might potentially have some value in
monitoring response to therapy. Atrophy of the
putamen, as demonstrated by MRI, exceeds
that of the caudate in mild disease (Harris et al.,
1992). Decreased basal ganglia volume has
been shown to correlate with presence of the
gene for HD in asymptomatic subjects (Ayl-
ward et al., 1994). Additionally, greater rate of
basal ganglia atrophy has been shown to corre-
late with earlier symptom onset (Aylward et al.,
1997).
Parkinson’s disease
The diagnosis of Parkinson’s disease rests
on clinical observation including core findings
of expressionless face, paucity of voluntary
movement, and resting tremor among others.
On MRI, reflecting the pathologic hallmark of
loss of pigmented cells in the substantia nigra,
especially in the pars compacta, the width of
the pars compacta is narrowed relative to con-
trols (Fig. A5.3.4; Duguid et al., 1986). The
observed narrowing of the pars compacta pre-
sumably relates to neuronal loss or may other-
wise reflect iron deposition.
The diagnostic challenge in Parkinson’s dis-
ease is generally one of separating it from nu-
merous other Parkinsonian syndromes, includ-
ing progressive supranuclear palsy, corticobasal
degeneration and the various forms of multiple
system atrophy. Occasionally, imaging may be
helpful in this regard. It should first be noted
that the finding on proton density weighted
images of pars compacta narrowing is not useful
in this regard as it has been reported in other
Parkinsonian syndromes including striatonigral
degeneration and progressive supranuclear
palsy (Stern et al., 1989). Separate findings for
the specific Parkinsonian syndromes are occa-
sionally observed and may be useful. In progres-
sive supranuclear palsy (PSP), some, but not all,
patients exhibit focal atrophy/signal abnormal-
ity of superior colliculi and/or periadequeductal
regions (Fig. A5.3.5; Savoiardo et al., 1989), a
finding not seen in typical Parkinson’s patients.
The diagnosis of PSP can be confused with that
of corticobasal degeneration (CBD), especially
at disease onset because atypical presentations
are not uncommon and the clinical syndromes
may be incompletely expressed (Davis et al.,
1985; Bergeron et al., 1996). Soliveri and col-
leagues reviewed the cognitive and MRI find-
Current Protocols in Magnetic Resonance Imaging
ings of these two clinically overlapping ill-
nesses. They found the MRI findings of mid-
brain atrophy in PSP and asymmetric fron-
toparietal atrophy in CBD to be the most useful
and consistent aids to clinical evaluation in
distinguishing the two diseases (Soliveri et al.,
1999).
The term multiple system atrophy refers to
the complex syndrome exhibiting various de-
grees of Parkinsonism (striatonigral degenera-
tion), ataxia (olivopontocerebellar degenera-
tion), and autonomic dysfunction (Shy-Drager
syndrome). Each of these disorders may occur
either in isolation or in combination. Nearly all
forms of multiple system atrophy exhibit focal
low T2-weighted signal in the putamen, de-
scribed as “slit-like” signal void (Fig. A5.3.6;
Savoiardo et al., 1989; Lang et al., 1994). The
signal loss is presumed due to iron deposition
related to neuronal degeneration and is best
demonstrated with the susceptibility weighting
afforded by a T2*-weighted gradient echo se-
quence. Signal loss in the putamen distin-
guishes patients with multiple system atrophy
from those with typical Parkinson’s disease and
age-matched controls (Olanow, 1992).
Olivopontocerebellar degeneration is charac- Figure A5.3.5 Patient with progressive supranuclear palsy. Characteristic focal marked atrophy
of the superior colliculi (black arrows) and periaqueductal region (white arrowheads) is present.
terized by marked neuronal loss in inferior
olives, pons, and cerebellum. In advanced
cases, inspection of T1-weighted sagittal and
T2-weighted transverse images reveals marked
atrophy of pons and cerbellar cortex, findings
helpful in diagnosis (Fig. A5.3.7; Savoiardo et
al., 1990).
Structural lesions of a variety of etiologies
may result in a Parkinsonian syndrome. Toxic
diseases resulting in focal basal ganglia lesions
may be associated with Parkinsonism. Focal
bilateral necrosis of the basal ganglia may result
from carbon monoxide poisoning, cyanide poi-
soning, methanol intoxication, among others,
and all may be associated with Parkinsonism
(Ley and Gali, 1983; Carella et al., 1988; Mess-
ing and Storch, 1988; Verslegers et al., 1988;
Rosenberg et al., 1989; Lee and Marsden, 1994;
Choi and Cheon, 1999; Sohn et al., 2000).
Metabolic disorders resulting in dense basal
ganglia calcification such as hypoparathyroid-
ism and pseudohypoparathyroidism may be
associated with various movement disorders
related to compromised function of the calci-
fied nuclei (Pearson et al., 1981). Parkinsonism
may be an unusual manifestation of hydro-
cephalus including both obstructive causes and
as a component of normal pressure hydro-
Figure A5.3.6 Patient with striatonigral degeneration. T2-weighted coronal image demonstrates
cephalus (Miodrag et al., 1987; Curran and Miscellaneous slit-like signal void in putamen (large arrows) resulting from iron depostion. In this case, some high
Brain Pathology Dementia signal was also seen adjacent to the patient’s left putamen (small arrows).
A5.3.13 A5.3.14
Supplement 6 Supplement 6 Current Protocols in Magnetic Resonance Imaging

A deep gray matter) atrophy.
B C
Figure A5.3.7 Profound atrophy of pons and cerebellum demonstrated by T2-weighted transverse (A) and T1-weighted
sagittal, midline (B) and para-midline (C) in this patient with olivopontocerebellar degeneration as a component of multiple
system atrophy.
Lang, 1994; Friedman, 1996; Krauss et al.,
1997). Even multiple deep grey infarcts may
result in Parkinsonism (Zijlmans et al., 1995).
Critical Parameters and
Troubleshooting
Motion artifacts may be a problem for nearly
any MRI examination and can be particularly
problematic in the population evaluated for
dementia. As always, proper care in placing and
securing the patient in the head coil is critical
Current Protocols in Magnetic Resonance Imaging
as is the interaction of the MRI technologist
with the patient. In cases in which such meas-
ures are insufficient, prudent conscious seda-
tion should be employed if at all possible. If
sedation fails or is not possible, then faster
imaging may be considered. As a first step, the
provided sequences may be shortened simply
by reducing the acquisition matrix (Nx, Ny).
Fast imaging techniques do exist and will be
described in a later unit. However, such tech-
niques are particularly limited in matters of
contrast and so may poorly demonstrate sub-
cortical ischemic injury and focal (especially
Proper contrast in fast FLAIR imaging is
important to the dementia protocol. As noted
previously, manufacturers vary in their imple-
mentation of this sequence, including the pa-
rameters employed. Utilization of FLAIR
should begin with the manufacturer’s standard
parameters. Beyond this, some modifications
may yield more appealing images with im-
proved T2-weighted lesion contrast. Use of a TR
that is too low (i.e., <7000 msec) compromises
both T2-weighted contrast and CSF suppres-
sion. The TE should be fairly long, at least 100
msec, again for reasons related to T2-weighted
contrast. The inversion time (TI) is dictated by
the T1 of CSF and for a 1.5-T magnet should be
between 2200 and 2400 msec. The reader is
referred to Rydberg et al. (1995) for a thorough
discussion of the considerations involved.
The choice of TE in the T2*-weighted se-
quence affects sensitivity to blood products.
The use of a longer TE allows more time for
spin dephasing thereby enhancing detection of
small areas of old blood. In some ways, this
choice reflects a trade-off between increasing
sensitivity to old blood and reducing distracting
skull base susceptibility artifacts. The use of
coronal plane for this sequence is designed to
limit uncertainty in recognition of skull base
artifacts. The recommended TE range of 20 to
25 msec is empirically derived. Susceptibility
effects are strongly related to magnetic field
strength and are significantly decreased even at
1.0 T compared to 1.5 T. Use of longer TE at
lower field strengths will only partially com-
pensate for the reduced susceptibility effect.
Anticipated Results
The exam should provide full imaging char-
acterization of brain parenchymal lesions in-
cluding cortical and subcortical infarctions,
encephalomalacia, as well as diffuse and focal
atrophy. To accomplish this end, several types
of sequences with various T1, T2, and suscepti-
bility weighting are employed, including T1-
weighted spin echo, fast FLAIR, T2-weighted
FSE or CSE, and T2*-weighted gradient echo.
Each of these is performed in the plane in which
it is most helpful. The coronal T2*-weighted
gradient echo sequence optimally demon-
strates the hemorrhagic nature of lesions and in
doing so may illuminate the diagnosis. The
addition of a sequence in the coronal plane
Miscellaneous better demonstrates the location of some le-
Brain Pathology Dementia sions and is particularly helpful in charac-
A5.3.15 A5.3.16
Supplement 7 Supplement 7
terizing encephalomalacia related to prior
trauma.
Literature Cited
Adams, R.D., Victor, M., and Ropper, A.H. 1997.
Principles of Neurology. McGraw-Hill, New
York.
Aylward, E.H., Brandt, J., Codori, A.M., Mangus,
R.S., Barta, P.E., and Harris, G.J. 1994. Reduced
basal ganglia volume associated with the gene
for Huntington’s disease in asymptomatic at-risk
persons. Neurology 44:823-828.
Aylward, E.H., Li, Q., Stine, O.C., Ranen, N., Sherr,
M., Barta, P.E., Bylsma, F.W., Pearlson, G.D.,
and Ross, C.A. 1997. Longitudinal change in
basal ganglia volume in patients with Hunt-
ington’s disease. Neurology 48:394-399.
Aylward, E.H., Anderson, N.B., Bylsma, F.W., Wag-
ster, M.V., Barta, P.E., Sherr, M., Feeney, J.,
Davis, A., Rosenblatt, A., Pearlson, G.D., and
Ross, C.A. 1998. Frontal lobe volume in patients
with Huntington’s disease. Neurology 50:252-
258.
Barclay, L.L., Linden, C., and Murtagh, R. 1992.
Medial temporal atrophy as a magnetic reso-
nance imaging marker for Alzheimer’s disease.
J. Neuroimaging 2:131-135.
Bergeron, C., Pollanen, M.S., Weyer, L., Black, S.E.,
and Lang, A.E. 1996. Unusual clinical presenta-
tions of cortical-basal ganglionic degeneration.
Ann. Neurol. 40:893-900.
Braffman, B.H. 2000. The aging brain and neurode-
generative disorders. In Neuroimaging: Clinical
and Physical Principles (R.A. Zimmerman, W.A.
Gibby, and R.F. Carmody, eds.) pp. 951-978.
Springer, New York.
Braffman, B.H., Grossman, R.I., Goldberg, H.I.,
Stern, M.B., Hurtig, H.I., Hackney, D.B., Bi-
laniuk, L.T., and Zimmerman, R.A. 1989. MR
imaging of Parkinson disease with spin-echo and
gradient-echo sequences. A.J.R. 152:159-165.
Carella, F., Grassi, M.P., Savoiardo, M., Contri, P.,
Rapuzzi, B., and Mangoni, A. 1988. Dystonic-
Parkinsonian syndrome after cyanide poisoning:
Clinical and MRI findings. J. Neurol. Neurosurg.
Psychiatry 51:1345-1348.
Chawluk, A. 1999. Neuroimaging of normal brain
aging and dementia. In Neuroimaging: A Com-
panion to Adams and Victor’s Principles of Neu-
rology (J.O. Greenberg, ed.) pp. 301-332.
McGraw-Hill, New York.
Choi, I.S. and Cheon, H.Y. 1999. Delayed move-
ment disorders after carbon monoxide poison-
ing. Eur. Neurol. 42:141-144.
Convit, A., De Leon, M.J., Tarshish, C., De Santi, S.,
Tsui, W., Rusinek, H., and George, A. 1997.
Specific hippocampal volume reductions in indi-
viduals at risk for Alzheimer’s disease. Neuro-
biol. Aging 18:131-138.
Cooper, B. 1994. Health care policy and planning
for dementia: An international perspective. In
Dementia and Normal Aging (F.A. Huppeit and
Current Protocols in Magnetic Resonance Imaging
 D.W. O’Conner, eds.) pp. 519-551. Cambridge
University Press, Cambridge.
Curran, T. and Lang, A.E. 1994. Parkinsonian syn-
dromes associated with hydrocephalus: Case re-
ports, a review of the literature, and pathophysi-
ological hypotheses. Mov. Disord. 9:508-520.
Davis, P.H., Bergeron, C., and McLachlan, D.R.
1985. Atypical presentation of progressive su-
pranuclear palsy. Ann. Neurol. 17:337-343.
de Leon, M.J., George, A.E., Stylopoulos, L.A.,
Smith, G., and Miller, D.C. 1989. Early marker
for Alzheimer’s disease: The atrophic hippocam-
pus. Lancet 2:672-673.
Duguid, J.R., De La Paz, R., and DeGroot, J. 1986.
Magnetic resonance imaging of the midbrain in
Parkinson’s disease. Ann. Neurol. 20:744-747.
Evans, D.A., Funkenstein, H.H., Albert, M.S.,
Scherr, P.A., Cook, N.R., Chown, M.J., Hebert,
L.E., Hennekens, C.H., and Taylor, J.O. 1989.
Prevalence of Alzheimer’s disease in a commu-
nity population of older persons. Higher than
previously reported. JAMA 262:2551-2556.
Flicker, C., Ferris, S.H., and Reisberg, B. 1991. Mild
cognitive impairment in the elderly: Predictors
of dementia. Neurology 41:1006-1009.
Friedman, J.H. 1996. Walking-induced Parkinson-
ism due to presumed idiopathic normal pressure
hydrocephalus. Mov. Disord 11: 99-101.
Frisoni, G.B., Beltramello, A., Weiss, C., Geroldi,
C., Bianchetti, A., and Trabucchi, M. 1996. Lin-
ear measures of atrophy in mild Alzheimer dis-
ease. A.J.N.R. 17:913-923.
Harris, G.J., Pearlson, G.D., Peyser, C.E., Aylward,
E.H., Roberts, J., Barta, P.E., Chase, G.A., and
Folstein, S.E. 1992. Putamen volume reduction
on magnetic resonance imaging exceeds caudate
changes in mild Huntington’s disease. Ann.
Neurol. 31:69-75.
Hollister, L.E. and Boutros, N. 1991. Clinical use of
CT and MR scans in psychiatric patients. J.
Psychiatry Neurosci. 16:194-198.
Jack, C.R. Jr., Petersen, R.C., Xu, Y.C., Waring, S.C.,
O’Brien, P.C., Tangalos, E.G., Smith, G.E., Ivnik,
R.J., and Kokmen, E. 1997. Medial temporal
atrophy on MRI in normal aging and very mild
Alzheimer’s disease. Neurology 49:786-794.
Jack, C.R. Jr., Petersen, R.C., Xu, Y.C., O’Brien,
P.C., Smith, G.E., Ivnik, R.J., Boeve, B.F., War-
ing, S.C., Tangalos, E.G., and Kokmen, E. 1999.
Prediction of AD with MRI-based hippocampal
volume in mild cognitive impairment. Neurology
52:1397-1403.
Kaye, J.A., Swihart, T., Howieson, D., Dame, A.,
Moore, M.M., Karnos, T., Camicioli, R., Ball,
M., Oken, B., and Sexton, G. 1997. Volume loss
of the hippocampus and temporal lobe in healthy
elderly persons destined to develop dementia.
Neurology 48:1297-1304.
Krauss, J.K., Regel, J.P., Droste, D.W., Orszagh, M.,
Borremans, J.J., and Vach, W. 1997. Movement
disorders in adult hydrocephalus. Mov. Disord.
12:53-60.
Current Protocols in Magnetic Resonance Imaging
Lang, A.E., Curran, T., Provias, J., and Bergeron, C.
1994. Striatonigral degeneration: Iron deposi-
tion in putamen correlates with the slit-like void
signal of magnetic resonance imaging. Can. J.
Neurol. Sci. 21:311-318.
Larson, E.B., Reifler, B.V., Sumi, S.M., Canfield,
C.G., and Chinn, N.M. 1986. Diagnostic tests in
the evaluation of dementia. A prospective study
of 200 elderly outpatients. Arch. Intern. Med.
146:1917-1922.
Lee, M.S. and Marsden, C.D. 1994. Neurological
sequelae following carbon monoxide poisoning
clinical course and outcome according to the
clinical types and brain computed tomography
scan findings. Mov. Disord. 9:550-558.
Ley, C.O. and Gali, F.G. 1983. Parkinsonian syn-
drome after methanol intoxication. Eur. Neurol.
22:405-409.
Messing, B. and Storch, B. 1988. Computer to-
mography and magnetic resonance imaging in
cyanide poisoning. Eur. Arch. Psychiatry Neurol.
Sci. 237:139-143.
Miodrag, A., Das, T.K., and Shepherd, R.J. 1987.
Normal pressure hydrocephalus presenting as
Parkinson’s syndrome. Postgrad. Med. J.
63:113-115.
Olanow, C.W. 1992. Magnetic resonance imaging in
Parkinsonism. Neurol. Clin. 10:405-420.
Pearson, D.W., Durward, W.F., Fogelman, I., Boyle,
I.T., and Beastall, G. 1981. Pseudohypoparathy-
roidism presenting as severe Parkinsonism. Post-
grad. Med. J. 57:445-447.
Rosenberg, N.L., Myers, J.A., and Martin, W.R.
1989. Cyanide-induced Parkinsonism: Clinical,
MRI, and 6-fluorodopa PET studies. Neurology
39:142-144.
Rutledge, J.N., Hilal, S.K., Silver, A.J., Defendini,
R., and Fahn, S. 1987. Study of movement dis-
orders and brain iron by MR. A.J.R. 149:365-
379.
Rydberg, J.N., Riederer, S.J., Rydberg, C.H., and
Jack, C.R. 1995. Contrast optimization of fluid-
attenuated inversion recovery (FLAIR) imaging.
Magn. Reson. Med. 34:868-877.
Savoiardo, M., Strada, L., Girotti, F., D’Incerti, L.,
Sberna, M., Soliveri, P., and Balzarini, L. 1989.
MR imaging in progressive supranuclear palsy
and Shy-Drager syndrome. J. Comput. Assist.
Tomogr. 13:555-560.
Savoiardo, M., Strada, L., Girotti, F., Zimmerman,
R.A., Grisoli, M., Testa, D., and Petrillo, R. 1990.
Olivopontocerebellar atrophy: MR diagnosis
and relationship to multisystem atrophy. Radiol-
ogy 174:693-696.
Savoiardo, M., Strada, L., Oliva, D., Girotti, F., and
D’Incerti, L. 1991. Abnormal MRI signal in the
rigid form of Huntington’s disease. J. Neurol.
Neurosurg. Psychiatry 54:888-891.
Scheltens, P., Leys, D., Barkhof, F., Huglo, D., We-
instein, H.C., Vermersch, P., Kuiper, M., Stein-
ling, M., Wolters, E.C., and Valk, J. 1992. Atro-
phy of medial temporal lobes on MRI in “prob-
able” Alzheimer’s disease and normal ageing:
Diagnostic value and neuropsychological corre-
lates. J. Neurol. Neurosurg. Psychiatry 55:967-
972.
Schoenberg, B.S., Kokmen, E., and Okazaki, H.
1987. Alzheimer’s disease and other dementing
illnesses in a defined United States population:
Incidence rates and clinical features. Ann.
Neurol. 22:724-729.
Shellock, F.G. and Kanal, E. 1996. Magnetic Reso-
nance Bioeffects, Safety and Patient Manage-
ment. Lippincott Raven, Philadelphia.
Simmons, J.T., Pastakia, B., Chase, T.N., and Shults,
C.W. 1986. Magnetic resonance imaging in
Huntington disease. A.J.N.R. 7:25-28.
Sohn, Y.H., Jeong, Y., Kim, H.S., Im, J.H., and Kim,
J.S. 2000. The brain lesion responsible for Park-
insonism after carbon monoxide poisoning.
Arch. Neurol. 57:1214-1218.
Soliveri, P., Monza, D., Paridi, D., Radice, D.,
Grisoli, M., Testa, D., Savoiardo, M., and Girotti,
F. 1999. Cognitive and magnetic resonance im-
aging aspects of corticobasal degeneration and
progressive supranuclear palsy. Neurology
53:502-507.
Squire, L.R. 1992. Memory and the hippocampus:
A synthesis from findings with rats, monkeys,
and humans. Psychol. Rev. 99:195-231.
Starkstein, S.E., Brandt, J., Bylsma, F., Peyser, C.,
Folstein, M., and Folstein, S.E. 1992. Neuropsy-
chological correlates of brain atrophy in Hunt-
ington’s disease: A magnetic resonance imaging
study. Neuroradiology 34:487-489.
Stern, M.B., Braffman, B.H., Skolnick, B.E., Hurtig,
H.I., and Grossman, R.I. 1989. Magnetic reso-
nance imaging in Parkinson’s disease and Park-
insonian syndromes. Neurology 39:1524-1526.
Miscellaneous
Brain Pathology Dementia
A5.3.17 A5.3.18
Supplement 6 Supplement 6
Verslegers, W., Van den Kerchove, M., Crols, R., De
Potter, W., Appel, B., and Lowenthal, A. 1988.
Methanol intoxication. Parkinsonism and de-
creased Met-enkephalin levels due to putaminal
necrosis. Acta. Neurol. Belg. 88:163-171.
Yock, D.H. 1995. Magnetic Resonance Imaging of
CNS Disease—A Teaching File. p. 281. Mosby-
Year Book, Inc. (cases 455 and 456).
Zijlmans, J.C., Thijssen, H.O., Vogels, O.J., Kremer,
H.P., Poels, P.J., Schoonderwaldt, H.C., Merx,
J.L., van ’t Hof, M.A., Thien, T., and Horstink,
M.W. 1995. MRI in patients with suspected vas-
cular Parkinsonism. Neurology 45:2183-2188.
Key References
Shellock and Kanal, 1996. See above.
Covers a number of important patient management
issues related to MR imaging, including recom-
mended safety procedures, a list of metallic implants
that have been tested for MR compatibility, and a
list of other sources on MR safety.
Terry, R.D., Katzman R., and Bick, K.L. 1999.
Alzheimer Disease, 2nd Edition. Lippincott Wil-
liams & Wilkins, Philadelphia.
A current review of the major aspects of Alzheimer’s
disease including clinical, epidemiologic, struc-
tural, imaging, chemical, genetic, and molecular.
Contributed by Danial K. Hallam
University of Washington
Seattle, Washington
Noriko Salamon-Muramaya
Northwestern University
Chicago, Illinois
Current Protocols in Magnetic Resonance Imaging

CHAPTER A6
Clinical fMRI
INTRODUCTION
linical functional MRI (fMRI), in the context of this chapter, refers specifically to
C the use of blood oxygenation level dependent (BOLD) contrast to detect localized
hemodynamic responses due to specific regional neuronal activity elicited by defined
cognitive tasks to which the patient is exposed in a controlled manner. The method was
first demonstrated in humans in the early 1990s in a research setting (Bandettini et al.,
1992; Kwong et al., 1992; Ogawa et al., 1992), but has rapidly found clinical applications
(Lee et al., 1999; Thulborn, 1999). At clinical field strengths of 1.5 T, the signal change
is small (1% to 3%) thereby requiring the use of image averaging during the cognitive
task. Thus, the stimulus paradigm has a block design, consisting of repetitive cycles of at
least two different stimulus conditions. These two conditions differ by the cognitive
function being examined. Both the baseline and the active conditions may last from 20
to 60 sec and as many as 10 cycles may be used. Images are acquired continuously across
all cycles. This block design allows the signal averaging that is essential to detect the
small signal changes induced by the paradigm at 1.5 T, although these changes can be
increased at higher field strengths (Gati et al., 1997; Thulborn 1999). The alternative
approach of event-related design is being developed in a research setting (Rosen et al.,
1998; Richter, 1999), but the lower signal-to-noise performance does not suit the short
duration of acquisition times that are important in clinical applications (Marquart et al.,
2000). An interesting approach, in which no paradigm is apparently required to demon-
strate connectivity between different regions of the brain, may have clinical applications
but will not be discussed further until clinical applications (UNIT A6.1) are reported from
multiple sites (Biswal et al., 1995).
The statistical processing of such data to derive the activation map can take many forms,
and no general agreement as to a common method has been reached within the research
community. Generally, for imaging data of adequate quality from a cooperative subject,
the processing method will have little impact on the clinical information sought in
presurgical planning. The available software packages have been reviewed elsewhere
(Gold et al., 1998). Some of this software is public domain, is supported with excellent
documentation, and continues to evolve (Friston et al., 1991; Cox 1996; Eddy et al., 1996;
Strupp, 1996). It is generally sufficient to use a simple Student t test to detect differences
in MR signal intensity between the two conditions. This information is presented as a
color scale, representing statistical significance, displayed over the original images or
superimposed over a high-resolution anatomic image coregistered to the original func-
tional images during acquisition.
Imaging is usually performed with a fast imaging sequence to maximize signal averaging
of as many images as possible in the shortest time. This is important for patient comfort
and to minimize task fatigue. Echo-planar imaging is the most common strategy, and is
supplied by most scanner manufacturers (Cohen and Weisskoff, 1991). Other methods
are also available including spiral imaging (Glover et al., 1996) and turboflash (Frahm et
al., 1992). Although the mechanism of the MR signal difference is a complex function of
localized changes in blood flow, blood volume, hematocrit, and hemoglobin oxygen
binding (Cohen, 1997; Buxton et al., 1998), the phenomenon depends on the changes in
magnetic susceptibility arising from changes in the tissue content of paramagnetic
Contributed by Keith R. Thulborn
Current Protocols in Magnetic Resonance Imaging (2001) A6.0.1-A6.0.3
Copyright © 2001 by John Wiley & Sons, Inc.
deoxyhemoglobin and diamagnetic oxyhemoglobin (Thulborn et al., 1982; Ogawa et al.,
1990). This is the basis for the improved sensitivity of fMRI at higher magnetic field, and
predicts the success of the newer very high field scanners (3.0 T) appearing on the clinical
market (Thulborn, 1999).
The interpretation of the small signal differences detected by fMRI requires that there are
no systematic changes in scanner performance over time. Routine quality assurance (UNIT
A6.2) avoids frustration for the technologist resulting from wasted scanner time, as well as
for the patient who must repeat a study if the first did not provide beneficial clinical
information. Some simple system performance parameters should be monitored routinely.
The clinical questions of fMRI are largely driven by the need to localize various functions.
This means that the host of research paradigms (UNIT A6.3) being used to investigate many
aspects of human cognition can be reduced to a simple, robust and informative subset of
tasks to map eloquent cortex.
LITERATURE CITED
Bandettini, P.A., Wong, E.C., Hinks, R.S., Tikofsky, R.S., and Hyde, J.S. 1992. Time course EPI of human
brain function during task activation. Magn. Reson. Med. 25:390-397.
Biswal, B., Yetkin, F.Z., Haughton, V.M., and Hyde, J.S. 1995. Functional connectivity in the motor cortex
of resting human brain using echo-planar MRI. Magn. Reson. Med. 34:537-541.
Buxton, R.B., Wong, E.C., and Frank, L.R. 1998. Dynamics of blood flow and oxygenation changes during
brain activation: The balloon model. Magn. Reson. Med. 39:855-864.
Cohen, M.S. 1997. A linear systems approach to the parametric analysis of fMRI time series. NeuroImage
6:93-103.
Cohen, M.S., and Weisskoff, R.M. 1991. Ultrafast imaging. Magn. Reson. Imaging 9:1-37.
Cox, R.W. 1996. AFNI: software for analysis and visualization of functional magnetic resonance
neuroimages. Comput. Biomed. Res. 29:162-173.
Eddy, W.F., Fitzgerald, M., Genovese, C., Mockus, A., and Noll, D.C. 1996. Functional Imaging Analysis
Software: Computational Olio. Proceedings in Computational Statistics. pp. 39-49. Physica-Verlag,
Heidelberg.
Frahm, J., Bruhn, H., Merboldt, K.D., Hanicke, W., and Math, D. 1992. Dynamic MR imaging of human
brain oxygenation during rest and photic stimulation. J. Magn. Reson. Imaging 2:501-505.
Friston, K.J., Frith, C.D., Liddle, P.F., and Frackowiak, R.S. 1991. Comparing functional (PET) images: The
assessment of significant change. J. Cereb. Blood Flow Metab. 11:690-699.
Gati, J.S., Menon, R.S., Ugurbil, K., and Rutt, B.K. 1997. Experimental determination of the BOLD field
strength dependence in vessels and tissue. Magn. Reson. Med. 38:296-302.
Glover, G.H., Lemieux, S.K., Drangova, M., and Pauly, J.M. 1996. Decomposition of inflow and blood oxygen
level-dependent (BOLD) effects with dual-echo spiral gradient-recalled echo (GRE) fMRI. Magn. Reson.
Med. 35:299-308.
Gold, S., Christian, B., Arndt, S., Zeien, G., Cizadlo, T., Johnson, D.L., Fiaum, M., and Andreasen, N.C.
1998. Functional MRI statistical software packages: A comparative analysis. Human Brain Mapp. 6:73-84.
Kwong, K.K., Belliveau, J.W., Chesler, D.A., Goldberg, I.E., Weisskoff, R.M., Poncelet, B.P., Kennedy, D.N.,
Hoppel, B.E., Cohen, M.S., Turner, R., Cheng, H.-M., Brady, T.J., and Rosen, B.R. 1992. Dynamic
magnetic resonance imaging of human brain activity during primary sensory stimulation. Proc. Natl. Acad.
Sci. U.S.A. 89:5675-5679.
Lee, C.C., Ward, H.A., Sharbrough, F.S., Meyer, F.B., March, W.R., Raffel, C., So, E.L., Cascino, G.D., Shin,
C., Xu, Y., Riederer, S.J., and Jack, C.R. 1999. Assessment of functional MR imaging in neurosurgical
planning. Am. J. Neuroradiol. 20:1511-1519.
Marquart, M., Birn, R., and Haughton, V. 2000. Single- and multiple-event paradigms for identification of
motor cortex activation. Am. J. Neuroradiol. 21:94-98.
Ogawa, S., Tank, D.W., Menon, R.S., Ellermann, J.M., Kim, S.-G., Merkle, H., and Ugurbil, K. 1992. Intrinsic
signal changes accompanying sensory stimulation: Functional brain mapping using MRI. Proc. Natl.
Acad. Sci. U.S.A. 89:5951-5955.
Ogawa, S., Lee, T.M., Nayak, A.S., and Glynn, P. 1990. Oxygenation-sensitive contrast in magnetic resonance
image of rodent brain at high fields. Magn. Reson. Med. 14:68-78.
Clinical fMRI Introduction
A6.0.1 A6.0.2
Supplement 1 Supplement 1 Current Protocols in Magnetic Resonance Imaging
Richter, W. 1999. High temporal resolution functional magnetic resonance imaging at very-high-field. Topics
Magn. Reson. Imaging 10:51-62.
Rosen, B.R., Buckner, R., and Dale, A.M. 1998. Event-related functional MRI: Past present and future. Proc.
Natl. Acad. Sci. U.S.A. 95:773-780.
Strupp, J.P. 1996. Stimulate: A GUI based fMRI analysis software package. NeuroImage 3:607.
Thulborn, K.R. 1999. Clinical rationale for very high field (3.0 Tesla) functional MR imaging. Topics Magn.
Reson. Imaging 10:37-50.
Thulborn, K.R., Waterton, J.C., Matthews, P.M., and Radda, G.K. 1982. Oxygenation dependence of
transverse relaxation time of water protons in whole blood at high field. Biochim. Biophy. Acta 714:265-
270.
KEY REFERENCES
Moonen, C. and Bandettini, P. (eds.) Medical Radiology: Diagnostic Imaging and Radiation Oncology.
Functional MRI. Springer-Verlag, Berlin, Germany.
Matthews, P. and Jezzard, P. (eds.) Oxford University Press, Oxford, England. In press.
Keith R. Thulborn
Current Protocols in Magnetic Resonance Imaging
Clinical Applications of fMRI UNIT A6.1
Presurgical mapping is the most pressing clinical application for functional MRI (fMRI).
Clinical examples have been described elsewhere in the literature (Lee et al., 1999;
Thulborn, 1999a). The diagnostic applications in psychiatric disease in which dysfunction
has no obvious structural basis are still research-based. An early example is the possible
role of fMRI in the diagnosis of probable Alzheimer’s disease (Thulborn et al., 2000).
Applications in stroke rehabilitation (Booth et al., 1999; Thulborn et al., 1999a), learning
disabilities, and neurological disorders remain to be developed. These future clinical
applications are likely to require longitudinal examinations and will require the develop-
ment of calibrated paradigms and a broad clinical experience with each population. These
experiences will be inserted into Current Protocols in Magnetic Resonance Imaging as
they become pertinent.
The clinical scenario in which fMRI is requested by a referring physician is usually after
detection of a lesion in the brain of a patient for whom surgical intervention is being
contemplated. The location of eloquent cortex relative to the lesion allows a surgical
approach that minimizes loss of cognitive abilities or allows the patient and surgeon to
be appropriately prepared to decide on the risk/benefit ratio of the anticipated surgical
procedure.
Due to the field-strength dependence of both signal-to-noise performance and the mag-
netic susceptibility effect of blood oxygenation level dependent (BOLD) contrast, fMRI
is best performed on the highest-field strength magnet available (Turner et al., 1993;
Thulborn et al., 1996, 1997; Gati et al., 1997). The author’s clinical fMRI experience is
with both clinical 1.5 and 3.0 T systems (Thulborn et al., 1996, 1999a), although similar
results may be expected from other manufacturers if quality assurance (see UNIT A6.2) is
adhered to. Because many images are required over a large volume of the brain,
multi-planar, echo-planar imaging (Cohen and Weisskoff, 1991) offers a high duty cycle
with good image quality. Gradient-echo sequences provide the greatest sensitivity for
BOLD contrast and are used routinely (Bandettini et al., 1992). The echo-time (TE) is
matched to the uncorrected transverse relaxation time (T2*) in the region of the brain of
interest and at the field strength being used (∼50 msec at 1.5 T, 25 msec at 3.0 T). The
spatial resolution is dependent on scanner capabilities, but smaller voxels produce greater
BOLD contrast by diminishing partial volume effects (Thulborn et al., 1997). Generally,
echo-planar images provide voxel dimensions of 3 × 3 × 5 mm3 at 1.5 T and 3 × 3 × 3
mm3 at 3.0 T. The higher signal-to-noise ratio (SNR) at higher field strength allows smaller
voxels with the concomitant gain in BOLD contrast.
IMAGING LANGUAGE FUNCTION BASIC
PROTOCOL
The paradigm for language mapping can take many forms. A block design can be used in
which central fixation is compared to visually presented sentences. Each sentence consists
of a simple or conjoined sentence followed by a simple question that can be answered as
either “TRUE” or “FALSE” by pressing one of two finger switches held in either hand.
There are 5 pairs of sentences and questions per 30 sec block, allowing about 6 sec for
reading comprehension and answering. As a visual presentation is used, visual acuity
must be assured. This is done prior to the imaging session during the demonstration of
the task to the patient. Visual acuity issues are addressed with MR-compatible spectacles
(MIT Optical). Further concerns about performing the task while in the scanner can be
addressed in a scanner simulator (Rosenberg et al., 1997). This also screens for unexpected
claustrophobia in a more relaxed situation than in the scanner room, where the time
pressure of a busy clinical schedule is more apparent.
Clinical fMRI Clinical fMRI
A6.0.3 Contributed by Keith R. Thulborn and Denise Davis A6.1.1
Current Protocols in Magnetic Resonance Imaging (2001) A6.1.1-A6.1.9
Supplement 1 Copyright © 2001 by John Wiley & Sons, Inc.
 The description of the study presented to the patient by the nursing staff and MR
technologist during the screening interview and paradigm training is an essential compo-
nent of establishing rapport with the patient. Such rapport is important to obtain maximum
cooperation from the patient. The need for no head motion must be emphasized. This can
be aided by encouraging the patient not to talk unless there is a problem. In the same way,
during the study in the scanner, the technologist must not ask questions of the patient that
require responses, but rather provide information as statements about the progress of the
study. Not all patients are good candidates for fMRI, as can be judged from the training
session. Patients with difficulty maintaining stationary head position, possibly due to
tremors, respiratory difficulties, or pain, or who are not fully attentive, possibly due to
medications, should not be considered for this type of study. Other issues that may be
contraindications are various metal implants, of which dental braces in children are a good
example. These appliances do not represent a hazard to a conventional MR examination,
but rather seriously degrade the quality of gradient echo imaging due to magnetic
susceptibility effects that can decrease signal intensity across large regions of the brain.
Table A6.1.1 lists the hardware requirements for the BOLD contrast fMRI component of
a comprehensive MRI examination. The radiofrequency coil should be a high perform-
ance volume coil for whole brain imaging. The gradient set should be capable of
single-shot, echo-planar imaging. The typical gradient strength for high quality echo-pla-
nar imaging is on the order of 40 mT/m with slew rates greater than 150 mT/m/sec. In
some cases, there are commercially available portable gradient inserts that can be
retrofitted to a scanner to provide such performance, if it is not available from the scanner
gradients (Binder et al., 1994). The temporal stability of signal intensity should typically
be better than 1% over 30 min (Thulborn, 1999b). The synchronization control system
for paradigm presentation is also useful for simplifying the role of the technologist in
synchronizing the paradigm and scanner (Thulborn et al., 1996). The use of a visor helps
to minimize head motion (Thulborn, 1999c); however, lower-resolution echo planar
imaging (EPI) scans (as described herein) are viable with gradient strengths of 25 mT/m
and slew rates of 75 mT/m/sec.
NOTE: The eligibility of the patient for fMRI must be considered carefully, and the
clinical history must be available to provide an appropriate fMRI protocol. Direct
discussion with the referring physician is optimal to minimize failed studies due to poor
selection of candidate patients. The cooperation of the patient with cognitive tasks within
the scanner can sometimes be gauged from the history. Indications of patient motion
should be asked about, including respiratory difficulties, pain, and tremors. Medications
can also interfere with and decrease BOLD contrast. Few medications have been evaluated
for this potential interference but this should be considered in the interpretation (Gollub
et al., 1998). The use of one of the robust primary sensory or motor paradigms serves as
Table A6.1.1 Equipment for fMRI
Magnet ≥1.5 T
Coil type Volume transmit/receive head coil
Gradient coil strength and slew rate ∼40 mT/m, >150 mT/m/sec
Motion cushions Useful
Peripheral gating Unnecessary except for monitoring
Cardiac gating Unnecessary
Respiratory gating Unnecessary
Respirator Contraindication
Clinical Oxygen No
Applications
of fMRI Use of contrast agents No
A6.1.2
Current Protocols in Magnetic Resonance Imaging
the control for such interference. The referring physician can be helpful in emphasizing
to the patient the importance of the functional information for surgical planning.
1. Interview the patient to ensure that no contraindications for MRI are present.
Generally standard screening forms are used for all patients scanned in a magnetic
resonance system.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(1996) for discussion of what implants may be safely scanned using magnetic resonance.
If a family member is to be in the scan room with the patient, that person must also be
screened for contraindications. The presence of a parent can aid studies on a child, but the
parent must be instructed to minimize talking to the child during the study.
2. For research studies: Explain the entire procedure to the patient and obtain written
informed consent on an appropriate form approved by the Institutional Review Board.
3. Demonstrate the paradigm.
After explaining the difference between a fMRI and conventional MRI examination, the
technologist should demonstrate the paradigm to the patient in the interview room. Have
the patient perform the paradigm on a computer screen. The technologist should assess
the performance of the patient to ensure that the task is appropriate to the skills of the
patient. Assess visual acuity at this time, as this situation best matches the scanner distances
and screen luminosity.
4. Have the patient change into hospital-supplied clothing after removing all jewelry
and mascara. Encourage the use of the bathroom facilities at this time to avoid
unnecessary disruption of the scanning schedule.
5. Address any concerns about the scanner environment (e.g., acoustical noise, confine-
ment), preferably in a simulator, in a relaxed manner. Let the patient exert some
control over this process by asking questions. Demonstrate the use of ear plugs.
This process is particularly appropriate to gain the cooperation of pediatric patients who
are often more curious than anxious.
6. If this study is to be used for stereotaxic localization of a lesion with respect to surface
head anatomy within the next 24 hr, apply skin markers to the head.
Functional studies are usually done significantly prior to surgery, when such markers
would be inappropriate.
7. Take the patient to the scanner, and ask the patient to lie comfortably in supine position
on the MR table, positioning in the head radiofrequency coil, thereby replicating the
experience in the simulator. With ear plugs, respiratory monitoring, and peripheral
monitoring devices in place, stabilize the head with soft cushions, landmarked on the
nasion. Place the visor over the face of the patient. Establish the head position with
the sighting device of the visor. If stimulation goggles are being used, position the
stimulation device so that the patient is comfortable. If only a screen is being used,
move the patient into the bore of the magnet and install the screen within the magnet
bore or at the end of the patient table.
8. Place the emergency stop device in one of the patient’s hands to allow the patient to
immediately indicate any major concern during scanning; otherwise, tell the patient
to communicate with the technologist over the intercom system if necessary.
Clinical fMRI
A6.1.3
Current Protocols in Magnetic Resonance Imaging Supplement 7
 9. Adjust the visual stimulus presentation system for focus by projecting an eye chart.
Use the eye chart to center the projected stimulus on the screen.
This step requires verbal interaction with the subject.
IMPORTANT NOTE: Prior to starting the study, the patient should be asked if he or she
is comfortable, and then, if so, told to maintain the current position. Unnecessary talking
is to be discouraged due to increased head motion that results. During the fMRI study, the
technologist should avoid asking questions of the subject but rather make statements that
inform but do not require verbal responses from the patient (see above).
If the patient is unable to follow the procedure up to this stage, they may not be an
appropriate candidate for fMRI. The technologist should then proceed with the remainder
of the conventional MRI examination.
Sequence 1: Rapid sagittal scout view
10. Optional: Some very high-field scanners (3 T) now have first- and second-order shim
gradients with appropriate rapid shimming software to improve the magnetic field
homogeneity over the head of each patient. This can be used as a separate and initial
acquisition if available.
11. Use a sagittal scout view for graphic prescription of the functional imaging plane in
transverse, coronal, or oblique planes (Table A6.1.2). If higher-order shimming is not
available, this sequence also performs a first order autoshim to improve magnet
homogeneity for the specific patient under study.
Sequence 2: Gradient-echo, echo-planar fMRI
12. Perform gradient-echo, echo-planar fMRI (sequence 2; Table A6.1.3).
Ultra-fast imaging after automated shimming and reference scan for ghost correction form
the basis for the functional study. Acquisition parameters will vary with scanner and
operating system, but Table A6.1.3 shows typical parameters.
IMPORTANT NOTE: The beginning of the paradigm presentation and the image acqui-
sition must be synchronized. This is most easily achieved with a synchronization control
system (Thulborn et al., 1996).
The duration of the acquisition is based on experience with patients and paradigms. 10
min is the upper limit of patient tolerance before head motion and fatigue counterbalance
improved statistical power. The rule of thumb is that at least 60 images per paradigm
condition are required to reach statistically reliable results for the language paradigm
described above, in most cooperative patients. The images for this paradigm are best
performed in the transverse plane and should cover the vertex of the brain to the bottom
of the temporal lobes. As fMRI becomes more routine, the current restrictions of duty cycle
and total number of images that can be acquired will be removed by the scanner vendors.
Specific pulse sequences to be used for fMRI are undergoing rapid development and will
be quickly outdated. For these reasons, specific sequences will not be given.
Sequence 3: High-resolution anatomical images
13. Table A6.1.4 shows typical acquisition parameters for a high-resolution dataset
showing the surface topography.

Table A6.1.2 describes typical acquisition parameters. As functional activation maps are of low spatial resolution, a high-resolution dataset can
be useful to the neurosurgeon to plan a surgical approach to a lesion avoiding not only
This acquisition should be of sufficient quality for recognizing anatomic landmarks rather eloquent cortex, but also structures such as veins and arteries. Although there are well
than of minimal duration. known differences in image distortions between echo-planar and conventional images,
these distortions are quantifiable and should be considered in the surgical planning
(Cohen, 1999).
Table A6.1.2 Parameters for Sagittal Acout Image at 1.5 and 3.0 T
Table A6.1.3 Parameters for fMRI at 1.5 and 3.0 T
Patient position Supine
Scan type Spin echo Patient position Supine
Imaging plane (orientation) Sagittal Scan type Gradient echo echo-planar
Central slice or volume center Lateral canthus of the eye Imaging plane (orientation) Transverse
Echo time (TE) Minimum full (or 16 msec at Central slice or volume center Lateral canthus of the eye
1.5 T, 14 msec at 3.0 T) Echo time (TE) 50 msec at 1.5 T, 25–50 msec at
Repeat time (TR) 500 msec at 1.5 T, 500–700 msec 3.0 T
at 3.0 T Repeat time (TR) 3000 msec
Flip angle (FA) 90° Flip angle (FA) 90°
Field of view (FOVx, FOVy) 240 mm, 240 mm Field of view (FOVx, FOVy) 200 mm, 200 mm
Resolution (Δx, Δy) 0.94 mm, 1.25 mm Resolution (Δx, Δy) 3.1 mm, 3.1 mm
Number of data points collected (Nx, Ny) 256, 192 Number of data points collected (Nx, Ny) 64, 64 (ramp sampling)
Display matrix (Dx, Dy) 256, 256 Display matrix (Dx, Dy) 64, 64
Slice thickness (Δz) 5 mm Slice thickness (Δz) 5 mm at 1.5 T, 3 mm at 3.0 T
Number of slices 12 Number of slices 12
Slice gap 1 mm Slice gap 1 mm
Number of acquisitions (Nacq) 1 Number of acquisitions (Nacq) 1
Swap read and phase encoding No Swap read and phase encoding Yes
Slice locations L36–R36 Read direction Left–right
Saturation pulses None Slice locations Graphic prescription, 12 slices
Scan time 1 min, 26 sec at 1.5 T, 2 min, 14 Saturation pulses None
Clinical sec at 3.0 T Scan time 6 min for 6 cycles of 1 min each
Applications
of fMRI Clinical fMRI

A6.1.4 A6.1.5
Supplement 7 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 5
 Table A6.1.4 Parameters for High-Resolution Anatomic Images at 1.5 and 3.0 T tions from the lesion and readily identified 2. Patient cooperation in terms of motion can
topographic landmarks. The positions of large be verified by viewing the images from the
Patient position Supine veins that may complicate the interpretation of functional study in an animation loop. Rapid
Scan type 3-D Spoiled Gradient Echo BOLD contrast (especially at 1.5 T but less head movements of greater than one voxel are
Imaging plane (orientation) Transverse important at 3.0 T; Thulborn et al., 1999b) are readily apparent as are slow drifts in head po-
Central slice or volume center Lateral canthus of the eye useful to indicate to the physician. sition during the study. Head motion is most
Echo time (TE) Minimum (∼5 msec) As registration errors can occur due to pa- often the source of failure even in apparently
tient motion and image artifacts, an estimate of cooperative patients. This problem is reduced
Repeat time (TR) 25 msec
these potential errors should be included. by comfortable positioning, head support, and
Flip angle (FA) 40°
For lesions that may perturb the hemody- the visor (Thulborn, 1999c).
Fields of view (FOVx, FOVy) 240 mm, 240r mm, with r = 3/4 namic coupling of neuronal activity, care 3. Cooperation with performing the para-
(rectangular field of view)
should be taken to ensure that the BOLD con- digm is also essential. The language paradigm
Resolution (Δx, Δy) 0.94 mm, 1.25 mm trast remains valid up to the edge of the lesion. responses should be verified to have high accu-
Number of data points collected (Nx, Ny) 256, 192r, with r = 3/4 (rectangular Confidence in this assumption may be im- racy and normal response times. Low accuracy
field of view) proved by performing additional studies such and fast responses may indicate guessing (e.g.,
Display matrix (Dx, Dy) 256, 256 as diffusion and perfusion MR studies as de- incorrect strategy, low motivation, inappropri-
Slice thickness (Δz) 1.5 mm scribed in Chapter A1. In regions in which ate match to patient skills). The control tasks
Number of slices 124 diffusion or perfusion measurements are abnor- should show good activation maps as patient
Slice gap 0 mm mal, it is best to assume that BOLD contrast cooperation with the task is virtually assured
Number of acquisitions (Nacq) 1 will be less sensitive and that no interpretation by definition.
Swap read and phase encoding No of function can be made reliably. This has not 4. Determine the time between the patient
Read direction Anterior–posterior proven to be a significant problem in clinical entering the scanner and the attempt at the
Slice locations Graphic prescription, 124 slices settings of arteriovenous malformations, tu- failed fMRI study. Loss of alertness after pro-
Saturation pulses None mors, or even subacute stroke; however, one longed times in the scanner does not improve
recent case report emphasizes the concern for activation maps.
Scan time 7 min, 44 sec
this potential pitfall (Holodny et al., 1999). 5. Ensure that the technologist or nurse has
There is significant variation in the extent of informed the patient appropriately and in detail
This high-resolution study allows the anterior and posterior commissures to be identified activation between different individuals. This about the nature of the study and that the patient
and covers the entire brain so that Talairach coordinates (Talairach and Tournoux, 1988) is not surprising given the wide variation ob- has understood the instructions. A new tech-
can be used if requested. served with behavioral measurements of cog- nologist may not have developed the same skill
nition; however the patterns of activation pre- set in establishing patient rapport as experi-
IMPORTANT NOTE: Following an fMRI study, further sequences would be expected to
sumably reflecting the large-scale neurocogni- enced MRI staff. A formal educational program
involve the conventional imaging to fully characterize the nature of the lesion for which surgical
planning is being done. The identification of large blood vessels is useful to exclude artifacts tive networks involved in cognitive processing for the staff may be required.
from veins in activation maps. The acquisition of a cerebral venogram can be useful for this seem to be reproducible (Mesulam, 1990, 6. Data management and analysis can some-
purpose. If contrast is to be given to characterize the lesion, this represents an opportunity to 1998). Knowledge of the nature of this network times lead to errors, especially if the procedures
perform a perfusion study to establish the regions of normal perfusion in which BOLD contrast in an appropriately matched control population are not used routinely. Practice and experience
can be reliably interpreted. Some perfusion studies are described in Chapter A1. is helpful for clear interpretation of activation result in efficiently performed studies with op-
maps in individual patients. timal results. A trial data set can be used to train
new or recalibrate previous operator skills
COMMENTARY Critical Parameters and when necessary.
Background Information paradigm as used for the language paradigm Troubleshooting
Functional imaging to map eloquent cortex discussed above. Otherwise, the data must be When an activation study fails to yield the Anticipated Results
is still in its infancy for clinical applications. moved off the scanner into a computer environ- expected results, a systematic approach to un- An activation map in which the areas of
The clinical requirement for a high success rate ment in which other public domain or commer- derstanding the source of the problem can be significant BOLD contrast are displayed in
on individual patients with a comprehensible cial software can be used to process the data. informative and help improve subsequent stud- color over the detailed anatomic images includ-
timely report is a challenge for any neuroradi- Many statistical methods have been discussed ies. If all the images have been obtained and ing the lesion of concern is the desirable result.
ology service. Success depends on careful qual- (Lange, 1999) and some have been made avail- used in the analysis without a satisfactory result Images through the entire brain would be best
ity assurance at all stages including patient able as software packages (Gold et al., 1998). then the following steps may be helpful. to show all anatomic landmarks for the surgeon.
preparation, paradigm selection, data acquisi- These packages usually allow the activation 1. Daily quality assurance (QA) should be The addition of markers if the patient is going
tion, and image analysis. Quality assurance and maps colored by statistical significance to be verified to ensure that scanner performance and immediately to the operation would help
paradigm design are treated in detail as separate superimposed over the high-resolution concomitant image quality has not degraded. If anatomic localization. This implies that the
units. anatomic images (Cox, 1996). image quality is below specifications, a service analysis is available within minutes and that the
The statistical processing of fMRI data can For clinical decisions based on fMRI results, call may be required. This is usually easily data can be transmitted immediately to the
be performed on some clinical scanners. In the areas of activation must be reported with checked and the least likely source of any prob- operating room for use on stereotaxically
some cases, the activation map can be observed respect to the location of the lesion. The Talai- lems. The patient study should not be done until guided surgical instrumentation. With well
Clinical
Applications in real time during the acquisition. The proc- rach coordinates for areas of activation are less daily QA has been verified. characterized paradigms and experience with
of fMRI essing is a simple t-test for a two-condition useful to the surgeon than distances and direc- Clinical fMRI

A6.1.6 A6.1.7
Supplement 5 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 MRI. (C. Moonen and P. Bandettini., eds.) pp.
137-148. Springer-Verlag, Berlin.
Cohen, M.S. and Weisskoff, R.M. 1991. Ultrafast
imaging. Magn. Reson. Imaging 9:1-37.
Cox, R.W. 1996. AFNI, software for analysis and
visualization of functional magnetic resonance
neuroimages. Comput. Biomed. Res. 29:162-
173.
Gati, J.S., Menon, R.S., Ugurbil, K., and Rutt, B.K.
1997. Experimental determination of the BOLD
field strength dependence in vessels and tissue.
Magn. Reson. Med. 38:296-302.
Gold, S., Christian, B., Arndt, S., Zeien, G., Cizadlo,
T., Johnson, D.L., Fiaum, M., and Andreasen,
N.C. 1998. Functional MRI statistical software
packages: A comparative analysis. Hum. Brain
Mapp. 6:73-84.
Gollub, R.L., Breiter, H.C., Kantor, H., Kennedy, D.,
Gastfriend, D., Mathew, R.T., Makris, N., Gui-
maraes, A., Riorden, J., Campbell, T., Foley, M.,
Hyman, S., Rosen, B., and Weisskoff, R.. 1998.
Cocaine decreases cortical cerebral blood flow
but does not obscure regional activation in func-
tional magnetic resonance imaging in human
subjects. J. Cereb. Blood. Flow Metab. 18:724-
734.
Holodny, A.I., Schulder, M., Lui, W.C., Maldjian,
J.A., and Kalnin, A.J. 1999. Decreased BOLD
functional MR activation of the motor and sen-
sory cortices adjacent to a glioblastoma multi-
forme: Implications for image-guided neurosur-
gery. Am. J. Neuroradiol. 20:609-612.
Lange, N. 1999. Statistical procedures for functional
MRI. In Medical Radiology: Diagnostic Imag-
ing and Radiation Oncology, Functional MRI.
(C. Moonen and P. Bandettini., eds.) pp. 301-
Figure A6.1.1 Representative activation map at 3.0 T for language paradigm, showing exten- 335. Springer-Verlag, Berlin.
sive network activation with areas of frontal eye fields (FEF, precentral sulcus), supplementary Lee, C.C., Ward, H.A., Sharbrough, F.S., Meyer,
eye fields (SEF, medial frontal cortex), prefrontal cortex (PFC), intraparietal sulcus (IPS), Broca’s F.B., March, W.R., Raffel, C., So, E.L., Cascino,
area (BA, left inferior frontal cortex), Wernicke’s area (WA, left superior temporal gyrus), and G.D., Shin, C., Xu, Y., Riederer, S.J., and Jack,
visual cortex (calcarine sulcus). This pattern is observed in adults irrespective of gender or C.R. 1999. Assessment of functional MR imag-
handedness. The right inferior frontal cortex also shows activation in the contralateral location ing in neurosurgical planning. Am. J. Neurora-
to Broca’s area as can sometimes be observed with Wernicke’s area. The color scale indicates diol. 20:1511-1519.
the t-test statistic for activation. The higher t-statistic (yellow) has greater statistical significance. Mesulam, M.-M. 1990. Large-scale neurocognitive
See color plate. networks and distributed processing for atten-
tion, language and memory. Annu. Neurol.
28:597-613.
interpretation of such data, these results are Binder, J.R., Rao, S.M., Hammeke, T.A., Yetkin, Mesulam, M.-M. 1998. From sensation to cognition.
achievable at a high success rate with appropri- F.Z., Jesmanowicz, A., Bandettini, P.A., Wong, Brain 121:1013-1052.
E.C., Estowski, L.D., Goldstein, M.D.,
ately chosen patients. The analysis of perfusion Rosenberg, D.R., Sweeney, J.A., Gillen, J.S.,
Haughton, V.M., and Hyde, J.S. 1994. Functional
and diffusion data should be included in the magnetic resonance imaging of human auditory Chang, S.Y., Varanelli, M.J., O’Hearn, K., Erb,
report to indicate in which areas of the brain a cortex. Annu. Neurol. 35:662-672. P.A., Davis, D., and Thulborn, K.R. 1997. Mag-
BOLD contrast interpretation can be confi- netic resonance imaging of children without se-
Booth, J.R., MacWhinney, B., Thulborn, K.R., dation: Preparation with simulation. J. Am. Acad.
dently rendered. Images for a normal individual Sacco, K., Voyvodic, J., and Feldman, H.M.
Child Adolesc. Psychiatry 36:853-59
performing the language paradigm are shown 1999. Functional organization of activation pat-
terns in children: Whole brain fMRI imaging Shellock, F.G. 1996. Pocket Guide to MR Proce-
in Figure 6.1.1.
during three different cognitive tasks. Prog. dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Neuro-Psychopharmacol. Biol. Psychiat.
Literature Cited 23:669-682.
Bandettini, P.A., Wong, E.C., Hinks, R.S., Tikofsky,
R.S., and Hyde, J.S. 1992. Time course EPI of Cohen, M.S. 1999. Echo-planar imaging and func-
Clinical human brain function during task activation. tional MRI. In Medical Radiology: Diagnostic
Applications Magn. Reson. Med. 25:390-397. Imaging and Radiation Oncology. Functional
of fMRI
A6.1.8
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
Talairach, J. and Tournoux, P. 1988. Co-planar
stereotaxic atlas of the human brain 3D propor-
tional system: An approach to cerebral imaging.
(G. Thieme, ed.) Springer-Verlag, New York.
Thulborn, K.R. 1999a. Clinical rationale for very
high field (3.0 Tesla) functional MR imaging.
Topics Magn. Reson. Imaging 10:37-50.
Thulborn, K.R. 1999b. Quality assurance in clinical
and research echo-planar functional MRI. In
Medical Radiology: Diagnostic Imaging and Ra-
diation Oncology, Functional MRI. (C. Moonen
and P. Bandettini, eds.) pp. 337-345. Springer-
Verlag, Berlin.
Thulborn, K.R. 1999c. Visual feedback to stabilize
head position for fMRI. Magn. Reson. Med.
41:1039-1043.
Thulborn, K.R., Davis, D., Erb, P., Strojwas, M., and
Sweeney, J.A. 1996. Clinical fMRI: Implemen-
tation and experience. NeuroImage 4:S101-
S107.
Thulborn, K.R., Chang, S.Y., Shen, G.X., and Voy-
vodic, J.T. 1997. High resolution echo-planar
fMRI of human visual cortex at 3.0 Tesla. NMR
Biomed. 10:183-90.
Thulborn, K.R., Carpenter, P.A., and Just, M.A.
1999a. Plasticity of language-related brain func-
tion during recovery from stroke. Stroke 30:749-
754.
Thulborn, K.R., Gindin, T., and Talagala, S.L.
1999b. High-resolution echo-planar fMRI meas-
ures localized cortical microvascular responses,
not just large venous drainage patterns. In Pro-
ceedings Seventh Scientific Meeting Interna-
tional Society of Magnetic Resonance in Medi-
cine. Philadelphia, PA, May 27.
Thulborn, K.R., Martin, C., and Voyvodic, J. 2000.
fMRI using a visually guided saccade paradigm
in Alzheimer’s disease. Am. J. Neuroradiol.
21:524-531.
Turner, R., Jezzard, P., Wen, H., Kwong, K.K., Bi-
han, D.L., Zeffiro, T. and Balaban, R.S. 1993.
Functional mapping of the human visual cortex
at 4 and 1.5 Tesla using deoxygenation contrast
EPI. Magn. Reson. Med. 29:277-279.
Key Reference
Moonen, C. and Bandettini, P. (eds.) 1999. Medical
Radiology: Diagnostic Imaging and Radiation
Oncology, Functional MRI. Springer-Verlag,
Berlin.
Comprehensive book on all issues related to fMRI
Contributed by Keith R. Thulborn and
Denise Davis
University of Illinois at Chicago
Chicago, Illinois
Clinical fMRI
A6.1.9
Quality Assurance for Clinical fMRI UNIT A6.2 Table A6.2.2 Stability of MR Signal Intensity for Echo-Planar Imaging at 1.5
and 3.0 T

The functional MRI (fMRI) procedure has several sources of variance that determine the Patient position NA
success of the examination. These include the scanner, patient, and paradigm. A full Scan type Gradient echo echo-planar
discussion of these items has been given elsewhere (Thulborn, 1999; Thulborn and Imaging plane (orientation) Transverse
Gisbert, 2000). As blood oxygenation level dependent (BOLD) contrast is a small effect, Central slice or volume center Center
high signal-to-noise performance is mandatory. Because the preparation of a functional Echo time (TE) 50 msec at 1.5 T, 25 msec at 3.0 T
activation map requires averaging multiple images over time, the scanner must produce Repeat time (TR) 2000 msec
high temporal stability of the signal intensity. There are many determinants of such Flip angle (FA) 90°
performance but not all possibilities need to be checked separately. An adequate approach Fields of view (FOVx, FOVy) 200 mm, 200 mm
has been to verify total system performance under the conditions of a functional MRI Resolution (Δx, Δy) 3.1 mm, 3.1 mm
study on a phantom. This testing is done daily prior to patient studies. Number of data points collected (Nx, Ny) 64, 64 (ramp sampling)
Display matrix (Dx, Dy) 64, 64
SCANNER STABILITY BASIC Slice thickness (Δz) 5 mm at 1.5 T, 3 mm at 3 T
PROTOCOL Number of slices 10 (521 images total)
The scanner must be able to produce many high-quality, temporally-stable images Slice gap 12.5 mm
continuously over many minutes. Daily assessment of stability and image quality can Number of acquisitions (Nacq) 1
ensure that fMRI studies are acquired under optimal conditions. Swap read and phase encoding No
Table A6.2.1 lists the hardware requirements for the BOLD contrast fMRI component of Slice locations Graphic prescription, 10 slices
a comprehensive MRI examination. The radiofrequency coil should be a high perform- Saturation pulses None
ance volume coil for whole brain imaging. The gradient set should be capable of Ramp sampling On
single-shot, echo-planar imaging. The typical gradient strength for high quality echo-pla- Scan time 17 min, 12 sec
nar imaging is on the order of 25 to 40 mT/m with slew rates of 75 to 150 mT/m/sec or
better. The limit of 512 images is a current limit on file data length on one scanner for a standard
echo-planar sequence and represents a convenient maximum acquisition length. This study
Materials gives day to day measures of signal stability, signal-to-noise performance, and ghosting
Spherical phantom in loader with doped water, 16-cm diameter levels.
3. Compare these measurements with baseline values obtained at the time of acceptance
Equipment setup
from the manufacturer. Deviations from baseline indicate the need for a service call.
1. Place the phantom on the scanner table. Adjust the table such that the laser light is at
the center of the phantom. Turn off the laser light and move the phantom to the center The reliability of fMRI results from a scanner showing fluctuations in these quality
of the magnet. assurance parameters cannot be assured.

Quality assurance sequence COMMENTARY

2. Perform scans using the protocol listed in Table A6.2.2.
Background Information be maintained. A scanner installed in an ill-pre-
There are many factors that determine over- pared site can be severely compromised. Hav-
Table A6.2.1 Equipment for fMRI all scanner performance. It is important to en- ing accepted a well installed scanner, perform-
sure that the initial installation has been done ance may degrade slowly over time. This can
Magnet 1.5 T or greater correctly and that manufacturer specifications happen for many reasons and it is important to
Coil type Volume transmit/receive head coil have been met. A complete report on the instal- separate scanner from sitting issues.
Gradient Set 25–40 mT/m, 75–150 mT/m/sec lation should be forthcoming from the vendor.
Motion cushions Useful This report will include magnet homogeneity Critical Parameters and
Peripheral gating Unnecessary except for monitoring specified in Hertz over different diameters of a Troubleshooting
Cardiac gating Unnecessary spherical volume. It will include radio fre- Many sources of potential problems reflect
Respiratory gating Unnecessary quency (RF) testing of the room at all potential the site rather than the scanner. This can be
Respirator Contraindication problem sources including the door, window, suggested as noise floor increases without
and penetration panel. These measurements changes in signal intensity. The RF seals on the
Oxygen No
should typically be better than 100 dB at the door and penetration panel are likely sources.
Use of contrast agents No
operating frequency of the scanner. Signal-to- Sometimes electrical circuits within the magnet
noise ratio, ghost-to-signal measurement, and room can become noise sources. Care should
Quality a stability performance, such as with this pro- be taken to never breach the wave guides with
Assurance for tocol, should be established close to acceptance conducting materials. Wire used to provide
Clinical fMRI Clinical fMRI of the scanner so that a performance history can electrical connections through the wave guides
Contributed by Keith R. Thulborn and Denise Davis A6.2.1 A6.2.2
Current Protocols in Magnetic Resonance Imaging (2001) A6.2.1-A6.2.4
Copyright © 2001 by John Wiley & Sons, Inc. Current Protocols in Magnetic Resonance Imaging
 for accessory equipment used by some investi-
101.0
gators in fMRI experiments are a notorious
source of this problem.
RF doors have variable quality and longev-
ity. Some pneumatically sealed doors have been
100.5 shown to maintain excellent performance for
years while others have degraded in months.
Careful maintenance of the site is a necessary
element of quality assurance.
100.0 The scanner is not without its own issues.
An increasing power requirement for determin-
ing the 90° nutation (flip) angle for the head RF
coil suggests a coil problem. A decreased signal
99.5
without a change in the noise figure or power
should be less than 1% over the 17 min acqui-
sition. Although not an absolute value, the sig-
nal-to-noise ratio for this scanner and method
of measurement is over 330. Similarly, the
ghosting level, reflecting intrinsic gradient per-
formance and eddy current correction, is under
2%. These parameters are useful for overall
assessment of the scanner electronics, gradient
performance, RF noise isolation of the scanner
room, and magnet homogeneity.
Literature Cited
Thulborn, K.R. 1999. Quality assurance in clinical
and research echo-planar functional MRI. In

Normalized signal intensity (%)

requirement may point to a preamplifier prob-
lem. Low signal also comes from poor magnet
99.0 homogeneity which can be caused by a simple
600 paper clip inadvertently flying into the magnet.
Many of the details of troubleshooting are best
left to the field service engineer; however,
someone has to be aware of the problem in order
500
to contact service, hopefully before patient care
is compromised. Daily quality assurance en-
sures timely response to potential problems.
400
Anticipated Results
Typical results are shown in Figure A6.2.1.
Typically, the peak-to-peak signal variation
300
Medical Radiology: Diagnostic Imaging and Ra-
diation Oncology. Functional MRI (C. Moonen
and P. Bandettini, eds.) pp. 337-345. Springer-
Verlag, Berlin.
Thulborn, K.R. and Gisbert, A. 2000. Clinical Ap-
plications of Mapping Neurocognitive Processes
in the Human Brain with Functional MRI. (P.
Matthews and P. Jezzard, eds.) Oxford Univer-
sity Press, Oxford. In press.
Contributed by Keith R. Thulborn and
Denise Davis
University of Illinois at Chicago
Chicago, Illinois

Signal-to-noise ratio (SNR)

200
2.5

2.0

1.5

Ghosting (%)
1.0

0.5

0.0
1 101 201 301 401 501

Image no.

Figure A6.2.1 Representative results of the quality assurance protocol for one day from a 1.5 T
scanner. (A) Normalized signal intensity over 512 images showing a peak-to-peak variation of less Quality
than 1%. (B) Signal-to-noise ratio over 512 images showing a mean value above 300. (C) Assurance for
Percentage ghosting over 512 images showing a mean value of less than 2%. Clinical fMRI Clinical fMRI

A6.2.3 A6.2.4
Current Protocols in Magnetic Resonance Imaging Supplement 7 Supplement 7 Current Protocols in Magnetic Resonance Imaging
Paradigms for Clinical fMRI
The block design, in which several cycles
of two cognitive different states are compared,
is the most useful form of cognitive paradigms
used clinically (Marquart et al., 2000). The
event-related design (Rosen et al., 1998; Rich-
ter, 1999), in which many cycles of short du-
ration stimuli (∼1 sec) are presented with
longer rest periods (∼15 sec), can be used if
the appropriate synchronization control sys-
tem and analysis software are available. The
single-event design has the advantage of al-
lowing each event to be examined and grouped
according to the specific behavioral response
to each event; however the transient blood
oxygenation level dependent (BOLD) effect
for each event is small, thereby requiring ex-
tensive image averaging to reach BOLD detec-
tion thresholds. Given the need for robust re-
sults in the shortest duration testing, the two-
condition block design is the simplest to
implement on clinical machines.
Given the neurosurgical concern with pres-
ervation of eloquent cortex, brain functions of
particular interest are primary sensory (e.g.,
visual, auditory) and motor functions, and high
level processing of language comprehension
and expression. Compromise of these func-
tions is usually readily apparent clinically and
the quality of life of the patient is severely
diminished. Functional MRI offers a means to
locate these functions, thereby allowing the
surgeon to plan on preserving these functions,
or prepare the patient appropriately. Only four
paradigms will be described as other para-
digms can be derived from the same principles
(Thulborn, 1999).
PARADIGMS
Basic Clinical Paradigm 1: Language
Comprehension
Various modifications of this task have been
described elsewhere (Just et al., 1996; Thul-
born et al., 1999a). A central fixation condition
(30 sec) is compared to a reading condition (30
sec) in which a simple sentence is read, fol-
lowed by answering a question as TRUE or
FALSE by depressing a finger switch. The
record of the answers provide a behavioral
measure of task performance (response time
and accuracy), which is useful to verify that
the task is being completed as requested.
Contributed by Keith R. Thulborn and Denise Davis
Current Protocols in Magnetic Resonance Imaging (2001) A6.3.1-A6.3.6
Copyright © 2001 by John Wiley & Sons, Inc.
UNIT A6.3
Basic Clinical Paradigm 2: Eye
Movement (Visually Guided Saccades)
This group of tasks has been described in
detail elsewhere (Luna et al., 1998; Luna and
Sweeney, 1999). The central fixation condition
(30 sec) is compared to a condition (30 sec) in
which the patient is asked to make saccadic eye
movements to a spot of light randomly appear-
ing at one of five locations spanning about 20°
along the horizontal meridian.
Basic Clinical Paradigm 3: Motor
Cortex (Finger-Thumb Apposition)
This task compares a finger-thumb apposi-
tion condition (30 sec) to a rest condition (30
sec). The pattern of finger thumb apposition can
be made unilateral or bilateral and vary from
very simple (index finger to thumb) to more
complex patterns. This is a robust activation
response usually requiring only a couple of
cycles in a cooperative patient. Even motivated
patients with compromised function and lim-
ited movement capability can produce useful
maps. The use of other body parts such as arms,
legs and feet tend to increase head motion
unless restricted in extent.
Basic Clinical Paradigm 4: Visual Figure A6.3.1 Representative activation map at 1.5 T for the language paradigm,
Cortex showing extensive network activation with areas of frontal eye fields (FEF, precentral
Primary and association visual cortex can be sulcus), supplementary eye fields (SEF, medial frontal cortex), intraparietal sulcus (IPS),
robustly activated by comparing central fixa- prefrontal cortex (PFC), Broca’s area (BA, left inferior frontal cortex), homologous Broca’s
tion in a dark field condition (30 sec) with a area (hBA, right inferior frontal cortex), Wernicke’s area (WA, left superior temporal gyrus),
flashing checkerboard condition (black and and visual cortex (V1, calcarine sulcus). This pattern is observed in adults irrespective of
gender or handedness. The color scale indicates the t-test statistic for activation. The
white, 8 Hz, 30 sec). The checkerboard stimu-
higher t-statistic (yellow) has greater statistical significance. See color plate.
lus can be presented as whole visual field,
alternating left-right hemifields or combina-
tions of upper-lower and left-right quadrant with perfusion and diffusion MRI studies. A No differences across gender or age parameters
fields. The robustness of this stimulus mini- provocative test of hemodynamic reserve could have been observed by the authors. This does
mizes the duration of these studies to a few be used such as with vasodilators (i.e., carbon not exclude the possibility that more detailed
minutes. dioxide or acetazolamide). Until there is greater studies of larger groups may reveal subtle vari-
clinical experience, such tests should be used ations in the future.
when presurgical mapping is contemplated.
COMMENTARY
Such cases usually require intravenous contrast Critical Parameters and
enhancement to characterize the lesion so that Troubleshooting
Background Information
The interpretation of activation maps from a quantitative perfusion study can be made with When the clinical question relates to map-
single patients in the setting of disease has some the same contrast bolus (Thulborn et al., ping language and the expected activation map
uncertainty. Test-retest data is usually not avail- 1999b). has not been derived, there is a logical set of
able. The disease process and medications used Having established the regions of the brain steps to take to understand the problem. Head
to treat it may disturb the normal hemodynamic in which an interpretation can be rendered, each motion is typically the problem, which can
coupling to local neuronal activity. A conserva- paradigm requires a pattern of activation in a easily be checked by viewing the images in an
tive interpretation of activation maps is applied matched normal population. The activation pat- animation loop on a computer screen. Alterna-
only to the regions of established normal physi- terns in normal groups of adults presented in tively, head motion can be quantified using
ology. These regions can be defined readily the Figures A6.3.1 to A6.3.4, which are pro- software with coregistration algorithms. Head
Paradigms for vided to give a basis for such interpretations. motion will increase in any patient who is
Clinical fMRI Clinical fMRI
A6.3.1 A6.3.2
Current Protocols in Magnetic Resonance Imaging
 Figure A6.3.2 Representative activation map at 1.5 T for the visually guided saccade
paradigm which is used as a control for the language paradigm. The areas of activation
(FEF, SEF, IPS, V1, MT/V5) are similar except for the absence of Broca’s (BA) and
Wernicke’s areas (WA), involved in language comprehension. This pattern is observed in
Figure A6.3.3 Representative activation map at 1.5 T for the motor paradigm with
both young and elderly adults. See color plate.
bilateral finger-thumb apposition in which motor (M1), supplementary motor area
(SMA), and somatosensory (S1) activation are observed. This pattern has been
uncomfortable in the scanner or who does not This may relate to medications or severe cere- reported in normal adults from multiple laboratories. See color plate.
understand the requirement for immobility dur- brovascular disease with compromised
ing scanning. This can be considered a failure hemodynamic reserve. In some circumstances, and IPS areas are activated in eye movement A representative activation map for the VGS
of patient preparation. An active educational a vasodilation perfusion study could be used to paradigms and appear to be related to eye move- paradigm at 1.5 T is shown in Figure A6.3.2.
process with the staff is important to build skills verify this possibility, if there is a clear medical ment during reading. This paradigm is particu- This trivial task can be performed by every
for performing fMRI on patients. If head mo- indication for this study. larly useful as it probes frontal, parietal, tem- cooperative patient and even demented patients
tion is not the problem, patient cooperation can poral, and occipital lobes in a single study. The (Thulborn et al., 2000). It produces robust ac-
be verified by examining the behavioral re- Anticipated Results duration of the paradigm can be less than 10 tivation in similar regions to the language para-
sponses. If there is a poor success rate for A representative activation map for the lan- min. For children or adults in whom reading digm except for Wernicke’s (language compre-
answers to questions on the simple sentences guage paradigm at 1.5 T is shown in Figure skills may be deficient, an auditory presentation hension) and Broca’s (language expression)
to be read, then the patient may have closed A6.3.1. Activation is observed in Wernicke’s has been used successfully (Binder et al., 1994; areas. This paradigm is a useful control study
their eyes or may have a visual acuity problem area (superior temporal gyrus, usually left Wessinger et al., 1997; Booth et al., 1999). for language mapping to ensure that the BOLD
that required correction with MR-compatible hemisphere and often bilateral), Broca’s area Because of the loud noise of the scanner, audi- contrast sensitivity is sufficient for the patient
spectacles. Quality assurance (UNIT A6.2) on the (inferior frontal lobe, usually left hemisphere tory presentation requires careful thought about of interest.
scanner can be checked, but if this is the prob- and often bilateral), frontal eye fields (FEF, use of MR-compatible earphones with ear A representative activation map for the mo-
lem the study should not have been performed. precentral sulcus bilaterally), supplementary plugs. Ear lobe compression by earmuffs can tor paradigm at 1.5 T is shown in Figure A6.3.3.
Examination of the visually-guided saccade eye fields (SEF, medial frontal cortex bilater- produce discomfort and increase motion. The activation patterns show both primary mo-
(VGS) paradigm results may indicate that the ally), intraparietal sulcus (IPS, bilaterally) and Paradigms for tor and somatosensory cortex activation along
patient may produce weak BOLD contrast. visual cortex (V1) bilaterally. The FEF, SEF, Clinical fMRI Clinical fMRI

A6.3.3 A6.3.4
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging

A the medial frontal cortex is also activated.
B 23:669-682.
Figure A6.3.4 Representative activation map at 1.5 T for the visual
paradigm in which a flashing checkerboard stimulus is shown to the (A)
left upper quadrant and (B) right upper quadrant. The clear contralateral
activation in the calcarine fissure indicates that the subject did maintain
central fixation for the entire task. Bilateral stimulation indicates lack of
central fixation. V5 (also termed MT) is the area of visual association
cortex involved in detection of visual motion. IPS and V1 are the intra-
parietal sulcus and primary visual cortex, respectively. See color plate.
Current Protocols in Magnetic Resonance Imaging
the precentral gyrus and postcentral gyrus, re-
spectively. Supplementary motor cortex along
Representative activation maps for the vis-
ual paradigm at 1.5 T are shown for visual
stimulation localized to the left and right upper
quadrant fields in Figure A6.3.4A and A6.3.4B,
respectively. The activated visual cortex along
the inferior bank of calcarine sulcus in the
occipital lobe is activated in the contralateral
hemisphere. Activation is also present in the
intraparietal sulcus and in the visual motion
area of MT/V5.
LITERATURE CITED
Binder, J.R., Rao, S.M., Hammeke, T.A., Yetkin,
F.Z., Jesmanowicz, A., Bandettini, P.A., Wong,
E.C., Estowski, L.D., Goldstein, M.D.,
Haughton, V.M., and Hyde, J.S. 1994. Functional
magnetic resonance imaging of human auditory
cortex. Annu. Neurol. 35:662-672.
Booth, J.R., MacWhinney, B., Thulborn, K.R.,
Sacco, K., Voyvodic, J., and Feldman, H.M.
1999. Functional organization of activation pat-
terns in children: Whole brain fMRI imaging
during three different cognitive tasks. Prog.
Neuro-Psychopharmacol. Biol. Psychiatry
Just, M.A., Carpenter, P.A., Keller, T.A., Eddy, W.F.,
and Thulborn, K.R. 1996. Brain activation
modulated by sentence comprehension. Science
274:114-116.
Luna, B., Thulborn, K.R., Strojwas, M.H., McCur-
tain, B.J., Berman, R.A., Genovese, C.R., and
Sweeney, J.A. 1998. Dorsal cortical regions sub-
serving visually-guided saccades in humans: A
fMRI study. Cereb. Cortex 8:40-47.
Luna, B. and Sweeney, J.A. 1999. Cognitive func-
tional magnetic resonance imaging at very-high-
Paradigms for
Clinical fMRI Clinical fMRI
A6.3.5 A6.3.6
field: Eye movement control. Topics Magn.
Reson. Imaging 10:3-15.
Marquart, M., Birn, R., and Haughton, V. 2000.
Single- and multiple-event paradigms for identi-
fication of motor cortex activation. Am. J.
Neuroradiol. 21:94-98.
Richter, W. 1999. High temporal resolution func-
tional magnetic resonance imaging at very-high-
field. Topics Magn. Reson. Imaging 10:51-62.
Rosen, B.R., Buckner, R.L., and Dale, A.M. 1998.
Event-related functional MRI: Past present and
future. Proc. Natl. Acad. Sci. U.S.A. 95:773-780.
Thulborn, K.R. 1999. Clinical rationale for very
high field (3.0 Tesla) functional MR imaging.
Topics Magn. Reson. Imaging 10:37-50.
Thulborn, K.R., Carpenter, P.A., and Just, M.A.
1999a. Plasticity of language-related brain func-
tion during recovery from stroke. Stroke 30:749-
754.
Thulborn, K.R., Gindin, T.S., Davis, D., and Erb, P.
1999b. Comprehensive MRI protocol for stroke
management: Tissue sodium concentration as a
measure of tissue viability in a non-human pri-
mate model and clinical studies. Radiology
139:26-34.
Thulborn, K.R., Martin, C., and Voyvodic, J. 2000.
fMRI using a visually guided saccade paradigm
in Alzheimer’s disease. Am. J. Neuroradiol.
21:524-531.
Wessinger, C.M., Buonocore, M.H., Kussmaul,
C.L., and Mangun, G.R. 1997. Tonotopy in hu-
man auditory cortex examined with functional
magnetic resonance imaging. Human Brain
Mapp. 5:18-25.
Contributed by Keith R. Thulborn and
Denise Davis
University of Illinois at Chicago
Chicago, Illinois
Current Protocols in Magnetic Resonance Imaging
Diffusion Tensor Imaging UNIT A6.4
Diffusion tensor imaging (DTI) can provide a wealth of information on water molecular BASIC
diffusion in biological tissues. Typical parameters that can be calculated from DTI include PROTOCOL
apparent diffusion coefficients (ADC), diffusion anisotropy indices, and principal diffu-
sion directions. These parameters can be used to reveal tissue structures, such as white-
matter fiber tracts in the human brain, and monitor their changes during disease
progression and regression. Presently, clinical applications of DTI are focused primarily
on the central nervous system (Pierpaoli et al., 1996), to study white-matter diseases,
neonate brain development, brain tumors, psychiatric disorders, and surgical planning of
brain lesions (Melhem et al., 2002). With the ongoing research and development, the scope
of DTI applications is expected to expand in the future.
Clinical DTI is typically requested by neurologists, oncologists, and neurosurgeons to
evaluate the involvement of white-matter tracts in neurological disorders, such as multiple
sclerosis, adrenoleukodystrophy, Krabbe’s disease, and brain neoplasms. Diffusion tensor
imaging can assist neurosurgeons in identifying the location of the white-matter fiber
tracts in patients with brain tumors, prior to surgical resection (Zhou et al., 2000). With
this knowledge, optimal surgical strategies can be designed to minimize risk while
maximizing the benefit of the surgical procedure. Following surgery, clinicians can also
use DTI to help evaluate the impact of the surgical procedure on fiber tracts and monitor
fiber tract reorganization during recovery.
Typically, DTI is added onto the conventional MRI examination. This add-on sequence
is performed prior to administration of the contrast agent in order to avoid the T2-short-
ening effect that may be introduced by the contrast agent. Although many pulse sequences
have been adapted to DTI, single-shot diffusion-weighted echo planar imaging (EPI) is
most popular for brain clinical applications due to its ultrafast speed and motion insensi-
tivity. Diffusion tensor sequences based on EPI are essentially the same as diffusion-
weighted EPI sequences, except that they are capable of applying the diffusion-weighting
gradient in six or more non-collinear directions (Basser et al., 1994; Basser and Pierpaoli,
1998). Single-shot EPI-based DTI pulse sequences are susceptible to artifacts caused by
tissue magnetic susceptibility variations and system imperfections, such as eddy currents,
gradient-induced vibration, and field inhomogeneities. While the susceptibility variations
are inherent to the imaged object, system imperfections can often be reduced through
calibrations and compensations.
Unlike conventional images (e.g., T1-weighted and T2-weighted images) that can be
viewed directly for diagnosis, original diffusion-weighted images for DTI must be
processed further to yield DTI maps prior to interpretation. Typically, the processed DTI
data are represented as a gray-scale scalar map based on a diffusion anisotropy index (e.g.,
fractional anisotropy or relative anisotropy; see Commentary) to assess the size, density,
and location of the fiber tracts. Color display can also be used to enhance image
visualization. For example, fibers in the left/right, anterior/posterior, and superior/inferior
directions can be color-coded with red, green, and blue, respectively, with fibers along
off-axis orientations represented by the mixture of the three colors (Pajevic and Pierpaoli,
1999). The intensity of each color-coded pixel denotes the amplitude of the selected
diffusion anisotropy index. A DTI data set can also be displayed as a vector map to
illustrate the direction of principal diffusion at each pixel. The vector maps are particularly
useful to track fibers in three dimensions (Xue et al., 1999; Basser et al., 2000).
Fiber-tracking techniques based on DTI vector maps are sometimes referred to as
tractography. Presently, they are predominantly employed in research applications.
Diffusion Tensor
Clinical fMRI
Contributed by Xiaohong Joe Zhou and Keith R. Thulborn A6.4.1
Current Protocols in Magnetic Resonance Imaging (2003) A6.4.1-A6.4.12
Copyright © 2003 by John Wiley & Sons, Inc. Supplement 11
Table A6.4.1 Equipment for DTI
Magnet ≥1.5 T
Coil type Volume transmit/receive head coil
Gradient coil strength Typically, the gradient strength is ≥20
mT/m
Gradient slew rate ≥100 T/m/sec
Readout receiver Bandwidth ≥250 kHz
Peripheral gating Unnecessary except for monitoring
Cardiac gating Unnecessary
Respiratory gating Unnecessary
Use of contrast agents No
For clinical DTI studies, the MRI system should be equipped with hardware capable of
single-shot diffusion-weighted EPI. Typically, this includes a fast-readout receiver with
full bandwidth no less than 250 kHz, a gradient subsystem with a slew-rate faster than
∼100 T/m/sec and a gradient amplitude at least 20 mT/m along each axis, and a magnet
of ≥1.5 T. Key hardware requirements for DTI are listed in Table A6.4.1. In addition to
the hardware requirements, software for DTI data acquisition and post-processing must
be properly installed. The DTI processing software may reside on the scanner so that raw
diffusion tensor images can be immediately processed and viewed on the same console.
Alternatively, the processing software can be installed on a separate computer networked
to the scanner. Following data acquisition, raw DTI data can be transferred from the
scanner to the computer and processed remotely. The resultant DTI maps can be viewed
and analyzed on the remote computer. Typically, image processing and analysis take 5 to
10 min after data acquisition. The total scan time for the procedure (i.e., scan time plus
processing time) is typically well under 20 min. The exact time depends on the details
(e.g, magnetic field strength and number of slices to cover the desired anatomy) of the
protocol as shown in Table A6.4.4.
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be needed for a particular patient, such as crash carts or oxygen.
Set up patient and equipment
1. Screen the patient to ensure that no contraindications for MRI such as cardiac
pacemakers or other implants containing ferromagnetic materials are present. The
screening procedure is the same as that used in conventional MRI brain scans with
special attention to metal materials (e.g., dentures) that may cause magnetic field
distortion. Also be sure to inquire if the patient has any health conditions that may
require the presence of special emergency equipment during the scanning procedure,
or necessitate any other precautions.
Generally, standard screening forms (APPENDIX 1) are used for all patients scanned in a
magnetic resonance system.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(1996) for discussion of what implants may be safely scanned using magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This person must also
be screened for contraindications and the absence of ferromagnetic metal objects on the
body or clothing. This person should also be instructed to minimize talking with the patient,
Imaging which may cause image misregistration problems due to patient motion.
A6.4.2
Supplement 11 Current Protocols in Magnetic Resonance Imaging
 2. For studies under a research protocol, explain the procedure to the patient, answer Table A6.4.2 Sequence Parameters for Rapid Three-Plane Scout Scan
any questions that may arise from the patient, and obtain written, informed consent
from the patient on a form approved by the institutional review board. Patient position Supine
Scan type Gradient echo
3. Have the patient remove all jewelry and change into a hospital-supplied gown to Imaging plane (orientation) Three orthogonal planes (transverse,
eliminate any metal that might be in the clothing. coronal, and sagittal)
Central slice or volume center Lateral canthus of the eye
4. Ask the patient to wash off mascara or other makeup to avoid local tissue heating and
Echo time (TE) Minimum (e.g., ∼1.6 msec)
image artifacts. Encourage the use of bathroom facilities at this time to avoid
Receiver bandwidth (RBW) ±16 kHz
unnecessary disruption during the examination.
Repeat time (TR) Minimum (e.g., 27 msec)
5. Bring the patient to the scanner room and explain the scan procedure to the patient, Flip angle (FA) 30°
including acoustic noise and scanner confinement to be experienced in the magnet. Fields of view (FOVx, FOVy) 240 mm, 240mm
Mention to the patients that it is normal to hear high-pitched noises for not more than Resolution (Δx, Δy) 0.94 mm, 1.88 mm
8 min at 1.5 T (the time can be reduced to ∼4 min at 3.0 T) and to feel minor rocking Number of data points collected (Nx, Ny) 256, 128
of the patient table (due to the diffusion-weighting gradients). Instruct the patient to Display matrix (Dx, Dy) 256, 256
hold as still as possible during scanning. If the patient is unable to hold still, provide Slice thickness (Δz) 5 mm
an appropriate medication for conscious sedation. To minimize artifacts, the patient Number of slices 3
should not talk and should avoid or minimize swallowing or other movement during Slice gap 5 mm
each scan (i.e., as long as the banging sound continues). Between scans, talking and Number of excitations (NEX) 1
swallowing are generally allowed, but should be avoided when comparative posi- Slice locations L10–R10; S10–I10; A10–P10 (when
tional studies are being performed; the patient should be informed when this is the centered at nasion)
case. Saturation pulses None
Scan time ∼11 sec
6. Ask the patient to lie on the table comfortably in a supine position. Set up any
triggering, gating, and monitoring equipment, if necessary for other portions of the
examination, although they are not required for DTI. Place a pillow or other support If the three-plane localization sequence is not available, a fast gradient echo sequence can
under the knees to increase patient comfort. be used instead to acquire a set of sagittal images (parameters are given in Table A6.4.3).
The images obtained in the scout scan will be used to determine the slice location and
7. Provide a pair of earplugs, earphones, or headphones to the patients and demonstrate coverage in the subsequent scan prescriptions (sequences 2 and 3).
how to wear the device to protect ears from the loud acoustic noise generated by the
gradient switching. Explain to the patient that wearing the ear-protecting device does Sequence 2: Diffusion tensor sequence
not impede his/her ability to communicate with the technologist through an intercom 11. Select transverse planes to avoid excessive image distortion especially at 1.5 T
system in the magnet. The patient may call out at any time if he or she feels it (Weisskoff et al., 1993). If diffusion tensor images in other planes are desired, acquire
necessary. interleaved multiple sets of 2-D transverse images as shown in Figure A6.4.1, and
8. Place an emergency stop device (e.g., a squeeze bulb) in the patient’s hand and instruct combine all transverse images to form a 3-D data set from which an arbitrary plane
the patient to use it only when a major concern arises during scanning, or if the orientation can be viewed. Acquisition parameters for each set of 2-D transverse
intercom system malfunctions. images are given in Table A6.4.4.

9. Position the patient’s head in the head-coil holder, padded with a soft cushion (if
available). Additional cushions may be used on both sides of the head to stabilize the
patient. Place the head coil around the patient’s head, use the landmark light to center
the patient on the nasion, and move the patient to the magnet center.
Generally, the head and neck are positioned horizontally (i.e., not tilted) and the neck and
head lie along the axis of the patient table. Other positions may be appropriate depending
on the needs or the condition of the patient.
Once this step is completed, so long as the patient does not move on the table, the table
itself can be moved and then returned to the same position without jeopardizing the
positioning of one scan relative to another.
Sequence 1: Rapid scout scan
10. Run a rapid scout scan to ensure correct positioning of the head in three dimensions
using a three-plane localization sequence (provided by most vendors) with the
parameters given in Table A6.4.2.
Clinical fMRI
At each slice location, the sequence will acquire N + 1 images with N diffusion-weighted
images all having the same b-value but different diffusion-weighting gradient directions
and an additional “base” image in which the diffusion-weighting gradient is set to zero.
All these images are required to calculate DTI maps. N must be at least 6; the recommended
value of N is ∼27, as shown in Table A6.4.4.
12. Whenever possible, use the minimum TE available on the scanner to reduce signal
loss due to T2 decay.
13. If the sequence contains an option for minimizing eddy current effects, use this option
when the diffusion-weighted images exhibit inconsistent geometric distortion, which
can be detected by paging through the N diffusion-weighted images at the same slice
location.
Inconsistent geometric distortion typically manifests itself as misalignment at the edge of
the imaged objects.
Diffusion Tensor
Imaging

A6.4.3 A6.4.4
Current Protocols in Magnetic Resonance Imaging Supplement 11 Supplement 11 Current Protocols in Magnetic Resonance Imaging
 Table A6.4.3 Sequence Parameters for Rapid Scout Scan in Sagittal Plane Table A6.4.4 Sequence Parameters for DTIa

Patient position Supine Patient position Supine

Scan type 2-D gradient echo Scan type Single-shot spin-echo echoplanar
Imaging plane (orientation) Sagittal Imaging plane (orientation) Transverse
Central slice or volume center Lateral canthus of the eye Central slice or volume center Lateral canthus of the eye
Echo time (TE) Minimum (e.g., a few milliseconds) Echo time (TE) Minimum (<100 msec)
Receiver bandwidth (RBW) ±16 kHz Echo train length (ETL) 80
Repeat time (TR) 34 msec Receiver bandwidth (RBW) ∼±200 kHz
Flip angle (FA) 30° Repeat time (TR) 4000 msec at 1.5 T; 5000 msec at 3.0 T
Fields of view (FOVx, FOVy) 240 mm, 240mm Flip angle (FA) 90°
Resolution (Δx, Δy) 0.94 mm, 1.88 mm Fields of view (FOVx, FOVy) 220 mm, 220 mm
Number of data points collected (Nx, Ny) 256, 128 Resolution (Δx, Δy) 1.72 mm, 1.72 mmb
Display matrix (Dx, Dy) 256, 256 Number of data points collected (Nx, Ny) 128, 80 (with ramp sampling)
Slice thickness (Δz) 5 mm Display matrix (Dx, Dy) 128, 128
Number of slices 3 Slice thickness (Δz) 5 mm
Slice gap 5 mm Number of slices ∼23 for full brain coverage or <23 for
Number of excitations (NEX) 1 covering only the suspected lesions
Swap read and phase encoding No Slice gap 1 mm
Slice locations L10–R10 (when centered at lateral Number of excitations (NEX) 2 (at 1.5 T); 1 (at 3.0 T)
canthus of the eye) Number of acquisitions (Nacq) 1 or 2, depending on total number of
Saturation pulses None slices; see the brief explanation below
Scan time ∼5 sec under “scan time,” and APPENDIX 2
Slice locations Graphic prescription based on a
sagittal localizer image for either full
or partial brain coverage
slice 1 slice 2 slice 3
Read direction Left–right
slice n
Saturation pulses None
1st set
b-value 500–1000 sec/mm2
Diffusion gradient directions 27 (evenly distributed in 3-D)
2nd set
Scan time 7 min, 28 sec for more than ∼13 slices,
slice 1 slice 2 slice 3 slice n and 3 min, 44 sec for fewer slices (at
slice direction 1.5 T); 4 min, 38 sec for more than
∼17 slices, and 2 min, 19 sec for fewer
slices (at 3.0 T)
Figure A6.4.1 Interleaved slice acquisition scheme to produce a pseudo 3-D data set. At least aThe parameters in Table A6.4.4 are based on the authors’ experience on GE Signa 1.5 T NV/i and 3.0 T
two sets of transverse slices are required. The second set of slices (solid gray areas) are centered VH/i scanners.
at the gaps (hatched areas) between the first set of slices (solid gray areas). bΔy represents the apparent resolution calculated by dividing FOV by D .
y y

14. After completing the prescription, inform the patient of the scan time for the sequence
and instruct the patient to hold still during scanning. Execute an EPI-based DTI
sequence using the parameters given in Table A6.4.4.
Sequence 3: High-resolution 3-D anatomical images
15. Prescribe and execute a 3-D gradient echo sequence to cover the entire brain
according to the parameters in Table A6.4.5.
As diffusion tensor images are of low spatial resolution, a high-resolution 3-D data set is
particularly useful for clinicians to relate the locations of the fiber tracts to the possible
lesions. Although there can be considerable differences in image distortion between
EPI-based diffusion tensor images and the 3-D gradient-echo images, image-registration
software such as AIR (automatic image registration; Woods et al., 1993) can be used to
register the two data sets and fuse them into a single set of images.
Diffusion Tensor
Clinical fMRI
Data processing
16. Using software provided by the scanner manufacturer or third party, process the
images acquired in step 14 to generate diffusion tensor images, such as grayscale
fractional or relative diffusion anisotropy maps, color-coded fiber direction maps,
and diffusion-trace (or ADC) maps (Sorensen et al., 1996).
For advanced applications, a vector map showing the principal diffusion directions can
also be generated if the processing software supports this option. These maps can be fused
with the high-resolution 3-D anatomical images (as well as functional activation maps, if
available) to provide a comprehensive data set to clinicians for evaluating the relationship
of fiber tracts with high-resolution anatomy and brain activation areas.
Imaging

A6.4.5 A6.4.6
Current Protocols in Magnetic Resonance Imaging Supplement 11 Supplement 11 Current Protocols in Magnetic Resonance Imaging
 Table A6.4.5 Sequence Parameters for High-Resolution 3-D Anatomic Imaginga
λ2
Patient position Supine
Scan type 3-D spoiled gradient echo z
Imaging plane (orientation) Transverse
D2
Central slice or volume center Lateral canthus of the eye
Echo time (TE) 4.2 msec
Repeat time (TR) Minimum (e.g., ∼11 msec) y λ3 D3
D1
Flip angle (FA) 25° λ1
Receiver bandwidth (RBW) ~±16 kHz x
Fields of view (FOVx, FOVy) 240 mm, 240 mm
Resolution (Δx, Δy) 0.94 mm, 1.25 mm
Number of data points collected (Nx, Ny) 256, 192 laboratory frame diffusion ellipsoid
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) ∼1.8 mm Figure A6.4.2 Diffusion ellipsoid in a laboratory reference frame. The lengths of the diffusion
Number of slices 124 ellipsoid axes (D1, D2, and D3) correspond to the eigenvalues given by Equation A6.4.2. The
Slice gap 0 mm directions of the axes correspond to eigenvectors (λ1, λ2, and λ3). The largest eigenvalue, D1, is
Number of excitations (NEX) 1 known as the principal diffusion coefficient, and its eigenvector, λ1, points to the principal diffusion
Slice locations Graphic prescription based on a direction.
sagittal localizer image; 124 slices to
cover the entire brain white-matter fiber tracts, one eigenvalue is
3
Saturation pulses None larger than the other two eigenvalues. The larg- 1 2
RA =
Scan time ∼4 min, 28 sec est eigenvalue and its associated eigenvector 3 Davg
∑ ( Di − Davg )
i =1
aThe parameters in Table A6.4.4 are based on GE Signa scanners. are referred to as the principal diffusion coeffi-
cient and the principal diffusion direction, re- Equation A6.4.3
spectively.
COMMENTARY To highlight the location of the fiber tracts, 3
2
Background Information Each element in the symmetric matrix in Equa- a scalar diffusion anisotropy image can be cal-
culated based on one of the anisotropy parame-
∑ ( Di − Davg )
Fiber tracts in the central nervous system tion A6.4.1 represents a diffusion coefficient i =1
ters, such as relative anisotropy (RA in Equa- FA = 1.5
consist of bundles of myelin-covered axons defined in a laboratory reference frame. Since 3
grouped together along certain pathways (Buhl there are six elements in the diffusion tensor, a tion A6.4.3), fractional anisotropy (FA in Equa-
tion A6.4.4), or volume ratio (VR in Equation
∑ Di2
and Lubke, 1989; Nauta and Feirtag, 1996). minimum of six diffusion-weighted images, i =1
Because of this directional linkage and connec- each acquired with a distinct diffusion gradient A6.4.5; Basser, 1995; Pierpaoli and Basser,
tivity, water molecules diffuse more freely direction, is required to fully characterize dif- 1996; Ulug and van Zijl, 1999). Relative an- Equation A6.4.4
along the fiber tracts than in other directions. fusion tensor at each spatial location. isotropy is the ratio of the standard deviation of
In contrast, diffusion in the other tissues, such The diffusion tensor elements in Equation the eigenvalues D1, D2, and D3 to their average D1 D2 D3
Davg and ranges from 0 to √ ⎯⎯2 . This index is VR = 3
as gray matter, is less directionally dependent. A6.4.1 are patient-orientation dependent. To Davg
By characterizing this diffusion anisotropy at eliminate this dependency, the matrix in Equa- linear over a wide range of anisotropy found in
each spatial location, the fiber tracts can be tion A6.4.1 can be diagonalized to Equation the human brain (Ulug and van Zijl, 1999).
distinguished from the other tissues and the A6.4.2. Fractional anisotropy describes the deviation of Equation A6.4.5
fiber orientations mapped in three dimensions the anisotropic diffusion from a presumed iso-
(Pierpaoli and Basser, 1996). ⎡ D1 0 0 ⎤ tropic diffusion process defined by Davg, nor- Davg = ( D1 + D2 + D3 ) / 3
In general, molecular diffusion in biological ⎢ ⎥ malized to the sum of the squared eigenvalues.
D′ = ⎢0 D2 0 ⎥ Fractional anisotropy ranges from 0 to 1 and is
tissues can be approximated by a second-rank ⎢ ⎥ Equation A6.4.6
tensor at each spatial location. This tensor can more sensitive than RA to low values of anisot-
⎢⎣ 0 0 D3 ⎥⎦
be expressed as the 3 × 3 matrix (Arfken, 1970) ropy. Unlike RA and FA, VR describes diffu- Characterization and visualization of the fi-
in Equation A6.4.1. sion anisotropy by calculating the volume ratio ber tract connectivity using DTI is still in its
Equation A6.4.2 of the diffusion ellipsoid to a hypothetical dif- infancy. The primary limitation is the rather low
⎡ Dxx Dxy Dxz ⎤ fusion sphere with the radius Davg. To map the spatial resolution due to insufficient signal-to-
⎢ ⎥ The elements D1, D2, and D3, in the above
fiber orientation and connectivity, the principal noise ratio (SNR) in the raw diffusion-weighted
D = ⎢ Dxy Dyy Dyz ⎥ matrix are known as the eigenvalues. Each
⎢ ⎥ eigenvector is typically employed to guide the images. At the low spatial resolution, partial
eigenvalue, Di (i = 1, 2, 3), corresponds to an
⎥ fiber-tracking process. volume effect can make the single tensor model
⎢D eigenvector, λi. The sum of the three eigenval-
⎣⎢ xz Dyz Dzz ⎥⎦ invalid. For example, when two fiber tracts with
ues is known as the trace of diffusion tensor.
different orientations coexist in a voxel, the
Equation A6.4.1 These relationships are shown in Figure A6.4.2.
Diffusion Tensor model based on a single tensor becomes inade-
For a diffusionally anisotropic tissue, such as Clinical fMRI Imaging

A6.4.7 A6.4.8
Current Protocols in Magnetic Resonance Imaging Supplement 11 Supplement 11 Current Protocols in Magnetic Resonance Imaging

Figure A6.4.3 When eddy current compensation or correction is sub-optimal, bright rims can be
observed at the edge of the diffusion anisotropy map (A). With improved eddy current compensation
or correction, this artifact can be effectively reduced (B).
quate. Clinical diffusion tensor images at 1.5 T
are typically acquired with an acquisition ma-
trix size of 128 × 128 or smaller on a field of
view (FOV) of ∼20 to 24 cm and a slice thick-
ness of ∼5 mm. The SNR can be effectively
improved at higher magnetic fields (Thulborn,
1999). Very high magnetic fields (i.e., ≥3 T),
coupled with multi-shot acquisition techniques
with motion correction (Butts et al., 1996; Mori
et al., 2001), can potentially lead to higher
spatial resolution in diffusion tensor images
within an acceptable time. New developments
in these areas will be updated as they become
clinically available.
Critical Parameters and
Troubleshooting
Achieving the highest possible SNR is piv-
otal for successful DTI studies. To reduce un-
necessary signal loss due to relaxation effects,
TE must be minimized at a given b-value, while
TR should be maximized within the total scan
time constraint. The minimally achievable TE
strongly depends on the gradient strength avail-
able on the scanner. The higher the gradient
strength, the shorter the minimum TE. The other
key parameter that influences SNR is the b-
value. Recent studies suggested that the optimal
b-value depends slightly on the number of dif-
f usio n- weig htin g gr ad ien t dir ections
(Poonawalla et al., 2000). When 27 directions
are used, for example, a b-value of ∼750
sec/mm2 produces diffusion tensor images of
excellent quality. The optimal number of direc-
Current Protocols in Magnetic Resonance Imaging
) *
tions, as well as the optimal gradient orienta- Figure A6.4.4 Two transverse diffusion tensor images based on fractional anisotropy from a patient with
tions, are continuously being investigated infiltrating glioblastoma multiforme (hollow arrows in both A and B). Both images were acquired on a GE
(Jones et al., 1999; Papadakis et al., 1999; Skare 1.5 T scanner with a single-shot EPI pulse sequence using the protocol given by Table A6.4.4. The solid
et al., 2000). The present consensus is to dis- arrows in A and B show the external capsule and splenium, respectively.
tribute the diffusion gradient directions as uni-
formly as possible across three dimensions. The
authors’ experience indicates that the quality of
diffusion anisotropy images improves as the
number of diffusion gradient directions in-
creases when the total scan time is held con-
stant. However, the gain in image quality di-
minishes after the number of directions exceeds
∼25.
When low SNR is observed in raw diffusion-
weighted images, the imaging protocol should
be carefully examined to ensure that there is no
change from the established protocol. Some
common sources of prescription errors include
reduced FOV, thinner slice thickness, increased
TE/TR ratio, and change in receiver bandwidth.
If no obvious parameter change is detected,
quality assurance procedures should be per-
formed on the scanner to determine the root
cause.
Diffusion tensor calculations assume that
the raw diffusion-weighted images are all prop-
erly registered. In practice, this is rarely the
case, as the diffusion-weighting gradient in
different directions can produce different eddy
currents, resulting in inconsistent image distor- Figure A6.4.5 (A) A fractional anisotropy map from a patient with multiple sclerosis. (B) A color-coded
tion as well as intensity loss among the raw fractional anisotropy map with red, green, and blue color representing fibers along the right/left, ante-
diffusion images (Zhou and Reynolds, 1997; rior/posterior, and superior/inferior directions, respectively. Regions with white-matter loss are indicated
Jezzard et al., 1998). This effect can be readily by the arrows. The images were acquired on a GE 3.0 T scanner with a single-shot EPI pulse sequence.
This black and white facsimile of the figure is intended only as a placeholder; for full-color version of figure
identified by a bright rim at the edge of the brain
Clinical fMRI go to http://www.interscience.wiley.com/c_p/colorfigures.htm.
A6.4.9 A6.4.10
Supplement 11 Supplement 11 Current Protocols in Magnetic Resonance Imaging
and/or an increased intensity in the gray-matter
region in diffusion anisotropy maps (Fig.
A6.4.3A). Although the eddy current–induced
problem can be reduced in certain pulse se-
quence designs, the best approach is to optimize
eddy current compensation on the scanner. This
can be achieved by a coordinated effort with
the field service engineer from the scanner
manufacturer. A pair of images before and after
eddy current correction is shown in Figure
A6.4.3.
Although DTI sequences based on single-
shot EPI are rather insensitive to intra-shot
motion within an image, excessive intra-shot
motion, as well as motion among different
diffusion-weighted images, can still degrade
the diffusion tensor image quality. Motion
often arises from two sources, patient bulk
motion and table vibration induced by the
strong diffusion-weighting gradient. The first
problem can be eliminated with improved pa-
tient cooperation (or sedation). A short-term
solution to the second problem is to decrease
the diffusion-weighting gradient amplitude.
This can be done by either prolonging the
diffusion gradient pulse width while keeping
the same b-value, or reducing the b-value and
maintaining the same gradient pulse width. The
former approach also results in a longer TE,
which degrades the SNR. Another solution is
to mechanically secure the table to improve
motion resistance. If DTI is performed as an
add-on sequence to a conventional brain exam,
the DTI study should be performed prior to
contrast agent injection to avoid any adverse
effect on the diffusion image caused by the
contrast agent.
Anticipated Results
Examples of anticipated images from DTI
are shown in Figures A6.4.4 and A6.4.5. The
images in Figure A6.4.4 were acquired on a GE
1.5 T Signa NV/i MRI scanner using the Basic
Protocol. The impact of infiltrating glioblas-
toma multiforme on the fiber tracts is clearly
visible on the right side of the brain. The major
fiber tracts, such as the external capsule (Fig.
A6.4.4A) and splenium (Fig. A6.4.4B), are
readily identified in the left hemisphere not
affected by the tumor mass. A number of U-fi-
bers are also resolved. These structures are not
readily visible in conventional T1-weighted or
T2-weighted images. Figure A6.4.5 shows two
DT images obtained from a 35-year-old female
patient with multiple sclerosis using a GE 3 T
Signa VH/i scanner. Loss of white-matter sub-
stance can be appreciated at superior corona
Current Protocols in Magnetic Resonance Imaging
radiata (Fig. A6.4.5A, shown by the arrow) in
the gray-scale FA map. The color-coded FA
map (Fig. A6.4.5B) provides additional infor-
mation on the orientation of the affected fiber
tract (shown by the arrow). The color scheme
is explained in the figure legend. The use of
color coding is of great value in differentiating
and identifying fibers that are of interest or
those close to each other.
Acknowledgments
The authors are grateful to Aziz Poonawalla
and Donna Shobat for their assistance in image
acquisitions. Dr. Zhou also thanks Drs. Norman
Leeds, Jeffrey Weinberg, Christof Karmonik,
and Bryan Mock for helpful discussions.
Literature Cited
Arfken, G. 1970. Mathematical Methods for Physi-
cists. Academic Press, New York.
Basser, P.J. 1995. Inferring microstructural features
and the physiological state of tissues from diffu-
sion-weighted images. NMR Biomed. 8:333-
344.
Basser, P.J. and Pierpaoli, C. 1998. A simplified
method to measure the diffusion tensor from
seven MR images. Magn. Reson. Med. 39:928-
934.
Basser, P.J., Mattiello, J., and LeBihan, D. 1994.
Estimation of the effective self-diffusion tensor
from the NMR spin echo. J. Magn. Reson. B
103:247-254.
Basser, P.J., Pajevic, S., Pierpaoli, C., Duda, J., and
Aldroubi, A. 2000. In vivo fiber tractography
using DT-MRI data. Magn. Reson. Med. 44:625-
632.
Buhl, E.H. and Lubke, L. 1989. Intracellular lucifer
injection in fixed brain slices combined with
retrograde tracing, light and electron micros-
copy. Neuroscience 28:3-16.
Butts, K., Pauly, K.J., deCrespigny, A., and Moseley,
M. 1996. Diffusion-weighted interleaved echo-
planar imaging with a pair of orthogonal naviga-
tor echoes. Magn. Reson. Med. 35:763-770.
Jezzard, P., Barnett, A.S., and Pierpaoli, C. 1998.
Characterization of and correction for eddy cur-
rent artifacts in echo planar diffusion imaging.
Magn. Reson. Med. 39:801-812.
Jones, D.K., Horsfield, M.A., and Simmons, A.
1999. Optimal strategies for measuring diffusion
in anisotropic systems by magnetic resonance
imaging. Magn. Reson. Med. 42:515-525.
Melhem, E.R., Mori, S., Mukundan, G., Kraut,
M.A., Pomper, M.G., and van Zijl, P.C.M. 2002.
Diffusion tensor MR imaging of the brain and
white matter tractography. Am. J. Roentgenol.
178:3-16.
Mori, S., Itoh, R., Zhang, J., Kaufmann, W.E., van
Zijl, P.C., Solaiyappan, M., and Yarowsky, P.
2001. Diffusion tensor imaging of the develop-
ing mouse brain. Magn. Reson. Med. 46:18-23.
Nauta, W.J.H. and Feirtag, M. 1996. Fundamental
Neuroanatomy. W.H. Freeman Co., New York.
Pajevic, S. and Pierpaoli, C. 1999. Color schemes to
represent the orientation of anisotropic tissues
from diffusion tensor data: Application to white
matter fiber tract mapping in the human brain.
Magn. Reson. Med. 42:526-540.
Papadakis, N.G., Xing, D., Huang, C.L., Hall, L.D.,
and Carpenter, T.A. 1999. A comparative study
of acquisition schemes for diffusion tensor im-
aging using MRI. J. Magn. Reson. 137:67-82.
Pierpaoli, C. and Basser, P.J. 1996. Toward a quan-
titative assessment of diffusion anisotropy.
Magn. Reson. Med. 36:893-906.
Pierpaoli, C., Jezzard, P., Basser, P.J., Barnett, A.,
and Di Chiro, G. 1996. Diffusion tensor MR
imaging of the human brain. Radiology 201:637-
648.
Poonawalla, A., Karmonik, C., and Zhou, X.J. 2000.
Optimization of b-value and gradient orientation
for diffusion tensor MRI. Proc. Intl. Soc. Magn.
Reson. Med. 8th Meeting 2:801.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Skare, S., Hedehus, M., Moseley, M.E., and Li, T.Q.
2000. Condition number as a measure of noise
performance of diffusion tensor data acquisition
schemes with MRI. J. Magn. Reson. 147:340-
352.
Sorensen, A.G., Buonanno, F.S., Gonzalez, R.G.,
Schwamm, L.H., Lev, M.H., Huang-Hellinger,
F.R., Reese, T.G., Weisskoff, R.M., Davis, T.L.,
Suwanwela, N., Can, U., Moreira, J.A., Copen,
W.A., Look, R.B., Finklestein, S.P., Rosen, B.R.,
and Koroshetz, W.J. 1996. Hyperacute stroke:
Evaluation with combined multisection diffusion-
weighted and hemodynamically weighted echo-
planar MR imaging. Radiology 199:391-401.
Diffusion Tensor
Clinical fMRI Imaging
A6.4.11 A6.4.12
Supplement 11 Supplement 11
Thulborn, K.R. 1999. Clinical rationale for very-
high-field (3.0 Tesla) functional magnetic reso-
nance imaging. Top. Mag. Reson. Imaging
10:37-50.
Ulug, A.M. and van Zijl, P.C. 1999. Orientation-in-
dependent diffusion imaging without tensor di-
agonalization: Anisotropy definitions based on
physical attributes of the diffusion ellipsoid. J.
Magn. Reson. Imaging 9:804-813.
Weisskoff, R.M., Cohen, M.S., and Rzedzian, R.R.
1993. Nonaxial whole-body instant imaging.
Magn. Reson. Med. 29:796-803.
Woods, R.P., Mazziotta, J.C., and Cherry, S.R. 1993.
MRI-PET registration with automated algo-
rithm. J. Comput. Assist. Tomogr. 17:536-546.
Xue, R., van Zijl, P.C., Crain, B.J., Solaiyappan, M.,
and Mori, S. 1999. In vivo three-dimensional
reconstruction of rat brain axonal projections by
diffusion tensor imaging. Magn. Reson. Med.
42:1123-1127.
Zhou, X. and Reynolds, H.G. 1997. Quantitative
analysis of eddy current effects on diffusion-
weighted EPI. Proc. Intl. Soc. Magn. Reson.
Med. 5th Meeting 3:1722.
Zhou, X.J., Leeds, N.E., Karmonik, C., and Mock,
B.J. 2000. Magnetic resonance tractography for
pre-surgical planning and post-surgical evalu-
ation. Proc. Intl. Soc. Magn. Reson. Med. 8th
Meeting 1:480.
Contributed by Xiaohong Joe Zhou
M.D. Anderson Cancer Center
Houston, Texas
Keith R. Thulborn
University of Illinois at Chicago
Chicago, Illinois
Current Protocols in Magnetic Resonance Imaging

CHAPTER A7
Head and Neck
INTRODUCTION
R imaging has represented a tremendous advance in the noninvasive evaluation of
M head and neck diseases. The ability to obtain high-resolution images of the orbits,
pharynx, skull base, larynx, and cranial nerves has greatly benefited the work-up of
abnormalities in these regions, and has been an important application of MR imaging.
Typically, evaluation of abnormalities in the head and neck requires careful attention to
detail and a thorough understanding of the underlying anatomy. It is a requisite for the
MR imaging technique used in the evaluation of these regions that it yield high-resolution
images, sufficient tissue contrast, and minimal motion artifacts. The following units
outline protocols that will assist in attaining these goals.
Jonathan S. Lewin
Cranial Nerve I UNIT A7.1
The olfactory bulbs and tracts mediate the sense of smell from the nasal cavity to the
brain. Unfortunately they are located in a precarious position for MR imaging, above the
air-filled nasal cavity and ethmoid sinuses at a bone-air-soft tissue interface. This creates
problems with susceptibility artifact. This issue, plus the very small size of the structures
to be studied and the superimposed eye motion artifact makes imaging of the olfactory
system a technical challenge.
IMAGING OF CRANIAL NERVE I BASIC
PROTOCOL
The olfactory bulb and tract is actually a second order neuron with the primary sensory
(ciliary) nerves in the ethmoid and nasal cavity vault. Lesions in the sinonasal cavity as
well as the anterior cranial fossa can affect the sense of smell (Li et al., 1993, 1994;
Yousem, 1993). The most common diseases to affect olfaction are sinonasal bacterial and
viral infections and neurodegenerative disorders such as multiple sclerosis, Alzheimer’s
disease, and Parkinson’s disease (Doty et al., 1998, 1999; Li et al., 1994). Primary entities
to affect the olfactory bulbs and tracts include traumatic shearing or contusional injuries,
congenital disorders such as Kallmann’s disease and holoprosencephaly, and tumors such
as meningiomas and olfactory neuroblastomas (Doty et al., 1997; Li et al., 1994; Yousem,
1993; Yousem et al., 1996a, 1996b). Because the olfactory bulbs and tracts are very tiny
structures oriented in an anteroposterior plane, they are best evaluated with thin-section
coronal scanning (Yousem et al., 1997a). However the proximity of these structures to
the air containing sinonasal cavity makes the use of gradient echo scanning prohibitive.
Imaging of the cranial nerves requires a focused approach based on clinical symptoma-
tology and signs. Thin section imaging and high resolution are essential components to
this evaluation. Table A7.1.1 lists the hardware necessary to perform the procedure, along
with appropriate parameters.
NOTE: Be sure that technicians and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.

Set up equipment and patient

1. Interview (screen) the patient to ensure that he or she has no contraindications such
as cardiac pacemakers or other implants containing ferromagnetic materials. Also be
sure to find out if the patient has any health conditions that may require the presence
of special emergency equipment during the scanning procedure, or necessitate any
other precautions.

Table A7.1.1 Equipment Parameters for Cranial

Nerve Imaging

Coil type 5-in. surface coil

Gradient coil strength 25 mT/m
Flow compensation pulse Yes
Peripheral gating N/A
Respiratory gating N/A
Respirator PRN
Oxygen PRN
Motion cushions Useful
Head and Neck Head and Neck

Contributed by Jonathan S. Lewin A7.0.1 Contributed by Robert W. Evers and David M. Yousem A7.1.1
Current Protocols in Magnetic Resonance Imaging (2002) A7.0.1 Current Protocols in Magnetic Resonance Imaging (2001) A7.1.1-A7.1.8
Copyright © 2002 by John Wiley & Sons, Inc. Supplement 6 Copyright © 2001 by John Wiley & Sons, Inc.
 Generally standard screening forms are used for all patients scanned in a magnetic
resonance system.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(1996) for discussion of what implants may be safely scanned using magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
2. If the procedure is a research protocol, have the patient sign any necessary consent
form.
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
Once this step has been performed, so long as the patient does not move on the table, the
table itself can be moved and then replaced in the same position as before without
jeopardizing the positioning of one scan relative to another.
10. If the patient is unable to hold still, provide an appropriate sedative.
Sequence 1: Coronal scout scan
11. Run coronal 2-D FMPSPGR (sequence 1; fast multi-planar spoiled gradient echo)
scout using the imaging sequence given in Table A7.1.2.
At 1.5 T, fat and water will precess in and out of phase respectively every 2.3 msec. Thus,
utilizing a TE of 4.6 msec will maximize the contributory signal within each voxel. The
primary reasoning for a quick gradient echo localizer is to assess the maximum dimension
that the 5-in. coil provides in regard to SNR (signal-to-noise ratio) and anatomic coverage.
We have found that the utilization of a T1 weighted FMPSPGR quickly provides the imaging
personnel with an above average survey with regard to coverage and SNR.

4. Have the patient wash off any mascara and other makeup to avoid local tissue heating Sequence 2: Sagittal T1-weighted spin echo scan
and image artifacts. 12. Run sagittal T1-weighted scan (sequence 2) using the imaging sequence given in Table
5. Inform the patient about what will occur during the procedure, what he or she will A7.1.3.
experience while in the magnet, and how to behave, including the following: Spatial saturation pulses and gradient moment nulling (flow compensation) are not
necessary. Ideally, one’s area of coverage should maximize the useful signal dimensions
a. If earphones or headphones are used to protect the ears from the loud sounds of the 5-in. surface coil.
produced by the gradients, the patient will be asked to wear these, but will be able
to communicate with you at any time during the imaging. Sequence 3: Coronal T1-weighted scan
b. The patient will be given a safety squeeze-bulb or similar equipment to request 13. Run coronal spin echo T1-weighted scan (sequence 3) using the imaging sequence
assistance at any time (demonstrate how this works). given in Table A7.1.4.
c. For good results the patient should not talk, and should avoid or minimize Saturation pulses and gradient moment nulling are not necessary. The sections should
swallowing or other movement, during each scan—i.e., as long as the banging begin at the nasion and extend posteriorly to the level of the optic chiasm, visualized on
sounds continue. Between scans, talking and swallowing are allowed in most the coronal “scout” image. Fast spin echo scanning may be used, however, it may be helpful
cases, but should be avoided when comparative positional studies are being to increase the matrix size or zero fill interpolate (ZIP) to more easily visualize small details.
performed; the patient will be informed when this is the case. Echo train lengths ≤4 should then be used.
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
6. Help the patient mount onto the table. Either before or right after the patient lies down,
set up any triggering devices or other monitoring equipment that is to be used. Table A7.1.2 Clinical Imaging Parameters for 2-D FMPSPGR Coronal Scout
Scan (Sequence 1)
7. Center the surface coil over the nasion (the region where the key information is
desired). Make sure that the head and neck are constrained to prevent motion, Patient position Supine
especially if high-resolution scans are to be run. Scan type Fast multi-planar spoiled gradient
echo
Generally the patient’s head is fixed so that the head is horizontal (not tilted) and the neck Imaging plane (orientation) Coronal
and head lie along the axis of the patient table; other positions may be appropriate Central slice or volume center Middle optic nerve
depending on the needs at hand.
Echo time (TE) 4.6 msec
Surface coil imaging is required to visualize the olfactory bulbs and tracts. A 5-in. round Receiver bandwidth (RBW) 16 kHz
general purpose coil can be used, suspended one inch over the nasion, without touching Repeat time (TR) 100 msec
the patient. Instructions to the patient to refrain from excessive eye or head movement are Flip angle (FA) 70°
critical. The gaze should be fixed on a single spot. Anatomic imaging with high-quality Field of view (FOVx, FOVy) 300 mm, 300 mm
spin echo T1-weighted scan is the mainstay of the evaluation. Field strength considerations
Resolution (Δx, Δy) 0.58 mm, 1.17 mm
are not critical. No monitoring is required.
Number of data points collected (Nx, Ny) 512, 256
8. If needed, place a pillow or other support under the knees to make the patient more Slice thickness (Δz) 5 mm
comfortable. Number of slices Variable (∼20)
Slice gap 1 mm
9. Use the centering light to center on the nasion in the center of the 5-in. round surface
Number of excitations (NEX) 1
coil and advance the patient into the center of the magnet.
Cranial Nerve I Scan time ∼1 min Head and Neck

A7.1.2 A7.1.3
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Sequence 4: Coronal T2-weighted scan Table A7.1.5 Clinical Imaging Parameters for Coronal T2 FSE Fat Saturationa
14. Run coronal fast spin echo (FSE) T2-weighted scan using the imaging sequence given
in Table A7.1.5. Patient position Supine
Scan type Fast spin echo
Saturation pulses and gradient moment nulling are not necessary. The sections should be Imaging plane (orientation) Coronal
the same as those for the T1-weighted scans. Acquisition time will be 4 to 5 min in length.
Central slice or volume center Middle optic nerve
Manual chemical fat saturation techniques should be employed to ensure a homogeneous
“fat suppressed” image (see Troubleshooting for additional information on the advantages Echo time (TE) 90–120 msec
of manual chemical saturation). Receiver bandwidth (RBW) 20 kHz
Echo train length (ETL) 12–32
Repeat time (TR) 3000–4000 msec
Table A7.1.3 Clinical Imaging Parameters for Sagittal T1 Flip angle (FA) 90°
Field of view (FOVx, FOVy) 160 mm, 160 mm
Patient position Supine Resolution (Δx, Δy) 0.63 mm, 0.63 mm
Scan type Spin echo Number of data points collected (Nx, Ny) 256, 256
Imaging plane (orientation) Sagittal Slice thickness (Δz) 2–3 mm
Central slice or volume center Middle optic nerve Number of slices Variable (∼20)
Echo time (TE) Minimum Slice gap 0 mm
Receiver bandwidth (RBW) 16 kHz Number of excitations (NEX) 2
Repeat time (TR) 400–600 msec Flow compensation Yes
Flip angle (FA) 90° ZIP 512 Yes, changes resolution to 0.31
Field of view (FOVx, FOVy) 160 mm, 160 mm mm by 0.31 mm
Resolution (Δx, Δy) 0.63 mm, 0.63 mm Extended dynamic range (EDR) Yes
Number of data points collected (Nx, Ny) 256, 256 Chemical saturation Yes
Slice thickness (Δz) 3 mm Slice series Interleaved
Number of slices Variable (∼20) Scan time ∼4–5 min
Slice gap 0 mm aFSE, fast spin echo.

Number of excitations (NEX) 1

Extended dynamic range (EDR) Yes COMMENTARY
Slice series Interleaved
Scan time ∼3 min Background Information general, to improve the overall SNR of an
It was not until the introduction of surface image, the operator has a great number of pa-
coils that adequate evaluation of the olfactory rameters to alter. Maximizing the signal re-
Table A7.1.4 Clinical Imaging Parameters for Coronal T1 Spin Echo bulbs and tracts became possible. The im- ceived within each voxel is one of the best ways
proved signal-to-noise ratio is what is needed to improve the image quality/SNR while main-
Patient position Supine to evaluate these structures, which measure ∼20 taining the spatial resolution.
Scan type Fast spin echo mm by 4 mm in size. Cutting the bulbs and
Imaging plane (orientation) Coronal tracts in cross-section, i.e., in the coronal plane, Critical Parameters
Central slice or volume center Middle optic nerve is necessary, as neither the sagittal nor trans- The most common reason why an inade-
Echo time (TE) Minimum verse plane is reliable in reproducibly visualiz- quate study is obtained is the presence of mo-
Receiver bandwidth (RBW) 16 kHz ing the olfactory apparatus. Because it is the tion artifact from eyeball movement and inho-
Echo train length (ETL) ≤4 morphology of the bulbs and tracts which is mogeneous fat suppression. Often the globes
Repeat time (TR) 400–600 msec critical rather than signal intensity, one should are at the same coronal plane as the bulbs and
Flip angle (FA) 90° rely most heavily on the T1-weighted scans. bulb-tract junction. The motion of the globes
These will tell if the bulbs and tracts are there, will cause artifacts in a transverse plane that
Field of view (FOVx, FOVy) 160 mm, 160 mm
if they are hypoplastic or encephalomalacic, obscures the bulbs and tracts. For this reason,
Resolution (Δx, Δy) 0.63 mm, 0.63 mm
and if they have been traumatized. one must tell the subject to fix his or her gaze
Number of data points collected (Nx, Ny) 256, 256 With the utilization of smaller fields of view on a single point, reduce blinking, and to refrain
Slice thickness (Δz) 2–3 mm (FOVs) to improve spatial resolution, there is from head movement. Having scan times as
Number of slices Variable (∼20) often a decrease in signal-to-noise ratio (SNR). short as possible without losing SNR is essen-
Slice gap 0 mm Taking the time to complete a dedicated exami- tial. For this reason, fast spin echo T2-weighted
Number of excitations (NEX) 2–3 nation with a surface coil will yield better imaging is advantageous.
Flow compensation Yes results for those structures close to the coil. The Incompletely fat-suppressed images will
Extended dynamic range (EDR) Yes smaller surface coil will tend to improve the impair and reduce overall confidence with a
Slice series Interleaved overall SNR because of its inherent ability to differential diagnosis and often confuse the
Cranial Nerve I Scan time 4–5 min provide more signal within a smaller voxel. In referring clinicians. Due to the inherent diffi- Head and Neck

A7.1.4 A7.1.5
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 culties with air/tissue interfaces in regard to fat
suppression, it is essential that the imaging
professional know some tricks of the trade to
overcome this disability. Beginning with an
automatic pre-scan and moving on to a manual
spectral suppression should help identify where
to accurately place the chemical saturation
pulse. If there is an improper centering of the
center frequency, typically the suppression
pulse that was intended to suppress fat will also
be at the wrong frequency.
Troubleshooting
It is important to make sure that the slice
locations and surface coil placement are opti-
mal. Improper “tilting” or centering of the 5-in.
surface coil may render one side of the anatomy
to appear to have a higher SNR than the other.
Helpful hints in regard to complete fat sup-
pression over your imaging volume include the
following. (1) Prior to beginning any examina-
tion, the patient should be required to remove
all make-up and remove all metallic objects
(i.e., dental work, earrings, hairpins, etc.). This
simplistic idea will pay off in the long run by
significantly improving the magnetic field ho-
mogeneity to the imaging volume that is being
evaluated. (2) Having a visual spectrum that
allows the user to identify the water peak and
the fat peak improves the chemical suppression
of the imaging volume on selective fat suppres-
sion series. However, even when utilizing
proper automatic and manual suppression tech-
niques to suppress fat, there are often additive
disabilities that lead to an overall poor image.
We are referring to the presence not only of the
air-tissue interface but also dental work that
cannot be removed. In the case of excessive
diamagnetic properties, a STIR (short tau in-
version recovery) would be in order. STIR takes
the “guesswork” out of the chemical saturation
process by utilizing the null point of fat to give
the clinician a “fat suppressed image” (see
“optional sequence” for STIR parameters in
lieu of FSE T2 with fat saturation in Table
A7.1.6).
In the event that a congenital cause for anos-
mia is suspected, one should perform an unen-
hanced scan through the brain to evaluate the
septum pellucidum (septo-optic dysplasia is
associated with aplasia of the olfactory bulbs)
and the cerebral hemispheres (to evaluate for
holoprosencephaly).
Anticipated Results
When imaging has been done appropriately,
Cranial Nerve I you should see the olfactory bulbs arise in a
A7.1.6
narrow groove on either side of the crista galli Table A7.1.6 Clinical Imaging Parameters for STIR “Optional
at the mid ethmoid level. Evidence of sinusitis Sequence” if Coronal T2 FSE Fat Saturation is Poora
or sinonasal masses should be apparent, par-
ticularly on the T2-weighted scans. You should Patient position Supine
be able to follow the bulbous bulbs to the Scan type Inversion recovery, spin echo
slit-like tracts as they proceed posteriorly in the Imaging plane (orientation) Coronal
olfactory sulcus lateral to the gyrus rectus. The Central slice or volume center Middle optic nerve
tracts enter the brain at the medial and lateral Echo time (TE) 50 msec
olfactory septal nuclei lateral to the optic chi- Receiver bandwidth (RBW) 20 kHz
asm. The full course of these structures should Echo train length (ETL) 20
be demonstrated. Repeat time (TR) 3000–4000 msec
Inversion time (TI) For 1.5 Tesla, 150 msec
Time Considerations
Flip angle (FA) 180°
This is a quick protocol once the proper
placement of the patient, coils, and slices are Field of view (FOVx, FOVy) 160 mm, 160 mm
determined. Extra time instructing the patient Resolution (Δx, Δy) 0.63 mm, 0.63 mm
on eye movement is worth the investment. If Number of data points collected (Nx, Ny) 256, 256
the head needs to be examined, allot time ac- Slice thickness (Δz) 2–3 mm
cordingly; you will have to switch coils. If there Number of slices Variable (∼20)
is a need to implement an additional sequence Slice gap 0 mm
such as an STIR, please understand that in order Number of excitations (NEX) 3–4
to maintain the same SNR as the FSE T2 you Flow compensation Yes
will need to decrease your effective TE and quite ZIP 512 Yes, changes resolution to
possibly may have to increase your NEX. Typi- 0.31 mm by 0.31 mm
cally STIRs will yield a lower overall SNR in Extended dynamic range (EDR) Yes
comparison to FSE T2 with all other parameters Chemical saturation Not needed
remaining consistent.
Slice series Interleaved
Scan time ∼5–6 min
Index of Terms aFSE, fast spin echo.
The clinical imaging instructions and termi-
nology utilized in this unit are primarily geared
towards General Electric equipment. An index utilization of 32-bit data processing as opposed b. NPW essentially doubles the phase en-
of terms is provided should bridge the gap of to the standard 16-bit processor. EDR in this coding steps (to maintain resolution);
vendor specific terminology. way should improve SNR and resolution, but c. With NPW, you must reduce the NEX by
Chemical saturation A technique that will utilize twice as much memory as a conven- half in order to maintain scan time.
applies an additional radiofrequency (RF) tional acquisition. Number of excitations (NEX) is a factor
pulse (at a desired distance from the center Flow compensation Or sometimes more that is utilized to calculate the overall scan time
frequency) to selectively suppress a tissue. This widely expressed as gradient moment nulling, and will directly effect the SNR. NEX is essen-
technique can be utilized to suppress the signal is a way in which the system places flowing or tially the number of times that data is sampled
from water, fat, or silicone. moving spins into “phase coherence” with sta- per acquisition. Note that increasing the NEX
Echo time (TE) The time that is measured tionary spins. to achieve overall better SNR is a rather ineffi-
from the initiation of the initial RF (radio fre- Inversion time (TI) With inversion recov- cient way to improve signal. Doubling the NEX
quency) pulse and the peak of the echo. ery pulse sequences, typically the inversion from 2 to 4 will only yield a 40% increase in
Echo train length (ETL) In fast spin echo time is the time from the first 180° RF pulse to SNR while it doubles scan time. Take note of
or turbo spin echo imaging, the ETL will actu- the center of the next 90° RF pulse. This inver- the formula for scan time:
ally equal the number of echoes prescribed per sion time will essentially “null” the desired
Scan time = (TR) × (no. of phase
TR. Successive 180° refocusing pulses are ap- tissue depending on how long or short the TI
encoding steps) × (NEX)
plied to “re-phase the dephasing” protons in an (inversion time) selected is, and the T1 relaxa-
effort to maximize the number of lines of k- tion time of the corresponding tissue. Receiver bandwidth (RBW) The range
space per TR. The formula for scan time in No phase wrap (NPW) will prevent wrap of frequencies that the MRI scanner is actually
relation to fast spin echo imaging and ETL is around artifacts (also known as aliasing arti- “tuned” to receive. This will directly affect the
as follows: facts) in the phase encoding direction. NPW overall SNR. This will not be done by increas-
should only be used when necessary because it ing or decreasing the signal, but rather there
Scan time = (TR) × (no. of phase
does all of the following: will be an increase and/or decrease in the
encoding steps/ETL) × (NEX)
a. NPW doubles the FOV in the phase en- amount of noise received relative to the altera-
Extended dynamic range (EDR) An im- coding direction; tion of the RBW. An increase in the RBW will
aging-enhancement tool that will allow the Head and Neck
A7.1.7
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 increase the range of frequencies that the scan-
ner will evaluate and thus decrease the overall
SNR. In comparison, utilizing a narrow band-
width should yield less noise and improve the
overall SNR. The relationship of the receiver
bandwidth and SNR can be thought of as in-
versely proportional to the square root of the
bandwidth.)
Rectangular field of view (REC FOV)
Asymmetric field of view (typically in the
phase encoding direction). REC FOV is typi-
cally utilized when a body part is longer in one
direction than another. By utilizing an asym-
metric FOV, the system will not collect a por-
tion of the data, thereby decreasing the scan-
ning time.
Repetition time (TR) The time in a pulse
sequence between successive excitation pulses.
Spatial saturation employs an additional
RF pulse to cause moving spins within a deter-
mined area to be selectively dephased. This
application will reduce motion from flow
and/or respiratory artifacts and will limit the
number of slices per TR in general. Since a
spatial saturation pulse employs additional RF,
which is to be deposited into the patient, special
attention should be focused on the SAR (spe-
cific absorption rate). Today all MRI scanners
have a program that internally monitors how
much RF can be applied over a given period of
time. This formula takes into account the pa-
tient’s body weight. This actual body weight
needs to be accurately input at all times for
patient safety. An inappropriate weight will
cause the improper limit of RF to be transmitted
relative to a safe period of time in which this is
to occur.
Tailored RF An imaging option that im-
proves image quality on fast spin echo (FSE)
sequences with relatively short TE’s. Tailored
RF will improve edge blurring by reducing
overall echo spacing.
ZIP 512/ZIP 1024 Better known as “zer-
ofill interpolation process,” this is a reconstruc-
tion algorithm that allows the user to scan at a
256 × 256 matrix and then the data are zero-
filled to a 512 × 512 matrix (or 1024 × 1024,
respectively).
ZIP 2/ZIP 4 Slice Zip essentially is also
a “zerofill interpolation process” that will cre-
ate additional slices through the interpolation
procedure. These slices are created with an
offset of 50% of the original imaging slice
locations.
Cranial Nerve I
A7.1.8
Literature Cited
Doty, R.L., Yousem, D.M., Pham, L.T., Kreshak,
A.A., Geckle, R., and Lee, W.W. 1997. Olfactory
dysfunction in patients with head trauma. Arch.
Neurol. 54:1131-1140.
Doty, R.L., Li, C., Mannon, L.J., and Yousem, D.M.
1998. Olfactory dysfunction in multiple sclero-
sis. Relation to plaque load in inferior frontal and
temporal lobes. Ann. N.Y. Acad. Sci. 855:781-
786.
Doty, R.L., Li, C., Mannon, L.J., and Yousem, D.M.
1999. Olfactory dysfunction in multiple sclero-
sis. Relation to longitudinal changes in plaque
numbers in central olfactory structures. Neurol-
ogy 53:880-882.
Li, C., Yousem, D.M., Hayden, R.E., and Doty, R.L.
1993. Olfactory neuroblastoma: MR evaluation.
Am. J. Neuroradiol. 14:1167-1172.
Li, C., Yousem, D.M., Doty, R.L., and Kennedy,
D.W. 1994. Neuroimaging in patients with olfac-
tory dysfunction. Am. J. Roentgenol. 162:411-
418.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Yousem, D.M. 1993. Imaging of sinonasal inflam-
matory disease: State of the art. Radiology
188:303-314.
Yousem, D.M., Geckle, R.J., Bilker, W., McKeown,
D., and Doty, R.L. 1996a. MR evaluation of
patients with congenital hyposmia or anosmia.
Am. J. Roentgenol. 166:439-444.
Yousem, D.M., Geckle, R.J., Bilker, W.B., McKe-
own, D.A., and Doty, R.L. 1996b. Post-traumatic
olfactory dysfunction: MR and clinical evalu-
ation. Am. J. Neuroradiol. 17:1171-1179.
Yousem, D.M., Geckle, R.J., Doty, R.L., and Bilker,
W.B. 1997a. Reproducibility and reliability of
volumetric measures of olfactory eloquent struc-
tures. Acad. Radiol. 4:264-269.
Yousem, D.M., Geckle, R.J., Bilker, W.B., Kroger,
H., and Doty, R.L. 1999b. Post-traumatic smell
loss: Relationship of psychophysical tests and
volumes of the olfactory bulbs and tracts and the
temporal lobes. Acad. Radiol. 6:264-272.
Yousem, D.M., Maldjian, J.A., Hummel, T., Alsop,
D.C., Geckle, R.J., Kraut, M.A., and Doty, R.L.
1999c. The effect of age on odor-stimulated
functional magnetic resonance imaging. Am. J.
Neuroradiol. 20:600-608.
Contributed by Robert W. Evers and
David M. Yousem
The Johns Hopkins Hospital
Baltimore, Maryland
Current Protocols in Magnetic Resonance Imaging
Cranial Nerves III to VI UNIT A7.2
Cranial nerves III, IV, V, and VI are small structures that travel in a reproducible manner
from the midbrain and pons to the cavernous sinus and then to the orbit. While there are
branches that course through other foramina of the skull, the emphasis in MR is to evaluate
the brainstem, the cavernous sinus, and the pericavernous regions for pathology. Because
the nerves run from a posterior to an anterior position, coronal scanning is ideal for seeing
the nerves in cross-section. Thin sections and contrast enhancement are required to best
visualize the diseases that affect these nerves (Castillo and Mukherji, 1996a, 1996b).
Terminology used in this unit is defined under the Index of Terms (see Commentary).
IMAGING OF CRANIAL NERVES III TO VI BASIC
PROTOCOL
The cranial nerves that supply the oculomotor muscles (III, IV, and VI) emanate from the
brainstem and proceed through the cavernous sinus to enter the superior orbital fissure.
Cranial nerve IV is unique in that it decussates in the midbrain and then exits the brainstem
from the posterior surface of the midbrain (all others remain ipsilateral and leave the
brainstem from its anterior surface). One branch of cranial nerve V, the ophthalmic branch
(V-1), also traverses the superior orbital fissure with the other cranial nerves while the
maxillary nerve (V-2) and mandibular nerve (V-3) escape to the head and neck via
foramina rotundum and ovale respectively. While cranial nerves III and V are routinely
visualized on standard brain scans, visualization of cranial nerves IV and VI requires a
dedicated effort. Fortunately, isolated lesions of these nerves are uncommon.
Table A7.2.1 lists the hardware necessary to perform the procedure, along with appropri-
ate parameters. Standard head coil imaging is sufficient to obtain adequate scans on these
cranial nerves. Anatomic imaging with high-quality fast spin echo T2-weighted scan with
fat suppression should yield dark nerves outlined by bright cerebrospinal fluid (CSF).
This is required to see the cisternal portions of these nerves. In the cavernous sinus, the
nerves are best seen as filling defects in the gadolinium-enhanced cavernous sinus walls
on coronal scans (Yousem et al., 1989, 1990a). Field strength considerations are not
critical, although field strength is directly proportional to signal-to-noise ratio (SNR).
Adjustments in maximizing overall SNR should be taken into account with small FOV’s
and reduced slice thicknesses. No monitoring is required.
Materials
Normal saline (0.9% NaCl), sterile
Gadolinium-based MR contrast agent (e.g., Magnevist, Omniscan, or Prohance)
NOTE: Be sure that technicians and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.
Table A7.2.1 Equipment Parameters for Cranial
Nerve Imaging
Coil type Head
Gradient coil strength 25 mT/m
Flow compensation pulse As needed
Peripheral gating NA
Motion cushions Useful
Use of contrast agents Yes
Head and Neck
Contributed by Robert W. Evers and David M. Yousem A7.2.1
Current Protocols in Magnetic Resonance Imaging (2001) A7.2.1-A7.2.12
Copyright © 2001 by John Wiley & Sons, Inc.
 Set up equipment and patient 9. Establish an intravenous line from which the contrast agent can be injected, and attach
1. Interview (screen) the patient to ensure that he or she has no contraindications such this line securely to the patient so that movement into or out of the magnet will not
as cardiac pacemakers or other implants containing ferromagnetic materials. Also be pull at the patient’s arm.
sure to find out if the patient has any health conditions that may require the presence
It is preferable to insert the line prior to imaging and to leave the patient in the magnet,
of special emergency equipment during the scanning procedure, or necessitate any with no intervening motion, between the scans run before contrast agent injection and those
other precautions. run after injection.
Generally, standard screening forms are used for all patients scanned in a magnetic 10. Use the centering light to center on the nasion and advance the patient into the center
resonance system. of the magnet.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
Once this step has been performed, so long as the patient does not move on the table, the
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
table itself can be moved and then replaced in the same position as before without
composition of the items, it is best to exclude patients with any metal implants; see Shellock
jeopardizing the positioning of one scan relative to another.
(1996) for discussion of what implants may be safely scanned using magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member, who can
11. If the patient is unable to hold still, provide an appropriate sedative.
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing. Sequence 1: Sagittal T1-weighted spin echo scan
Although a pilot scan can be acquired, it is not necessary here, because the first scan
2. If the procedure is a research protocol, have the patient sign any necessary consent covers almost the entire brain in a sagittal fashion. These sagittal images can then be used
form. as localizers for the application of saturation pulses when the transverse images are
3. Have the patient remove all jewelry and change into a gown to eliminate any metal acquired.
that might be found in clothing.
12. Run sagittal T1-weighted scan using the imaging sequence given in Table A7.2.2.
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating Spatial saturation pulses are applied inferiorly within the field of view. This will signifi-
and image artifacts. cantly reduce the propagation of blood flow–related artifacts and will demonstrate the full
course of the third cranial nerve seen emanating from the midbrain and heading to the
5. Inform the patient about what will occur during the procedure, what he or she will
cavernous sinus. Since aneurysms may cause cranial nerve III (and VI) deficits, this scan
experience while in the magnet, and how to behave, including the following: may demonstrate vascular compression of the nerve to best advantage.
a. If earphones or headphones are used to protect the ears from the loud sounds
produced by the gradients, the patient will be asked to wear these, but will be able Sequence 2: Transverse fast spin echo (FSE) scan
to communicate with you at any time during the imaging. 13. Run transverse fast spin echo (turbo spin echo) T2-weighted scan using the imaging
sequence given in Table A7.2.3.
b. The patient will be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this works).
c. For good results, the patient should not talk, and should avoid or minimize Table A7.2.2 Magnetic Resonance Imaging of Cranial Nerves III
swallowing or other movement, during each scan—i.e., as long as the banging to VI (Sequence 1; T1 Sagittal)
sounds continue. Between scans, talking and swallowing are allowed in most
cases, but should be avoided when comparative positional studies are being Patient Position Supine
performed; the patient will be informed when this is the case. Scan type Spin echo
Imaging plane (orientation) Sagittal
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
Central slice or volume center Mid-brain
6. Have the patient mount onto the table. Either before or right after the patient lies Echo time (TE) Minimum
down, set up any triggering devices or other monitoring equipment that is to be used. Receiver bandwidth (RBW) 16 kHz
Repeat time (TR) 400–600 msec
7. Center the patient’s head within the head coil and use the laser light to ensure Flip angle (FA) 90°
symmetry (for most brain scans, placing the center landmark at the eyebrow level Fields of view (FOVx, FOVy) 240 mm, 240 mm
allows full coverage of the relevant regions of anatomy). Make sure that the head and Resolution (Δx, Δy) 0.94 mm, 1.25 mm
neck are constrained to prevent motion, especially if high-resolution scans are to be Number of data points collected (Nx, Ny) 256, 192
run. Slice thickness (Δz) 5 mm
Generally, the patient’s head is fixed so that the head is horizontal (not tilted) and the neck Number of slices 28
and head lie along the axis of the patient table; other positions may be appropriate Slice gap 0 mm
depending on the needs at hand. Number of excitations (NEX) 1
Spatial saturation Inferior
8. If needed, place a pillow or other support under the knees to make the patient more
Slice series Interleaved
Cranial Nerves comfortable.
III to VI Scan time ∼2 min Head and Neck

A7.2.2 A7.2.3
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A7.2.3 Magnetic Resonance Imaging of Cranial Nerves III to VI Table A7.2.4 Magnetic Resonance Imaging of Cranial Nerves III to VI
(Sequence 2; T2 FSE Transverse)a (Sequence 3; FLAIR Transverse)a

Patient position Supine Patient position Supine

Scan type Fast spin echo Scan type FLAIR
Imaging plane (orientation) Transverse Imaging plane (orientation) Transverse
Central slice or volume center Mid-brain Central slice or volume center Mid-brain
Echo time (TE) 90–120 msec Echo time (TE) 133 msec
Receiver bandwidth (RBW) 20 kHz Receiver bandwidth (RBW) 16 kHz
Echo train length (ETL) 12–32 Echo train length (ETL) 12–32
Repeat time (TR) 3000–4000 msec Repeat time (TR) 8800 msec
Flip angle (FA) 90° Inversion time (TI) 2200 msec
Fields of view (FOVx, FOVy) 240 mm, 240 mm Flip angle (FA) 180°
Resolution (Δx, Δy) 0.94 mm, 0.94 mm Fields of view (FOVx, FOVy) 240 mm, 240 mm
Number of data points collected (Nx, Ny) 256, 256 Resolution (Δx, Δy) 0.94 mm, 0.94 mm
Slice thickness (Δz) 3 mm Number of data points collected (Nx, Ny) 256, 256
Number of slices Variable (∼20) Slice thickness (Δz) 5 mm
Slice gap 0 mm Number of slices Variable (∼20)
Number of excitations (NEX) 2 Slice gap 0 mm
Flow compensation Yes Number of excitations (NEX) 1
ZIP 512 Yes, changes resolution to 0.47 Flow compensation Yes
mm by 0.47 mm ZIP 512 Yes, changes resolution to 0.47
Chemical saturation Yes, fat mm by 0.47 mm
Spatial saturation Inferior Tailored RF Yes
Slice series Interleaved Spatial saturation Inferior
Scan time ∼4–5 min Slice series Interleaved
aFSE, fast spin echo. Scan time ∼4–5 min
aFLAIR, fluid attenuation inversion recovery.

Echo train length is 12 to 32, depending on slices allowed and minimum echo spacing.
Saturation pulses inferiorly and gradient moment nulling (flow compensation) are applied. Table A7.2.5 Magnetic Resonance Imaging of Cranial Nerves III to VI
The sections should be from the top of the midbrain to the mid medulla to encompass the (Sequence 4; Echo Planar/Diffusion Weighted Image Transverse)
origins of the cranial nerves. We now routinely apply frequency selective fat suppression
pulses to produce a better dynamic range of contrast and to null the skull base fat. This is Patient position Supine
of particular advantage when studying lesions or structures that traverse fatty areas such Scan type Echo planer/diffusion weighted
as the orbit (for cranial nerves III, IV, V-1, and VI) and the skull base foramina (for V-2, image
V-3). Acquisition time will be 4 to 5 min in length. For details on improvements in fat Imaging plane (orientation) Transverse
suppression techniques, see Troubleshooting. Central slice or volume center Mid-brain
Echo time (TE) Minimum
Sequence 3: Transverse FLAIR scan Receiver bandwidth (RBW) 62 kHz
14. Run transverse fluid attenuation inversion recovery (FLAIR) scan using the imaging Repeat time (TR) 10,000 msec
sequence given in Table A7.2.4. Flip angle (FA) 90°
Echo train length is 12 to 32 depending on slices allowed and minimum echo spacing. Fields of view (FOVx, FOVy) 240 mm, 240 mm
Saturation pulses inferiorly, tailored radiofrequency pulses, and gradient moment nulling Resolution (Δx, Δy) 1.88 mm, 1.88 mm
(flow compensation) are applied. The sections should be through the entire brain. Acqui- Number of data points collected (Nx, Ny) 128, 128
sition time will be 4 to 5 min in length. Since demyelinating disorders are a common cause Slice thickness (Δz) 5 mm
of cranial neuropathies, one needs to utilize FLAIR imaging because of its ability to display Number of slices Variable (∼10)
white matter lesions in the brainstem and supratentorial compartment in patients who have Slice gap 0 mm
cranial nerve III to VI deficits.
Number of excitations (NEX) 1
Slice series Interleaved
Sequence 4: Transverse echo planar/diffusion weighted scan
Scan time 1 min
15. Run transverse echo planar/diffusion weighted imaging using the imaging sequence
given in Table A7.2.5.
Cranial Nerves
III to VI Head and Neck

A7.2.4 A7.2.5
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 The minimum diffusion weighting should have a “b” value of at least 1000 sec/mm2. These Table A7.2.7 Magnetic Resonance Imaging of Cranial Nerves III
sections should be applied through the entire brain, and aid in the detection of acute stroke to VI (Sequence 6; Thin Transverse T1 Post-Contrast)
processes.
Patient position Supine
Sequence 5: Transverse T2-weighted spin echo scan Scan type Spin echo
16. Run transverse fast spin echo (turbo spin echo) thin cut imaging should also be Imaging plane (orientation) Transverse
employed using the imaging sequence given in Table A7.2.6. Central slice or volume center Mid-brain
Echo time (TE) Minimum
Selective chemical (fat) suppression should be utilized to decrease the signal from skull-
Receiver bandwidth (RBW) 16 kHz
base fat.
Repeat time (TR) 400–600 msec
Sequence 6: Transverse post-contrast T1-weighted scan Flip angle (FA) 90°
17. Leave the patient in the magnet, inject the contrast agent, and flush the line with 10 Fields of view (FOVx, FOVy) 200 mm, 200 mm
ml saline. Resolution (Δx, Δy) 0.78 mm, 0.78 mm
Number of data points collected (Nx, Ny) 256, 256
A dose of 0.1 mmol/kg of contrast agent is usually given. Slice thickness (Δz) 3 mm
18. Run transverse post-gadolinium T1-weighted scans using the imaging sequence given Number of slices Variable (∼20)
in Table A7.2.7. Slice gap 0 mm
Number of excitations (NEX) 2–3
Saturation pulses inferiorly placed and gradient moment nulling (flow compensation) are Flow compensation Yes
applied. Again these scans should be focused on the brainstem and cavernous sinus region Spatial saturation Inferior
and thus will also include parts of the orbit (though a separate orbit protocol may be
Slice series Interleaved
necessary to best visualize orbital pathology). It is imperative to utilize an inferior spatial
saturation pulse in conjunction with gradient moment nulling. Although the addition of Scan time ∼4 min
gradient moment nulling may slightly increase your overall effective TE, we have not found
this slight increase to be of any real clinical significance in altering the T1 of the tissue
examined. Table A7.2.8 Magnetic Resonance Imaging of Cranial Nerves III to VI
(Sequence 7; T1 Coronal Post-Contrast)

Patient position Supine

Table A7.2.6 Magnetic Resonance Imaging of Cranial Nerves III to VI Scan type Fast spin echo
(Sequence 5; Thin T2 FSE with Fat Saturation)
Imaging plane (orientation) Coronal
Central slice or volume center Mid-brain
Patient position Supine
Echo time (TE) Minimum
Scan type Fast spin echo
Receiver bandwidth (RBW) 16 kHz
Imaging plane (orientation) Transverse
Echo train length (ETL) ≤4
Central slice or volume center Mid-brain
Repeat time (TR) 400–600 msec
Echo time (TE) 90–120 msec
Flip angle (FA) 90°
Receiver bandwidth (RBW) 20 kHz
Field of view (FOVx, FOVy) 200 mm, 200 mm
Echo train length (ETL) 12–32
Resolution (Δx, Δy) 0.78 mm, 0.78 mm
Repeat time (TR) 3000–4000 msec
Number of data points collected (Nx, Ny) 256, 256
Flip angle (FA) 90°
Slice thickness (Δz) 3 mm
Fields of view (FOVx, FOVy) 200 mm, 200 mm
Number of slices Variable (∼20)
Resolution (Δx, Δy) 0.78 mm, 0.78 mm
Slice gap 0 mm
Number of data points collected (Nx, Ny) 256, 256
Number of excitations (NEX) 2–3
Slice thickness (Δz) 3 mm
Flow compensation Yes
Number of slices Variable (∼20)
ZIP 512 Yes, changes resolution to 0.39
Slice gap 0 mm
mm by 0.39 mm
Number of excitations (NEX) 2
Chemical saturation Fat saturation
Flow compensation Yes
Spatial saturation Inferior
ZIP 512 Yes, changes resolution to 0.39
Slice series Interleaved
mm by 0.39 mm
Scan time ∼4 min
Chemical saturation Fat saturation
Spatial saturation Inferior
Slice series Interleaved
Cranial Nerves
III to VI Scan time ∼4 min Head and Neck

A7.2.6 A7.2.7
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Sequence 7: Coronal post-contrast T1-weighted scan Table A7.2.9 Magnetic Resonance Imaging of Cranial Nerves III to VI
19. Run coronal post-gadolinium spin echo T1-weighted scan using the imaging sequence (Sequence 8; Optional MOTSA MRA)a
given in Table A7.2.8.
Patient position Supine
Saturation pulses inferiorly placed and gradient moment nulling (flow compensation) is Scan type 3-D gradient echo
encouraged. The sections should begin at the posterior brainstem margin (where cranial Imaging plane (orientation) Transverse
nerve IV leaves) and extend anteriorly to include the orbits. The field of view should be at Central slice or volume center Mid-brain
least 20 cm, to include parts of the skull base and head and neck supplied by the branches
Echo time (TE) 6.9 msec
of the trigeminal nerve. Frequency-selective fat suppression is useful. The sequence can be
performed as a fast spin echo T1-weighted scan, but the matrix should be ZIPed to 512 by Receiver bandwidth (RBW) 16 kHz
512, no more than an echo train length of 4 should be used, and fat suppression should be Repeat time (TR) 35 msec
applied. Acquisition time will be 2 to 4 min in length. Flip angle (FA) 30°
Field of view (FOVx, FOVy) 240 mm, 180 mm
ALTERNATE DEMYELINATING ETIOLOGIES: ANEURYSMS Resolution (Δx, Δy) 0.47 mm, 0.70 mm
PROTOCOL Number of data points collected (Nx, Ny) 512, 256
If demyelinating etiologies for the cranial nerve deficits are considered, then a more Slice thickness (Δz) 1.2 mm
extensive T2-weighted evaluation of the brain may be indicated. The post-gadolinium Number of slices Variable (∼50)
scans should be extended to cover the whole brain. If one identifies lesions that enhance Slice gap 0
and lesions that do not, it would imply a polyphasic disease such as multiple sclerosis. Number of excitations (NEX) 1
Flow compensation Yes
As indicated previously, aneurysms of nearby vessels may impinge on cranial nerves,
ZIP 512 Yes, changes resolution to 0.23
thereby causing deficits. In the case of cranial nerve III, these are usually posterior mm by 0.35 mm
communicating artery aneurysms although the nerve also courses between the posterior ZIP 2 Yes
cerebral and superior cerebellar arteries. Cranial nerve V may be impacted by branches Extended dynamic range (EDR) Yes
of the basilar artery as well and lead to trigeminal neuralgia (“tic doulereux”; Chong, Chemical saturation Fat saturation
1996; Hutchins et al., 1989). Cavernous sinus aneurysms may lead to cranial nerve VI Spatial saturation Superior
palsies, as this nerve is in close proximity to the cavernous carotid artery. For these Magnetization transfer Yes
reasons, one should scrutinize the spin echo scans for any signs of unusual flow voids that Slice series Interleaved
might indicate an aneurysm, and have a very liberal threshold for performing an MR Scan time ∼8 min
angiogram of the intracranial vessels. We have found that the utilization of a MOTSA aUtilize ramped RF pulse in the direction of arterial blood flow. Utilize an overlap of ∼10 mm in
(multiple overlapping thin slice acquisition) proves to be a very predictable and homoge- between adjacent slabs (completed as needed).
neous time-of-flight technique. With the addition/employment of MOTSA, we often find
it necessary to ramp the RF pulse across the imaging volume.
COMMENTARY
Sequence 8: MRA (MR angiography) scan Background Information should be evaluated for a source of such drop
An optional MRA using the imaging sequence given in Table A7.2.9 to evaluate aneu- There are a plethora of lesions that may metastases.
rysms and/or other pathologies may be required in conjunction with conventional imag- cause cranial nerve III to VI pathology, and With the advent of a ramped RF pulse and
ing. The decision to complete the optional MRA should be made prior to the therefore these studies often require more the utilization of MOTSA, the resultant
administration of intravenous contrast medium. A 3-D vascular time-of-flight (TOF) clinical input and physician monitoring to tai- angiogram is very homogeneous and there is
lor the evaluation to the clinical and imaging an overall reduction in the apparent dephasing
utilizing a technique known as MOTSA (multiple overlapping thin slice acquisition)
findings. A pupil-sparing third-nerve palsy of blood across the imaging volume. “Ramped
typically provides the clinician with a very good assessment of the vascular anatomy. A
might suggest an aneurysm, so one must add RF” actually adjusts the flip angle of the pulse
TR of 35 msec, TE = 6.9 msec should be used so that fat and water are out of phase with an MR angiography (MRA). Add an optic sequence to take into account the normal
one another. This “out of phase” TE should reduce the contribution of orbital fat signal neuritis history to the mix and the study be- dephasing pattern of blood as it traverses
overlapping structures such as the ophthalmic arteries. Take note that the TE will vary with comes a search for demyelinating plaques. through the transverse imaging slab. Thus, this
field strength in order to have fat and water out of phase. At 1.5 T, every 2.3 msec, fat and Add pain and one might be looking at a vas- should minimize an artifact that typically oc-
water go out of phase and back in phase, respectively. The slice thickness is 1.2 mm with cular compression syndrome. The key here is curs during 3-D TOF imaging in which there is
a ZIP 2 (50% overlap slice interpolation) and an imaging matrix of 512 by 256 that is to evaluate the individual slices of the MRA sufficient SNR on one aspect of the slab, and,
ZIPed to 512 by 512. A maximum field of view (FOV) should be (24 cm, 18 cm) and it searching for bright vessels impacting dark as you get to the opposite side there is a drastic
is suggested to utilize magnetization transfer and a ramped RF pulse as imaging options nerves. If there are multiple cranial nerve defi- drop in SNR. Evaluation of the raw data is often
to improve the angio quality. cits, one might be dealing with neurofibroma- critical, especially where vascular compression
tosis, subarachnoid seeding on nerve roots, is a possible etiology for the symptoms. The
sarcoidosis, Lyme disease, granulomatous in- addition of the MRA should be made prior to
fections, or other conditions, all of which re- the administration of intravenous contrast me-
quire greater scrutiny of the post-contrast dium (MRA is the subject of UNIT B7.3).
Cranial Nerves
III to VI scans (Yousem et al., 1990b). The whole brain Angiograms that are done in the reverse order Head and Neck

A7.2.8 A7.2.9
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 7
 will yield an overall appearance of not having
a suppressed venous flow. This can make it
difficult to access subtle lesions/pathologies
that are simply overlapped by large prominent
veins. If the decision to perform an MR
angiogram is made after contrast medium has
already been given, it may be advantageous to
the physician interpreting the exam to have the
patient return the next day for a non-contrast-
enhanced MRA of the area in question. If the
patient is not able to return for the conventional
TOF (time-of-flight) angiography, then the im-
aging professional should make attempts to
utilize an angiographic technique that does not
employ TOF effects. Phase-contrast angiogra-
phy (PCA) employs techniques that are unique
because it is based on velocity-induced phase
shifts to produce an image. Thus your success
with PCA depends largely on selecting the
correct Venc (velocity encoding) within the ves-
sel you are attempting to image. One major
difference that you will note between TOF
imaging and PCA imaging is the significant
improvement in background suppression with
PCA. PCA employs two bipolar gradients in
which there is essentially a subtraction of signal
from stationary tissue. Only flowing spins
against the gradient will be demonstrated.
Finally, one should be aware that sellar and
clival masses commonly present with cranial
nerve VI symptoms (Yousem et al., 1989,
1990a). Sagittal post-gadolinium scans may be
useful to demonstrate the meningeal or extrac-
ranial extent of such lesions as chordomas,
meningiomas, and nasopharyngeal carcinomas
that may lead to such symptoms.
Critical Parameters
The accurate prescription of the slice loca-
tions is the most important step in this process.
If one uses the sagittal scan and prescribes
transverse slices from the top of the midbrain
to the mid-medulla, one should have appropri-
ate coverage. Cranial nerve VI courses from the
pontomedullary junction superiorly to enter the
inferior petrosal sinus, crosses Dorello’s canal,
and then runs within the cavernous sinus. One
must be able to see all these segments well. In
the end it is usually the post-gadolinium scans
that are the most revealing, so one should take
the time to repeat these if they are of poor
quality.
Troubleshooting
One of the most difficult tasks to accomplish
Cranial Nerves is homogenous fat saturation within your im-
III to VI
A7.2.10
Supplement 7
aging volume. Typically the overall fat suppres-
sion is at its worst when the imaging volume is
adjacent to an air/tissue interface and local
magnetic field inhomogeneities (i.e., secon-
dary to metallic dental work). Although the
scanners of today can often complete this task
automatically, there is no better way of improv-
ing your fat suppressed image than by manually
adjusting the center frequency and the position
of the RF pulse that will impose the fat suppres-
sion.
When imaging post contrast, more often
than not there will be a significant increase in
flow artifacts over your imaging volume. This
will make it difficult to assess very subtle le-
sions. What the authors have found is that, with
the simultaneous addition of gradient moment
nulling and inferior spatial saturation pulses, a
drastic improvement of the image quality and
directly decreased motion across the anatomy
have been achieved. It should be noted that with
the addition of gradient moment nulling, there
typically will be an increase in the minimum
TE that may slightly (in theory) adjust the T1
weighting of your image. The authors have not
found this slight increase in the effective TE to
be of any real clinical significance.
When there is a need to perform an MRA to
exclude aneurysm and you are utilizing a tech-
nique such as MOTSA, take note of two very
real pitfalls that can significantly hamper your
results. First make sure that the ramped RF
pulse is in the direction that the flow is going.
“Go with the flow” is an easy way to remind
one’s self which way to utilize this unique tool.
If it is prescribed in the opposite direction, it
will give results that are not conducive to good
angiographic imaging. The next thing to re-
member while utilizing an MOTSA technique
is the key word “overlapping.” There must be
an adequately applied overlap between adja-
cent slabs in order to insure an even flowed
appearance. Adequate overlapping of ∼10 mm
between slabs is usually sufficient.
Anticipated Results
You should always see cranial nerves III and
V in the cisterns. Cranial nerve IV may be seen
as a fine line emanating from behind the mid-
brain and traversing the ambient cistern. The
foramen ovale and rotundum are usually out-
lined by enhancing perneural vascular plexi.
The cavernous sinus should enhance brightly
along with the pituitary stalk. If it is not enhanc-
ing, check your intravenous access, as you may
have infiltrated the contrast dye.
Current Protocols in Magnetic Resonance Imaging
Time Considerations
Since the post-contrast and the T2-weighted
scans are the most valuable in this scenario, you
can skimp on some signal averages if necessary
to complete the uncooperative patient’s study.
Sequences such as fast recovery, fast spin echo
have been developed to create T2 contrast with
much shorter TR values. These sequences in-
corporate additional RF pulses that will drive
up longitudinal magnetization. This allows the
user to “tip” the protons back into the transverse
plane in a relatively shorter period of time,
thereby decreasing the overall imaging time
without sacrificing T2 contrast.
Index of Terms
The clinical imaging instructions and termi-
nology utilized in this unit are primarily geared
towards General Electric equipment. An Index
of Terms is provided that should bridge the gap
of vendor specific terminology.
Chemical saturation A technique that
applies an additional radiofrequency (RF)
pulse (at a desired distance from the center
frequency) to selectively suppress a tissue. This
technique can be utilized to suppress the signal
from water, fat, or silicone.
Echo time (TE) The time that is measured
from the initiation of the initial RF (radio fre-
quency) pulse and the peak of the echo.
Echo train length (ETL) In fast spin echo
or turbo spin echo imaging, the ETL will actu-
ally equal the number of echoes prescribed per
TR. Successive 180° refocusing pulses are ap-
plied to “rephase the dephasing” protons in an
effort to maximize the number of lines of k-
space per TR. The formula for scan time in
relation to Fast Spin Echo imaging and ETL is
as follows:
Scan time = (TR) × (no. of phase
encoding steps/ETL) × (NEX)
Extended dynamic range (EDR) An im-
aging-enhancement tool that will allow the
utilization of 32-bit data processing as opposed
to the standard 16-bit processor. EDR in this
way should improve SNR and resolution, but
will utilize twice as much memory as a conven-
tional acquisition.
Flow compensation Sometimes more
widely expressed as gradient moment nulling,
this is a way in which the system places flowing
or moving spins into “phase coherence” with
stationary spins.
Inversion time (TI) With inversion recov-
ery pulse sequences, typically the inversion
time is the time from the first 180° RF pulse to
Current Protocols in Magnetic Resonance Imaging
the center of the next 90° RF pulse. This inver-
sion time will essentially “null” the desired
tissue depending on how long or short the TI
(inversion time) selected is, and the T1 relaxa-
tion time of the corresponding tissue.
No phase wrap (NPW) Will prevent wrap
around artifacts (also known as aliasing arti-
facts) in the phase encoding direction. NPW
should only be used when necessary because
of the following:
a. NPW doubles the FOV in the phase direc-
tion.
b. NPW essentially doubles the phase en-
coding steps (to maintain resolution).
c. With NPW, you must reduce the NEX by
half in order to maintain scan time.
Number of excitations (NEX) is a factor
that is utilized to calculate the overall scan time
and will directly effect the SNR. NEX is essen-
tially the number of times that data are sampled
per acquisition. Note that increasing the NEX
to achieve overall better SNR is a rather ineffi-
cient way to improve signal. Doubling the NEX
from 2 to 4 will only yield a 40% increase in
SNR while it doubles scan time. Take note of
the formula for scan time:
Scan time = (TR) × (no. of phase
encoding steps) × (NEX)
Receiver bandwidth (RBW) The range
of frequencies that the MRI scanner is actually
“tuned” to receive. This will directly affect the
overall SNR. This will not be done by increas-
ing or decreasing the signal, but rather there
will be an increase and/or decrease in the
amount of noise received relative to the altera-
tion of the RBW. An increase in the RBW will
increase the range of frequencies that the scan-
ner will evaluate and thus decrease the overall
SNR. In comparison, utilizing a narrow band-
width should yield less noise and improve the
overall SNR. The relationship of the receiver
bandwidth and SNR can be thought of as in-
versely proportional to the square root of the
bandwidth.
Rectangular field of view (REC FOV)
Asymmetric field of view (typically in the
phase encoding direction). REC FOV is typi-
cally utilized when a body part is longer in one
direction than another. By utilizing an asym-
metric FOV, the system will not collect a por-
tion of the data, thereby decreasing the scan-
ning time.
Repetition time (TR) The time in a pulse
sequence between successive excitation pulses.
Spatial saturation employs an additional
RF pulse to cause moving spins within a deter- Head and Neck
A7.2.11
 mined area to be selectively dephased. This
application will reduce motion from flow
and/or respiratory artifacts and will limit the
number of slices per TR in general. Since a
spatial saturation pulse employs additional RF,
which is to be deposited into the patient, special
attention should be focused on the SAR (spe-
cific absorption rate). Today all MRI scanners
have a program that internally monitors how
much RF can be applied over a given period of
time. This formula takes into account the pa-
tient’s body weight. This actual body weight
needs to be accurately input at all times for
patient safety. An inappropriate weight will
cause the improper limit of RF to be transmitted
relative to a safe period of time in which this is
to occur.
Tailored RF An imaging option that im-
proves image quality on fast spin echo (FSE)
sequences with relatively short TE’s. Tailored
RF will improve edge blurring by reducing
overall echo spacing.
ZIP 512/ZIP 1024 Better known as “zer-
ofill interpolation process,” this is a reconstruc-
tion algorithm that allows the user to scan at a
256 × 256 matrix and then the data are zero-
filled to a 512 × 512 matrix (or 1024 × 1024
respectively).
ZIP 2/ZIP 4 Slice zip essentially is also a
“zerofill interpolation process” that will create
additional slices through the interpolation pro-
cedure. These slices are created with an offset
of 50% of the original imaging slice locations.
Cranial Nerves
III to VI
A7.2.12
Literature Cited
Castillo, M. and Mukherji, S.K. 1996a. Magnetic
resonance imaging of cranial nerves IX, X, XI,
and XII. Top. Magn. Reson. Imaging 8:180-186.
Castillo, M. and Mukherji, S.K. 1996b. MRI of
enlarged dorsal ganglia, lumbar nerve roots, and
cranial nerves in polyradiculoneuropathies.
Neuroradiology 38:516-520.
Chong, V.F. 1996. Trigeminal neuralgia in naso-
pharyngeal carcinoma. J. Laryngol. Otol.
110:394-396.
Hutchins, L.G., Harnsberger, H.R., Hardin, C.W.,
Dillon, W.P., Smoker, W.R., and Osborn, A.G.
1989. The radiologic assessment of trigeminal
neuropathy. AJR Am. J. Roentgenol. 153:1275-
1282.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Yousem, D.M., Atlas, S.A., Grossman, R.I., Sergott,
R.C., Savino, P.I., and Bosley, T.C. 1989. MR of
Tolosa-Hunt syndrome. Am. J. Neuroradiol.
10:1181-1184.
Yousem, D.M., Arrington, J.A., Kumar, A.J., and
Bryan, R.N. 1990a. Bright lesions on sel-
lar/parasellar T1-weighted scans. Clin. Imaging
14:99-105.
Yousem, D.M., Patrone, P.L., and Grossman, R.I.
1990b. Leptomeningeal metastases: MR evalu-
ation. J. Comput. Assist. Tomogr. 14:255-261.
Contributed by Robert W. Evers and
David M. Yousem
The Johns Hopkins Hospital
Baltimore, Maryland
Current Protocols in Magnetic Resonance Imaging
Cranial Nerves VII To VIII UNIT A7.3
Currently a debate is raging as to whether, with the high-resolution scanning currently
available, one needs to perform contrast-enhanced scans to exclude vestibular schwan-
nomas in patients with sensorineural hearing loss (Allen et al., 1996; Chakeres and
Schmalbrock, 1997; Curtin, 1997; El-Gammal, 1997; Hudgins, 1998; Jackler, 1996). This
is a frequent indication for ordering studies in the elderly, and cost-containment consid-
erations are often cited in the move to eliminate extra sequences and contrast agent
expenses. Cranial nerves VII and VIII are well seen outlined by cerebrospinal fluid (CSF)
in most patients and are the nerves most commonly involved with neurogenic neoplasms.
The reasons why contrast-enhanced scans are still being performed are explained in this
unit, but revolve around pathologies other than schwannomas of the nerves.
IMAGING OF CRANIAL NERVES VII TO VIII BASIC
PROTOCOL
Cranial nerves VII and VIII are usually affected by lesions in the cerebellopontine angle,
internal auditory canal, and temporal bone. It is true, however, that cranial nerve VII then
leaves the temporal bone to course through the parotid gland to innervate the muscles of
facial expression. Therefore, though the typical evaluation of these nerves focuses on their
intracranial and intraosseous portions, one must scrutinize the sagittal and coronal scans
for the parotid glands, where lesions affecting the facial nerve may reside. In the case of
congenital sensorineural hearing loss, a dedicated temporal bone study with high-resolu-
tion T2-weighted scans may be contemplated.
Imaging of the cranial nerves requires a focused approach based on clinical symptoma-
tology and signs. Thin-section imaging, high resolution, and contrast enhancement are
essential components to this evaluation.
Table A7.3.1 lists the hardware necessary to perform the procedure, along with appropri-
ate parameters. Standard head coil imaging is sufficient to obtain adequate scans on the
intracranial portions of these cranial nerves. Anatomic imaging with high-quality fast spin
echo T2-weighted scan with the dark nerves outlined by bright CSF is required to see the
portions of these nerves coursing through the cerebellopontine (CP) angle and internal
auditory canal (IAC; Marsot-Dupuch et al., 1999; Iwayama et al., 1999). Gadolinium-en-
hanced scans are critical to the complete evaluation of these nerves, though the evaluation
may be limited to sections through the lower pons and medulla, including the full range
of the temporal bone. These scans should be centered on the internal auditory canal. No
monitoring is required.
If a congenital source of hearing loss is the basis of the study, then one must know the
scanner well enough to know whether surface coils will be required to obtain an adequate
thin section evaluation of the cochlea, vestibule, semicircular canals, cochlear viaduct,
Table A7.3.1 Equipment Parameters for Cranial
Nerve Imaging
Coil type Head
Gradient coil strength 25 mT/m
Flow compensation pulse Yes, as needed
Peripheral gating Not appropriate
Motion cushions Useful
Use of contrast agents Yes
Head and Neck
Contributed by Robert W. Evers and David M. Yousem A7.3.1
Current Protocols in Magnetic Resonance Imaging (2001) A7.3.1-A7.3.13
Copyright © 2001 by John Wiley & Sons, Inc.
 and vestibular aqueduct. Most 1.0 T or higher magnets will have adequate signal-to-noise
ratio to allow the thin section scanning (2 mm or less) that is required for this indication.
These scans will be performed at 512 by 512 matrices. At lower field strengths, or to
optimize signal-to-noise ratio at higher field strengths, placing 5-in. round general-pur-
pose surface coils over the ears may be advantageous. However for the traditional “rule
out acoustic schwannoma,” these extra preparations are not necessary.
Materials
Normal saline (0.9% NaCl), sterile
Gadolinium-based contrast agent (e.g., Magnevist, Omniscan, or Prohance)
Set up equipment and patient
1. Interview (screen) the patient to ensure that he or she has no contraindications such
as cardiac pacemakers or other implants containing ferromagnetic materials. Also be
sure to find out if the patient has any health conditions that may require the presence
of special emergency equipment during the scanning procedure, or necessitate any
other precautions.
Generally standard screening forms are used for all patients scanned in a magnetic
resonance system.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(1996) for discussion of what implants may be safely scanned using magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
2. If the procedure is a research protocol, have the patient sign any necessary consent
form.
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating
and image artifacts.
5. Inform the patient about what will occur during the procedure, what he or she will
experience while in the magnet, and how to behave, including the following:
a. If earphones or headphones are used to protect the ears from the loud sounds
produced by the gradients, the patient will be asked to wear these, but will be able
to communicate with you at any time during the imaging.
b. The patient will be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this works).
c. For good results the patient should not talk, and should avoid or minimize
swallowing or other movement, during each scan—i.e., as long as the banging
sounds continue. Between scans, talking and swallowing are allowed in most
cases, but should be avoided when comparative positional studies are being
performed; the patient will be informed when this is the case.
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
6. Help the patient mount onto the table. Either before or right after the patient lies down,
set up any triggering devices or other monitoring equipment that is to be used.
Cranial Nerves
VII To VIII
A7.3.2
Current Protocols in Magnetic Resonance Imaging
7. Center the patient’s head within the head coil and use the laser light to ensure
symmetry. For most brain scans, placing the center landmark at the eyebrow level
allows full coverage of the relevant regions of anatomy. Make sure that the head and
neck are constrained to prevent motion, especially if high-resolution scans are to be
run.
Generally the patient’s head is fixed so that the head is horizontal (not tilted) and the neck
and head lie along the axis of the patient table; other positions may be appropriate
depending on the needs at hand.
8. If needed, place a pillow or other support under the knees to make the patient more
comfortable.
9. Establish an intravenous line from which the contrast agent can be injected, and attach
this line securely to the patient so that movement into or out of the magnet will not
pull at the patient’s arm.
It is preferable to insert the line prior to imaging and to leave the patient in the magnet,
with no intervening motion, between the scans run before contrast agent injection and those
run after injection.
10. Use the centering light to position the patient and put him or her into the center of
the magnet.
Once this step has been performed, so long as the patient does not move on the table, the
table itself can be moved and then replaced in the same position as before without
jeopardizing the positioning of one scan relative to another.
11. If the patient is unable to hold still, provide an appropriate sedative.
Sequence 1: Sagittal T1-weighted spin echo scan
Although a pilot scan can be acquired, it is not necessary here, because the first scan
covers almost the entire brain in a sagittal fashion. These sagittal images can then be used
as localizers for the application of saturation pulses when the transverse images are
acquired.
12. Run sagittal T1-weighted scan (see Table A7.3.2).
Saturation pulses are inferiorly placed. Often this scan best demonstrates the course of the
2nd genu and descending intramastoid portion of the facial nerve outlined by the air-filled
mastoid bone. The exit from the stylomastoid foramen is well depicted, and the parotid
glands are seen in their entirety when 5 mm interleaved scans are prescribed ear to ear.
One can also assess the degree to which the temporal bones are fat-filled (marrow fat) and
whether one needs to place fat-suppression pulses for post-gadolinium studies.
Sequence 2: Transverse fast spin echo (FSE) scan
13. Run transverse fast spin echo T2-weighted scan (see Table A7.3.3).
Echo train length 12-32 depending on slices allowed and minimum echo spacing. Satura-
tion pulses inferiorly and gradient moment nulling (flow compensation) should be applied.
One now routinely applies frequency selective fat suppression pulses to produce a better
dynamic range of contrast and to null the skull base and temporal bone fat. Acquisition
time will be 3 minutes in length. Since hearing abnormalities may be affected along the
full course of the auditory system, this study and the FLAIR that follows encompass the
whole brain.
Sequence 3: Transverse FLAIR scan
14. Run transverse fluid attenuation inversion recovery (FLAIR) scan (see Table A7.3.4).
Head and Neck
A7.3.3
Current Protocols in Magnetic Resonance Imaging
 Table A7.3.2 Sagittal T1-Weighted Spin Echo Scan (Sequence 1)a Table A7.3.4 Transverse Fluid Attention Inversion Recovery (FLAIR) Scan
(Sequence 3)a
Patient position Supine
Scan type Spin echo Patient position Supine
Imaging plane (orientation) Sagittal Scan type Inversion recovery fast spin echo
Central slice or volume center Mid-brain Imaging plane (orientation) Transverse
Echo time (TE) Minimum Central slice or volume center Mid-brain
Receiver bandwidth (RBW) 16 kHz Echo time (TE) 133 msec
Repeat time (TR) 400–600 msec Receiver bandwidth (RBW) 16 kHz
Flip angle (FA) 90° Echo train length (ETL) 12–32
Fields of view (FOVx, FOVy) 240 mm, 240 mm Repeat time (TR) 8800 msec
Resolution (Δx, Δy) 0.94 mm, 1.25 mm Inversion time (TI) 2200 msec
Number of data points collected (Nx, Ny) 256, 192 Flip angle (FA) 180°
Slice thickness (Δz) 5 mm Fields of view (FOVx, FOVy) 240 mm, 240 mm
Number of slices 28 Resolution (Δx, Δy) 0.94
Slice gap 0 mm Number of data points collected (Nx, Ny) 256, 256
Number of excitations (NEX) 1 Slice thickness (Δz) 5 mm
Spatial saturation Inferior Number of slices Variable (∼20)
Slice series Interleaved Slice gap 0 mm
Scan time ∼2 min Number of excitations (NEX) 1
aThis sequence covers entire brain. Flow compensation Yes
ZIP 512 Yes, changes resolution to 0.47
mm by 0.47 mm
Table A7.3.3 Transverse Fast Spin Echo T2-Weighted scan (Sequence 2)a,b
Tailored RF Yes
Spatial saturation Inferior
Patient position Supine
Slice series Interleaved
Scan type Fast spin echo
Scan time ∼4–5 min
Imaging plane (orientation) Transverse
aThis sequence covers entire brain.
Central slice or volume center Mid-brain
Echo time (TE) 90–120 msec
Receiver bandwidth (RBW) 20 kHz
Echo train length (ETL) 12–32 Echo train length 12-32 depending on slices allowed and minimum echo spacing. Satura-
Repeat time (TR) 3000–4000 msec tion pulses inferiorly placed, tailored radiofrequency pulses and gradient moment nulling
Flip angle (FA) 90° (flow compensation) are applied. The sections should be through the entire brain. Acqui-
Fields of view (FOVx, FOVy) 240 mm, 240 mm sition time will be 4 to 5 minutes in length.
Resolution (Δx, Δy) 0.94 mm, 1.25 mm Demyelinating disorders may lead to facial nerve and auditory symptoms, so one should
Number of data points collected (Nx, Ny) 256, 192 use a sequence that does well at displaying the white matter lesions in the brainstem,
Slice thickness (Δz) 5 mm infratentorial, and supratentorial compartments.
Number of slices Variable
Slice gap 0 mm Sequence 4: Transverse echo planar/diffusion weighted scan
Number of excitations (NEX) 2 15. Run transverse echo planar/diffusion weighted imaging (see Table A7.3.5).
Flow compensation Yes
The minimum diffusion weighting should have a “b” value of at least 1000 s/mm2. These
ZIP 512 Yes, changes resolution to 0.47 sections should be applied through the entire brain and aid in the detection of acute stroke
mm by 0.63 mm processes.
Chemical saturation Yes, fat
Spatial saturation Inferior Sequence 5: Transverse T2-weighted spin echo scan
Slice series Interleaved When there is a question of sensorineural hearing loss, one must evaluate the VII and VIII
Scan time ∼3 min nerves with a very high-resolution imaging technique
aThis sequence covers entire brain.
bTo improve the overall fat suppression at air tissue interfaces, complete the manual fat suppression to 16. Run a transverse heavily T2-weighted fast spin echo (see Table A7.3.6).
selectively suppress fat.
One should employ fat saturation techniques, gradient moment nulling (flow compensa-
tion), and inferior placement of a spatial saturation pulse.
Cranial Nerves
VII To VIII Head and Neck

A7.3.4 A7.3.5
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A7.3.5 Transverse Echo Planar/Diffusion Weighted Imaging (Sequence Table A7.3.7 Transverse Post-Gadolinium T1-Weighted Scan (Sequence 6)
4)a
Patient position Supine
Patient position Supine Scan type Spin echo
Scan type Echo planar/diffusion weighted Imaging plane (orientation) Transverse
image Central slice or volume center Internal auditory canals
Imaging plane (orientation) Transverse Echo time (TE) Minimum
Central slice or volume center Mid-brain Receiver bandwidth (RBW) 16 kHz
Echo time (TE) Minimum Repeat time (TR) 400–600 msec
Receiver bandwidth (RBW) 62 kHz Flip angle (FA) 90°
Repeat time (TR) 10,000 msec Fields of view (FOVx, FOVy) 180 mm, 180 mm
Flip angle (FA) 90° Resolution (Δx, Δy) 0.70 mm, 0.70 mm
Fields of view (FOVx, FOVy) 240 mm, 240 mm Number of data points collected (Nx, Ny) 256, 256
Resolution (Δx, Δy) 1.88 mm, 1.88 mm Slice thickness (Δz) 3 mm
Number of data points collected (Nx, Ny) 128, 128 Number of slices Variable (∼20)
Slice thickness (Δz) 5 mm Slice gap 0 mm
Number of slices Variable (∼10) Number of excitations (NEX) 2–3
Slice gap 0 mm Flow compensation Yes
Number of excitations (NEX) 1 Spatial saturation Inferior
Slice series Interleaved Slice series Interleaved
Scan time 1 min Scan time ∼5–7 min
aThis sequence covers entire brain.

Table A7.3.8 Coronal Post-Gadolinium Spin Echo T1-Weighted Scan

Table A7.3.6 Question of Sensorineural Hearing Loss (Transverse (Sequence 7)a
Heavily T2-Weighted Fast Spin Echo; Sequence 5)a
Patient position Supine
Patient position Supine
Scan type Fast spin echo
Scan type Fast spin echo
Imaging plane (orientation) Coronal
Imaging plane (orientation) Transverse
Central slice or volume center Mid-brain
Central slice or volume center Internal auditory
canals Echo time (TE) Minimum
Echo time (TE) 200 msec Receiver bandwidth (RBW) 16 kHz
Receiver bandwidth (RBW) 24 kHz Echo train length (ETL) ≤4
Echo train length (ETL) 24 Repeat time (TR) 400–600 msec
Repeat time (TR) 4000–5000 msec Flip angle (FA) 90°
Flip angle (FA) 90° Fields of view (FOVx, FOVy) 180 mm, 180 mm
Fields of view (FOVx, FOVy) 180 mm, 180 mm Resolution (Δx, Δy) 0.70 mm, 0.70 mm
Resolution (Δx, Δy) 0.35 mm, 0.35 mm Number of data points collected (Nx, Ny) 256, 256
Number of data points collected (Nx, Ny) 512, 512 Slice thickness (Δz) 3 mm
Slice thickness (Δz) 2 mm Number of slices Variable
Number of slices Variable (∼20) Slice gap 0 mm
Slice gap 0 mm Number of excitations (NEX) 2–3
Number of excitations (NEX) 4 Flow compensation Yes
Flow compensation Yes ZIP 512 Yes, changes resolution to 0.35
mm by 0.35 mm
Chemical saturation Yes, fat
Chemical saturation Yes, fat
Spatial saturation Inferior
Spatial saturation Inferior
Slice series Interleaved
Slice series Interleaved
Scan time ∼4 min
Scan time ∼2–4 min
aTo improve the overall fat suppression at air tissue interfaces, complete the manual fat
aTo improve the overall fat suppression at air tissue interfaces, complete the manual fat suppression to
suppression to selectively suppress fat.
selectively suppress fat.

Cranial Nerves
VII To VIII Head and Neck

A7.3.6 A7.3.7
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Sequence 6: Transverse post-contrast T1-weighted scan Table A7.3.9 2D Fast Spin Echo (Sequence 8)a,b
17. Leave the patient in the magnet, inject the contrast agent, and flush the line with 10
ml saline. Patient position Supine
Scan type Fast spin echo
A dose of 0.1 mmol/kg of contrast agent is usually given. Imaging plane (orientation) Transverse
18. Run transverse post-gadolinium T1-weighted scan (see Table A7.3.7) through the Central slice or volume center Mid-brain
internal auditory canals. Echo time (TE) 180 msec
Receiver bandwidth (RBW) 32 kHz
Saturation pulses inferiorly placed and gradient moment nulling (flow compensation) Echo train length (ETL) 24
should be applied. These scans are performed without fat suppression (as opposed to the
Repeat time (TR) 3000-5000 msec
coronal scans that follow) so as to visualize the rare lipomas or hemorrhagic lesions that
can affect the CP angle, IAC, and/or the temporal bone. One must be able to translate the Flip angle (FA) 90°
transverse scans to the coronal scans so as to interpret the suppressed and nonsuppressed Fields of view (FOVx, FOVy) 240 mm, 240 mm
images accurately. If necessary , one can perform multiformatting to produce transverse Resolution (Δx, Δy) 0.94 mm, 0.94 mm
pre-contrast studies from the sagittal 5-mm contiguous T1-weighted first sequence and Number of data points collected (Nx, Ny) 256, 256
correlate the two. For some, it may be easier to perform transverse precontrast studies and Slice thickness (Δz) 1.5 mm
eliminate this potential pitfall. If this is done, the parameters should not change from pre- Number of slices Variable
to post-contrast studies, other than single average scans for precontrast sequence. Slice gap 0 mm
Number of excitations (NEX) 4–6
Sequence 7: Coronal post-contrast T1-weighted scan
Flow compensation Yes
19. Run coronal post-gadolinium spin echo T1-weighted scan (see Table A7.3.8).
ZIP 512 or ZIP 1024 Yes, changes resolution to 0.47
Frequency-selective fat suppression should be applied. Spatial saturation pulses inferiorly mm, 0.47 mm
placed and gradient moment nulling (flow compensation) are encouraged. The sections ZIP 2 Not appropriate
should include the whole of the temporal bone, brainstem, and the majority of the parotid ZIP 4 Not appropriate
gland.The sequence can be performed as a fast spin echo T1-weighted scan, but the matrix Tailored RF No
should be ZIPped to 512 by × 512, no more than an echo train length of 4 should be used, Extended dynamic range (EDR) Yes
and fat suppression should be applied.
No phase wrap (NPW) No
Chemical saturation Yes
ALTERNATE IMAGING OF CRANIAL NERVES VII TO VIII WHERE NON-NEOPLASTIC Spatial saturation Inferior
PROTOCOL LESIONS ARE SUSPECTED Scan time 8 min
aThis sequence covers entire brain.
Demyelinating Disorders bTo improve the overall fat suppression at air tissue interfaces, complete the manual fat suppression to
If demyelinating etiologies for the cranial nerve deficits are considered, then a more selectively suppress fat.
extensive T2-weighted evaluation of the brain may be indicated. The post-gadolinium
scans should be extended to cover the whole brain. If one identifies lesions that enhance
and lesions that do not, it would imply a polyphasic disease such as multiple sclerosis. spin echo scans, they do allow thinner slice partitions, but scan times can get prolonged
with larger matrices, even with ZIPping.
Congenital Sensorineural Hearing Loss
As indicated previously, a more dedicated T2-weighted study is required for those Sequence 8: 2D fast spin echo
instances when congenital causes of sensorineural hearing loss are considered. In this 1. Run sequence 8 according to Table A7.3.9.
case, high-resolution imaging should include slice thicknesses of less than 2 mm and
matrices of 512 by 512. An advantage of using a head coil for these studies is that both COMMENTARY
ears can be imaged simultaneously with the brainstem. Often, surface coil imaging
produces inadequate evaluation of the exit zones of the nerves, due to signal drop-off. Background Information quality T2-weighted sequences, the post-con-
Because vestibular schwannomas are not the trast studies are critical to demonstrate labyrin-
While many institutions used a 3D-gradient echo sequence that allows sub-millimeter
only cause of hearing loss or vertigo, one should thitis, neuritis, subarachnoid seeding, enhance-
slice partitions (some as thin as 0.5 mm to 0.7 mm), others are using multi-slice 2D fast not perform a high-resolution T2-weighted scan ment of the vestibular aqueduct (seen often in
spin echo scanning. We prefer 2D fast spin echo scans with 1.5-mm thick images using (such as the one above) to screen for sensor- Menière’s disease), labyrinthine schwan-
a head coil to provide the best compromise of coverage, signal to noise, and scan timing ineural hearing loss without also performing nomas, temporal bone schwannomas, and some
(see Table A7.3.9). We use a TR of 3000 to 5000 msec, TE = 80 msec, ETL = 24, 1.5mm post-contrast T1-weighted scans (Fulbright et intracanalicular schwannomas (Ginsberg et al.,
contiguous slices, 256 by 256 matrices ZIPped to 512 by 512 or 1024 by 1024, with 4-6 al., 1995; Fitzgerald and Mark, 1998; Hudgins, 1996; Haught et al., 1998).
NEX. The in-plane matrices (number of data points collected) of the 3D GRASS (Gradient 1998; Sartoretti-Schefer et al., 1999). Although Hemifacial spasm is thought to be caused by
Recalled Acquisition in the Steady State) scans are usually limited to 256 by 256, though it is probably true that over 90% of vestibular vascular compression of the facial nerve. An
this is compensated by the use of smaller fields of view and surface coils images. The schwannomas may be identified with high MRA (MR angiography) (see Table A7.3.10)
Cranial Nerves
VII To VIII localization for these scans may require extra scout images. As for multi-slice 2D fast Head and Neck

A7.3.8 A7.3.9
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A7.3.10 MR Angiography to See Vascular Compression of Facial Nervea,b
Patient position
Scan type
Imaging plane (orientation)
Central slice or volume center
Echo time (TE)
Receiver bandwidth (RBW)
Repeat time (TR)
Flip angle (FA)
Fields of view (FOVx, FOVy)
Resolution (Δx, Δy)
Number of data points collected (Nx, Ny)
Slice thickness (Δz)
Number of slices
Slice gap
Number of excitations (NEX)
Flow compensation
ZIP 512
ZIP 2
Extended dynamic range (EDR)
Chemical saturation
Spatial saturation
Magnetization transfer
Slice series
Scan time
aT sagittal, FSE transverse, FLAIR transverse, and EPI/DWI transverse sequences cover entire brain.
1
bTo improve the overall fat suppression at air tissue interfaces, complete the manual fat suppression to
selectively suppress fat.
should be added to the protocol centered at the
internal auditory canal. One should scrutinize
the raw data files to see the bright vessel im-
pacting on the low signal intensity facial nerve.
There are many philosophies on how to
screen or evaluate the patient with sensorineu-
ral hearing loss. In many instances the protocol
will depend on the patient’s age, the ancillary
clinical findings, and the presence of other
medical illnesses. Once again, an ounce of
conversation with the patient or clinician may
save a pound of scan time (or gadolinium).
If there are multiple cranial nerve deficits or
bilateral hearing deficits, one might be dealing
with neurofibromatosis type II, subarachnoid
seeding on nerve roots, sarcoidosis, Lyme dis-
ease, granulomatous infections, etc., all of
which require greater scrutiny of the post-con-
trast scans (Ginsberg et al., 1996; Sartoretti-
Schefer et al., 1999; Smith et al., 1997). One
might need to evaluate the whole brain for a
source of drop metastases.
Cranial Nerves
VII To VIII
A7.3.10
Supine
3-D gradient echo
Transverse
Internal auditory canals
6.9 msec
16 kHz
35 sec
30°
240 mm, 240 mm
0.47 mm, 0.70 mm
512, 256
1.0 mm
Variable
0 hr.
1 Often, one of the most difficult tasks to
Yes
Yes, changes resolution to 0.23
mm by 0.35 mm
Yes
Yes
Fat saturation
Superior
Yes
Interleaved
∼8 min
Critical Parameters
The patient must be reasonably symmetric
in the head coil so that both internal auditory
canals are in the same plane. This is critical for
the high-resolution T2-weighted scans, where
little coverage is afforded by the longer scan
times from the 512 by 512 matrix. The employ-
ment of multiple quick scouts will often posi-
tion the anatomy perfectly. The authors have
utilized a T1-weighted FMPSPGR (fast multi-
planar spoiled gradient echo: see Table A7.1.2)
to quickly localize the IACs in patients who
were difficult to position symmetrically. Adjust
the imaging parameters to be TR = 100 msec,
TE = minimum, θ (flip angle) = 70°, FOVx =
FOVy = 20 cm, slice thickness 4 mm with no
gap, 512 phase encoding steps and 256 readout
points. It may be necessary to run a few of these
quick “scouts” in order to get the proper align-
ment. One can start with a transverse image and
then oblique the following scouts from there.
Marginal time employed with this will yield
Current Protocols in Magnetic Resonance Imaging
very good results, especially when comparing
anatomic details coming in and out of plane on
different slices. These techniques help produce
a quicker and more confident interpretation.
Troubleshooting
Obtaining coronal scout images and
obliquing the transverse scans to ensure both
IACs are in plane together may be helpful steps.
One can also use the transverse T2-weighted
brain images to determine the optimal center
point for the transverse slices. If not having a
pre-contrast T1-weighted scan is problematic
and the reformatted sagittal does not suffice,
then one should just bite the bullet and call the
patient back for an unenhanced scan after 12
master is complete fat saturation within the
imaging volume. Typically, the overall fat sup-
pression is at its worst when the imaging vol-
ume is adjacent to an air tissue interface causing
local magnetic field inhomogeneities (i.e., sec-
ondary to metallic dental work). Although the
scanners of today can often complete this task
automatically, there is no better way of improv-
ing the fat suppressed image then by manually
adjusting the center frequency and the position
of the RF pulse that will impose the fat suppres-
sion.
While imaging post contrast, it is likely that
flow artifacts over the imaging volume will
significantly increase, making very subtle le-
sions difficult to assess. The authors have found
that, with the simultaneous addition of gradient
moment nulling and inferior spatial saturation
pulses, the image quality improved drastically
and the motion across the anatomy decreased
However, with the addition of gradient moment
nulling, an increase in the minimum TE will
typically result, and may change the T1 weight-
ing of the image. We have not found this slight
increase in the effective TE to be of any real
clinical significance.
Anticipated Results
The post-contrast scans through the IACs
will demonstrate nearly all vestibular schwan-
nomas. As for the facial nerve, it is normal to
see its geniculate ganglion horizontal tympanic
portion vertical descending intramastoid por-
tion enhance. The intracanalicular and labyrin-
thine portions prior to the first genu of the nerve
should not enhance or else it is pathologic.
The high-resolution T2-weighted scans
should show the endolymph and perilymph in
the cochlea, labyrinth, cochlear aqueduct, ves-
Current Protocols in Magnetic Resonance Imaging
tibular aqueduct and semicircular canals as
bright structures amidst the dark temporal
bone. The turns of the cochlea and the size of
the aqueducts should be scrutinized.
Time Considerations
Since the post-contrast scans are the most
valuable in this scenario, one can skimp on
some signal averages, if necessary, to complete
an uncooperative patient’s study. Some fast
recovery fast spin echo sequences create T2
contrast with much shorter three digit TR val-
ues, and could also be used to speed the patient
along.
Index of Terms
The clinical imaging instructions and termi-
nology utilized in this unit are primarily geared
towards General Electric equipment. An Index
of Terms is provided that should bridge the gap
of vendor specific terminology.
Chemical saturation A technique that
applies an additional radiofrequency (RF)
pulse (at a desired distance from the center
frequency) to selectively suppress a tissue. This
technique can be utilized to suppress the signal
from water, fat, or silicone.
Echo time (TE) The time that is measured
from the initiation of the initial RF (radio fre-
quency) pulse and the peak of the echo.
Echo train length (ETL) In fast spin echo
or turbo spin echo imaging, the ETL will actu-
ally equal the number of echoes prescribed per
TR. Successive 180° refocusing pulses are ap-
plied to “rephase the dephasing” protons in an
effort to maximize the number of lines of k-
space per TR. The formula for scan time in
relation to Fast Spin Echo imaging and ETL is
as follows:
Scan time = (TR) × (no. of phase
encoding steps/ETL) × (NEX)
Extended dynamic range (EDR) An im-
aging-enhancement tool that will allow the
utilization of 32-bit data processing as opposed
to the standard 16-bit processor. EDR in this
way should improve SNR and resolution, but
will utilize twice as much memory as a conven-
tional acquisition.
Flow compensation Sometimes more
widely expressed as gradient moment nulling,
this is a way in which the system places flowing
or moving spins into “phase coherence” with
stationary spins.
Inversion time (TI) With inversion recov-
ery pulse sequences, typically the inversion
time is the time from the first 180° RF pulse to Head and Neck
A7.3.11
 the center of the next 90° RF pulse. This inver-
sion time will essentially “null” the desired
tissue depending on how long or short the TI
(inversion time) selected is, and the T1 relaxa-
tion time of the corresponding tissue.
No phase wrap (NPW) will prevent wrap
around artifacts (also known as aliasing arti-
facts) in the phase encoding direction. NPW
should only be used when necessary because
of the following:
a. NPW doubles the FOV in the phase direc-
tion.
b. NPW essentially doubles the phase en-
coding steps (to maintain resolution).
c. With NPW, you must reduce the NEX by
half in order to maintain scan time.
Number of excitations (NEX) is a factor
that is utilized to calculate the overall scan time
and will directly effect the SNR. NEX is essen-
tially the number of times that data are sampled
per acquisition. Note that increasing the NEX
to achieve overall better SNR is a rather ineffi-
cient way to improve signal. Doubling the NEX
from 2 to 4 will only yield a 40% increase in
SNR while it doubles scan time. Take note of
the formula for scan time:
Scan time = (TR) × (no. of phase
encoding steps) × (NEX)
Receiver bandwidth (RBW) The range
of frequencies that the MRI scanner is actually
“tuned” to receive. This will directly affect the
overall SNR. This will not be done by increas-
ing or decreasing the signal, but rather there
will be an increase and/or decrease in the
amount of noise received relative to the altera-
tion of the RBW. An increase in the RBW will
increase the range of frequencies that the scan-
ner will evaluate and thus decrease the overall
SNR. In comparison, utilizing a narrow band-
width should yield less noise and improve the
overall SNR. The relationship of the receiver
bandwidth and SNR can be thought of as in-
versely proportional to the square root of the
bandwidth.
Rectangular field of view (REC FOV)
Asymmetric field of view (typically in the
phase encoding direction). REC FOV is typi-
cally utilized when a body part is longer in one
direction than another. By utilizing an asym-
metric FOV, the system will not collect a por-
tion of the data, thereby decreasing the scan-
ning time.
Repetition time (TR) The time in a pulse
sequence between successive excitation pulses.
Cranial Nerves Spatial saturation employs an additional
VII To VIII RF pulse to cause moving spins within a deter-
A7.3.12
mined area to be selectively dephased. This contrast-enhanced MR studies. AJNR Am. J. cian’s perspective. AJNR Am. J. Neuroradiol.
application will reduce motion from flow Neuroradiol. 19:1433-1436. 17:1226-1228.
and/or respiratory artifacts and will limit the Fulbright, R.K, Erdum, E., Sze, G., and Byrne, T. Marsot-Dupuch, K., Dominguez-Brito, A., Ghasli,
number of slices per TR in general. Since a 1995. Cranial nerve enhancement in the Guil- K., and Chouard, CH. 1999. CT and MR findings
lain-Barré syndrome. AJNR Am. J. Neuroradiol. of Michel anomaly: Inner ear aplasia. AJNR Am.
spatial saturation pulse employs additional RF,
16:923-925. J. Neuroradiol. 20:281-284.
which is to be deposited into the patient, special
Ginsberg, L.E., De Monte, F., and Gillenwater, A.M. Sartoretti-Schefer, S., Kollias, S., and Valavanis, A.
attention should be focused on the SAR (spe-
1996. Greater superficial petrosal nerve: Anat- 1999. Ramsay Hunt syndrome associated with
cific absorption rate). Today all MRI scanners omy and MR findings in perineural tumor brain stem enhancement. AJNR Am. J. Neurora-
have a program that internally monitors how spread. AJNR Am. J. Neuroradiol. 17:389-393. diol. 20:278-80.
much RF can be applied over a given period of Haught, K., Hogg, J.P., Killeffer, J.A., Voelker, J.L., Shellock, F.G. 1996. Pocket Guide to MR Proce-
time. This formula takes into account the pa- and Schochet, S.S.J. 1998. Entirely intracanali- dures and Metallic Objects. Lippincott-Raven,
tient’s body weight. This actual body weight cular meningioma: Contrast-enhanced MR find- Philadelphia.
needs to be accurately input at all times for ings in a rare entity. AJNR Am. J. Neuroradiol. Smith, M.M., Thompson, J.E., Thomas, D., Castillo,
19:1831-1833. M., Carrier, D., Mukherji, S.K., and Gilliam, D.
patient safety. An inappropriate weight will
cause the improper limit of RF to be transmitted Hudgins, P.A. 1998. Inner ear imaging: More than 1997. Choristomas of the seventh and eighth
rule out acoustic. AJNR Am. J. Neuroradiol. cranial nerves. AJNR Am. J. Neuroradiol.
relative to a safe period of time in which this is
19:1807-1808. 18:327-329.
to occur.
Iwayama, E., Naganawa, S., Ito, T., Fukatsu, H.,
Tailored RF An imaging option that im-
Ikeda, M., Ishigeki, T., and Ichinose, N. 1999.
proves image quality on fast spin echo (FSE) High-resolution MR cisternography of the cere- Contributed by Robert W. Evers and
sequences with relatively short TE’s. Tailored bellopontine angle: 2D versus 3D fast spin-echo David M. Yousem
RF will improve edge blurring by reducing sequences. AJNR Am. J. Neuroradiol. 20:889-
895.
The Johns Hopkins Hospital
overall echo spacing. Baltimore, Maryland
ZIP 512/ZIP 1024 Better known as “zer- Jackler, R.K. 1996. Cost-effective screening for
ofill interpolation process,” this is a reconstruc- acoustic neuroma with unenhanced MR: a clini-
tion algorithm that allows the user to scan at a
256 × 256 matrix and then the data are zero-
filled to a 512 × 512 matrix (or 1024 × 1024
respectively).
ZIP 2/ZIP 4 Slice ZIP essentially is also
a “zerofill interpolation process” that will cre-
ate additional slices through the interpolation
procedure. These slices are created with an
offset of 50% of the original imaging slice
locations.
Literature Cited
Allen, R.W., Harnsberger, H.R., Shelton, C., et al.
1996. Low-cost high-resolution fast spin-echo
MR of acoustic schwannoma: An alternative to
enhanced conventional spin-echo MR? AJNR
Am. J. Neuroradiol. 17:1205-1210.
Bigelow, D.C., Eisen, M.D., Smith, P.G., Yousem,
D.M., Levine, R.S., Jackler, R.K., Kennedy,
D.W., and Kotapka, M.J. 1998. Lipomas of the
internal auditory canal and cerebellopontine an-
gle. Laryngoscope 108:1459-1469.
Chakeres, D.W. and Schmalbrock, P. 1997. MR
techniques for the internal auditory canal. AJNR
Am. J. Neuroradiol. 18:1394-1395.
Curtin, H.D. 1997. Rule out eighth nerve tumor:
Contrast-enhanced T1-weighted or high-resolu-
tion T2-weighted MR? AJNR Am. J. Neuroradiol.
18:1834-1838.
El-Gammal, T. 1997. MR techniques for the internal
auditory canal. AJNR Am. J. Neuroradiol.
18:1393-1394.
Fitzgerald, D.C. and Mark, A.S. 1998. Sudden hear-
ing loss: Frequency of abnormal findings on
Head and Neck
A7.3.13
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
Cranial Nerves IX To XII
Most lesions that affect the lower cranial nerves are centered at the jugular foramen, where
cranial nerves IX, X, and XI leave the calvarium (Castillo and Mukherji, 1996; Rubinstein
et al., 1995). Because of the presence of the jugular vein in the same location, a strategy
based on differentiating neurogenic lesions (e.g., schwannomas, meningiomas) from
vascular ones (e.g., paragangliomas, high-riding jugular bulbs, aneurysms) must be
employed. Cranial nerve XII lesions are usually schwannomas or secondary invasions
from skull-base neoplasms (Thompson and Smoker, 1994; Karpati et al., 1998; Russo et
al., 1997). This sometimes means performing MR arteriograms or venograms in addition
to the standard fare of high-resolution and contrast-enhanced scans. These nerves have
predictable courses and symptom complexes, which make demonstration of pertinent
lesions very gratifying.
IMAGING OF CRANIAL NERVES IX-XII
Cranial nerves IX to XII are rarely affected by pathology compared with cranial nerves
III, V, VII, and VIII. Nonetheless, their evaluation is challenging, since lesions of these
nerves span the gamut from intracranial to extracranial sites. In the case of the vagus
nerve, one may need to scan the full neck and chest to ascertain the source of the vagal
symptoms. The glossopharyngeal (cranial nerve IX) and vagus nerve (cranial nerve X)
are more commonly affected by diseases than the spinal accessory nerve (cranial nerve
XI, which receives input from cervical nerve rootlets) and hypoglossal nerve (cranial
nerve XII; Castillo and Mukherji, 1996; Chong and Fan, 1996; Ortiz and Reed, 1995).
While IX to XI traverse the jugular foramen (where the jugular vein is the dominant
structure, replete with its accompanying flow artifacts) cranial nerve XII exits via the
smaller hypoglossal canal at the foramen magnum. Imaging of these cranial nerves
requires a focused approach based on clinical symptomatology and signs. Thin slice
imaging, high resolution, and contrast enhancement are essential components to this
evaluation.
Table A7.4.1 lists the hardware necessary to perform the procedure, along with appropri-
ate parameters. Standard head coil imaging is sufficient to obtain adequate scans on the
intracranial portions of these cranial nerves. Anatomic imaging with high-quality fast spin
echo T2-weighted scan with the dark nerves outlined by bright cerebrospinal fluid (CSF)
are required to see the portions of these nerves coursing to the jugular foramen and
hypoglossal canal. Gadolinium-enhanced scans are critical to the complete evaluation of
these nerves, though the evaluation may be limited to sections through the lower pons and
medulla. No monitoring is required. It may be necessary to continue the evaluation into
the head and neck region with a neck coil in the case of vagus nerve lesions. Particularly
Table A7.4.1 Equipment Parameters for Cranial
Nerve Imaging
Coil type Head
Gradient coil strength 25 mT/m
Flow compensation pulse Yes
Respirator If needed
Oxygen If needed
Motion cushions Useful
Use of contrast agents Yes
Contributed by Robert W. Evers and David M. Yousem
Current Protocols in Magnetic Resonance Imaging (2001) A7.4.1-A7.4.11
Copyright © 2001 by John Wiley & Sons, Inc.
UNIT A7.4 if a source of a vocal cord paralysis is being pursued, then a coil that covers the brainstem,
the neck, and even the upper mediastinum to the aortic arch may be required. For upper
vagus or glossopharyngeal or hypoglossal masses, the coverage afforded by most head
coils is sufficient, provided the patient’s vertex is placed at the very top of the head coil.
Materials
Normal saline (0.9% NaCl), sterile
Gadolinium-based MR contrast agent (e.g., Magnevist, Omniscan, or Prohance)
Set up equipment and patient
1. Interview (screen) the patient to ensure that he or she has no contraindications such
as cardiac pacemakers or other implants containing ferromagnetic materials. Also be
sure to find out if the patient has any health conditions that may require the presence
of special emergency equipment during the scanning procedure, or necessitate any
other precautions.
BASIC Generally standard screening forms are used for all patients scanned in a magnetic
PROTOCOL resonance system.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(1996) for discussion of what implants may be safely scanned using magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
2. If the procedure is a research protocol, have the patient sign any necessary consent
form.
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating
and image artifacts.
5. Inform the patient about what will occur during the procedure, what he or she will
experience while in the magnet, and how to behave, including the following:
a. If earphones or headphones are used to protect the ears from the loud sounds
produced by the gradients, the patient will be asked to wear these, but will be able
to communicate with you at any time during the imaging.
b. The patient will be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this works).
c. For good results the patient should not talk, and should avoid or minimize
swallowing or other movement, during each scan—i.e., as long as the banging
sounds continue. Between scans, talking and swallowing are allowed in most
cases, but should be avoided when comparative positional studies are being
performed; the patient will be informed when this is the case.
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
6. Help the patient mount onto the table. Either before or right after the patient lies down,
set up any triggering devices or other monitoring equipment that is to be used.
7. Center the patient’s head within the head coil and use the laser light to ensure
symmetry. For most brain scans, placing the center landmark at the eyebrow (nasion)
Cranial Nerves level allows full coverage of the relevant regions of anatomy.
Head and Neck IX To XII
A7.4.1 A7.4.2
Current Protocols in Magnetic Resonance Imaging
 Generally the patient’s head is fixed so that the head is horizontal (not tilted) and the neck Sequence 4: Transverse echo planar/diffusion weighted scan
and head lie along the axis of the patient table; other positions may be appropriate 15. Run transverse echo planar/diffusion weighted imaging (see Table A7.4.5).
depending on the needs at hand.
The minimum diffusion weighting should have a “b” value of at least 1000 sec/mm2.
8. If needed, place a pillow or other support under the knees to make the patient more These slices should be applied through the entire brain and aid in the detection of acute
comfortable. stroke processes.
9. Establish an intravenous line from which the contrast agent can be injected, and attach
this line securely to the patient so that movement into or out of the magnet will not Table A7.4.2 Sagittal T1-Weighted Spin Echo Scan (Sequence 1)a
pull at the patient’s arm.
Patient position Supine
It is preferable to insert the line prior to imaging and to leave the patient in the magnet,
with no intervening motion, between the scans run before contrast agent injection and those Scan type Spin echo
run after injection. Imaging plane (orientation) Sagittal
Central slice or volume center Mid-brain
10. Use the centering light to position the patient and put him or her into the center of Echo time (TE) Minimum
the magnet. Receiver bandwidth (RBW) 16 kHz
Once this step has been performed, so long as the patient does not move on the table, the Repeat time (TR) 400–600 msec
table itself can be moved and then replaced in the same position as before without Flip angle (FA) 90°
jeopardizing the positioning of one scan relative to another. Fields of view (FOVx, FOVy) 240 mm, 240 mm
11. If the patient is unable to hold still, provide an appropriate sedative. Resolution (Δx, Δy) 0.94 mm, 1.25 mm
Number of data points collected (Nx, Ny) 256, 192
Sequence 1: Sagittal T1-weighted spin echo scan Slice thickness (Δz) 5 mm
Although a pilot scan can be acquired, it is not necessary here because the first scan covers Number of slices 28
almost the entire brain in a sagittal fashion. These sagittal images can then be used as Slice gap 0 mm
localizers for the application of saturation pulses when the transverse images are acquired. Number of excitations (NEX) 1
12. Run sagittal T1-weighted scan (See Table A7.4.2). Spatial saturation Inferior
Slice series Interleaved
Saturation pulses inferiorly are applied. This scan is critical to visualizing the jugular Scan time ∼2 min
foramen and the jugular vein. Because the saturation pulses applied produce a nice flow
aThis sequence covers entire brain.
void in the jugular foramen, one may be able to exclude a glomus jugulare with this
sequence. Since the transverse scans that follow may be confusing due to turbulent flow
simulating a jugular foramen mass, the sagittal scans often are the most reliable in this Table A7.4.3 Transverse Fast Spin Echo Scan (Sequence 2)a
regard. When positioned properly, these scans may also be the ones that demonstrate
carotid body tumors that may produce vagus nerve or sympathetic plexus symptoms. Patient position Supine
Scan type Fast spin echo
Sequence 2: Transverse fast spin echo (FSE) scan Imaging plane (orientation) Transverse
13. Run transverse fast spin echo (or turbo spin echo) T2-weighted scans (see Table Central slice or volume center Mid-brain
A7.4.3) through the entire brain. Echo time (TE) 90–120 msec
Echo train length 12-32 depending on slices allowed and minimum echo spacing. Receiver bandwidth (RBW) 20 kHz
Echo train length (ETL) 12–32
Saturation pulses are applied inferiorly and gradient moment nulling is applied. One now Repeat time (TR) 3000–4000 msec
routinely applies frequency selective fat suppression pulses to produce a better dynamic Flip angle (FA) 90°
range of contrast and to null the skull base, jugular foramen, and temporal bone fat.
Fields of view (FOVx, FOVy) 240 mm, 240 mm
Acquisition time will be 3 to 4 min in length. The lower cranial nerves are well outlined in
the bright CSF. Resolution (Δx, Δy) 0.94 mm, 1.25 mm
Number of data points collected (Nx, Ny) 256, 192
Sequence 3: Transverse FLAIR scan Slice thickness (Δz) 5 mm
14. Run transverse fluid attenuation inversion recovery (FLAIR) scan (see Table A7.4.4). Number of slices Variable (∼20)
Slice gap 0 mm
Echo train length 12-32 depending on slices allowed and minimum echo spacing. Satura- Number of excitations (NEX) 2
tion pulses inferiorly, tailored radiofrequency pulses and gradient moment nulling are
Flow compensation Yes
applied. The slices should be through the entire brain. Acquisition time will be 4 to 5 min
in length. ZIP 512 Yes
Chemical saturation Yes, fat
Demyelinating disorders may lead to lower cranial nerve symptoms. Although FLAIR does Spatial saturation Inferior
well at displaying white matter lesions in the brainstem and supratentorial compartment, Slice series Interleaved
recent studies have suggested that the T2-weighted sequence remains the most reliable Cranial Nerves
Head and Neck IX To XII Scan time 3–4 min
means of evaluating the posterior fossa white matter and cortex.
aThis sequence covers entire brain.
A7.4.3 A7.4.4
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
Sequence 5: 2-D time of flight scan Table A7.4.6 2-D Time of Flight Scan (Sequence 5)a
16. Perform 2-D time of flight scan (see Table A7.4.6).
Patient position Supine
Because of the flow artifact associated with the jugular vein in the jugular foramen, it may
Scan type 2-D gradient echo
be helpful to perform an MR venographic study through the posterior fossa and the skull
base. This can be relegated to a 2-D time-of-flight (TOF) mode (see Table A7.4.6) where Imaging plane (orientation) Coronal
Central slice or volume center Skull base
Table A7.4.4 Transverse FLAIR Scan (Sequence 3)a Echo time (TE) Minimum
Receiver bandwidth (RBW) 16 kHz
Patient position Supine
Repeat time (TR) 30 msec
Scan type Inversion recovery fast spin echo
Flip angle (FA) 60°
Imaging plane (orientation) Transverse
Fields of view (FOVx, FOVy) 240 mm, 180 mm
Central slice or volume center Mid-brain
Resolution (Δx, Δy) 0.47 mm, 0.70 mm
Echo time (TE) 133 msec
Number of data points collected (Nx, Ny) 512, 256
Receiver bandwidth (RBW) 16 kHz
Slice thickness (Δz) 2 mm
Echo train length (ETL) 12–32
Number of slices Variable (∼40)
Repeat time (TR) 8800 msec
Slice gap 0 mm
Inversion time (TI) 2200 msec
Number of excitations (NEX) 1
Flip angle (FA) 180°
Flow compensation Yes
Fields of view (FOVx, FOVy) 240 mm, 240 mm
Spatial saturation Inferior
Resolution (Δx, Δy) 0.94 mm, 0.94 mm
Slice series Interleaved
Number of data points collected (Nx, Ny) 256, 256
Scan time ∼5–6 min
Slice thickness (Δz) 5 mm aIn order to suppress the arterial phase and accentuate the venogenic phase, the
Number of slices Variable (∼20) saturation pulse should be a “walking” also known as a “traveling” saturation pulse.
Slice gap 0 mm The saturation pulse should move with the actual imaging slice, thereby increasing its
Number of excitations (NEX) 1 efficacy in overall arterial suppression.
Flow compensation Yes
inferior saturation pulses are applied and downward flow is optimally demonstrated. Once
ZIP 512 Yes (changes resolution to 0.47
again, viewing the raw data provides one with the visualization of the “opacified” vessel
mm, 0.47 mm)
and the suspected “mass.” Since most glomus jugulare paragangliomas grow into the
Tailored RF Yes
jugular vein, the presence of a “filling defect” amidst the jugular vein flow would suggest
Spatial saturation Inferior this diagnosis. Lesions outside the vein may be schwannomas. If this 2-D TOF is to be
Slice series Interleaved utilized, it should be used prior to the administration of any contrast medium.
Scan time ∼4–5 min
Another very useful imaging tool to assess the jugular vein and visualize the vessel and
aThis sequence covers entire brain.
suspected “mass,” would be the addition of a 3-D fast gradient echo pulse sequence with
a double dose of contrast medium. This rapid imaging technique should give the dynamic
Table A7.4.5 Transverse Echo Planar/Diffusion Weighted Scan (Sequence 4)a evaluation of an arterial through venous phase of vessel filling, as well as multi-planar
reconstruction. In addition, there will be an overall improved SNR (signal to noise ratio)
Patient position Supine with respect to conventional 2-D imaging techniques.
Scan type Echo planar/diffusion weighted
image Sequence 6: Transverse post-contrast T1-weighted scan
Imaging plane (orientation) Transverse 17. Leave the patient in the magnet, inject the contrast agent, and flush the line with 10
Central slice or volume center Mid-brain ml saline.
Echo time (TE) Minimum
A dose of 0.2 mmol/kg of contrast agent is usually given.
Receiver bandwidth (RBW) 62 kHz
Repeat time (TR) 10,000 msec 18. Run transverse post-gadolinium T1-weighted scan (see Table A7.4.7).
Flip angle (FA) 90° A 256 by 256 matrix through the pons, medulla and foramen magnum is used. Saturation
Fields of view (FOVx, FOVy) 240 mm, 240 mm pulses inferiorly and gradient moment nulling (flow compensation) are applied.
Resolution (Δx, Δy) 1.88 mm, 1.88 mm
Pre-gadolinium scans are often not necessary and only add time. However, if in a pinch,
Number of data points collected (Nx, Ny) 128, 128 due to the suspicion of a lesion that could be bright precontrast, simply reconstruct the
Slice thickness (Δz) 5 mm sagittal 5 mm contiguous T1-weighted images from the first sequence into a transverse
Number of slices Variable (∼10) plane and correlate the two. For some, it may be easier to perform transverse pre-contrast
Slice gap 0 mm studies and eliminate this potential pitfall. If this is done, the parameters should not change
Number of excitations (NEX) 1 from pre- to post-contrast studies, other than single average scans for pre-contrast series.
Slice series Interleaved
Cranial Nerves
Scan time ∼1 min Head and Neck IX To XII
aThis sequence covers entire brain.
A7.4.5 A7.4.6
Current Protocols in Magnetic Resonance Imaging Supplement 6 Supplement 6 Current Protocols in Magnetic Resonance Imaging
 Table A7.4.7 Transverse Post-Contrast T1-Weighted Scan Table A7.4.9 T2*-Weighted Gradient Echo Scan (Sequence 9)
(Sequence 6)
Patient position Supine
Patient position Supine Scan type Gradient echo
Scan type Spin echo Imaging plane (orientation) Transverse
Imaging plane (orientation) Transverse Central slice or volume center Mid-brain
Central slice or volume center Mid-brain Echo time (TE) 20 msec
Echo time (TE) Minimum Receiver bandwidth (RBW) 16 kHz
Receiver bandwidth (RBW) 16 kHz Repeat time (TR) 500 msec
Repeat time (TR) 400–600 msec Flip angle (FA) 20°
Flip angle (FA) 90° Fields of view (FOVx, FOVy) 240 mm, 180 mm
Fields of view (FOVx, FOVy) 180 mm, 180 mm Resolution (Δx, Δy) 0.94 mm, 0.94 mm
Resolution (Δx, Δy) 0.70 mm, 0.70 mm Number of data points collected (Nx, Ny) 256, 192
Number of data points collected (Nx, Ny) 256, 256 Slice thickness (Δz) 5 mm
Slice thickness (Δz) 3 mm Number of slices Variable (∼20)
Number of slices Variable (∼20) Slice gap 0 mm
Slice gap 0 mm Number of excitations (NEX) 1
Number of excitations (NEX) 2–3 Spatial saturation Inferior
Flow compensation Yes Slice series Interleaved
Spatial saturation Inferior Scan time ∼3 min
Slice series Interleaved
Scan time ∼5–7 min
Sequence 7: Coronal post-contrast T1-weighted scan
19. Run coronal post-gadolinium fast spin echo T1-weighted scan (see Table A7.4.8).
Table A7.4.8 Coronal Post-Contrast T1-Weighted Scan (Sequence 7) Frequency selective fat suppression pulses are applied. Saturation pulses inferiorly placed
in conjunction with gradient moment nulling (flow compensation) are encouraged. The
Patient position Supine slices should include the whole of the temporal bone and brainstem. This T1-weighted
Scan type Fast spin echo sequence is being performed as a fast spin echo T1-weighted scan, so be sure to utilize an
Imaging plane (orientation) Coronal echo train length not to exceed 4 and make an attempt to ZIP the display matrix to 512 by
Central slice or volume center Mid-brain 512. Acquisition time will be 2 to 4 min in length.
Echo time (TE) Minimum
Sequence 9: T2*-weighted gradient echo scan for hemosiderosis
Receiver bandwidth (RBW) 16 kHz
20. If hemosiderosis is contemplated, perform a T2*-weighted gradient echo scan through
Echo train length (ETL) ≤4
the posterior fossa (see Table A7.4.9).
Repeat time (TR) 400–600 msec
Flip angle (FA) 90° This allows better detection of blood products than fast spin echo techniques (Pribitkin et
Fields of view (FOVx, FOVy) 180 mm, 180 mm al., 1994).
Resolution (Δx, Δy) 0.70 mm, 0.70 mm
Number of data points collected (Nx, Ny) 256, 256 ALTERNATE IMAGING OF CRANIAL NERVES IX TO XII WHERE NON-NEOPLASTIC
Slice thickness (Δz) 3 mm PROTOCOL LESIONS ARE SUSPECTED
Number of slices Variable
If demyelinating etiologies for the cranial nerve deficits are considered, then the post-
Slice gap 0 mm
gadolinium scans should be extended to cover the whole brain. If one identifies lesions
Number of excitations (NEX) 2
that enhance and lesions that do not, it would imply a polyphasic disease such as multiple
Flow compensation Yes
sclerosis.
ZIP 512 Yes (changes resolution to 0.35
mm, 0.35 mm) 1. Repeat sequence 6 according to Table A7.4.7. Change slice thickness to 5 mm to
Chemical saturation Yes, fat cover the entire brain.
Spatial saturation Inferior 2. Repeat sequence 7 according to Table A7.4.8. Change slice thickness to 5 mm in order
Slice series Interleaved to cover the entire brain.
Scan time ∼2–4 min

Cranial Nerves
Head and Neck IX To XII

A7.4.7 A7.4.8
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
COMMENTARY
Background Information
The vagus nerve is the longest-running of
the cranial nerves. Its recurrent laryngeal nerve
branch loops under the aortic arch on the left
and the subclavian artery on the right. There-
fore, in the evaluation of vocal cord paralysis,
the vagus nerve’s origin from the medulla, its
egress through the pars vasculosa of the jugular
foramen, its course within the carotid sheath to
the take-off of these vessels, and the course of
the recurrent laryngeal nerve looping under the
subclavian artery (right) or aorta (left) and as-
cending in the tracheoesophageal groove, must
be examined.
If there are multiple cranial nerve deficits or
bilateral hearing deficits, one might be dealing
with neurofibromatosis type II, subarachnoid
seeding on nerve roots, sarcoidosis, Lyme dis-
ease, granulomatous infections, etc., all of
which require greater scrutiny of the post-con-
trast scans. The whole brain might need to be
evaluated for a source of drop metastases.
Critical Parameters
The problems associated with jugular flow
artifacts are the biggest hurdle in creating diag-
nostic images in this regions. Too often the
turbulence causes soft tissue signal in the jugu-
lar foramen, and the slow flow seen as “en-
hancement” on the post-contrast studies only
complicates matters. Good quality T1-weighted
scans may be the lifesaver. As for the vagus
nerve, the issues related to this evaluation re-
volve around on how low the scanning must go.
Often the chest images are distorted due to
respiratory and/or pulsation artifact.
Troubleshooting
Again, MR venograms will clarify most
issues as they relate to the jugular foramen. If
one is able to obtain a history of palatal palsy,
uvular deviation, or other pharyngeal sympto-
matology, it may point to an upper vagus nerve
lesion, and the need to extend the scan into the
lower neck and/or chest may be obviated. To
get good-quality images of the left recurrent
laryngeal nerve at the aortic arch, a chest pro-
tocol with respiratory gating may be required.
If this is a consideration, a dedicated chest MRI
may be in order, and techniques will have to be
employed such as cardiac gating and respira-
tory compensation/triggering. In order to ac-
centuate the visualization of the laryngeal nerve
at the aortic arch, one of the best pulse se-
quences to utilize would be a “double inversion
Current Protocols in Magnetic Resonance Imaging
recovery” better known as black blood imaging
(see Table A7.4.10). This technique allows the
user to obtain a “black blood” image (which is
fantastic for vascular anatomy) while maintain-
ing a parenchymal fat plane. This fat plane
enables the radiologist to essentially roadmap
the nerve from origin-to-pathology. We like to
think of this “black blood” technique as a snap-
shot image that acquires sequential slices in a
breath hold and cardiac gated style. When util-
izing this pulse sequence, ensure that the tech-
nique is done prior to the administration of
contrast medium. Post-contrasted black blood
imaging will often leave the user with less than
desirable results.
Anticipated Results
One should be able to see cranial nerves IX
and X coursing together from the medulla to
the jugular foramen. The spinal accessory nerve
is virtually never separately discerned, and the
hypoglossal nerve is inconstantly visualized.
Schwannomas of these nerves will be evident
on post-contrast studies.
Time Considerations
In chasing vagus nerve lesions, one may be
forced to employ a head, neck, and chest coil.
The time demands on switching the coils and
repositioning the patient may be prohibitive,
and the case should be charged accordingly. In
most instances the post-contrast transverse
scans are the most valuable, so reducing the
NEX or employing fast spin echo T1-weighted
scans may be economical.
Index of Terms
The clinical imaging instructions and termi-
nology utilized in this unit are primarily geared
towards General Electric equipment. An Index
of Terms is provided that should bridge the gap
of vendor specific terminology.
Chemical saturation A technique that
applies an additional radiofrequency (RF)
pulse (at a desired distance from the center
frequency) to selectively suppress a tissue. This
technique can be utilized to suppress the signal
from water, fat, or silicone.
Echo time (TE) The time that is measured
from the initiation of the initial RF (radio fre-
quency) pulse and the peak of the echo.
Echo train length (ETL) In fast spin echo
or turbo spin echo imaging, the ETL will actu-
ally equal the number of echoes prescribed per
TR. Successive 180° refocusing pulses are ap-
Table A7.4.10 Black Blood (Double Inversion Recovery) Pulse Sequencea
Patient position Supine
Scan type Fast spin echo
Imaging plane (orientation) Transverse
Central slice or volume center Chest
Echo time (TE) 30–60 msec
Receiver bandwidth (RBW) 32 kHz
Echo train length (ETL) 32
Repeat time (TR) Minimum
Inversion time (TI) Auto (∼600 msec)
Flip angle (FA) 180°
Fields of view (FOVx, FOVy) 240 mm, 240 mm
Resolution (Δx, Δy) 0.94 mm, 1.25 mm
Number of data points collected (Nx, Ny) 256, 192
Slice thickness (Δz) 5 mm
Number of slices Variable
Slice gap 2 mm
Number of excitations (NEX) 1
ZIP 512 Yes (changes resolution to 0.47
mm, 0.63 mm)
Slice series Interleaved
ECG gating Yes
Scan time ∼5–6 min
aFor this sequence, cardiac gating is essential. The minimum T to be utilized will be calculated by
R
the imaging system. Typically, the image may look proton density weighted because the TRs are
relatively long (1500 msec to 1800 msec) and the TEs are generally short (30 msec to 60 msec). The
inversion time is typically set by the MR system. The system calculates the “inversion time” needed
to suppress the overall signal from blood based on the heart rate. This inversion time will change slightly
from patient to patient. A neurovascular coil is used in this sequence when the imaging position is neck
or above. A torse array coil is used when the imaging position is chest.
plied to “rephase the dephasing” protons in an sion time will essentially “null” the desired
effort to maximize the number of lines of k- tissue depending on how long or short the TI
space per TR. The formula for scan time in (inversion time) selected is, and the T1 relaxa-
relation to Fast Spin Echo imaging and ETL is tion time of the corresponding tissue.
as follows: Multi-phase is an imaging option that al-
lows the user to run a pulse sequence over and
Scan time = (TR) × (no. of phase
over again. Its application is primarily utilized
encoding steps/ETL) × (NEX)
in contrast-enhanced angiography when at-
Extended dynamic range (EDR) An im- tempting to “catch” an arterial phase of contrast
aging-enhancement tool that will allow the media, followed by a mixed arterial and venous
utilization of 32-bit data processing as opposed filling phase. Multi-phase allows continual
to the standard 16-bit processor. EDR in this scanning of a desired area independent of the
way should improve SNR and resolution, but reconstruction phase.
will utilize twice as much memory as a conven- No phase wrap (NPW) will prevent wrap
tional acquisition. around artifacts (also known as aliasing arti-
Flow compensation Sometimes more facts) in the phase encoding direction. NPW
widely expressed as gradient moment nulling, should only be used when necessary because
this is a way in which the system places flowing of the following:
or moving spins into “phase coherence” with a. NPW doubles the FOV in the phase direc-
stationary spins. tion.
Inversion time (TI) With inversion recov- b. NPW essentially doubles the phase en-
ery pulse sequences, typically the inversion coding steps (to maintain resolution).
Cranial Nerves time is the time from the first 180° RF pulse to c. With NPW, you must reduce the NEX by
Head and Neck IX To XII the center of the next 90° RF pulse. This inver- half in order to maintain scan time.
A7.4.9 A7.4.10
Supplement 12 Supplement 12 Current Protocols in Magnetic Resonance Imaging
 Number of excitations (NEX) is a factor
that is utilized to calculate the overall scan time
and will directly effect the SNR. NEX is essen-
tially the number of times that data are sampled
per acquisition. Note that increasing the NEX
to achieve overall better SNR is a rather ineffi-
cient way to improve signal. Doubling the NEX
from 2 to 4 will only yield a 40% increase in
SNR while it doubles scan time. Take note of
the formula for scan time:
Scan time = (TR) × (no. of phase
encoding steps) × (NEX)
Receiver bandwidth (RBW) The range
of frequencies that the MRI scanner is actually
“tuned” to receive. This will directly affect the
overall SNR. This will not be done by increas-
ing or decreasing the signal, but rather there
will be an increase and/or decrease in the
amount of noise received relative to the altera-
tion of the RBW. An increase in the RBW will
increase the range of frequencies that the scan-
ner will evaluate and thus decrease the overall
SNR. In comparison, utilizing a narrow band-
width should yield less noise and improve the
overall SNR. The relationship of the receiver
bandwidth and SNR can be thought of as in-
versely proportional to the square root of the
bandwidth.
Rectangular field of view (REC FOV)
Asymmetric field of view (typically in the
phase encoding direction). REC FOV is typi-
cally utilized when a body part is longer in one
direction than another. By utilizing an asym-
metric FOV, the system will not collect a por-
tion of the data, thereby decreasing the scan-
ning time.
Repetition time (TR) The time in a pulse
sequence between successive excitation pulses.
Spatial saturation employs an additional
RF pulse to cause moving spins within a deter-
mined area to be selectively dephased. This
application will reduce motion from flow
and/or respiratory artifacts and will limit the
number of slices per TR in general. Since a
spatial saturation pulse employs additional RF,
which is to be deposited into the patient, special
attention should be focused on the SAR (spe-
cific absorption rate). Today all MRI scanners
have a program that internally monitors how
much RF can be applied over a given period of
time. This formula takes into account the pa-
tient’s body weight. This actual body weight
needs to be accurately input at all times for
patient safety. An inappropriate weight will
cause the improper limit of RF to be transmitted
Current Protocols in Magnetic Resonance Imaging
relative to a safe period of time in which this is
to occur.
Tailored RF An imaging option that im-
proves image quality on fast spin echo (FSE)
sequences with relatively short TE’s. Tailored
RF will improve edge blurring by reducing
overall echo spacing.
ZIP 512/ZIP 1024 Better known as “zer-
ofill interpolation process,” this is a reconstruc-
tion algorithm that allows the user to scan at a
256 × 256 matrix and then the data are zero-
filled to a 512 × 512 matrix (or 1024 × 1024
respectively).
ZIP 2/ZIP 4 Slice ZIP essentially is also
a “zerofill interpolation process” that will cre-
ate additional slices through the interpolation
procedure. These slices are created with an
offset of 50% of the original imaging slice
locations.
Literature Cited
Castillo, M. and Mukherji, S.K. 1996. Magnetic
resonance imaging of cranial nerves IX, X, XI,
and XII. Top. Magn. Reson. Imaging 8:180-186.
Chong, V.F. and Fan, Y.F. 1996. Jugular foramen
involvement in nasopharyngeal carcinoma. J.
Laryngol. Otol. 110:987-90.
Karpati, R.L., Loevner, L.A., Cunning, D.M.,
Yousem, D.M., Li, S., and Weber, R.S. 1998.
Synchronous hypoglossal nerve and sympa-
thetic nervous system plexus schwannomas. AJR
Am. J. Roentgenol. 171(6):1505-1507.
Ortiz, O. and Reed, L. 1995. Spinal accessory nerve
schwannoma involving the jugular foramen.
AJNR Am. J. Neuroradiol. 16:986-989.
Pribitkin, E.A., Rondinella, L., Rosenberg, S., and
Yousem, D.M. 1994. Superficial siderosis of the
central nervous system: An underdiagnosed
cause of sensorineural hearing loss and ataxia.
Am. J. Otol. 15:415-418.
Rubinstein, D., Burton, B.S., and Walker, A.L. 1995.
The anatomy of the inferior petrosal sinus, glos-
sopharyngeal nerve, vagus nerve, and accessory
nerve in the jugular foramen. AJNR Am. J.
Neuroradiol. 16:185-194.
Russo, C.P., Smoker, W.R., and Weissman, J.L.
1997. MR appearance of trigeminal and hypo-
glossal motor denervation. AJNR Am. J.
Neuroradiol. 18:1375-1383.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Thompson, E.O. and Smoker, W.R. 1994. Hypo-
glossal nerve palsy: a segmental approach. Ra-
diographics 14:939-958.
Contributed by Robert W. Evers and
David M. Yousem
The Johns Hopkins Hospital
Baltimore, Maryland
The Orbit and Optic Nerves UNIT A7.5
Magnetic resonance imaging (MRI) of the orbits and optic nerves has developed into the
“gold standard” of imaging modalities for the evaluation of many soft tissue abnormalities
of the orbit (see Mafee et al., 1996). Its advantage is a result of its superior soft tissue
contrast, which allows for the excellent evaluation of fat, muscle, nervous tissue, cerebro-
spinal fluid, and blood vessels. Computed tomography (CT) remains the modality of
choice for evaluation of the bony structures of the orbit. MRI is also more flexible,
allowing for multiplanar imaging that is not possible with CT. For these reasons, CT and
MRI can be complementary, especially in situations where one encounters entities such
as sinus pathology, which bridges the air spaces, bone, orbit, and brain. Ultrasound also
has many advantages for evaluation of the globes, but it does not provide the more
universal views of the orbit and periorbital area that are possible with MRI.
IMAGING OF THE ORBITS AND BRAIN BASIC
PROTOCOL
In most situations it is important not to focus strictly on the orbits, but to also image the
adjacent structures, including the face, scalp, brain, sinuses, and blood vessels—in
particular the cavernous sinus. Trying to determine the location of pathology based on
clinical symptoms alone is frequently not very accurate, even after a detailed clinical
exam. For example, a patient may present with diplopia, which suggests an extraocular
muscle disorder, but the true etiology may be a dural arteriovenous fistula of the posterior
fossa, with secondary venous congestion of the cavernous sinus. In a case where the
pathology is not where it is anticipated to exist, focusing the exam exclusively on the
orbits would overlook the true underlying pathology. There are a few indications where
a very focused, detailed exam of the orbit is indicated, in isolation from other structures.
This most commonly occurs when there is a question of a primary disorder of the globe,
such as choroidal melanoma. Even in this case, however, it may be necessary to image
the brain, in order to exclude and characterize more distant metastatic disease.
A standard head coil (Table A7.5.1) is usually utilized for a general orbital examination,
which allows for a fast and simple setup. The exam is similar to that used for standard
imaging of the head, but the focus of the exam is on the orbit and some of the pulse
sequences used for the orbital exam are not routinely used for brain imaging. Table A7.5.1
lists the hardware necessary to perform the procedure, along with appropriate parameters.
Table A7.5.1 Equipment Parameters for Orbit Imaging
Coil type Head coil or small diameter (10 cm),
circular, phased array surface coil
Fat-saturation pulse Preferable
Gradient coil strength 20 mT/m
Cardiac gating No
Peripheral gating No
Respiratory gating No
Respirator No
Oxygen No
Motion cushions Helpful
Use of contrast agents Yes
Susceptibility-reducing eye pads Optional
Head and Neck Head and Neck
A7.4.11 Contributed by Donald William Chakeres and Eric C. Bourekas A7.5.1
Current Protocols in Magnetic Resonance Imaging (2001) A7.5.1-A7.5.13
Copyright © 2001 by John Wiley & Sons, Inc.
 Materials cases, but should be avoided when comparative positional studies are being
Normal saline (0.9% NaCl), sterile performed; the patient will be informed when this is the case.
Extravascular contrast agent (e.g., Magnevist, Omniscan, or Prohance) d. Nevertheless, the patient may call out at any time if he or she feels it necessary.

Set up equipment and patient 6. Help the patient mount onto the table. Either before or right after the patient lies down,
1. Interview (screen) the patient to ensure that he or she has no contraindications such set up any triggering devices or other monitoring equipment that is to be used.
as cardiac pacemakers or other implants containing ferromagnetic materials. Also be 7. Instruct the patient to try to avoid scanning with their eyes throughout the procedure
sure to find out if the patient has any health conditions that may require the presence because this movement will lead to motion artifacts.
of special emergency equipment during the scanning procedure, or necessitate any Since it is essentially impossible to have the patient fix their vision point for a long period
other precautions. of time, the best advice to give them is to keep their eyes closed. If the patient sleeps, there
In the case of an orbital exam, it is particularly important to confirm that the patient does may be rapid eye movements that can cause artifacts (Fig. A7.5.2).
not have metallic fragments or implants in the eyes (see Kelly et al., 1986; Lagouros et al.,
1987; Otto et al., 1992). If there is a history of a possible metallic foreign body, then plain
radiographs are recommended as an initial study. Metallic foreign bodies in the orbit are
considered an absolute contraindication to MR, as there have been reports of blindness as
a result. Although CT is more sensitive in detecting smaller metallic foreign bodies, and is A B
advocated by some, plain films are adequate. There are a few metal implants now used in
the eye, particularly with cataract surgery.
Generally standard screening forms are used for all patients scanned in a magnetic
resonance system.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(1996) for discussion of what implants may be safely scanned using magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
2. If the procedure is a research protocol, have the patient sign any necessary consent Figure A7.5.1 Magnetic susceptibility artifacts. Image (A) is a transverse T2-weighted image. Image (B) is a transverse
form. proton density image. Both images demonstrate magnetic susceptibility artifact from mascara distorting the globes. There
is loss of signal and “smearing” of the image.
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
One should double-check for any metallic objects.
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating
and image artifacts.
Mascara and other makeup commonly produce magnetic susceptibility artifacts (Fig.
A7.5.1). This is particularly important in the evaluation of the orbit. Some tattoos will also
produce artifacts (Sacco et al., 1987; Herrick et al., 1997) and possible heating.
5. Inform the patient about what will occur during the procedure, what he or she will
experience while in the magnet, and how to behave, including the following:
a. If earphones or headphones are used to protect the ears from the loud sounds
produced by the gradients, the patient will be asked to wear these, but will be able
to communicate with you at any time during the imaging. Many systems have
handsets that can be used to alert the technologists if there is a problem.
If the patient does not utilize a headset then earplugs are necessary.
b. The patient will be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this works).
Figure A7.5.2 Sagittal T1-weighted image phase encoded artifacts
c. For good results the patient should not talk, and should avoid or minimize (ghosting). This is a sagittal image of the orbit demonstrating a choroidal
swallowing or other movement, during each scan—i.e., as long as the banging melanoma. Note the phase encoding artifact (arrows) of the muscle cone.
sounds continue. Between scans, talking and swallowing are allowed in most Since the globe cannot be fixed, motion artifacts are common. This patient
The Orbit and demonstrates a sclera band and an invasive melanoma of the choroids
Optic Nerves along the superior posterior portion of the globe. Head and Neck

A7.5.2 A7.5.3
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 For orbital imaging, it is best not to use visual stimuli because the stimuli will encourage Table A7.5.3 Parameters for Transverse T2-Weighted Images (Sequence 2)
eye motion.
8. Center the head in the head coil or adjust the surface coil. Patient position Supine
Scan type Spin echo
It may be important to place sponges around the patient to help support the head in a
Imaging plane (orientation) Transverse
comfortable position and to suppress motion during the scanning process. If the patient is
not in a comfortable position he or she will most likely move during the exam. Central slice or volume center Laser light centered on nose
Echo time (TE) 90 msec
9. If needed, place a pillow or other support under the knees to make the patient more Repeat time (TR) 5000 msec
comfortable. Flip angle (FA) 90°
10. Use the centering light to position the patient’s nose and put him or her into the center Fields of view (FOVx, FOVy) 240 mm, 240 mm
of the magnet. Resolution (Δx, Δy) 0.94 mm, 0.94 mm
Once this step has been performed, so long as the patient does not move on the table, the Number of data points collected (Nx, Ny) 256, 256
table itself can be moved and then replaced in the same position as before without Display matrix (Dx, Dy) 256, 256
jeopardizing the positioning of one scan relative to another. Slice thickness (Δz) 3–5 mm
Number of slices 20
11. If the patient is unable to hold still, provide an appropriate sedative.
Slice gap 1–2 mm
Number of acquisitions (Nacq) 1–2
Sequence 1: Sagittal T1-weighted series
Swap read and phase encoding Yes
12. Run T1-weighted, short TE, short TR, spin-echo sequence (Table A7.5.2).
Saturation pulses Not applicable
The field of view must be large (24 cm) so that portions of the head, and particularly the Scan time ∼2 min
orbits, are not excluded from the image. In general, there is no great advantage to making
the field of view very small, since wrap-around (aliasing) artifacts will degrade the image
quality when utilizing a head coil. This sequence allows for good visualization of the fat Short tau inversion recovery (STIR) coronal images can also be acquired to help suppress
spaces in the orbits and the resulting image obtained can be used as a planning image for the orbital fat (Table 7.5.10). This sequence highlights the optic nerves. This type of
the other sequences to follow. sequence can help differentiate non-enhancing white matter lesions from contrast enhance-
ment.
Sequence 2: T2-weighted imaging 14. Let the patient know you are ready, and begin the scan. Remind the patient not to
13. Run transverse T2-weighted imaging (sequence 2; Table A7.5.3). move the eyes.
Using the images generated in sequence 1, transverse images are planned to include most
of the sinuses and brain. Set the imaging parameters as shown in Table A7.5.3. This Sequence 3: High resolution transverse T1-weighted contrast enhanced images
sequence will make abnormalities with high water content and long T2 more conspicuous, 15. Remove the patient from the scanner (the patient should not move on the table).
a finding that is of value in characterizing the vitreous and the sinuses. Since the images Establish an intravenous line from which the contrast agent can be injected, and attach
are quite T2-weighted, the fat structures are partially suppressed. this line securely to the patient so that movement into or out of the magnet will not
pull at the patient’s arm. Move the patient back into the scanner.
Table A7.5.2 Parameters for Sagittal T1-Weighted Images (Sequence 1) It is preferable to insert the line prior to imaging and to leave the patient in the magnet,
with no intervening motion between the scans run before contrast agent injection and those
Patient position Supine run after injection.
Scan type Spin echo
Imaging plane (orientation) Sagittal 16. Leaving the patient in the magnet, inject the contrast agent, flush the i.v. line with 10
Central slice or volume center Laser light centered on nose ml saline and run transverse T1-weighted high-resolution scan (sequence 3; Table
Echo time (TE) As short as possible A7.5.4).
Repeat time (TR) 500 msec A dose of 0.1 mmol/kg of contrast agent is usually given.
Flip angle (FA) 90°
Sequence 3 is a high-resolution T1-weighted series that highlights the fat spaces of the
Fields of view (FOVx, FOVy) 240 mm, 240 mm
bones and orbits (Fig. A7.5.3). The arterial blood vessels (carotid, ophthalmic, circle of
Resolution (Δx, Δy) 0.94 mm, 0.94 mm
Willis, and the internal maxillary branches) can all be seen as well. Small structures within
Number of data points collected (Nx, Ny) 256, 256 the orbit, including some of the cranial nerves and the veins, can also be visualized as
Display matrix (Dx, Dy) 256, 256 lower-signal-intensity regions.
Slice thickness (Δz) 3–5 mm
Since these images are a three-dimensional acquisition and have high resolution of high
Number of slices 15
quality, it is possible to post-process them into any other orientation plane or surface
Slice gap 1–2 mm images.
Number of acquisitions (Nacq) 1–2
Swap read and phase encoding Not applicable If this type of high-resolution three-dimensional pulse sequence is not possible, then a
The Orbit and Saturation pulses Not applicable proton-density or T1-weighted spin-echo sequence with relatively thin slices would be an
Optic Nerves appropriate imaging alternative. Head and Neck
Scan time 2–4 min
A7.5.4 A7.5.5
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 5
 Table A7.5.4 Parameters for Transverse T1-Weighted High Resolution Images
(Sequence 3)

Patient Position Supine

Scan type 3-dimensional gradient echo
Imaging plane (orientation) Transverse
Central slice or volume center Laser light centered on nose
Echo time (TE) As short as possible
Repeat time (TR) 50 msec
Flip angle (FA) 30°
Fields of view (FOVx, FOVy) 240 mm, 240 mm
Resolution (Δx, Δy) 0.94 mm, 0.94 mm
Number of data points collected (Nx, Ny) 256, 256
Display matrix (Dx, Dy) 512, 256
Slice thickness (Δz) 1.5–2 mm
Number of slices 28–37
Slice gap 0 mm
Number of acquisitions (Nacq) 1
Swap read and phase encoding Yes
Saturation pulses Not applicable
Scan time ∼6–7 min

Sequence 4: Transverse T1-weighted contrast enhanced imaging with fat saturation

17. Obtain slices from the anterior margins of the orbit back into the posterior fossa Figure A7.5.3 High-resolution transverse T1-weighted en-
(sequence 4; Table A7.5.5). hanced image. This is a 3-D sequence and can be reformat-
ted in any orientation since the image thickness is only 1.5
This sequence takes advantage of the combination of fat saturation and contrast enhance- mm and the matrix is 512. This sequence is used to create
ment (see Barakos et al., 1991; Anzai et al., 1992; Amano et al., 1997). Since there are high-quality-resolution images. The higher the initial quality of
many fatty spaces in and around the orbit, recognition of contrast enhancement on the image, the better the reconstruction. This sequence is
T1-weighted images may not be possible without fat saturation, since the fat is already high optimal for evaluation of cranial nerves intracranially, prior to
in signal intensity. Fat saturation makes the contrast enhancement more conspicuous. This their entering the orbit. The image is similar to a T1-weighted
is particularly important when imaging infiltrative lesions of the orbit. Fat saturation also spin echo sequence, but much higher resolution. Note the
makes the normally enhancing structures such as the extraocular muscles stand out (Fig. good definition of the gray and white matter, the cerebrospinal
A7.5.4). fluid (CSF) spaces, and the blood vessels.

It is essential to use extra pre-scanning in order to determine the correct frequency to

saturate fat.
If there is anything causing severe degradation of the image quality, such as metal artifact
from dental hardware, it is recommended that the imaging sequence be repeated without
fat saturation in order to ensure a good quality exam.
Sequence 5: Coronal T1-weighted contrast enhanced imaging with fat saturation
18. Run coronal T1-weighted fat saturated contrast enhanced sequence (Fig. A7.5.5) to
evaluate the superior-inferior extension of pathology (sequence 5; Table A7.5.6).
Coronal imaging is also advantageous when evaluating the optic nerve and the extra-ocu-
lar muscles, since they are sectioned perpendicular to their courses. The transverse images
have more problems with partial volume averaging and with the structures looping in and
out of the section plane. Visualization of the cranial nerves in the cavernous sinus is
possible, and it is also important to identify the carotid arteries.
The Orbit and
Optic Nerves
DEDICATED HIGH-RESOLUTION SURFACE COIL ORBITAL STUDY ALTERNATE
If the goal of the examination is to completely focus on the orbital contents, then the use PROTOCOL
of a high-resolution surface coil can be in order (see Atlas et al., 1987). Dedicated surface
coils for the orbits are available, but more general-purpose coils used for small body parts
such as the wrist and temporal mandibular joint (TMJ) can be used. A higher signal-to-
noise ratio is possible with phased array coils. Because the sequences used for this type
of exam are of very high resolution, unless the advantage of the surface coils (in terms of
signal to noise) is greater than the loss of signal resulting from the high resolution and
small field of view, this type of evaluation is not advantageous. The procedure below is
most useful on high-field systems.
Sequence 6: Sagittal, transverse, and coronal T1-weighted imaging
1. Perform sagittal, transverse, and coronal T1-weighted scan (sequence 6; Table
A7.5.7).
Since the pathology may be very small (a few millimeters in dimension) it is advantageous
to acquire all three imaging orientation planes (Table A7.5.7). Because of partial volume
averaging, it may not be possible to see a small choroidal lesion on more than one slice,
and it is frequently impossible to predict which plane will be best for visualization.
Head and Neck

A7.5.6 A7.5.7
Supplement 5 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A7.5.5 Parameters for Transverse T1-Weighted Images Sequence 7: Spin-echo T2-weighted sequence
(Sequence 4) 2. Perform spin-echo T2-weighted scan (sequence 7; Table A7.5.8).
Patient position Supine The imaging orientation plane that best demonstrates the lesion on the T1-weighted images
is also used for the T2-weighted images (Table A7.5.8). When evaluating choroidal masses,
Scan type Spin echo
it is important to characterize the relaxation properties of the mass, since melanomas have
Imaging plane (orientation) Transverse odd signal characteristics that are frequently quite different from other tumors: they have
Central slice or volume center Laser light centered on nose much shorter T1 relaxation times as a result of the paramagnetic properties of melanin.
Echo time (TE) As short as possible Documenting odd signal behavior can help to differentiate a melanoma from another
Repeat time (TR) 500 msec
Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 240 mm, 240 mm
Resolution (Δx, Δy) 0.94 mm, 0.94 mm A B
Number of data points collected (Nx, Ny) 256, 256
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 3–5 mm
Number of slices 20
Slice gap 1–2 mm
Number of acquisitions (Nacq) 1–2
Swap read and phase encoding Yes
Saturation pulses Fat saturation
Contrast enhancement Yes
Scan time 2–4 min

Figure A7.5.5 Coronal T1-weighted fat saturation images obtained with a surface coil. Image (A) is
centered just behind the globe. Note the lack of enhancement of the optic nerve, in contrast to the
extraocular muscles. Image (B) is centered through the globe and allows detailed imaging of the globe.
Note the intense enhancement of the nasal cavity mucosa.

Table A7.5.6 Parameters for Coronal T1-Weighted Contrast Enhanced Images

with Fat Saturation (Sequence 5)

Patient position Supine

Scan type Spin echo
Imaging plane (orientation) Coronal
Central slice or volume center Laser light centered on nose
Echo time (TE) As short as possible
Repeat time (TR) 500 msec
Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 240 mm, 240 mm
Resolution (Δx, Δy) 0.94 mm, 0.94 mm
Number of data points collected (Nx, Ny) 256, 256
Figure A7.5.4 Transverse T1-weighted contrast enhanced Display matrix (Dx, Dy) 256, 256
image of the head with fat suppression. Note the intense Slice thickness (Δz) 3–5 mm
normal contrast enhancement of the extraocular muscles and Number of slices 20
the lack of signal from the orbital fat which allows for evalu- Slice gap 1–2 mm
ation of the orbital contents. The suppression of the fat in-
Number of acquisitions (Nacq) 1–2
creases the signal intensity differences between the contrast
enhanced structures and the other orbital structures, particu- Swap read and phase encoding No
larly the optic nerve. Without fat saturation the contrast en- Saturation pulses Fat saturation
hancement is “lost” in the fat signal. Imaging the entire head Contrast enhancement Yes
The Orbit and
Optic Nerves also allows for the evaluation of intracranial structures, which Scan time 2–4 min Head and Neck
may account for vision changes (occipital lobe lesions).
A7.5.8 A7.5.9
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 tumor, such as a lymphoma, metastases, or a hemangioma. It is possible to assess the T2 Table A7.5.9 Parameters for T1-Weighted Contrast Enhanced Images
properties of the pathology more accurately with a four-echo train than with one or two (Sequence 8)
TE images alone.
Patient position Supine
Sequence 8: T1-weighted contrast-enhanced sequence Scan type Spin echo
3. Perform T1-weighted contrast-enhanced scan (sequence 8; Table A7.5.9). Imaging plane (orientation) Best plane from prior exams
Contrast-enhanced high-resolution surface coil imaging (Figs. A7.5.4 and A7.5.5) can help Central slice or volume centr Laser light centered on nose
to characterize small lesions and potentially make them more conspicuous (Table A7.5.9). Echo time (TE) As short as possible
Repeat time (TR) 500 msec
Flip angle (FA) 90°
Table A7.5.7 Parameters for Sagittal, Transverse, anad Coronal T1-Weighted
Imaging (Sequence 6) Fields of view (FOVx, FOVy) 240 mm, 240 mm
Resolution (Δx, Δy) 0.94 mm, 0.94 mm
Patient position Supine Number of data points collected (Nx, Ny) 256, 256
Scan type Spin echo Display matrix (Dx, Dy) 256, 256
Imaging plane (orientation) Sagittal, transverse, and coronal Slice thickness (Δz) 3 mm
Central slice or volume center Laser light centered on nose Number of slices 20
Echo time (TE) As short as possible Slice gap 1 mm
Repeat time (TR) 500 msec Number of acquisitions (Nacq) 1
Flip angle (FA) 90° Swap read and phase encoding Swap if transverse plane is used
Fields of view (FOVx, FOVy) 240 mm, 240 mm Saturation pulses If desired
Resolution (Δx, Δy) 0.94 mm, 0.94 mm Contrast enhancement Yes
Number of data points collected (Nx, Ny) 256, 256 Scan time ∼2–4 min
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 3 mm
Table A7.5.10 Parameters for STIR-Weighted Contrast Enhanced Images
Number of slices 20 (Sequence 9)
Slice gap 1–2 mm
Number of acquisitions (Nacq) 1–2 Patient position Supine
Swap read and phase encoding Swap if transverse plane is used Scan type Short tau inversion recovery, spin
Saturation pulses Not applicable echo
Scan time ∼2–4 min Imaging plane (orientation) Best plane from prior exams
Central slice or volume center Laser light centered on nose
Echo time (TE) As short as possible
Table A7.5.8 Parameters for T2-Weighted Image in the Plane Best
Demonstrating the Pathology (Sequence 7)
Repeat time (TR) 1600 msec
Inversion time (TI) 90 msec
Patient position Supine Flip angle (FA) 180°
Scan type Spin echo Fields of view (FOVx, FOVy) 240 mm, 240 mm
Imaging plane (orientation) Best plane from the result of Resolution (Δx, Δy) 0.94 mm, 0.94 mm
sequence 6 Number of data points collected (Nx, Ny) 256, 256
Central slice or volume center Laser light centered on nose Display matrix (Dx, Dy) 256, 256
Echo time (TE) 30, 60, 90, 120 msec Slice thickness (Δz) 3 mm
Repeat time (TR) 5000 msec Number of slices 20
Flip angle (FA) 90° Slice gap 1 mm
Fields of view (FOVx, FOVy) 240 mm, 240 mm Number of acquisitions (Nacq) 1
Resolution (Δx, Δy) 0.94 mm, 0.94 mm Swap read and phase encoding Swap if transverse plane is used
Number of data points collected (Nx, Ny) 256, 256 Saturation pulses If desired
Display matrix (Dx, Dy) 256, 256 Contrast enhancement Yes
Slice thickness (Δz) 3 mm Scan time ∼7 min
Number of slices 20
Slice gap 1 mm
Number of acquisitions (Nacq) 1–2
Swap read and phase encoding Swap if transverse plane is used
Saturation pulses Not applicable
The Orbit and
Optic Nerves Scan time ∼2 min Head and Neck

A7.5.10 A7.5.11
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Fat saturation can be added, but will have the usual limitations. Motion of the globe and inhomogeneous field across the tissue. There- Anzai, Y., Lufkin, R.B., Jabour, B.A., and Hanafee,
blood flow-related artifacts become more troublesome with these types of sequences. fore, when fat saturation is utilized, there may W.N. 1992. Fat-suppression failure simulating
pathology on frequency-selective fat-suppres-
be regions in the tissue that are accidentally
sion images of the head and neck. AJNR Am. J.
Sequence 9: Short tau inversion recovery (STIR) weighted contrast enhanced suppressed and parts of the fat that are not. In Neuroradiol. 13(3):879-884.
sequence this case, there may be an appearance of false
Atlas, S.W., Bilaniuk, L.T., Zimmerman, R.A., et al.
4. Perform Short tau inversion recovery (STIR) weighted contrast enhanced scan enhancement. If there is question about the true 1987. Orbit: Initial experience with surface coil
(sequence 9; Table A7.5.10). source of the signal, it is advisable to acquire spin-echo imaging at 1.5 T. Radiology 164:501-
images without saturation pulses. 509.
On systems that do not have fat saturation, short tau inversion recovery (STIR) sequences Chemical shift artifacts can obscure detail, Barakos, J.A., Dillon, W.P., and Chew, W.M. 1991.
can be of value (Table A7.5.10). This is true for lower-field magnets in particular. This particularly at the margins of the extraocular Orbit, skull base and pharynx: Contrast en-
sequence suppresses the fat by choosing an inversion time that brings the signal to fat near muscles and the globe. The best way to avoid hanced fat suppression MR imaging. Radiology
null (signal void). This technique is not specific for fat though, but rather for suppression 179:191-198.
this problem is to use fat saturation techniques.
of a range of short T1 tissues. In this sequence, the fat will have a lower signal than usual, Herrick, R.C., Hayman, L.A., Taber, K.H., et al.
enhancing the contrast difference with other tissues. 1997. Artifacts and pitfalls in MR imaging of the
Anticipated Results or bit: A clinical review. Radiographics
The goal of the exam is to identify and 17(3):707-724.
COMMENTARY characterize all of the orbital structures and
Kelly, W.M., Paglen, P.G., Pearson, J.A., et al. 1986.
their adjacent connections. Since these struc- Ferromagnetism of intraocular foreign body
Background Information are important, since pathology may be present tures can have a wide variation in their signal causes unilateral blindness after MR study.
Orbital pathology can be very diverse, so it even when the nerve’s normal configuration is properties, it is necessary to use multiple imag- AJNR Am. J. Neuroradiol. 7:243-245.
is difficult to predict which technique will be unaltered, as in the case of optic neuritis. Se- ing sequences. It should be possible to follow Lagouros, P.A., Langer, B.G., Peyman, G.A., et al.
ideal for any one patient. Pathology embedded quences should evaluate the intracranial seg- the optic nerve from the globe to the optic 1987. Magnetic resonance imaging and in-
within the orbital fat is frequently easier to ments of the optic nerves, cranial nerves, cere- radiations. Each extraocular muscle should be traocular foreign bodies. Arch. Ophthalmol.
evaluate, since the fat acts as a natural contrast brospinal fluid (CSF) spaces (optic nerve 105:551-553.
identifiable and the lacrimal glands should be
agent and outlines the pathology (which fre- sheath, Meckel’s cave, prepontine cistern, and both visible and symmetric. The fatty spaces of Mafee, M.F., Ainbinder, D., Afshani, E., and Mafee,
quently has lower signal and longer T1 relaxa- the suprasellar cistern) and the vascular spaces R. 1996. The eye, imaging of the globe, orbit,
the pterygopalatine fissure, superior orbital fis-
tion times), though this may not be true for (cavernous sinus, carotid artery). Pathology and visual pathways. Neuroimaging Clin. North
sure, and inferior orbital fissure should be vis- Am. 6:29-59.
orbital lipomas. Lipomas usually occur in the frequently bridges these structures, as in the ible. In addition, the superior orbital fissure,
anterior portion of the orbit so it is easier to case of perineural spread of tumor. Otto, P.M., Otto, R.A., Virapongse, C., et al. 1992.
foramen rotundum, and optic canal should be Screening test for detection of metallic foreign
recognize the pathology. Thyroid ophthal- Despite its relatively simple configuration, well demonstrated, including their associated objects in the orbit before magnetic resonance
mopathy is frequently associated with infiltra- the globe is difficult to image because of the nerves and vessels. It is also important to ex- imaging. Invest Radiol 27:308-311.
tion of the fat without much of a change in small size of its components and its tendency clude pathology of the cavernous sinus and Sacco, D.C., Steiger, D.A., Bellon, E.M., et al. 1987.
signal properties, in which case one sees prop- to move. Surface coil and fat saturation imaging carotid arteries. Artifacts caused by cosmetics in MR imaging of
tosis without an apparent mass. One can usually may be helpful. If there is pathology, its complete extent the head. AJR Am. J. Roentgenol. 148:1001-
differentiate hemorrhage in the orbit from fat 1004.
should be fully evaluated. In the case of the
by its signal characteristics on T2 or gradient Critical Parameters and orbit, it is possible that pathology may track all Shellock, F.G. 1996. Pocket Guide to MR Proce-
echo imaging. Iron products in the form of Troubleshooting dures and Metallic Objects. Lippincott-Raven,
the way back to the occipital poles. Since the
met-hemoglobin can have high signal on T1- Suppression of globe motion is almost im- Philadelphia.
superior inferior extension of pathology can
weighted imaging, but frequently demonstrate possible, which is unfortunate, since it can also be very broad, it may also be important to
high signal properties on T2-weighted images severely mar the quality of an exam, particu- Key References
image all the way back to the foramen magnum. Mafee et al., 1996. See above.
as well. More acute or chronic hemorrhage larly if the exam is focused on the globe. Fast
(hemosiderin) can demonstrate very low signal image sequences may help suppress the effects This reference is a general review article focused on
Acknowledgements imaging of the orbit and visual system.
due to the marked T2 shortening that is associ- of motion, but these sequences frequently have The authors would like to thank Linda Chak-
ated with these strongly magnetic components. lower resolution and simply do not have good eres for the editing of the manuscript.
Usually, the orbital fat clearly outlines pa- enough image quality to answer some clinical
thology of the extraocular muscles. One can see questions. Contributed by Donald William Chakeres
Literature Cited and Eric C. Bourekas
infiltration of the muscles without enlargement If the patient has metallic hardware, fat satu- Amano, Y., Amano, M., and Kumazaki, T. 1997. Ohio State University College of Medicine
on noncontrast T2-weighted or contrast-en- ration imaging may be impossible. In fact, Normal contrast enhancement of extraocular and Public Health
hanced T1-weighted imaging. routine imaging of the orbit may be impossible muscles: Fat-suppressed MR findings. AJNR
Am. J. Neuroradiol. 18:161-164.
Columbus, Ohio
Pathology of the paranasal sinuses is usually in some patients who have dental hardware, and
obvious since the air spaces of the sinuses even patients with no dental hardware can have
should be seen as signal voids. Inspissated bulk magnetic susceptibility artifacts. Mag-
secretions may mimic the appearance of air netic susceptibility pads can be placed over the
spaces, with no signal visible. This appearance eyes to limit artifacts.
will be contradictory to findings seen on CT or The magnetic field within a patient is not
plain radiographs. constant; it varies with the composition of the
Evaluation of the optic nerve is very impor- tissue being studied. The susceptibility differ-
The Orbit and
Optic Nerves tant. Therefore, contrast-enhanced sequences ence between air and soft tissues leads to an Head and Neck

A7.5.12 A7.5.13
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
Imaging of the Paranasal Sinuses UNIT A7.6 Table A7.6.1 Equipment Parameters

Coil type Circularly polarized head coil (or

While the preferred initial imaging modality of sinonasal disease has traditionally been phased-array neck coil, if available)
computed tomography (CT), magnetic resonance imaging (MRI) has demonstrated Gradient coil strength 30 mT/m (or whatever the system permits)
several unique benefits for examination of the sinonasal cavities. Specifically, MRI, in Cardiac gating No
most instances, can better distinguish between soft tissues when compared with CT. The Peripheral gating For safety only
distinction of soft tissues becomes most clinically relevant when evaluating for possible Respiratory gating No
sinonasal tumors and when inspecting for extension of tumor or inflammation into the Respirator If required by patient
orbit or intracranial spaces. It is in these areas that MRI has found its major application Oxygen If required by patient
in sinus imaging. This unit presents MRI techniques for imaging the sinuses with (see Motion cushions Useful
Basic Protocol 2) and without (see Basic Protocol 1) contrast enhancement. The parame- Contrast agent Used in Basic Protocol 2
ters given here are derived from experience at 1.5 T and may need to be slightly altered
depending on the field strength and the equipment manufacturer.
find out if the patient has any health conditions that may require the presence of the
special emergency equipment during the scanning procedure or any other necessary
IMAGING OF THE PARANASAL SINUSES BY MRI WITHOUT CONTRAST BASIC precautions.
ENHANCEMENT PROTOCOL 1
Generally, standard screening forms (APPENDIX 1) are used for all patients scanned in a
Magnetic resonance imaging scans can be performed at a range of field strengths. magnetic resonance system.
Generally, the high signal-to-noise ratio obtained at high-field strengths allows one to
scan either faster or with higher resolution when compared with lower field strengths. The presence of any ferromagnetic metals may be a health hazard to the patient when he
The sequences described within this unit are based on the authors’ experience with a or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
Siemens 1.5 T Symphony scanner, but are expected to be equally applicable to machines
and Crues (1998), Shellock and Kanal (1998a,b), Shellock and Shellock (1998a,b) and
from other manufacturers. Shellock (2001) for a discussion of what implants may be scanned safely using magnetic
resonance.
Scanning a patient or volunteer is a team effort between technologists, nurses, and
physicians. In most cases, the technologist assumes the lead role in acquiring the Patients may be accompanied into the magnet room by a friend or family member who can
appropriate images. sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
The following set of sequences comprises the preferred method of imaging the sinuses.
2. If the procedure is a research protocol, have the patient sign any necessary consent
The usual sequences employed include transverse T1-weighted spin echo and fat-saturated
forms.
T2-weighted fast or turbo spin-echo (FSE) images without contrast agent administration.
Unenhanced MRI may be sufficient for diagnosis in uncomplicated inflammatory disease 3. Have the patient remove all jewelry and change into a gown to eliminate any metal
if there is no suspicion that the sinonasal pathology has violated the skull base. However, that may be found in clothing.
contrast enhancement is frequently necessary to assess whether a tumor or more extensive
sinonasal inflammation has violated the skull base. In cases of suspected skull-base 4. Have the patient wash off any mascara or other make up to avoid local tissue heating
invasion, contrast-enhanced imaging (see Basic Protocol 2) is helpful due to its ability to and image artifacts.
better define the intracranial margins of a lesion or to detect perineural spread of a tumor. 5. Inform the patient about what will occur during the procedure, what he or she will
If a contrast agent is not used, coronal T1-weighted spin echo and T2-weighted fast experience while in the magnet, as well as how to behave. Issues to discuss include
spin-echo scans should also be obtained. The duration for the imaging process including the following:
all preparations will be ∼35 min if no contrast agent is administered. All of these sequences
are non-breathholding. a. If earplugs or headphones are used to protect the ears from loud sounds produced
by the gradients, the patient will be asked to wear these, but will be able to
Table A7.6.1 lists the hardware necessary to perform the examination along with the communicate with you at any time during the imaging.
appropriate parameters. The available gradient strength will depend on the scanner, and b. The patient will be given a safety squeeze bulb or similar equipment to request
the echo times given below may have to be varied accordingly. assistance at any time (demonstrate how this works).
NOTE: Be sure that technologists and nurses have immediate access to any emergency c. For good results, the patient should not talk, and should avoid or minimize
equipment that may be relevant to a given study, or that may be needed for a particular swallowing or other movements during each scan—i.e., as long as the banging
patient, such as crash carts or oxygen. sounds continue. Between scans, talking and swallowing are allowed in most
cases, but should be avoided when comparative positional studies are being
Set up patient and equipment performed; the patient will be informed when this is the case.
1. Interview the patient to ensure that he or she has no contraindications such as cardiac d. Nevertheless, the patient may call out at any time that he or she feels it necessary.
pacemakers or other implants containing ferromagnetic materials. Also, be sure to Imaging of the
Head and Neck Paranasal Sinuses
Contributed by Daniel T. Boll, Michael Coffey, and Jonathan S. Lewin A7.6.1 A7.6.2
Current Protocols in Magnetic Resonance Imaging (2002) A7.6.1-A7.6.12
Copyright © 2002 by John Wiley & Sons, Inc. Supplement 7 Supplement 7 Current Protocols in Magnetic Resonance Imaging
 6. Have the patient lie down on the table. Either before or right after the patient lies Table A7.6.3 Primary Clinical Imaging Parameters for Sequence 2 (T2-Weighted
down, set up any triggering devices or other monitoring equipment that is to be used. Transverse Fat-Saturated TSE)

7. Center the patient in a head coil at the region where the key information is desired. Patient position Supine
Make sure that the head is constrained to prevent motion. Scan type Turbo spin echo
Imaging plane (orientation) Transverse
Generally, the patient’s head is fixed so that the head is horizontal (not tilted) along the
axis of the MRI table. A dedicated head coil is employed. Central slice or volume center Centered on nasion
Echo time (TE) 62 msec
8. If needed, a pillow may be placed under the knees to make the patient more Echo train length (ETL) 7
comfortable. Repeat time (TR) 4000 msec
Flip angle (FA) 150°a
9. Use the centering light centered on the nasion of the patient and put him or her into
Fields of view (FOVx, FOVy) 180 mm, 180 mm
the center of the magnet.
Resolution (Δx, Δy) 0.70 mm, 0.70 mm
Once this step has been performed, so long as the patient does not move on the table, the Number of data points collected (Nx, Ny) 256, 256
table itself can be moved and then replaced in the same position as before without Slice thickness (Δz) 4 mm
jeopardizing the positioning of one scan relative to another. Number of slices 25
10. If the patient is unable to hold still, provide an appropriate sedative. Number of slabs 1
Slice gap 0.8 mm
Sequence 1: The pilot scan Number of acquisitions (Nacq) 2
11. To validate the patient’s position, run a scout scan to ensure correct localization of Swap read and phase encoding No
subsequent data set acquisitions using the imaging parameters given in Table A7.6.2. Read direction Right–left
Fat suppression Yes
This sequence usually consists of the three orthogonal planes to allow localization. The
Scan time 302 sec
images are also often used later to set up total coverage for the volume of interest.
aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this sequence

is 90°.
Sequence 2: Transverse T2-weighted, fat-saturated turbo spin echo (TSE)
12. Display all the coronal, sagittal, and transverse scout images in split window mode
on the scan monitor. 14. Position the slices from the scout images to cover the whole sinonasal region, from
the tip of the nose to the posterior part of the sphenoid sinus in the transverse plane,
13. Change the imaging parameters to those listed in Table A7.6.3.
from the superior margin of the frontal sinus to the palate in the sagittal plane, and
covering both lateral margins of the maxillary sinuses. Run sequence 2 according to
Table A7.6.3.
Table A7.6.2 Primary Clinical Imaging Parameters for Sequence 1 (Pilot Scan) An example is shown in Fig. A7.6.1.
Patient position Supine Sequence 3: Transverse T1-weighted spin echo (SE)
Scan type Gradient echo 15. Copy the imaging parameters such as number of slices, field of view, thickness of
Imaging plane (orientation) Transverse, coronal, and sagittal slices, and center of slices from the performed transverse T2-weighted fat-saturated
Central slice or volume center Laser light centered on nasion
sequence (sequence 2) and change the imaging parameters to those listed in Table
Echo time (TE) ≤5 msec
A7.6.4. Run sequence 3 according to Table A7.6.4.
Repeat time (TR) ≤20 msec
Flip angle (FA) 40° An example is shown in Fig. A7.6.2.
Fields of view (FOVx, FOVy) 280 mm, 280 mm
Resolution (Δx, Δy) 2.19 mm, 2.19 mm
Sequence 4: Coronal T1-weighted spin echo
Number of data points collected (Nx, Ny) 128,128
16. Display the transverse images (use images from sequence 2 or 3) on the scan monitor.
Slice thickness (Δz) 10 mm 17. Change the imaging parameters to those listed in Table A7.6.5. Position slices to
Number of slices 1 per orientation center of the T1- or T2-weighted spin echo image ensuring that the sinuses are fully
Slice gap Not applicable covered. Run sequence 4 according to Table A7.6.5.
Number of acquisitions (Nacq) 1 An example is shown in Fig. A7.6.3.
Swap read and phase encoding No
Read direction Anterior–posterior for the transverse Sequence 5: Coronal T2-weighted, fat-saturated turbo spin echo
images; Left–right for the coronal 18. Copy the imaging parameters such as number of slices, field of view, thickness of
image; and cranial–caudal for the
slices and center of slices from the performed coronal T1-weighted sequence (se-
sagittal image
Imaging of the quence 4) and change the imaging parameters to those listed in Table A7.6.6. Run
Scan time 9 sec
Head and Neck Paranasal Sinuses sequence 5 according to Table A7.6.6.
A7.6.3 A7.6.4
Current Protocols in Magnetic Resonance Imaging Supplement 7 Supplement 7 Current Protocols in Magnetic Resonance Imaging
Figure A7.6.1 This T2-weighted TSE image shows an opacification (arrow) of the right maxillary
sinus.
Figure A7.6.2 This T1-weighted SE image shows the maxillary sinus of an 8-month-old male.

Table A7.6.4 Primary Clinical Imaging Parameters for Sequence 3 (Transverse

Table A7.6.5 Primary Clinical Imaging Parameters for Sequence 4 (Coronal
T1-Weighted Spin Echo)
T1-Weighted Spin Echo)
Patient position Supine
Patient position Supine
Scan type Spin echo
Scan type Spin echo
Imaging plane (orientation) Transverse
Imaging plane (orientation) Coronal
Central slice or volume center Centered on nasion
Central slice or volume center Centered on nasion
Echo time (TE) 10 msec
Echo time (TE) 10 msec
Repeat time (TR) 532 msec
Repeat time (TR) 532 msec
Flip angle (FA) 90°
Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 180 mm, 180 mm
Fields of view (FOVx, FOVy) 180 mm, 180 mm
Resolution (Δx, Δy) 0.70 mm, 0.70 mm
Resolution (Δx, Δy) 0.70 mm, 0.70 mm
Number of data points collected (Nx, Ny) 256, 256a
Number of data points collected (Nx, Ny) 256, 256a
Slice thickness (Δz) 4 mm
Slice thickness (Δz) 4 mm
Number of slices 25
Number of slices 19
Number of slabs 1
Number of slabs 1
Slice gap 0.8 mm
Slice gap 0.8 mm
Number of acquisitions (Nacq) 2
Number of acquisitions (Nacq) 2
Swap read and phase encoding No
Swap read and phase encoding No
Read direction Right–left
Read direction Right–left
Scan time 242 sec
Scan time 242 sec
aThe true number of collected phase encoding lines is 224, a reduction due to a partial Fourier factor
aThe true number of collected phase encoding lines is 224, a reduction due to a partial Fourier factor
of 7/8.
of 7/8.

Imaging of the
Head and Neck Paranasal Sinuses

A7.6.5 A7.6.6
Current Protocols in Magnetic Resonance Imaging Supplement 9 Supplement 9 Current Protocols in Magnetic Resonance Imaging
 Table A7.6.7 Primary Clinical Imaging Parameters for Sequence 6 (Transverse
T1-Weighted Fat-Saturated Spin Echo)

Patient position Supine

Scan type Spin echo
Imaging plane (orientation) Transverse
Central slice or volume center Centered on nasion
Echo time (TE) 10 msec
Repeat time (TR) 500 msec
Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 180 mm, 180 mm
Resolution (Δx, Δy) 0.70 mm, 0.70 mm
Number of data points collected (Nx, Ny) 256, 256a
Slice thickness (Δz) 4 mm
Number of slices 25 (total)
Number of slabs 2
Slice gap 0.8 mm
Number of excitations (NEX) 2b
Number of acquisitions (Nacq) 2
Swap read and phase encoding No
Read direction Right–left
Fat suppression Yes
Scan time 390 sec
aThe true number of collected phase encoding lines is 192, a reduction due to a partial Fourier factor

of 3/4.
bThe number of cocatenation is set to be 2. This means that only half of the total slices will be excited

during a given repeat time.

Figure A7.6.3 This T1-weighted SE image shows a small retention cyst or polyp (arrow) within
the left maxillary sinus. BASIC IMAGING OF THE PARANASAL SINUSES BY MRI WITH CONTRAST
PROTOCOL 2 ENHANCEMENT
Table A7.6.6 Primary Clinical Imaging Parameters for Sequence 5 (Coronal In cases of suspected skull base, intracranial, intraorbital, or perineural extension of
T2-Weighted Fat-Saturated TSE)
sinonasal pathology, the administration of a contrast agent is advised. In addition, some
Patient position Supine
have suggested a role for post-contrast images if a solid mass cannot be distinguished
Scan type Turbo spin echo
from a cystic lesion on the non-contrast-enhanced examination, as characteristic enhance-
Imaging plane (orientation) Coronal
ment of a cyst wall can often be appreciated. If a contrast agent is employed, transverse
Central slice or volume center Centered on nasion
and coronal T1-weighted, fat-suppressed, contrast-enhanced sequences should be ob-
Echo time (TE) 62 msec
tained. If a contrast agent is administered, the duration of the imaging process including
Echo train length (ETL) 7 all preparations will be ∼50 min. All these sequences are non-breathholding.
Repeat time (TR) 4000 msec Materials
Flip angle (FA) 150°a
Normal saline (0.9% NaCl), sterile
Fields of view (FOVx, FOVy) 180 mm, 180 mm
Gadolinium-based MR contrast agent (e.g., Magnevist, Omniscan, or Prohance)
Resolution (Δx, Δy) 0.70 mm, 0.70 mm
Number of data points collected (Nx, Ny) 256, 256 1. Perform Basic Protocol 1, steps 1 to 6.
Slice thickness (Δz) 4 mm
2. Establish an intravenous line from which the contrast agent can be injected, and attach
Number of slices 19
this line securely to the patient so that movement into or out of the magnet will not
Number of slabs 1
Slice gap 0.8 mm
pull at the patient’s arm.
Number of acquisitions (Nacq) 2 It is preferable to insert the line prior to imaging and to leave the patient in the magnet, so
Swap read and phase encoding No that there is no intervening motion between the scans run before contrast agent injection
and those run after injection.
Read direction Right–left
Fat suppression Yes 3. Perform Basic Protocol 1, steps 7 to 10.
Scan time 302 sec 4. Perform Basic Protocol 1, sequences 1, 2, and 3.
aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this sequence
5. Administer the contrast agent. A dose of 0.2 mmol/kg gadolinium is usually given at
is 90°. Imaging of the
Head and Neck Paranasal Sinuses a rate of 1 ml/sec.

A7.6.7 A7.6.8
Current Protocols in Magnetic Resonance Imaging Supplement 9 Supplement 9 Current Protocols in Magnetic Resonance Imaging
 Figure A7.6.5 This T1-weighted SE image with fat saturation shows contrast enhancement of the
mucous membranes (arrow) of the opacified right maxillary sinus, same patient as in Figure A7.6.4.

6. Perform the following sequences 3 min post-injection.

Figure A7.6.4 This T1-weighted SE image with fat saturation shows contrast enhancement of the Sequence 6: Transverse T1-weighted, fat-saturated spin echo
mucous membranes (arrow) of the opacified right maxillary sinus. 7. Copy the imaging parameters such as number of slices, field of view, thickness of
slices, and center of slices from the performed transverse T1-weighted sequence
Table A7.6.8 Primary Clinical Imaging Parameters for Sequence 7 (Coronal (sequence 3) and change the imaging parameters to those listed in Table A7.6.7. Run
T1-Weighted Fat-Saturated Spin Echo) sequence 6 according to Table A7.6.7.
Patient position Supine An example is shown in Figure A7.6.4.
Scan type Spin echo Sequence 7: Coronal T1-weighted, fat-saturated spin echo
Imaging plane (orientation) Coronal 8. Copy the imaging parameters such as number of slices, field of view, thickness of
Central slice or volume center Centered on nasion slices, and center of slices from the performed coronal T1-weighted sequence (se-
Echo time (TE) 10 msec quence 4) and change the imaging parameters to those listed in Table A7.6.8. Run
Repeat time (TR) 723 msec sequence 7 according to Table A7.6.8.
Flip angle (FA) 90°
An example is shown in Fig. A7.6.5.
Fields of view (FOVx, FOVy) 180 mm, 180 mm
Resolution (Δx, Δy) 0.70 mm, 0.70 mm
Number of data points collected (Nx, Ny) 256, 256a COMMENTARY
Slice thickness (Δz) 4 mm in their sinuses demonstrated on MRI do not
Background Information
Number of slices 19 The examination of the anatomy and pathol- have symptoms classically attributed to sinusi-
Number of slabs 1 ogy of paranasal sinuses, as seen by MRI as tis (Cooke and Hadley, 1991).
Slice gap 0.8 mm well as CT, shows a wide variety of findings in The most common inflammatory abnor-
Number of acquisitions (Nacq) 2 a range of 15% to 40% of all adults that undergo malities noted on imaging studies include
Swap read and phase encoding No imaging (Conner et al., 1989). Some of the opacified sinus cavities, intrasinus air-fluid lev-
Read direction Right–left findings can be correlated to acute upper respi- els and polyps, mucoperiosteal thickening, iso-
Fat suppression Yes ratory infection and allergic episodes with a lated increased signal of the mucosal lining,
Scan time 281 sec typical seasonal pattern (Moore et al., 1986). mucus retention cysts, and fungal concretions
aThe true number of collected phase encoding lines is 192, a reduction due to a partial Fourier factor But most of the abnormalities in the paranasal (Zinreich, 1990). Abnormalities were most
sinuses were unrelated to the patients’ present- commonly seen in the ethmoid sinuses fol-
of 3/4.
ing problems (Moser et al., 1991) and are, lowed by the maxillary cavities, and, to a lesser
therefore, often reported as incidental findings extent, in the sphenoid and frontal sinuses
on scans performed for indications other than (Wani et al., 2001).
the evaluation of a paranasal sinus pathology. Findings of high significance on paranasal
Imaging of the Thus, many people with inflammatory changes sinus studies are mastoiditis as a known cause
Head and Neck Paranasal Sinuses

A7.6.9 A7.6.10
Current Protocols in Magnetic Resonance Imaging Supplement 9 Supplement 9 Current Protocols in Magnetic Resonance Imaging
of lateral venous sinus thrombosis (Fink and
McAuley, 2002), and sphenoid mucoceles. Al-
though benign, sphenoid mucoceles may in-
volve the orbit and cause acute restrictive
ophthalmoplegia, proptosis, and reduced visual
acuity (Conner et al., 1989).
Malignant pathologies involving the
paranasal sinuses are rare, but nasopharyngeal
carcinoma may spread to the paranasal sinuses.
Maxillary sinus infiltration by this tumor is
most common but contrast-enhanced MRI al-
lows differentiating tumor from inflammatory
changes in all sinuses (Chong et al., 1998). MR
imaging is also indicated for the evaluation of
primary tumors of the sinuses and nasal cavity,
including squamous cell carcinoma, esthe-
sioneuroblastoma, lymphoma, and inverted
papilloma. In the presence of an aggressive
tumor, particular attention should be directed
toward dural thickening or invasion, periorbital
invasion, or perineural extension. In addition,
the pterygopalatine fossa should be examined
for evidence of posterior direct or perineural
extension in the presence of maxillary sinus or
hard palate involvement by tumor.
Magnetic resonance imaging has proven to
be a useful tool for imaging the paranasal si-
nuses and nasal cavities. Of particular interest
in this region is the ability to distinguish lesions
that are highly cellular with little free water,
demonstrating intermediate signal on T2-
weighted images, most likely representing tu-
mor, from those lesions that have significant
amounts of serous and mucinous secretions,
very high in signal on T2-weighted images, that
more likely represent an inflammatory or post-
obstructive process containing predominantly
free water (Hasso and Lambert, 1994).
Furthermore, the multiplanar capabilities of
MRI allow identifying tumor margins demon-
strating the extent of the disease and specifically
the relationship to key landmarks critical for fur-
ther surgical treatment (Weissman et al., 1990).
CT remains the examination of choice for
evaluation of the thin cortical bones of the
osteomeatal complex, and for visualization of
osteolytic regions with high sensitivity and spa-
tial resolution (Weissman et al., 1990). Another
indication for CT is for evaluation of osteomas
involving the paranasal sinuses (Beahm et al.,
1990).
Critical Parameters and
Troubleshooting
The high magnetic field strength of 1.5 T
magnets and T2-weighted spin echo techniques
reveal a remarkably intense signal from abnor-
Current Protocols in Magnetic Resonance Imaging
mal tissue in the paranasal sinuses. Inflamma-
tory disease in the maxillary, sphenoid, eth-
moid, and frontal sinuses is detected and dem-
onstrated with greater clarity than any other
available technique. The additional administra-
tion of gadolinium as a contrast agent allows
detection and characterization of malignant
processes and their margins, along with detec-
tion of extension and infiltration when fat-satu-
rated T1-weighted spin echo techniques are
applied (Moore et al., 1986).
The majority of the pathological processes
can be adequately imaged by transverse T1- and
T2-weighted pulse sequences. When more tis-
sue specific information is required, additional
coronal and sagittal planes are helpful for tu-
mors that infiltrate the skull base or are sur-
rounded by air-filled sinuses (Beahm et al.,
1990).
High spatial resolution is desired so that
small areas of signal variation mostly due to
edema are not missed. It is also important to
center the slice-selecting planes in the trans-
verse, coronal, and sagittal orientation to per-
form a comparison with the opposite side in
order to identify any area of abnormality and
differentiate between pathology and anatomi-
cal variation.
Anticipated Results
A variety of findings within the paranasal
sinuses can be identified utilizing magnetic
resonance imaging and computed tomography.
The primary role of CT is in visualizing the
thin cortical bones of the osteomeatal complex
in order to precisely identify bony abnormali-
ties with high sensitivity and spatial resolution.
MRI is considered by many the imaging
modality of choice for evaluating any soft-tis-
sue lesion of the paranasal sinuses. These le-
sions may range from rather less significant
importance for the patient with no further treat-
ment such as mild sinus inflammation, polyps,
mucus retention cysts, mucoperiosteal thicken-
ing, or simple inflammatory increased signal of
the epithelial lining, to less common lesions
such as fungal concretions, sphenoid mucoce-
les, and infections such as mastoiditis, which
may result in further even life threatening com-
plications.
Although a specific diagnosis can be sug-
gested in most cases by other means of visuali-
zation, only MRI has the potential to demon-
strate the extent of the disease with high preci-
sion.
The visualization of malignancies such as
nasopharyngeal carcinoma, sinus carcinoma,
lymphoma, esthesioneuroblastoma, or nasal
cavity tumors can be accomplished with a
maximum degree of spatial resolution by MRI
and thereby help to identify routes of tumor
spread and key surgical landmarks for further
therapy.
Literature Cited
Beahm, E., Teresi, L., Lufkin, R., and Hanafee, W.
1990. MR of the paranasal sinuses. Surg. Radiol.
Anat. 12:203-208.
Chong, V.F., Fan, Y.F., and Khoo, J.B. 1998. Com-
puted tomographic and magnetic resonance im-
aging findings in paranasal sinus involvement in
nasopharyngeal carcinoma. Ann. Acad. Med.
Singapore 27:800-804.
Conner, B.L., Roach, E.S., Laster, W., and Georgitis,
J.W. 1989. Magnetic resonance imaging of the
paranasal sinuses: Frequency and type of abnor-
malities. Ann. Allergy 62:457-460.
Cooke, L.D. and Hadley, D.M. 1991. MRI of the
paranasal sinuses: Incidental abnormalities and
their relationship to symptoms. J. Laryngol.
Otol. 105:278-281.
Fink, J.N. and McAuley, D.L. 2002. Mastoid air
sinus abnormalities associated with lateral ve-
nous sinus thrombosis: Cause or consequence?
Stroke 33:290-292.
Hasso, A.N. and Lambert, D. 1994. Magnetic reso-
nance imaging of the paranasal sinuses and nasal
cavities. Top Magn. Reson. Imaging 6:209-223.
Moore, J., Potchen, M., Waldenmaier, N., Sierra, A.,
and Potchen, E.J. 1986. High-field magnetic
resonance imaging of paranasal sinus inflamma-
tory disease. Laryngoscope 96:267-271.
Moser, F.G., Panush, D., Rubin, J.S., Honigsberg,
R.M., Sprayregen, S., and Eisig, S.B. 1991. In-
Imaging of the
Head and Neck Paranasal Sinuses
A7.6.11 A7.6.12
Supplement 7 Supplement 7
cidental paranasal sinus abnormalities on MRI
of the brain. Clin. Radiol. 43:252-254.
Shellock, F.G. 2001. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Shellock, F.G. and Crues, J.V. 1998. Aneurysm
clips: Assessment of magnetic field interaction
associated with a 0.2-T extremity MR system.
Radiology 208:407-409.
Shellock, F.G. and Kanal, E. 1998a. Aneurysm clips:
Evaluation of MR imaging artifacts at 1.5 T.
Radiology 209:563-566.
Shellock, F.G. and Kanal, E. 1998b. Yasargil aneu-
rysm clips: Evaluation of interactions with a
1.5-T MR system. Radiology 207:587-591.
Shellock, F.G. and Shellock, V.J. 1998a. Cranial
bone flap fixation clamps: Compatibility at MR
imaging. Radiology 207:822-825.
Shellock, F.G. and Shellock, V.J. 1998b. Spetzler
titanium aneurysm clips: Compatibility at MR
imaging. Radiology 206:838-841.
Wani, M.K., Ruckenstein, M.J., and Parikh, S. 2001.
Magnetic resonance imaging of the paranasal
sinuses: Incidental abnormalities and their rela-
tionship to patient symptoms. J. Otolaryngol
30:257-262.
Weissman, J.L., Tabor, E.K., and Curtin, H.D. 1990.
Magnetic resonance imaging of the paranasal
sinuses. Top. Magn. Reson. Imaging 2:27-38.
Zinreich, S.J. 1990. Paranasal sinus imaging. Oto-
laryngol Head Neck Surg. 103:863-868.
Contributed by Daniel T. Boll, Michael
Coffey, and Jonathan S. Lewin
University Hospitals of Cleveland/
Case Western Reserve University
Cleveland, Ohio
Current Protocols in Magnetic Resonance Imaging
Lymph Node Staging in the Neck UNIT A7.7 Table A7.7.1 Equipment Parameters

Coil type Circularly polarized head coil (or

Examination of the anatomy and pathology of the lymph nodes of the neck as seen by phased-array neck coil, if available)
magnetic resonance imaging (MRI) shows a wide variety of findings and thereby presents Gradient coil strength 30 mT/m (or whatever the system permits)
a challenging task for the radiologist. Patients with a broad range of clinical presentations Cardiac gating No
and disease states are imaged with MRI due to its rapidly evolving improvements in spatial Peripheral gating For safety only
and contrast resolution, the reduction of artifacts with this technique, and the development Respiratory gating No
of new contrast agents (van den Brekel, 2000). Respirator If required by patient
Oxygen If required by patient
Although the reasons for performing MRI are numerous, the MR imaging sequences and
Motion cushions Useful
protocols for the evaluation of lymph nodes are straightforward and allow initial staging
Use of contrast agents Necessary
of disease and evaluation of the success of therapy.

IMAGING OF THE LYMPH NODES USING MRI
Magnetic resonance imaging scans can be performed at a range of field strengths. In
general, the high signal-to-noise ratio (SNR) obtained at high field strengths allows one
to scan either faster or with higher resolution compared to lower field strengths. The
sequences described within this unit are based on the authors’ experience using a Siemens
1.5 T Symphony scanner, but are expected to be equally applicable to scanners from other
manufacturers.
Scanning a patient or volunteer is a team effort of the technologists, nurses, and physi-
cians. In most cases, the technologist assumes the lead role in acquiring the appropriate
images.
The set of sequences described in the following protocol comprises the preferred
method of imaging the soft tissues of the neck. The usual sequences employed include
transverse T1-weighted spin echo and fast spin-echo (FSE) or turbo spin-echo (TSE)
fat-saturated T2-weighted images acquired prior to contrast administration. Contrast
enhancement is necessary to assess the extent of any metastatic spread and to
determine if the cancerous or diseased lymph node has infiltrated any adjacent
structures, most importantly arteries and veins, as well as adjacent nerves. In addition,
a better distinction of cystic lesions and abscesses can often be achieved, compared
to non-contrast examinations.
All of the sequences in this protocol are non-breath-hold examinations; however, the
patient should be advised not to swallow or move his or her head while the sequences
are running. The duration for the imaging process, including all preparations, is ∼30
min.
Table A7.7.1 lists the hardware necessary to perform the procedure along with appropriate
parameters. The available gradient strength will depend on the scanner, and the echo times
given later in other tables will have to be varied accordingly (the smaller the gradient
strength, the longer the echo time for a particular scan).
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.
Materials
Normal saline (0.9% NaCl), sterile
Gadolinium-based MR contrast agent (e.g., Magnevist, Omniscan, or Prohance)
Contributed by Daniel T. Boll, Andrik J. Aschoff, and Jonathan S. Lewin
Current Protocols in Magnetic Resonance Imaging (2003) A7.7.1-A7.7.9
Copyright © 2003 by John Wiley & Sons, Inc.
BASIC Set up patient and equipment
PROTOCOL 1. Interview the patient to ensure that he or she has no contraindication such as cardiac
pacemakers or other implants containing ferromagnetic materials. Also, be sure to
find out if the patient has any health conditions that may require the presence of
special emergency equipment during the scanning procedure, or that might necessi-
tate any other precautions.
Generally, standard screening forms (APPENDIX 1) are used for all patients scanned in a
magnetic resonance system.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
and Kanal (1998), Shellock and Crues (1998), and Shellock and Shellock (1998) for
discussion of what implants may be safely scanned using magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
2. If the procedure is a research protocol, have the patient sign any necessary consent
forms.
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in the clothing.
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating
and image artifacts.
5. Inform the patient about what will occur during the procedure, what he or she will
experience while in the magnet, as well as how to behave. Issues to discuss include
the following:
a. If earphones or headphones are used to protect the ears from the loud sounds
produced by the gradients, the patient will be asked to wear these, but will be able
to communicate with you at any time during the imaging.
b. The patient will be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this works).
c. For optimum results, the patient should not talk, and should avoid or minimize
swallowing or other movement during each scan—i.e., as long as the banging
sounds continue. Between scans, talking and swallowing are allowed in most
cases, but should be avoided when comparative positional studies are being
performed; the patient will be informed when this is the case.
Lymph Node d. Nevertheless, the patient may call out at any time if he or she feels it is necessary.
Staging in the
Head and Neck Neck
A7.7.1 A7.7.2
Supplement 9 Supplement 9 Current Protocols in Magnetic Resonance Imaging
 6. Have the patient climb onto the table. Either before or right after the patient lies down, Table A7.7.3 Primary Clinical Imaging Parameters for Sequence 2 (Transverse
set up any triggering devices or other monitoring equipment that is to be used. T2-Weighted Fat-Saturated TSE)

7. Establish an intravenous line through which the contrast agent can be injected. Attach Patient position Supine
this line securely to the patient so that movement into or out of the magnet will not Scan type Turbo spin echo
pull at the patient’s arm. Imaging plane (orientation) Transverse
Central slice or volume center Epiglottis
It is preferable to insert the line prior to imaging and to leave the patient in the magnet so
that there is no intervening motion between the scans run before contrast agent injection Echo time (TE) 71 msec
and those run after injection. Echo train length (ETL) 11
Repeat time (TR) 6000 msec
8. Center the patient in a head or neck coil at the region where the key information is Flip angle (FA) 150°a
desired. Make sure that the head and neck are constrained to prevent motion, Fields of view (FOVx, FOVy) 192 mm, 153 mm
especially if high-resolution scans are being performed. Resolution (Δx, Δy) 0.75 mm, 0.75 mm
Generally, the patient’s head is fixed so that it is horizontal (not tilted) and the neck and Number of data points collected (Nx, Ny) 256, 204
head lie along the long axis of the scanner table. Most scanners have a dedicated neck coil. Slice thickness (Δz) 5 mm
Number of slices 30
9. If needed, place a pillow or other support under the patient’s knees to make him or Slice gap 1.5 mm
her more comfortable. Number of acquisitions (Nacq) 2
10. Use the centering light to position the patient’s nasion and put him or her into the Read direction Right–left
center of the magnet. Fat suppression Yes
Scan time 233 sec
Once this step has been performed, so long as the patient does not move on the table, the aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this sequence
table itself can be moved and then replaced in the same position as previously without
is 90°.
jeopardizing the positioning of one scan relative to another (usually to within ±1 mm
accuracy).
11. If the patient is unable to hold still, provide an appropriate sedative.

Sequence 1: The pilot scan

12. To validate the patient’s position, run the system’s pilot scan to ensure the correct
location of the neck in three dimensions, using the imaging parameters given in Table
A7.7.2.

Table A7.7.2 Primary Clinical Imaging Parameters for Sequence 1 (Pilot Scan)

Patient position Supine

Scan type Gradient echo
Imaging plane (orientation) Transverse, coronal, and sagittal
Central slice or volume center Laser light centered on the larynx
Echo time (TE) ≤5 msec
Repeat time (TR) ≤20 msec
Flip angle (FA) 40°
Fields of view (FOVx, FOVy) 256 mm, 160 mm
Resolution (Δx, Δy) 1.0 mm, 1.0 mm
Number of data points collected (Nx, Ny) 256, 160 Figure A7.7.1 Mass of the oropharynx on transverse T2-weighted, fat-saturated, MR image
Slice thickness (Δz) 5 mm (arrowhead). Bilateral normal-sized internal jugular chain lymph nodes (arrows) are visible, without
Number of slices 1 per orientation signs of metastatic spread.
Slice gap Not applicable
Number of acquisitions (Nacq) 1
Scan time 9 sec

Lymph Node
Staging in the
Head and Neck Neck

A7.7.3 A7.7.4
Current Protocols in Magnetic Resonance Imaging Supplement 9 Supplement 9 Current Protocols in Magnetic Resonance Imaging
 Table A7.7.4 Primary Clinical Imaging Parameters for Sequence 3 (T1-Weighted Table A7.7.5 Primary Clinical Imaging Parameters for Sequence 4 (T1-Weighted
Transverse) Transverse Fat-Saturated)

Patient position Supine Patient position Supine

Scan type Spin echo Scan type Spin echo
Imaging plane (orientation) Transverse Imaging plane (orientation) Transverse
Central slice or volume center Epiglottis Central slice or volume center Epiglottis
Echo time (TE) 12 msec Echo time (TE) 12 msec
Repeat time (TR) 660 msec Repeat time (TR) 600 msec
Flip angle (FA) 90° Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 192 mm, 156 mm Fields of view (FOVx, FOVy) 192 mm, 156 mm
Resolution (Δx, Δy) 0.75 mm, 0.75 mm Resolution (Δx, Δy) 0.75 mm, 0.75 mm
Number of data points collected (Nx, Ny) 256, 208 Number of data points collected (Nx, Ny) 256, 208
Slice thickness (Δz) 4 mm Slice thickness (Δz) 5 mm
Number of slices 25 Number of slices 30
Slice gap 1.2 mm Slice gap 1.5 mm
Number of acquisitions (Nacq) 2 Number of excitations (NEX) 2a
Read direction Right–left Number of acquisitions (Nacq) 2
Scan time 279 sec Read direction Right–left
Fat suppression Yes
Scan time 506 sec
aThe number of concatenation is set to be 2. This means that only half of the total slices will be excited

during a given repeat time.

Figure A7.7.2 Imaging of a normal-sized cervical lymph node (arrow) by a transverse T1-
weighted SE sequence.

This sequence usually consists of three orthogonal planes to allow exact localization of the
bilateral cervical lymph nodes. These images are often also used later to determine where Figure A7.7.3 Imaging of a normal cervical lymph node (arrow) by a T1-weighted contrast
to place the saturation pulses and to confirm total coverage of the volume of interest. enhanced, fat-saturated SE sequence.

Sequence 2: Transverse T2-weighted, fat-saturated turbo spin echo (TSE)

13. Display the coronal, sagittal, and transverse scout images in split-window mode on
the scan monitor. Change imaging parameters to those listed in Table A7.7.3. Position
the slices to cover the area from the posterior fossa to the thoracic inlet in the coronal
and sagittal planes, as well as the anterior, posterior, and lateral margins of the neck
in the transverse plane. Let the patient know you are ready, and begin the scan. Lymph Node
Staging in the
An example image is shown in Figure A7.7.1. Head and Neck Neck

A7.7.5 A7.7.6
Current Protocols in Magnetic Resonance Imaging Supplement 9 Supplement 9 Current Protocols in Magnetic Resonance Imaging
Sequence 3: Transverse T1-weighted spin echo (SE)
14. Bring the sequence for a T1-weighted SE scan up on the console. Set the imaging
parameters as shown in Table A7.7.4. parotid gland
15. Use the History function to select slice positions from the previous sequence,
Sequence 2 (T2-weighted, transverse, fat-saturated TSE). Let the patient know you
are ready, and begin the scan.
su b
An example image is shown in Figure A7.7.2. su b m e n
ma tal n
nd
ibu odes
Sequence 4: Transverse T1-weighted, fat saturated spin echo (SE), post-contrast lar
no
d es

upper
16. Copy the slice parameters, such as number of slices, field of view, dimension of slices,
and center of slices from the performed T1-transverse sequence 3 and change the
imaging parameters to those listed in Table A7.7.5. larynx

des
y
prelaryngeal nodes

middle

ry no
arter
17. Before proceeding, administer the contrast agent.
pretracheal nodes

es s o
bral
A dose of 0.2 mmol/kg gadolinium is usually given at a rate of 1 ml/sec. paratracheal nodes

l acc
lower ce

verte
18. Wait 3 min before proceeding. Inform the patient of the 3-min delay.
rv

spina
19. Let the patient know you are ready, and begin the scan. icle

An example image is shown in Figure A7.7.3.

clavicle
20. Take the patient out of the magnet. upper
media
stinal
nodes

COMMENTARY
Background Information
Imaging may be performed for the evalu-
ation of any unknown neck mass, but the use
of imaging to detect and evaluate primary or
metastatic malignancy within lymph nodes is
common in the diagnosis and staging of head
and neck disease (Sakai et al., 2000). In chil-
dren, lymphadenopathy is the most common
solid neck mass, with lymphoma ranking third
among all the solid neck masses occurring in
childhood. In adults in the United States, lym-
phoma represents the seventh leading cause of
cancer deaths (Daehnert, 1999).
Unreliable criteria for differentiating benign
from malignant causes of lymphadenopathy,
such as increased anteroposterior diameter,
prominent calcifications, or marginal contrast
enhancement in computed tomography (CT),
are being replaced by the successful depiction
of the soft tissue microstructure inside the
lymph nodes (van den Brekel et al., 1996) by
means of MR imaging.
The evolving improvements in spatial reso-
lution, the reduction of artifacts, and the devel-
opment of new contrast agents allow anatomi-
cal visualization to precisely assess the extent
of any cancerous spread or infiltration into
adjacent structures, most importantly arteries
or veins, as well as into adjacent nerves. These
routes of extranodal tumor extension are his-
tologically found in 40% of lymph nodes with
Figure A7.7.4 Seven anatomic subsites of lymph nodes in the neck. Level I contains the submen-
diameters less than 2 cm, which would thereby tal and the submandibular nodes. Level II includes the upper deep cervical chain nodes. Level III
be considered normal by size criteria (Cum- includes the middle deep cervical chain nodes, and analogous Level IV includes the lower deep
mings, 1993). In 75% of lymph nodes larger cervical chain nodes. Level V comprises the spinal accessory nodes and the transverse cervical
than 3 cm, extranodal tumor spread is present. chain nodes. Level VI contains the pretracheal, the prelaryngeal, and the paratracheal nodes. In
Furthermore MR imaging has the potential Level VII the upper mediastinal lymph nodes can be found. Figure is based on Harnsberger (1995).
to precisely show anatomic landmarks that
make nodal levels reproducible. The lymph
tico-venous pathways. If cancer cells invade the In addition, MR imaging can be utilized for
nodes in the neck are divided into seven specific
reticular and medullary tissue of the lymphatic percutaneous MRI-guided biopsies of lymph
anatomic subsites. The level concept is driven
nodes, a blockage of the lymph flow occurs, nodes, using dedicated open scanners accom-
by the fact that the extent and level of cervical
leading to an increase of interstitial fluid. Vary- panied by the development of fast gradient echo
node involvement by metastatic tumor is prog-
ing amounts of interstitial fluid and infiltrated pulse sequences to secure the diagnosis by
nostically important (Harnsberger, 1995; Fig.
medullary meshwork results in variable MR histology (Lewin et al., 2000).
A7.7.4).
appearances; however, lymph nodes such as
The imaging findings complement the
these are all typically referred to as necrotic Critical Parameters and
physical examination, and the imaging-based
(Som, 1987). Troubleshooting
classification provides the radiologist with
Infections of bacterial, mycobacterial, or The high magnetic field strength of 1.5 T
clinically acceptable guidelines for classifying
viral cause are other common reasons for lym- magnets and T1-weighted spin echo sequences
the cervical nodes and communicating the find-
phadenopathy. Further indications are granu- with fat saturation and contrast enhancement,
ings to clinicians (Som et al., 2000).
lomatous conditions such as sarcoidosis, as as well as T2-weighted fast spin echo or turbo
The most common indication for perform-
well as primary and secondary involvement in spin echo techniques, reveal a remarkably in-
ing MRI of the soft tissue of the neck in adults
lymphoma, e.g., Hodgkin’s as well as non- tense and precise signal from the lymph nodes.
is a strong suspicion or staging of head and neck
Hodgkin’s lymphomas. Other metastatic neo- High spatial resolution is desired so that
cancer, especially carcinomas of the upper
plasms originating from breast and lung are small areas of signal variation are not missed.
aerodigestive tract (Lenz et al., 1993). In addi-
also common causes of diffuse metastatic en- The exact anatomical visualization of arteries
tion to metastatic squamous carcinoma of the
largement of cervical lymph nodes. Conditions and veins, as well as of neural structures such
upper aerodigestive tract, the differential diag-
such as sinus histiocytosis and eosinophilic as the cervical and brachial plexus, allows a
nosis of enlarged cervical lymph nodes also
Lymph Node granuloma also occasionally occur (Kaji et al., precise diagnosis of whether any cancerous
includes the lymphatico-lymphatic spread as
Staging in the 1997). formation has invaded those organs, which are
well as the metastatic dispersion via lympha- Head and Neck Neck

A7.7.7 A7.7.8
Current Protocols in Magnetic Resonance Imaging Supplement 9 Supplement 9 Current Protocols in Magnetic Resonance Imaging
most vulnerable to the spread of cancer. More-
over, it is important to center the planes in order
to perform a comparison with the opposite side
of the neck, to identify any area of abnormality,
and to differentiate between pathology and ana-
tomical variation.
Anticipated Results
A variety of findings concerning the lymph
nodes is possible utilizing magnetic resonance
imaging. To ensure that the complete anatomi-
cal region containing the cervical lymph path-
ways is imaged, the neck should be visualized
from the skull base to the thoracic inlet, thereby
clearly indicating the body of the hyoid bone,
the cricoid arch, the top of the manubrium, and
the back edge of the submandibular gland as
important landmarks. Posteriorly, the back
edge of the sternocleidomastoid muscle, the
lateral posterior edge of the anterior scalene
muscle, and the anterior edge of the trapezius
muscle should be included in the field of view
and should contain both the internal carotid and
common carotid arteries and the internal jugu-
lar vein.
Anteriorly, the clavicle, the medial margin
of the anterior belly of the digastric muscle, and
the mylohoid muscle are utilized as important
landmarks.
Although in most cases a specific diagnosis
can be suggested by other means of visualiza-
tion, only MRI has the potential to demonstrate
the extent of disease with high precision, thanks
to the high spatial resolution, thereby allowing
competitive and comparable MR imaging,
which is crucial to the staging process in esti-
mating the expected success of therapy.
Literature Cited
Cummings, B.J. 1993. Radiation therapy and the
treatment of the cervical lymph nodes. In Otolar-
yngology Head and Neck Surgery, 2nd ed. (B.J.
Cummings and J.M. Fredrickson, eds.) pp. 1626-
1648. Mosby, St. Louis.
Daehnert, W. 1999. Chest Disorders. In Radiology
Review Manual, 3rd ed. (W. Daehnert, ed.) pp.
418-420. Williams and Wilkins, Baltimore.
Current Protocols in Magnetic Resonance Imaging
Harnsberger, H.R. 1995. Lymph node division by
levels. In Head and Neck Imaging, 4th ed. (H.R.
Harnsberger, ed) pp. 291-294. Mosby, St. Louis.
Kaji, A.V., Mohuchy, T., and Swartz, J.D. 1997.
Imaging of cervical lymphadenopathy. Semin.
Ultrasound CT MR 18:220-249.
Lenz, M., Kersting-Sommerhoff, B., and Gross, M.
1993. Diagnosis and treatment of the N0 neck in
carcinomas of the upper aerodigestive tract: Cur-
rent status of diagnostic procedures. Eur. Arch.
Otorhinolaryngol. 250:432-438.
Lewin, J.S., Nour, S.G., and Duerk, J.L. 2000. Mag-
netic resonance image-guided biopsy and aspi-
ration. Top. Magn. Reson. Imaging 11:173-183.
Sakai, O., Curtin, H.D., Romo, L.V., and Som, P.M.
2000. Lymph node pathology: Benign prolifera-
tive, lymphoma, and metastatic disease. Ra-
diol.Clin.North Am. 38:979-998.
Shellock, F.G. and Crues, J.V. 1998. Aneurysm
clips: Assessment of magnetic field interaction
associated with a 0.2-T extremity MR system.
Radiology 208:407-409.
Shellock, F.G. and Kanal, E. 1998. Aneurysm clips:
Evaluation of MR imaging artifacts at 1.5 T.
Radiology 209:563-566.
Shellock, F.G. and Shellock, V.J. 1998. Cranial bone
flap fixation clamps: Compatibility at MR imag-
ing. Radiology 207:822-825.
Som, P.M. 1987. Lymph nodes of the neck. Radiol-
ogy 165:593-600.
Som, P.M., Curtin, H.D., and Mancuso, A.A. 2000.
Imaging-based nodal classification for evalu-
ation of neck metastatic adenopathy. Am. J.
Roentgenol. 174:837-844.
van den Brekel, M.W. 2000. Lymph node metasta-
ses: CT and MRI. Eur. J. Radiol. 33:230-238.
van den Brekel, M.W., Castelijns, J.A., and Snow,
G.B. 1996. Imaging of cervical lymphade-
nopathy. Neuroimaging Clin. N. Am. 6:417-434.
Contributed by Daniel T. Boll, Andrik J.
Aschoff, and Jonathan S. Lewin
University Hospitals of Cleveland and
Case Western Reserve University
Cleveland, Ohio
Head and Neck
A7.7.9
Supplement 9
CHAPTER A8
Extradural Spine
INTRODUCTION
echnical advancements in both hardware and software have improved the speed at
T which, progressively, imaging of the spine at ever higher resolutions can be under-
taken with reproducible success. If success is loosely defined as the ability of an imaging
modality to accurately diagnose clinically relevant disease, then MRI has achieved that
goal.
One of the major successes of modern intravenous (IV) contrast–enhanced MRI is its
ability in most cases to focus a disease process to a particular spinal compartment:
intramedullary, intradural-extramedullary, extradural (i.e., epidural), and mixed. This is
an important concept because this knowledge will assist in narrowing the differential-di-
agnostic possibilities and help direct specific additional diagnostic procedures—e.g.,
needle biopsy (aspiration) or surgical therapies. While other chapters will elaborate the
findings in each one of these compartments, the present chapter will discuss extradural
disease of the spine (Table A8.0.1).
UNIT A8.1 will outline the protocols that will enable the analysis of the normal and
pathologic intervertebral disc. Specifically, terminology for normal and degenerated disc
are defined in order to allow reliable visualization of disc pathology.
UNIT A8.2discusses the protocols that will define the various subtypes of spinal stenosis
and the protocols utilized to diagnose each of them. These protocols will allow the medical
imaging specialist to determine the level, region, and degree of severity of the stenosis
present.
UNIT A8.3details the protocols that will allow the evaluation of spondylosis deformans or
degenerative osteophyte formation. These protocols will reveal the presence, level, and
severity of peridiscal, uncovertebral, and posterior spinal facet (zygoapophyseal) joint
osteophytosis and their effect on the central spinal canal and spinal (intervertebral) neural
foramina.
UNIT A8.4 outlines the protocols used for demonstrating infectious and noninfectious spinal
inflammation. These protocols will allow the visualization of the focus and extent of the
inflammatory processes, as well as the presence or absence of involvement of the central
spinal canal and perispinal soft tissues.
UNIT A8.5 details the protocols best utilized to reveal the presence and extent of primary or
metastatic neoplasia involving the spine. In addition, related pathologic changes, such as
pathologic fractures and epidural tumor extension with spinal cord/cauda equina com-
pression, are also well analyzed with these techniques.
UNIT A8.6 describes the protocols used to evaluate the patient with either acute or remote
spinal trauma. These techniques will allow the determination of the extent of injury and
the tissues involved, and thereby will enhance therapeutic planning.
UNIT A8.7
has some overlap with protocols of other units in this chapter. Many disease
processes may cause compression of the spinal cord and/or cauda equina. The protocols
Extradural Spine
Contributed by J. Randy Jinkins and David D. Stark A8.0.1
Current Protocols in Magnetic Resonance Imaging (2003) A8.0.1-A8.0.3
Copyright © 2003 by John Wiley & Sons, Inc. Supplement 9
 Table A8.0.1 Differential Diagnosis of Extradural Spinal Lesions in this unit will enable the medical imaging specialist to pinpoint the degree and extent
of disease in order to assist image-guided surgical decompressive procedures in minimiz-
Developmental ing permanent neurologic deficit.
Lipoma
Extradural arachnoid cyst UNIT A8.8 covers the protocols judged best for analyzing the postoperative spine for
Developmental central spinal canal stenosis degenerative disease. These protocols will allow the definition and differentiation of the
Vascular various types of pathology that may be present and responsible for recurrent signs and
Vertebral hemangioma with pathologic vertebral expansion or epidural extension symptoms following spinal surgery.
Primary epidural hemangioma In summary, the MRI protocols in this chapter are aimed at enabling the medical
Traumatic imagining specialist to acquire images that will define the general type of pathology
Epidural hematoma present, its location and extent in three planes, and the effect of the pathologic process in
Foreign body (penetrating trauma) intra- and perispinal tissues. In this way, the imaging physician will be best able to assist
Traumatic epidural air the referring clinical physician in attempting to minimize permanent neurologic deficit,
Traumatic intervertebral disc herniation and in enhancing timely positive patient outcome.
Retropulsed vertebral fracture fragments
Inflammatory
Primary epidural phlegmonous (i.e., solid soft tissue) inflammation or abscess formation (e.g., J. Randy Jinkins and David D. Stark
bacterial, fungal)
Primary epidural infectious granuloma formation (e.g., tuberculosis)
Secondary epidural extension of inflammatory process (e.g., phlegmon, abscess, granuloma)
from adjacent spondylitis/discitis
Primary epidural parasitic cyst formation (e.g., cysticercus, echinococcus)
Neoplastic
Epidural extension of primary neoplastic spinal disease (e.g., myeloma/plasmacytoma, giant
cell tumor, eosinophilic granuloma, aneurysmal bone cyst, osteochondroma, osteoblastoma,
chordoma, osteosarcoma, chondrosarcoma, Ewing’s sarcoma, leukemia)
Epidural extension of hematogenously disseminated metastatic neoplastic spinal disease (e.g.,
carcinoma, sarcoma, lymphoma)
Primary epidural metastatic neoplastic disease
Schwannoma
Neurofibroma
Epidural extension of paraspinous neoplastic disease (e.g., lung tumor, renal tumor,
fibrosarcoma, neuroblastoma)
Degenerative
Intervertebral disc herniation
Epidural gas from degenerated (“vacuum”) intervertebral disc
Spinal osteophyte formation (e.g., vertebral body posterior facet joint)
Ligamentum flavum ossification
Ossification of the posterior longitudinal ligament
Degenerative spinal stenosis
Acquired
Extramedullary hematopoiesis
Amyloidosis
Postoperative epidural fibrosis
Iatrogenic extradural injection (e.g., from myelogram, epidural anesthesia)
Epidural lipomatosis
Paget’s disease of bone with vertebral expansion

Introduction Extradural Spine

A8.0.2 A8.0.3
Supplement 9 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 9
Herniated Intervertebral Disc
One of the most significant impacts of magnetic resonance (MR) has been its ability to
exquisitely depict normal and pathologic anatomy of the spine. Direct acquisitions
acquired in multiple planes, coupled with the ability to study the spine with different T1-
and T2-weighted images, have enabled the critical assessment of the spinal column and
its contents in ways not previously available to the medical imaging specialist. The
development of contrast media has further extended the capabilities of MR imaging of
the spine by improving its sensitivity, and sometimes its specificity, in certain categories
of pathology.
CONVENTIONAL AND FAST SPIN ECHO
While certain pathologic conditions may be assessed utilizing specialized imaging
acquisitions (e.g., gradient recalled echo: cervical spine), in general the normal signal
characteristics of the skeletal components of the vertebral column typically are best
illustrated on the basis of spin echo techniques (Karnaze et al., 1987; Modic et al., 1984;
Murayama et al., 1990; Van Dyke et al., 1989; Tsruda et al., 1990). In part because of low
relative water content, the intervertebral discs are normally seen to be hypointense on
T1-weighted images relative to adjacent vertebral bone marrow signal, becoming rela-
tively hyperintense on T2-weighted images. Upon aging, this MR signal pattern progres-
sively changes because of disc degeneration, with the disc becoming yet more hypointense
on T1-weighted images and reversing to become hypointense on T2-weighted images
(Edelman et al., 1985; Hedberg et al., 1988; Holtas et al., 1987; Masaryk et al., 1988;
Robertson et al., 1991; Ross et al., 1987).
The following sequences comprise the preferred protocol for high-field MR machines.
On some machines alternate gradient recalled echo acquisitions may be more desirable.
In addition, occasionally intravenous (i.v.) contrast-enhanced alternate protocols may be
useful to evaluate the integrity of the blood-nerve barrier in certain cases of suspected
compressive radiculopathy, or in cases of epidural masses of ambiguous etiology (e.g.,
epidural neoplasia/infection versus intervertebral disc fragment).
Table A8.1.1 lists the hardware necessary to perform the procedure, along with appropri-
ate parameters. The available gradient strength will depend on the scanner, and the echo
times (i.e., TE) given below in subsequent tables will be varied accordingly (the smaller
the gradient strength, the longer the echo time for a particular scan). See Jinkins (1993a,b),
Ross et al. (1989), Wasserstrom et al. (1993), and Yamashita et al. (1994).
This entire protocol should take 35 to 45 min to complete.
Table A8.1.1 Equipment Parameters for Spine Imaging in Cases of
Intervertebral Disc Disease
Coil type Cervical, thoracic, lumbar: phased array
surface coil (or other depending upon
machine compatibility and availability)
Peripheral gating Thoracic spine only (optional)
Respiratory gating Thoracic spine only (optional)
Flow compensation pulse Any level (optional)
Use of contrast agents Optional (see Alternate Protocol)
Contributed by J. Randy Jinkins and David D. Stark
Current Protocols in Magnetic Resonance Imaging (2001) A8.1.1-A8.1.11
Copyright © 2001 by John Wiley & Sons, Inc.
UNIT A8.1 NOTE: Be sure that technicians and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.
Set up patient and equipment
1. Interview (screen) the patient to ensure that he or she has no contraindications such
as cardiac pacemakers or other implants containing ferromagnetic materials. Also be
sure to find out if the patient has any health conditions that may require the presence
of special emergency equipment during the scanning procedure, or necessitate any
other precautions.
Generally, standard screening forms are used for all patients scanned in a magnetic
BASIC resonance system.
PROTOCOL
The presence of any ferromagnetic materials may be a health hazard to the patient when
he or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any implants; see Shellock (1996)
for discussion of what implants may be safely scanned using magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
2. If the procedure is a research protocol, have the patient sign any necessary consent
form.
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
4. Inform the patient about what will occur during the procedure, what he or she will
experience while in the magnet, and how to behave, including the following:
a. If earphones or headphones are used to protect the ears from the loud sounds
produced by the gradients, the patient will be asked to wear these, but will be able
to communicate with you at any time during the imaging.
b. The patient will be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this works).
c. For good results, the patient should not talk, and should avoid or minimize
swallowing or other movement, during each scan—i.e., as long as the banging
sounds continue. Between scans, talking and swallowing are allowed in most
cases, but should be avoided when comparative positional studies are being
performed; the patient will be informed when this is the case.
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
5. Have the patient mount onto the table. Either before or right after the patient lies
down, set up any triggering devices or other monitoring equipment that is to be used.
6. Center the coil over the region where the key information is desired.
Make sure that the body is constrained to prevent motion, especially if high-resolution scans
are to be run.
7. If needed, place a pillow or other support under the knees to make the patient more
comfortable.
8. Use the centering light to position the patient (cervical spine: thyroid cartilage;
Herniated thoracic spine: nipple line; lumbar spine: iliac crests) and put him or her into the
Intervertebral center of the magnet.
Extradural Spine Disc
A8.1.1 A8.1.2
Current Protocols in Magnetic Resonance Imaging
 Table A8.1.2 Primary Clinical Imaging Parameters for Sequence 1 (Pilot Scan) Table A8.1.3 Primary Clinical Imaging Parameters for Sequence 2
(T1-Weighted Image)
Patient position Supine
Scan type Gradient echo Patient position Supine
Imaging plane (orientation) Transverse Scan type Conventional spin echo
Central slice or volume center Laser light centered—cervical Imaging plane (orientation) Sagittal
spine: thyroid cartilage; thoracic Central slice or volume center Centered on:
spine: nipple line; lumbar spine: Cervical: 3rd cervical vertebra
iliac crests Thoracic: 6th thoracic vertebra
Echo time (TE) As short as possible Lumbar: 3rd lumbar vertebra
Repeat time (TR) As short as possible Echo time (TE) 10 msec
Flip angle (FA) 15° Repeat time (TR) 500 msec
Fields of view (FOVx, FOVy) Cervical: 240 mm, 240 mm Flip angle (FA) 90°
Thoracic: 340 mm, 340 mm Fields of view (FOVx, FOVy) Cervical: 240 mm, 240 mm
Lumbosacral: 280 mm, 280 mm Thoracic: 320 mm, 320 mm
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm Lumbosacral: 280 mm, 280 mm
Thoracic: 1.25 mm, 1.25 mm (may use rectangular field of view,
Lumbosacral: 1.09 mm, 1.09 mm e.g., half or three-quarter field] if
Number of data points collected (Nx, Ny) 256, 256 available, or tailor to region of
Display matrix (Dx, Dy) 256, 256 interest)
Slice thickness (Δz) 5 mm Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm
Slice gap Not applicable Thoracic: 1.25 mm, 1.25 mm
Number of acquisitions (Nacq) 1 Lumbrosacral: 1.09 mm, 1.09 mm
Scan time ∼10 sec Number of data points collected (Nx, Ny) 256, 256
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) Cervical: 3 mm
Once this step has been performed, so long as the patient does not move on the table, the Thoracic: 3 mm
table itself can be moved and then replaced in the same position as before without Lumbar: 4 mm
jeopardizing the positioning of one scan relative to another. Number of slices As many as needed to cover the
region of interest
9. If the patient is unable to hold still, provide an appropriate sedative.
Slice gap Cervical: 0.5 mm
Thoracic: 1 mm
Sequence 1: Rapid positioning pilot Lumbar: 1 mm
10. To validate the patient’s position, run the system’s pilot (or scout) scan (sequence 1) Number of acquisitions (Nacq) 4
to ensure correct location of the neck in three dimensions, using the imaging sequence Flow compensation Yes (if available)
given in Table A8.1.2 or similar parameters. Saturation pulses Yes; anterior
This sequence usually consists of three orthogonal planes to allow subsequent localization. cervical/thoracic/lumbar slabs to
The images are often also used later to determine where to place the saturation pulses and saturate larynx/vessels/heart
to set up total coverage of the volume of interest. Slice series Left to right or the reverse
depending on preference
Sequence 2: Sagittal T1-weighted conventional spin echo Scan time ∼8 min
11. Set the imaging parameters as shown in Table A8.1.3.
12. Use the pilot image to locate the spine in three dimensions to ensure coverage of the Sequence 4: Transverse T1-weighted conventional spin echo
region of interest (e.g., cervical, thoracic, lumbosacral spine). 16. Using the midline sagittal T1-weighted image acquired in sequence 2, set the trans-
verse acquisition parameters as follows:
13. Let the patient know you are ready and begin the scan.
a. Cervical spine: stacked images from C1 through C7-T1.
Sequence 3: Sagittal T2-weighted fast spin echo b. Thoracic spine: stacked images through levels of interest.
14. Review the pilot scans and set the saturation pulse so that it is correctly placed anterior
c. Lumbosacral spine: 5 slices each, angled to the plane of the intervertebral disc at
to the slab of interest.
L3-4, L4-5, and L5-S1; one slice each, angled to the intervertebral disc at L1-2
15. Run sequence 3 according to Table A8.1.4. and L2-3.
Herniated 17. Supplement additional slices according to visible disease present or to clinical query.
Intervertebral
Extradural Spine Disc 18. Run the sequence according to Table A8.1.5.
A8.1.3 A8.1.4
Current Protocols in Magnetic Resonance Imaging Supplement 2 Supplement 2 Current Protocols in Magnetic Resonance Imaging
 Table A8.1.4 Primary Clinical Imaging Parameters for Sequence 3 Table A8.1.5 Primary Clinical Imaging Parameters for Sequence 4
(T2-Weighted Image, FSE)a (T1-Weighted Image)

Patient position Supine Patient position Supine

Scan type Fast spin echo Scan type Conventional spin echo
Imaging plane (orientation) Sagittal Imaging plane (orientation) Transverse
Central slice or volume center Centered on area of interest (as in Central slice or volume center Centered on the area of interest
sequence 2, Table A8.1.3) (as in sequence 2, Table A8.1.3)
Echo time (TE) 100 msec Echo time (TE) 10 msec
Echo train length (ETL) 8 Repeat time (TR) 500 msec
Repeat time (TR) 4000 msec Flip angle (FA) 90°
Flip angle (FA) 90° Fields of view (FOVx, FOVy) As in sequence 2 (Table A8.1.3)
Fields of view (FOVx, FOVy) As in sequence 2 (Table A8.1.3) Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm
Resolution (Δx, Δy) Cervical: 0.47 mm, 0.47 mm Thoracic: 1.25 mm, 1.25 mm
Thoracic: 0.63 mm, 0.63 mm Lumbosacral: 1.09 mm, 1.09 mm
Lumbosacral: 0.55 mm, 0.55 mm Number of data points collected (Nx, Ny) 256, 256
Number of data points collected (Nx, Ny) 512, 512 Display matrix (Dx, Dy) 256, 256
Display matrix (Dx, Dy) 512, 512 Slice thickness (Δz) Cervical: 3 mm
Slice thickness (Δz) Table A8.1.3 Thoracic: 3–5 mm
Number of slices Varies with spinal level Lumbar: 4 mm
Slice gap Table A8.1.3 Number of slices Varies with spinal level
Number of acquisitions (Nacq) 1 Slice gap Cervical: 1 mm
Flow compensation Yes (if available) Thoracic: 1–2 mm
Lumbar: 1 mm
Saturation pulses Yes; anterior
cervical/thoracic/lumbar slabs Number of acquisitions (Nacq) 2
anteriorly to saturate Slice locations See Basic Protocol, step 16
larynx/vessels/heart Saturation pulses No
Fat suppression Yes Scan time ∼4 min
Slice series Left to right or the reverse
depending on preference
Scan time ∼4 min ALTERNATE CONTRAST-ENHANCED MRI ACQUISITIONS
PROTOCOL 2
aFSE, fast spin echo. In some situations, such where the findings of an MRI scan do not match the clinical
findings, the use of intravenous (i.v.) paramagnetic contrast material may be indicated
Sequence 5: Transverse T2-weighted fast spin echo (Jinkins, 1993a,b; Ross et al., 1989; Wasserstrom et al., 1993; Yamashita et al., 1994).
19. Using the midline T1-weighted image acquired in sequence 2, repeat the setup as in Materials
Table A8.1.6.
Normal sterile saline (0.9% NaCl)
20. Run sequences according to Table A8.1.6. Intravenous MRI contrast agent (e.g., Magnevist, Omniscan, or Prohance)

Set up patient and equipment

GRADIENT RECALLED ECHO SEQUENCES ALTERNATE
PROTOCOL 1
1. Using the same equipment, perform patient setup as in the Basic Protocol.
With some machines, or according to preferences, gradient recalled echo acquisitions
may be used in the sagittal and/or transverse planes to clearly distinguish between discs 2. Establish an intravenous line from which the contrast agent can be injected, and attach
and soft tissue and to clarify the spinal neural foramen in the cervical region (Murayama this line securely to the patient so that movement into or out of the magnet will not
et al., 1990; Van Dyke et al., 1989). pull at the patient’s arm.
It is preferable to insert the line prior to imaging and to leave the patient in the magnet,
Sequence 6: Sagittal gradient recalled echo with no intervening motion between the scans run before contrast agent injection and those
1. Run sequence 6 according to Table A8.1.7. run after injection.

Sequence 7: Transverse gradient recalled echo Perform pilot scan

2. Run sequence 7 according to Table A8.1.8. 3. Run a rapid three-plane positioning pilot scan (sequence 1; see Basic Protocol, step
10).
Herniated
Intervertebral
Extradural Spine Disc

A8.1.5 A8.1.6
Current Protocols in Magnetic Resonance Imaging Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging
 Table A8.1.6 Primary Clinical Imaging Parameters for Sequence 5 Table A8.1.8 Primary Clinical Imaging Parameters for Sequence 7 (T2* GRE)a
(T2-Weighted Image, FSE)a
Patient position Supine
Patient position Supine Scan type Gradient echo
Scan type Fast spin echo Imaging plane (orientation) Transverse
Imaging plane (orientation) Transverse Central slice or volume center Centered on the area of interest
Central slice or volume center Centered on the region of interest (as in sequence 2, Table A8.1.3)
(as in sequence 2, Table A8.1.3) Echo time (TE) 15 msec
Echo time (TE) 100 msec Repeat time (TR) 500 msec
Echo train length (ETL) 8 Flip angle (FA) 10° to 20°
Repeat time (TR) 4000 msec Fields of view (FOVx, FOVy) As in sequence 2 (Table A8.1.3)
Flip angle (FA) 90° Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm
Fields of view (FOVx, FOVy) As in sequence 2 (Table A8.1.3) Thoracic: 1.25 mm, 1.25 mm
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm Lumbosacral: 1.09 mm, 1.09 mm
Thoracic: 1.25 mm, 1.25 mm Number of data points collected (Nx, Ny) 256, 256
Lumbosacral: 1.09 mm, 1.09 mm Display matrix (Dx, Dy) 256, 256
Number of data points collected (Nx, Ny) 256, 256 Slice thickness (Δz) 3–4 mm
Display matrix (Dx, Dy) 256, 256 Number of slices Varies with level
Slice thickness (Δz) As in sequence 4 (Table A8.1.5) Slice gap 1 mm or less
Number of slices Varies with spinal level Number of acquisitions (Nacq) 1
Slice gap As in sequence 4 (Table A8.1.5) Slice locations See Basic Protocol, step 16
Number of acquisitions (Nacq) 2 Flow compensation Yes (if available)
Slice locations See Basic Protocol, step 16) Saturation pulses Yes
Flow compensation Yes (if available) Scan time ∼2 min
Saturation pulses No aT * GRE, T * gradient recalled echo.
2 2
Scan time ∼4 min
aFSE, fast spin echo.
Sequences 8 and 9: Contrast agent enhanced T1-weighted spin echo
4. Leaving the patient in the magnet, inject the contrast agent, flush the line with 10 ml
Table A8.1.7 Primary Clinical Imaging Parameters for Sequence 6 (T2* GRE)a saline, and then immediately run sagittal (see Basic Protocol, steps 11 to 13; sequence
8) and transverse (see Basic Protocol, steps 16 to 18; sequence 9) T1-weighted image
Patient position Supine sequences according to Tables A8.1.3 and A8.1.5, respectively (sequences 2 and 4 in
Scan type Gradient echo the Basic Protocol).
Imaging plane (orientation) Sagittal
An i.v. dose of 0.1 mmol/kg of contrast agent is usually given.
Central slice or volume center Centered on the area of interest
(as in sequence 2, Table A8.1.3)
Echo time (TE) 15 msec COMMENTARY
Repeat time (TR) 500 msec
Background Information bulge is broad-based in the transverse plane and
Flip angle (FA) 10° to 20°
The description of disc disease, above and may be circumferential (i.e., nondirectional),
Fields of view (FOVx, FOVy) As in sequence 2 (Table A8.1.3) beyond that of simple degeneration, is an area anterior, posterior, or lateral (i.e., directional). A
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm of diagnosis that suffers because of a lack of disc protrusion is a focal annular contour change
Thoracic: 1.25 mm, 1.25 mm uniform terminology. To be fair; there are many in the transverse plane; some authors have de-
Lumbosacral: 1.09 mm, 1.09 mm reasonable terms; however, their use is irregular, fined focal as less than 25% of the disc periphery.
Number of data points collected (Nx, Ny) 256, 256 overlapping and inconsistent. These terms Disc protrusions may be projected into any one
Display matrix (Dx, Dy) 256, 256 should be chosen carefully to reflect both the or more than one circumferential direction(s);
Slice thickness (Δz) 3 mm pathology as well as to fit the imaging proce- when protrusions are found in more than one
Number of slices Varies with spinal level dure. Successful, appropriate usage of terms for direction, the term “multidirectional disc pro-
Slice gap 1 mm or less spinal intervertebral disc disease will benefit the trusion” is used. A disc herniation may be tran-
Number of acquisitions (Nacq) 3 radiologist, the referring clinician, and ulti- sannular, transchondral, or transosseous with
mately the patient. A pragmatic system should regard to the path of egress. A transannular disc
Flow compensation Yes (if available)
be used that attempts, above all, to define terms extrusion is a disc fragment that has partially or
Saturation pulses Yes
according to what is actually clearly observed completely left the confines of the intervertebral
Scan time ∼6 min
on medical images: only three basic root terms disc via the peripheral annulus fibrosus. There
aT * GRE, T * gradient recalled echo. Herniated
2 2 can be used to describe annular contour changes are three types of transannular herniations: con-
Intervertebral
Extradural Spine Disc of disc bulge, protrusion, and herniation. A disc tiguous, sequestered, and transdural.

A8.1.7 A8.1.8
Current Protocols in Magnetic Resonance Imaging Supplement 2 Supplement 2 Current Protocols in Magnetic Resonance Imaging

Figure A8.1.1 Sagittal section showing allocation of transverse sections
from midline sagittal section of lumbosacral spine. Note the anterior spatial
saturation band (asterisks).
Contiguous disc herniations are focal ex-
tradural disc masses that remain in contact with
the parent disc and that have a polypoid shape
on sagittal sectioning; the diameter of the poly-
poid disc mass is usually larger than that of the
parent intervertebral disc, thus giving the ab-
normality its polypoid configuration. Seques-
tered disc herniations are disc fragments that
have migrated epidurally, either cranially, or
caudally, for a variable distance from the parent
disc. Transdural disc herniations are rare oc-
currences accounting for only 0.3% of all tran-
sannular disc herniations; the lumbar area is the
level most commonly affected. They are be-
lieved to be caused by preceding epidural scar-
ring (e.g., prior surgery, earlier disc extrusion,
previous trauma) resulting in dural adhesions
that promote rupture of the extruded disc di-
rectly into the thecal sac through a tear in the
dura/arachnoid matter. In addition to the tran-
sannular route of disc herniation, herniations
may enter into or through the cartilaginous
vertebral end plate or through the vertebral
body itself: a transchondral disc herniation is
also termed a Schmod node; a transosseous disc
herniation is also called a limbus vertebra. Note
that the simple, unmodified term disc hernia-
tion can be an imprecise, nebulous term that has
many varying meanings to too many physicians
(Taveras, 1989). Therefore it can be misleading.
Current Protocols in Magnetic Resonance Imaging
Barring internal derangement reaching the disc
surface (i.e., annular tears), the broad set of
terms defined here specifically describe virtu-
ally every pathologic condition involving the
borders of the intervertebral disc.
Critical Parameters and
Troubleshooting
Cerebrospinal fluid (CSF) flow, cardiac, la-
ryngeal, body wall and other sources of motion
can produce artifacts that can on occasion sig-
nificantly degrade the images. Proper spatial
(e.g., prevertebral) saturation pulses and some-
times flow compensation pulses and/or car-
diac/respiratory gating can reduce these arti-
facts significantly. In many instances, these
artifacts may be difficult or impossible to over-
come from patient to patient.
Anticipated Results
The goal of studying the intervertebral disc
is to diagnose clinically relevant disease of this
structure. While the terminology is at present
somewhat ambiguous from one medical center
to another and from physician to physician, the
images obtained utilizing these MRI sequences
will nevertheless yield reproducible results
enabling reliable visualization of disc pathol-
ogy. Some examples are shown in Figure
A8.1.1 and Figure A8.1.2.
Extradural Spine
A8.1.9
A B
C
Figure A8.1.2 Intervertebral disc herniation associated with enhancing radiculitis. (A) Sagittal T2-weighted (TR =
4000 msec, TE = 90 msec) section shows a large L5-S1 posterior intervertebral disc herniation (arrow). (B) Transverse
T1-weighted (TR = 500 msec, TE = 10 msec) section at L5-S1 shows the disc herniation (asterisk) extending toward
the right side. (C) Intravenous gadolinium–enhanced transverse T1-weighted (TR = 500 msec, TE = 10 msec) section
at L5-S1 shows several enhancing nerve roots (arrow) intrathecally.
Jinkins, J.R. 1993a. Magnetic resonance imaging of
Literature Cited
Edelman, R.R., Shoukimas, G.M., Stark, D.D., et al. benign nerve root enhancement in the unoper-
1985. High-resolution surface coil imaging of ated and postoperative lumbosacral spine.
lumbar disk disease. Am. J. Neuroradiol. 6:479- Neuroimag. Clin. North Am. 3:525-541.
485. Jinkins, J.R. 1993b. MR of enhancing nerve roots in
Hedberg, M.C., Drayer, B.P., Flom, A., Hodak, J.A., the unoperated lumbosacral spine. Am. J.
and Bird, C.R. 1988. Gradient echo (GRASS) Neuroradiol. 14:193-202.
MR imaging in cervical radiculopathy. Am. J. Karnaze, M.G., Gado, M.H., Sartor, K.J., and
Neuroradiol. 9:145-151. Hodges, F.J., III. 1987. Comparison of MR and
Holtas, S., Nordstrom, C.-H., Larsson, E.-M., and CT myelography in imaging the cervical and
thoracic spine. Am. J. Neuroradiol. 8:983-989.
Herniated Petterson, H. 1987. MR imaging of intradural
Intervertebral disk herniation. J. Assist. Comput. Tomogr.
Disc 11:353-356.
A8.1.10
Current Protocols in Magnetic Resonance Imaging
Masaryk, T.J., Ross, J.S., Modic, M.T., Boumphrey,
F., Bohlman, H., and Wilber, G. 1988. High-reso-
lution MR imaging of sequestered lumbar inter-
vertebral disks. Am. J. Neuroradiol. 9:351-358.
Modic, M.T., Pavlicek, W., Weinstein, M.A., Boum-
phrey, F., Ngo, F., Hardy, R., and Duchesneau,
P.M. 1984. Magnetic resonance imaging of in-
tervertebral disk disease. Radiology 152:103-
111.
Murayama, S., Numaguchi, Y., and Robinson, A.E.
1990. The diagnosis of herniated intervertebral
disks with MR imaging: A comparison of gradi-
ent-refocused echo and spin-echo pulse se-
quences. Am. J. Neuroradiol. 11:17-22.
Robertson, J.H., Meroni, R.M., Aprill, C.N., and
Smith, R.D. 1991. CT and MRI scans in thoracic
intradural disc herniation. Neuroradiology
33(S):331-332.
Ross, J.S., Perez-Reyes, N., Masaryk, T.J.,
Bohlman, H., and Modic, M.T. 1987. Thoracic
disk herniation. MR imaging. Radiology
165:511-515.
Ross, J.S., Modic, M.T., Masaryk, T.J., Carter, J.,
Marcus, R.E., and Bohlman, H. 1989. Assess-
ment of extradural degenerative disease with
Gd-dTPA enhanced MR imaging: Correlation
with surgical and pathologic findings. Am. J.
Neuroradiol. 10:11243-1249.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Taveras, J.M. 1989. Herniated intervertebral disk: A
plea for a more uniform terminology. Am. J.
Neuroradiol. 10:1283-1284.
Tsruda, J.S., Norman, D., Dillon, W., Newton, T.H.,
and Mills, D.G. 1990. Three-dimensional gradi-
ent recalled MR imaging as a screening tool for
the diagnosis of cervical radiculopathy. Am. J.
Neurol. 10:1263-1271.
Van Dyke, C., Ross, J.S., Tkach, I., Masaryk, T.J.,
and Modic, M.T. 1989. Gradient-echo MR im-
aging of the cervical spine: evaluation of ex-
tradural disease. Am. J. Neuroradiol. 10:627-
632.
Wasserstrom, R., Mamourian, A.C., Black, J.F., and
Lehman, R.A.W. 1993. Intradural lumbar disk
fragment with ring enhancement on MR. Am. J.
Neuroradiol. 14:401-404.
Yamashita, K., Hinoshima, K., and Kurata, A. 1994.
Gadolinium-DTPA-enhanced magnetic reso-
nance imaging of a sequestered lumbar interver-
tebral disc and its correlation with pathologic
findings. Spine 19:479-482.
Contributed by J. Randy Jinkins and
David D. Stark
Downstate Medical Center
State University of New York
Brooklyn, New York
Spinal Canal Stenosis UNIT A8.2
Spinal stenosis is defined generally as a narrowing of one or more of the following: the
central canal, foramina, and lateral recesses of the lumbar spine. Specifically, this stenosis
can be classified into three types as defined by location: (a) stenosis of the central spinal
canal, (b) stenosis of the intervertebral spinal neural foramen or foramina, and (c) stenosis
of the lateral recess(es) of the central spinal canal.
CONVENTIONAL AND FAST SPIN ECHO SEQUENCES BASIC
PROTOCOL
Although there is some clinical debate as to whether computed tomography (CT) or MR
is superior in diagnosing stenosis of the spinal canal, most MR-imaging physicians feel
that T2-weighted, fat-suppressed transverse and sagittal fast spin echo imaging is the better
method (Garfin et al., 1992; Hasegawa et al., 1993; Jinkins, 1999; Kent et al., 1992;
Schnebel et al., 1989). Nevertheless, some clinicians still prefer conventional myelogra-
phy. At present, only using conventional myelography can complete obstruction to
water-soluble contrast medium flow be demonstrated, which is sometimes present in
severe central spinal canal stenosis. In the future, perhaps intrathecal gadolinium-en-
hanced MRI will replace more conventional X-ray techniques, when and if the FDA
approves this route of administration.
The following sequences comprise the preferred protocol for high-field MR machines.
On some machines alternate gradient recalled echo acquisitions may be more desirable.
In addition, occasionally intravenous (i.v.) contrast enhanced alternate protocols may be
useful to evaluate the integrity of the blood-nerve barrier in certain cases of suspected
compressive radiculopathy, or clinically ambiguous claudication possibly of spinal origin.
Table A8.2.1 lists the hardware necessary to perform the procedure, along with appropri-
ate parameters. The available gradient strength will depend on the scanner, and the echo
times (i.e., TE) given below in other tables will be varied accordingly (the smaller the
gradient strength, the longer the echo time for a particular scan).
This entire protocol should take 35 to 45 min to complete.
NOTE: Be sure that technicians and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.

Set up patient and equipment

1. Interview (screen) the patient to ensure that he or she has no contraindications such
as cardiac pacemakers or other implants containing ferromagnetic materials. Also be
sure to find out if the patient has any health conditions that may require the presence
of special emergency equipment during the scanning procedure, or necessitate any
other precautions.

Table A8.2.1 Equipment Parameters for Spine Imaging in Cases of

Spinal Canal Stenosis

Coil type Cervical, thoracic, lumbar: phase array

Peripheral gating
Respiratory gating
Flow compensation pulse
Extradural Spine
surface coil (or other depending upon
machine compatibility and availability)
Thoracic spine only (optional)
Thoracic spine only (optional)
Any level (optional)
Extradural Spine

A8.1.11 Contributed by J. Randy Jinkins and David D. Stark A8.2.1

Current Protocols in Magnetic Resonance Imaging (2001) A8.2.1-A8.2.9
Current Protocols in Magnetic Resonance Imaging Copyright © 2001 by John Wiley & Sons, Inc. Supplement 2
 Generally, standard screening forms are used for all patients scanned in a magnetic
resonance system.
The presence of any ferromagnetic materials may be a health hazard to the patient when
he or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any implants; see Shellock (1996)
for discussion of what implants may be safely scanned using magnetic resonance.
5. Have the patient mount onto the table. Either before or right after the patient lies
down, set up any triggering devices or other monitoring equipment that is to be used.
6. Center the coil over the region where the key information is desired.
Make sure that the body is constrained to prevent motion, especially if high-resolution scans
are to be run.

Patients may be accompanied into the magnet room by a friend or family member, who can 7. If needed, place a pillow or other support under the knees to make the patient more
sit in the room during the scan and comfort the patient as needed. This companion must comfortable.
be screened as well to ensure the absence of loose metal objects on the body or clothing. 8. Use the centering light to position the patient and put him or her into the center of
2. If the procedure is a research protocol, have the patient sign any necessary consent the magnet (cervical spine: thyroid cartilage; thoracic spine: nipple line; lumbar
form. spine: iliac crests).
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
Table A8.2.3 Primary Clinical Imaging Parameters for Sequence 2 (T1-Weighted
4. Inform the patient about what will occur during the procedure, what he or she will Image)
experience while in the magnet, and how to behave, including the following:
Patient position Supine
a. If earphones or headphones are used to protect the ears from the loud sounds Scan type Conventional spin echo
produced by the gradients, the patient will be asked to wear these, but will be able Imaging plane (orientation) Sagittal
to communicate with you at any time during the imaging. Central slice or volume center Centered on:
b. The patient will be given a safety squeeze-bulb or similar equipment to request Cervical: the third cervical vertebra
assistance at any time (demonstrate how this works). Thoracic: the sixth thoracic vertebra
c. For good results, the patient should not talk, and should avoid or minimize Lumbar: the third lumbar vertebra
swallowing or other movement, during each scan—i.e., as long as the banging Echo time (TE) 10 msec
sounds continue. Between scans, talking and swallowing are allowed in most Repeat time (TR) 500 msec
cases, but should be avoided when comparative positional studies are being Flip angle (FA) 90°
performed; the patient will be informed when this is the case. Fields of view (FOVx, FOVy) Cervical: 240 mm, 240 mm
d. Nevertheless, the patient may call out at any time if he or she feels it necessary. Thoracic: 320 mm, 320 mm
Lumbosacral: 280 mm, 280 mm
(may use rectangular field of view,
Table A8.2.2 Primary Clinical Imaging Parameters for Sequence 1 (Pilot Scan) e.g., half or three-quarter field if
available, or tailor to region of
Patient position Supine interest)
Scan type Gradient echo Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm
Imaging plane (orientation) Transverse Thoracic: 1.25 mm, 1.25 mm
Central slice or volume center Laser light centered at cervical Lumbosacral: 1.09 mm, 1.09 mm
spine: thyroid cartilage; thoracic Number of data points collected (Nx, Ny) 256, 256
spine: nipple line; lumbar spine:
Display matrix (Dx, Dy) 256, 256
iliac crests
Slice thickness (Δz) Cervical: 3 mm
Echo time (TE) As short as possible
Thoracic: 3mm
Repeat time (TR) As short as possible
Lumbar: 1mm
Flip angle (FA) 15°
Number of slices 10, or as many as needed to cover the
Fields of view (FOVx, FOVy) Cervical: 240 mm, 240 mm
region of interest
Thoracic: 320 mm, 320 mm
Slice gap Cervical: 0.5 mm
Lumbosacral: 280 mm, 280 mm
Thoracic: 1 mm
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm
Lumbar: 1 mm
Thoracic: 1.25 mm, 1.25 mm
Number of acquisitions (Nacq) 2
Lumbosacral: 1.09 mm, 1.09 mm
Flow compensation Yes (if available)
Number of data points collected (Nx, Ny) 256, 256
Saturation pulses Yes; anterior
Display matrix (Dx, Dy) 256, 256 cervical/thoracic/lumbar slab to
Slice thickness (Δz) 5 mm saturate larynx/vessels
Slice gap Not applicable Slice series Left to right or the reverse depending
Number of acquisitions (Nacq) 1 on preference
Spinal Canal Scan time ∼10 sec Scan time ∼4 min
Stenosis Extradural Spine

A8.2.2 A8.2.3
Supplement 2 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 2
 Table A8.2.4 Primary Clinical Imaging Parameters for Sequence 3 Table A8.2.5 Primary Clinical Imaging Parameters for Sequence 4
(T2-Weighted Image, FSEa) (T1-Weighted Image)

Patient position Supine Patient position Supine

Scan type Fast spin echo Scan type Conventional spin echo
Imaging plane (orientation) Sagittal Imaging plane (orientation) Transverse
Central slice or volume center Centered on area of interest (as in Central slice or volume center Centered on the area of interest
sequence 2, Table A8.2.3) (as in sequence 2, Table A8.2.3)
Echo time (TE) 100 msec Echo time (TE) 10 msec
Echo train length (ETL) 8 Repeat time (TR) 500 msec
Repeat time (TR) 4000 msec Flip angle (FA) 90°
Flip angle (FA) 90° Fields of view (FOVx, FOVy) As in sequence 2 (Table A8.2.3)
Fields of view (FOVx, FOVy) As in sequence 2 (Table A8.2.3) Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm
Resolution (Δx, Δy) Cervical: 0.47 mm, 0.47 mm Thoracic: 1.25 mm, 1.25 mm
Thoracic: 0.63 mm, 0.63 mm Lumbosacral: 1.09 mm, 1.09 mm
Lumbosacral: 0.55 mm, 0.55 mm Number of data points collected (Nx, Ny) 256, 256
Number of data points collected (Nx, Ny) 512, 512 Display matrix (Dx, Dy) 256, 256
Display matrix (Dx, Dy) 512, 512 Slice thickness (Δz) Cervical: 3 mm
Slice thickness (Δz) As in sequence 2 (Table A8.2.3) Thoracic: 3–5 mm
Number of slices Varies with spinal level Lumbar: 4 mm
Slice gap As in sequence 2 (Table A8.2.3) Number of slices Varies with spinal level
Number of acquisitions (Nacq) 1 Slice gap Cervical: 1 mm
Flow compensation Yes (if available) Thoracic: 1–2 mm
Saturation pulses Yes; anterior Lumbar: 1 mm
cervical/thoracic/lumbar slabs to Number of acquisitions (Nacq) 2
saturate larynx/vessels/heart Slice location See Basic Protocol, step 16
Fat suppression Yes Flow compensation Yes (if available)
Slice series Left to right or the reverse Saturation pulses No
depending on preference Scan time ∼4 min
Scan time ∼4 min
aFSE: fast spin echo.

Sequence 3: Sagittal T2-weighted fast spin echo

Once this step has been performed, so long as the patient does not move on the table, the
table itself can be moved and then replaced in the same position as before without
jeopardizing the positioning of one scan relative to another.
9. If the patient is unable to hold still, provide an appropriate sedative.
Sequence 1: Rapid positioning pilot
10. To validate the patient’s position, run the system’s pilot (or scout) scan to (sequence
1) ensure correct location of the neck in three dimensions, using the imaging sequence
given in Table A8.2.2 or similar parameters.
This sequence usually consists of three orthogonal planes to allow subsequent localization.
The images are often also used later to determine where to place the saturation pulses and
to set up total coverage of the volume of interest.
Sequence 2: Sagittal T1-weighted conventional spin echo
11. Set the imaging parameters as shown in Table A8.2.3.
12. Use the pilot image to locate the spine in three dimensions to ensure coverage of the
region of interest (e.g., cervical, thoracic, lumbosacral spine).
13. Let the patient know you are ready and begin the scan.
Spinal Canal
Stenosis
14. Review the pilot scans and ensure that the saturation pulse is correctly placed anterior
to the slab of interest.
15. Run sequence 3 according to Table A8.2.4.
Sequence 4: Transverse T1-weighted conventional spin echo
16. Using the midline sagittal T1-weighted image acquired in sequence 2, set the
transverse acquisition parameters as follows:
a. Cervical spine: stacked images from C1 through C7-T1.
b. Thoracic spine: stacked images through levels of interest.
c. Lumbosacral spine: 5 slices each, angled to the plane of the intervertebral disc at
L3-4, L4-5, and L5-S1; one slice each, angled to the intervertebral disc at L1-2
and L2-3.
17. Supplement additional slices according to visible disease present or to clinical query.
18. Run the sequence according to Table A8.2.5.
Sequence 5: Transverse T2-weighted fast spin echo
19. Using the midline T1-weighted image acquired in sequence 2, repeat the setup as in
Table A8.2.6.
20. Run sequences according to Table A8.2.6.
Extradural Spine

A8.2.4 A8.2.5
Supplement 2 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 2
 Table A8.2.6 Primary Clinical Imaging Parameters for Sequence 5
(T2-Weighted Image, FSEa)
A B
Patient position Supine
Scan type Fast spin echo
Imaging plane (orientation) Transverse
Central slice or volume center Centered on the region of interest
(as in sequence 2, Table A8.2.3)
Echo time (TE) 100 msec
Echo train length (ETL) 8
Repeat time (TR) 4000 msec
Flip angle (FA) 90°
Fields of view (FOVx, FOVy) As in sequence 2, Table A8.2.3
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm
Thoracic: 1.25 mm, 1.25 mm
Lumbosacral: 1.09 mm, 1.09 mm
Number of data points collected (Nx, Ny) 256, 256
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) As in sequence 4 (Table A8.2.5)
Number of slices Varies with spinal level
Slice gap As in sequence 4 (Table A8.2.5)
Number of acquisitions (Nacq) 2
Slice location See Basic Protocol, step 16
Flow compensation Yes (if available)
Saturation pulses No
Scan time ∼4 min Figure A8.2.1 Cervical spinal stenosis. (A) Sagittal T2-weighted (TR = 4000 msec, TE = 100 msec) image
aFSE: fast spin echo.
shows multilevel stenosis of the central spinal canal. (B) Magnified view of (A) shows hyperintensity (arrows)
within the cervical spinal cord indicating myelomalacia.

COMMENTARY
Background Information
The etiology of central stenosis of the lum-
bosacral spinal canal may be either develop-
mental or acquired in nature, although most
cases presenting in adulthood represent a com-
bination of these factors. Pre-existent develop-
mental circumferential narrowing of the spinal
canal coupled with hypertrophic/redundant de-
generative changes of the intervertebral discs,
posterior spinal facet joints, and related spinal
structures/tissues results in a constriction of the
thecal sac and its contents. This may engender
a clinical syndrome known as “neurogenic or
spinal claudication.” Although it can be diffi-
cult on occasion to differentiate on a clinical
basis between neurogenic and vascular forms
of claudication, classically the former consists
of lower extremity pain that is both poorly
localized/characterized and is often associated
with leg numbness and/or weakness. These
signs and symptoms are frequently bilateral and
are usually preceded for months or years by
intractable low back pain. Neurogenic claudi-
cation is typically exacerbated by walking or
Spinal Canal standing erect and is relieved somewhat by
Stenosis
spinal flexion. A common presenting complaint
is also uni- or bilateral radiating lower extrem-
ity radicular pain. Other clinical findings such
as unusual or poorly defined lower extremity
paresthesias, penile erections on walking, and
sexual impotence are seen with lesser fre-
quency. The cause of this clinical syndrome is
believed to be the result of chronic or intermit-
tent progressive compression of the cauda
equina at the level(s) of central spinal canal
stenosis.
Pathophysiologically, the fundamental neu-
ral abnormalities include direct repeated local
nerve root trauma, focal nerve root ischemia,
inhibition of free access of the nerve root to the
nutritional effects of free-flowing CSF (cere-
brospinal fluid), obstruction of axoplasmic
flow within the nerve root, axon demyelina-
tion/remyelination, and extensive axon degen-
eration/regeneration. It has also been shown
that constriction of the lumbar canal, together
with repetitive motion induced by physical
flexion and extension, produces a stretching of
nerve roots and thereby results in dramatic
neural elongation and redundancy both above
and/or below the central spinal canal stenosis.
This nerve root redundancy on conventional
myelographic inspection may have a serpentine
configuration intradurally that simulates, and
can be confused with, vascular structures (i.e.,
vascular malformation). It has been postulated
that the major factor in the intrathecal enhance-
ment on i.v. gadolinium-enhanced MRI seen on
both sides of severe central lumbar canal
stenosis represents pathologic disruption of the
blood-nerve barrier within compressed, elon-
gated, and redundant nerve roots. Among other
things, it has been previously shown that benign
blood-nerve barrier disruption occurs in re-
sponse to mechanical trauma, Wallerian degen-
eration, and sterile inflammation. A contribu-
tory mechanism of intrathecal enhancement in
such severe cases of central spinal stenosis may
involve neural and perineural vascular engorge-
ment and stagnation [i.e., of the great radicular
vein(s)] that accompanies nerve root constric-
tion, thus resulting in prominent enhancement
of the vasa nervorum. From a practical stand-
point, the presence of intrathecal enhancement
in this setting signifies unequivocal underlying
neural pathology associated with spinal disease
of a stenotic nature in cases of clinical claudi-
cation of ambiguous etiology (Jinkins, 1993).
Critical Parameters and
Troubleshooting
CSF flow, cardiac, laryngeal, body wall and
other sources of motion can produce artifacts
that may, on occasion, significantly degrade the
images. Proper spatial (e.g., prevertebral) satu-
ration pulses and sometimes cardiac/respira-
tory gating can reduce these artifacts signifi-
cantly. In some instances, these artifacts may
be difficult or impossible to easily overcome
from patient to patient.
Anticipated Results
The goal of studying the spine in cases of
clinically suspected spinal stenosis is to deter-
mine the level(s), region(s), and degree(s) of
severity of the stenosis. The images acquired
utilizing these MRI sequences will allow the
most accurate assessment of the central spinal
canal, lateral recesses, and spinal neural foram-
ina that is available to the medical imaging
Extradural Spine

A8.2.6 A8.2.7
Supplement 2 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 2
 Schnebel, B., Kingston, S., Watkins, R., and Dillon,
W. 1989. Comparison of MRI to contrast CT in
the diagnosis of spinal stenosis. Spine 14:332-
Contributed by J. Randy Jinkins and
A B David D. Stark
337.
Downstate Medical Center
Shellock, F.G. 1996. Pocket Guide to MR Proce- State University of New York
dures and Metallic Objects. Lippincott-Raven,
Philadelpia.
Brooklyn, New York

Figure A8.2.2 Lumbar spinal stenosis. (A) Sagittal T1-weighted (TR = 500 msec, TE = 10 msec)
image shows degenerative anterolisthesis of L4 on L5 (arrow). (B) Sagittal T2-weighted (TR = 4000
msec, TE = 100 msec) image shows stenosis of the central spinal canal (arrow). (C) Parasagittal
T1-weighted (TR = 500 msec, TE = 10 msec) image shows partial collapse of the intervertebral disc
(open arrow) and stenosis of the spinal neural foramen (solid arrow).

specialist today. Some examples are shown in
Figure A8.2.1 and Figure A8.2.2.
Literature Cited
Garfin, S.R., Rydevik, B.L., Lipson, S.J., et al. 1992.
Spinal stenosis. In The Spine, 3rd ed. (R.H.
Rothman, F.A. Simeone, eds.) pp. 791-857. W.B.
Saunders, Philadelphia.
Hasegawa, T., An, H.S., and Haughton, V.M. 1993.
Imaging anatomy of the lateral lumbar spinal
canal. Semin. Ultrasound CT MRI 14(6):404-
Spinal Canal 413.
Stenosis
Jinkins, J.R. 1993. Gd-DTPA enhanced MR of the
lumbar spinal canal in patients with claudication.
J. Assist. Comput. Tomogr. 17:555-562.
Jinkins, J.R. 1999. MR evaluation of stenosis involv-
ing the neural foramina lateral recesses and cen-
tral canal of the lumbosacral spine. Neuroimag.
Clin. North Am. 7:493-511.
Kent, D.L., Haynor, D.R., Larson, E.B., and Deyo,
R.A. 1992. Diagnosis of lumbar spinal stenosis
in adults; a metaanalysis of the accuracy of CT,
MR, and myelography. Am. J. Roentgenol.
158:1135-1144.
Extradural Spine

A8.2.8 A8.2.9
Supplement 2 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 2
Spondylosis Deformans UNIT A8.3 6. Center the coil over the region where the key information is desired.
Make sure that the body is constrained to prevent motion, especially if high-resolution scans
Spondylosis deformans principally consists of osteophytosis at the borders of the inter- are to be run.
vertebral discs. Degeneration of the posterior spinal facet (zygapophyseal) joints and 7. If needed, place a pillow or other support under the knees to make the patient more
cervical uncovertebral joints are often included under the general label of “spondylosis.” comfortable.
By middle age, these changes are almost universally present throughout the population,
and may therefore be the cause of low back and radicular pain (Resnick, 1985; Schneck, 8. Use the centering light to position patient (cervical spine: thyroid cartilage; thoracic
1985). spine: nipple line; lumbar spine: iliac crests) and put him or her into the center of the
magnet.
CONVENTIONAL AND FAST SPIN ECHO ACQUISITIONS BASIC Once this step has been performed, so long as the patient does not move on the table, the
PROTOCOL table itself can be moved and then replaced in the same position as before without
In general, the signal characteristics of the skeletal components of the vertebral column
jeopardizing the positioning of one scan relative to another.
typically are best illustrated on the basis of spin-echo techniques. The margins of the
osteophytosis associated with spondylosis deformans are generally well defined utilizing 9. If the patient is unable to hold still, provide an appropriate sedative.
these acquisitions.
Sequence 1: Rapid positioning pilot
Table A8.3.1 lists the hardware necessary to perform the procedure, along with appropri- 10. To validate the patient’s position, run the system’s pilot (or scout) scan (sequence 1)
ate parameters. The available gradient strength will depend on the scanner, and the echo to ensure correct location of the neck in three dimensions, using the imaging sequence
times (e.g., TE) given in other tables will vary accordingly (i.e., the smaller the gradient given in Table A8.3.2 or similar parameters.
strength, the longer the echo time for a particular scan).
This sequence usually consists of three orthogonal planes to allow subsequent localization.
NOTE: Be sure that technicians and nurses have immediate access to any emergency The images are often also used later to determine where to place the saturation pulses and
equipment that may be relevant to a given study, or that may be needed for a particular to set up total coverage of the volume of interest.
patient, such as crash carts or oxygen.
Sequence 2: Sagittal T1-weighted conventional spin echo
Set-up patient and equipment 11. Set the imaging parameters as shown in Table A8.3.3.
1. Interview (screen) the patient to ensure that he or she has no contraindications such
as cardiac pacemakers or other implants containing ferromagnetic materials. Also be
Table A8.3.2 Primary Clinical Imaging Parameters for Sequence 1 (Pilot Scan)
sure to find out if the patient has any health conditions that may require the presence
of special emergency equipment during the scanning procedure, or necessitate any Patient position Supine
other precautions. Scan type Gradient echo
2. If the procedure is a research protocol, have the patient sign any necessary consent Imaging plane (orientation) Transverse, sagittal, coronal
form. Central slice or volume center Laser light centered at cervical
spine: thyroid cartilage; thoracic
3. Have the patient remove all jewelry and change into a gown to eliminate any metal spine: nipple line; lumbar spine:
that might be found in clothing. iliac crests
Echo time (TE) As short as possible
4. Inform the patient about what will occur during the procedure, what he or she will Repeat time (TR) As short as possible
experience while in the magnet, and how to behave. Flip angle (FA) 15°
5. Have the patient mount onto the table. Either before or right after the patient lies Fields of view (FOVx, FOVy) Cervical: 240 mm, 240 mm
down, set up any triggering devices or other monitoring equipment that is to be used. Thoracic: 340 mm, 340 mm
Lumbosacral: 280 mm, 280 mm
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm
Table A8.3.1 Equipment Parameters for Spine Imaging in Cases of
Spondylosis Deformans
Thoracic: 1.25 mm, 1.25 mm
Lumbosacral: 1.09 mm, 1.09 mm
Coil type Cervical, thoracic, lumbar: phase array Number of data points collected (Nx, Ny) 256, 256
surface coil (or other depending upon Display matrix (Dx, Dy) 256, 256
machine compatibility and availability) Slice thickness (Δz) 5 mm
Flow compensation Any level (optional) Number of slices 3
Slice gap Not applicable
Peripheral gating Thoracic spine only (optional)
Number of acquisitions (Nacq) 1
Respiratory gating Thoracic spine only (optional)
Scan time ∼10 sec
Use of contrast agents No Spondylosis
Extradural Spine Deformans
Contributed by J. Randy Jinkins and David D. Stark A8.3.1 A8.3.2
Current Protocols in Magnetic Resonance Imaging (2001) A8.3.1-A8.3.10
Copyright © 2001 by John Wiley & Sons, Inc. Supplement 1 Supplement 1 Current Protocols in Magnetic Resonance Imaging
 Table A8.3.3 Primary Clinical Imaging Parameters for Sequence 2 Table A8.3.4 Primary Clinical Imaging Parameters for Sequence 3
(T1-Weighted Image) (T2-Weighted Image, FSE)a

Patient position Supine Patient position Supine

Scan type Conventional spin echo Scan type Fast spin echo
Imaging plane (orientation) Sagittal Imaging plane (orientation) Sagittal
Central slice or volume center Centered on: Central slice or volume center Centered on area of interest (as in
Cervical: 3rd cervical vertebra Sequence 2, Table A8.3.3)
Thoracic: 6th thoracic vertebra Echo time (TE) 100 msec
Lumbar: 3rd lumbar vertebra Echo train length (ETL) 8
Echo time (TE) 10 msec Repeat time (TR) 4000 msec
Repeat time (TR) 500 msec Flip angle (FA) 90°
Flip angle (FA) 90° Fields of view (FOVx, FOVy) As in Sequence 2, Table A8.3.3
Fields of view (FOVx, FOVy) Cervical: 240 mm, 240 mm Resolution (Δx, Δy) Cervical: 0.47 mm, 0.47 mm
Thoracic: 320 mm, 320 mm Thoracic: 0.63 mm, 0.63 mm
Lumbosacral: 280 mm, 280 mm Lumbosacral: 0.55 mm, 0.55 mm
(may use rectangular field of view, Number of data points collected (Nx, Ny) 512, 512
e.g., half or three-quarter field, if Display matrix (Dx, Dy) 512, 512
available, or tailor to region of Slice thickness (Δz) Cervical: 3 mm
interest) Thoracic: 3 mm
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm Lumbar: 1 mm
Thoracic: 1.25 mm, 1.25 mm Number of slices Varies with spinal level
Lumbosacral: 1.09 mm, 1.09 mm Slice gap Cervical: 0.5 mm
Number of data points collected (Nx, Ny) 256, 256 Thoracic: 1 mm
Display matrix (Dx, Dy) 256, 256 Lumbar: 1 mm
Slice thickness (Δz) Cervical: 3 mm Number of acquisitions (Nacq) 1
Thoracic: 3 mm Flow compensation Yes (if available)
Lumbar: 1 mm Saturation pulses Yes; anterior
Number of slices As many as needed to cover cervical/thoracic/lumbar slabs
region of interest anteriorly to saturate
Slice gap Cervical: 0.5 mm larynx/vessels/heart
Thoracic: 1 mm Fat suppression Yes: chemical saturation or STIR
Lumbar: 1 mm (short tau inversion recovery)
Number of acquisitions (Nacq) 4 Slice series Left to right or the reverse
Flow compensation Yes (if available) depending on preference
Saturation pulses Yes; anterior Scan time ∼4 min
cervical/thoracic/lumbar slabs to aFSE: fast spin echo.
saturate larynx/vessels/heart
Slice series Left to right or the reverse Sequence 4: Transverse T1-weighted conventional spin echo
depending on preference
16. Using the midline sagittal T1-weighted image acquired in sequence 2, set the trans-
Scan time ∼8 min
verse acquisition parameters as follows:
a. Cervical spine: 3 to 4 slices each, angled to plane of individual intervetebral discs,
12. Use the pilot image to locate the spine in three dimensions to ensure coverage of the or stacked images throughout C1-T1.
region of interest (e.g., cervical, thoracic, lumbosacral spine). b. Thoracic spine: stacked images through levels of interest.
13. Let patient know you are ready, and begin the scan. c. Lumbosacral spine: 5 slices each, angled to the plane of the intervertebral disc at
L3-4, L4-5, and L5-S1; one or more slices each, angled to the intervertebral disc
Sequence 3: Sagittal T2-weighted fast spin echo at L1-2 and L2-3.
14. Review the pilot scans and ensure that the saturation pulse is correctly placed anterior 17. Supplement additional slices according to visible disease present or to clinical query.
to above the slab of interest.
18. Run sequence 4 according to Table A8.3.5.
15. Run sequence 3 according to Table A8.3.4.
Spondylosis
Extradural Spine Deformans

A8.3.3 A8.3.4
Current Protocols in Magnetic Resonance Imaging Supplement 1 Supplement 1 Current Protocols in Magnetic Resonance Imaging
 Table A8.3.5 Primary Clinical Imaging Parameters for Sequence 4 Table A8.3.6 Primary Clinical Imaging Parameters for Sequence 5
(T1-Weighted Image) (T2-Weighted Image, FSE)a

Patient position Supine Patient position Supine

Scan type Conventional spin echo Scan type Fast spin echo
Imaging plane (orientation) Transverse Imaging plane (orientation) Transverse
Central slice or volume center Centered on area of interest (as in Central slice or volume center Centered on region of interest (as
Sequence 2, Table A8.3.3) in Sequence 2)
Echo time (TE) 10 msec Echo time (TE) 100 msec
Repeat time (TR) 500 msec Echo train length (ETL) 8
Flip angle (FA) 90° Repeat time (TR) 4000 msec
Fields of view (FOVx, FOVy) As in Sequence 2, Table A8.3.3 Flip angle (FA) 90°
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm Fields of view (FOVx, FOVy) As in Sequence 2
Thoracic: 1.25 mm, 1.25 mm Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm
Lumbrosacral: 1.09 mm, 1.09 mm Thoracic: 1.25 mm, 1.25 mm
Number of data points collected (Nx, Ny) 256, 256 Lumbosacral: 1.09 mm, 1.09 mm
Display matrix (Dx, Dy) 256, 256 Number of data points collected (Nx, Ny) 256, 256
Slice thickness (Δz) Cervical: 3–4 mm Display matrix (Dx, Dy) 256, 256
Thoracic: 3–8 mm Slice thickness (Δz) Cervical: 3 mm
Lumbar: 4 mm Thoracic: 3 mm
Number of slices Varies with spinal level Lumbar: 1 mm
Slice gap Cervical: 1 mm Number of slices Varies with spinal level
Thoracic: 1–2 mm Slice gap Cervical: 0.5 mm
Lumbar: 1 mm Thoracic: 1 mm
Number of acquisitions (Nacq) 2 Lumbar: 1 mm
Slice locations See text (Basic Protocol, step 16) Number of acquisitions (Nacq) 2
Flow compensation Yes (if available) Slice locations See text (Basic Protocol, step 16)
Saturation pulses No Flow compensation Yes (if available)
Scan time ∼4 min Saturation pulses No
Scan time ∼4 min
aFSE: fast spin echo.

Sequence 5: Transverse T1-weighted fast spin echo Sequence 7: Sagittal gradient recalled echo
19. Using the midline T1-weighted image acquired in Sequence 2, repeat the setup as in 1. Run sequence 7 according to Table A8.3.8.
Table A8.3.6.
20. Run sequence 5 according to Table A8.3.6. Sequence 8: Transverse gradient recalled echo
2. Run sequence 8 according to Table A8.3.9.
Sequence 6: Transverse T2-weighted fast spin echo, fat suppressed (optional)
21. Using the midline T1-weighted image acquired in sequence 2, repeat the setup as in COMMENTARY
Table A8.3.7.
Background Information posteriorly and posterolaterally, may produce
22. Run sequence 6 according to Table A8.3.7. Spondylosis deformans is represented by or contribute to the various forms of spinal
osteophytes that form in response to degenera- stenosis, i.e., central spinal canal stenosis, lat-
tive protrusive disc alteration along the ring eral recess spinal canal stenosis, and neural
GRADIENT RECALLED ECHO ACQUISITIONS ALTERNATE
margins of the vertebral bodies at or near the foramen stenosis (Fig. A8.3.1).
PROTOCOL
With some machines, or according to preferences, gradient recalled echo acquisitions discovertebral margin. The bony vertebral ex- Osteoarthrosis of the apophyseal posterior
may be used in the sagittal and/or transverse planes to clearly distinguish between discs crescencies (i.e., osteophytes) appear and grow spinal (facet) joints of the spine consist of
and soft tissue and to clarify the spinal neural foramen in the cervical region (Tsuruda et from the point of traction. Initially, the growth erosion or complete denudation of the articular
al., 1990; Van Dyke et al., 1989; Yousem et al., 1991). is horizontal but becomes more vertical as the cartilage, joint-space narrowing, bony eburna-
process proceeds along the course of the out- tion, and osteophytosis. This may be accompa-
wardly protruding disc. These osteophytes nied by intra-articular osteocartilaginous loose
form predominantly in the anterior and lateral bodies and hyperplasia of the synovial mem-
Spondylosis aspects of the spine. Spondylosis, when present brane. A combination of factors may result in
Extradural Spine Deformans

A8.3.5 A8.3.6
Current Protocols in Magnetic Resonance Imaging Supplement 1 Supplement 1 Current Protocols in Magnetic Resonance Imaging
 Table A8.3.7 Primary Clinical Imaging Parameters for Sequence 6 Table A8.3.8 Primary Clinical Imaging Parameters for Sequence 7 (T2*GRE)a
(T2-Weighted Image, FSE, Fat Suppressed)
Patient position Supine
Patient position Supine Scan type Gradient echo
Scan type Fast spin echo Imaging plane (orientation) Sagittal
Imaging plane (orientation) Transverse Central slice or volume center Centered on the region of interest
Central slice or volume center Centered on region of interest (as (as in Sequence 2, Table A8.3.3)
in Sequence 2) Echo time (TE) 15 msec
Echo time (TE) 100 msec Repeat time (TR) 500 msec
Echo train length (ETL) 8 Flip angle (FA) 10° to 20°
Repeat time (TR) 4000 msec Fields of view (FOVx, FOVy) As in Sequence 2
Flip angle (FA) 90° Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm
Fields of view (FOVx, FOVy) As in Sequence 2 Thoracic: 1.25 mm, 1.25 mm
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm Lumbosacral: 1.09 mm, 1.09 mm
Thoracic: 1.25 mm, 1.25 mm Number of data points collected (Nx, Ny) 256, 256
Lumbrosacral: 1.09 mm, 1.09 mm
Display matrix (Dx, Dy) 256, 256
Number of data points collected (Nx, Ny) 256, 256
Slice thickness (Δz) 3 mm
Display matrix (Dx, Dy) 256, 256
Number of slices Varies with spinal level
Slice thickness (Δz) Cervical: 3 mm
Slice gap 1 mm or less
Thoracic: 3 mm
Number of acquisitions (Nacq) 3
Lumbar: 1 mm
Flow compensation Yes (if available)
Number of slices As many as needed to cover
Saturation pulses Yes
region of interest
Scan time ∼6 min
Slice gap Cervical: 0.5 mm
aT *GRE: T * gradient recalled echo.
Thoracic: 1 mm 2 2

Lumbar: 1 mm
Number of acquisitions (Nacq) 2
Slice locations See text (Basic Protocol, step 16) Table A8.3.9 Primary Clinical Imaging Parameters for Sequence 8 (T2*GRE)a
Flow compensation Yes (if available)
Saturation pulses No Patient position Supine
Fat suppression Yes, fat saturation or STIR (short Scan type Gradient echo
tau inversion recovery) Imaging plane (orientation) Transverse
Scan time ∼4 min Central slice or volume center Centered on the region of interest
(as in Sequence 2, Table A8.3.3)
Echo time (TE) 15 msec
narrowing of the spinal neural foramen and Luschka.” These articulations are not evident Repeat time (TR) 500 msec
entrapment/compression of the transiting spi- in the cervical spine in all individuals, and when Flip angle (FA) 10° to 20°
nal nerve root and, at some spinal levels (e.g., present are not found at all levels. Typically Fields of view (FOVx, FOVy) As in Sequence 2, Table A8.3.3
lumbar spine), the dorsal root ganglion. Those they are observed in the caudal five cervical Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm
factors that produce neural foramen stenosis vertebral bodies (i.e., C3 to C7). These articu- Thoracic: 1.25 mm, 1.25 mm
include a reduction in height of the neural lations are formed from bony ridges that extend Lumbosacral: 1.09 mm, 1.09 mm
foramen due to degenerative narrowing of the in a cranial direction from the superolateral Number of data points collected (Nx, Ny) 256, 256
apophyseal joint space and intervertebral disc margin of the vertebral body; these ridges are Display matrix (Dx, Dy) 256, 256
space, hyperplasia/redundance of the synovial termed the uncinate or lunate processes. The Slice thickness (Δz) 3–4 mm
membrane of the apophyseal joint, facet joint uncinate processes form a modified articulation
Number of slices Varies with spinal level
osteophyte formation, and posterolateral disco- with the inferolateral surface of the suprajacent
Slice gap 1 mm or less
genic encroachment (e.g., intervertebral disc vertebral body. The modification is produced
Number of acquisitions (Nacq) 1
bulge/protrusion/extrusion, discogenic verte- by fibrillation and fissuring of the marginal
bral osteophytosis). These alterations may also fibers of the interposed anulus fibrosus; the Slice locations See text (Basic Protocol, step 16)
contribute to concomitant stenosis of the cen- lining of the uncinate articulations are the car- Flow compensation Yes (if available)
tral spinal canal and the lateral recesses of the tilaginous end plates of the respective vertebral Saturation pulses Yes
central spinal canal. One other phenomenon bodies. Although it simulates synovium-lined Scan time ∼2 min
that may cause stenosis of the neural foramina articulation, strictly speaking, this is not what aT *GRE: T *gradient recalled echo.
2 2
in the cervical region is osteoarthrosis of the it is. These articulations develop postnatally
Spondylosis
uncovertebral articulations, or “joints of and undergo degenerative changes later in life. Extradural Spine Deformans

A8.3.7 A8.3.8
Current Protocols in Magnetic Resonance Imaging Supplement 1 Supplement 1 Current Protocols in Magnetic Resonance Imaging
 A B B

Figure A8.3.1 Cervical spondylosis. (A) Sagittal T1-weighted (TR = 500 msec, TE = 10 msec) image
shows diffuse cervical intervertebral disc space narrowing and posterior disc protrusions at multiple levels.
(B) Sagittal T2-weighted (TR = 4000 msec, TE = 100 msec) image with fat suppression shows the
degenerative anterior and posterior protrusive changes at multiple levels.

Because they are cartilaginous articulations,
degeneration of the uncovertebral articulations
is referred to as osteoarthrosis. Uncovertebral
articulation osteoarthrosis consists of narrow-
ing of the articular space, erosion/denudation
of the articular cartilage, and marginal osteo-
phytosis. As noted above, this process may
contribute to the narrowing of the spinal neural
foramina at the levels of the cervical spine at
which uncovertebral articulations exist.
Musculoligamentous and vertebral marrow
edema may occur in association with these
degenerative changes. These alterations can
only be observed properly using MRI (Fig.
A8.3.2).
Critical Parameters and
Troubleshooting
Cerebrospinal fluid (CSF) flow, cardiac, la-
ryngeal, body wall, and other sources of motion
can produce artifacts that can on occasion sig-
nificantly degrade the images. Proper spatial
(e.g., prevertebral) saturation pulses and some-
times flow compensation pulses and/or car-
diac/respiratory gating can reduce these arti-
facts significantly. In many instances these ar-
tifacts may be difficult or impossible to easily
overcome from patient to patient.
Degenerative musculoligamentous and ver-
tebral marrow edema are only properly demon-
strated on MRI using fat suppressed techniques
coupled with T2-weighted imaging. The sagit-
tal plane should be acquired in this manner; a
similar acquisition may also be acquired in the
transverse plane if desired.
Anticipated Results
The goal of studying the spine in patients
with suspected advanced degenerative spondy-
losis is to reveal and document the presence,
level, and severity of peridiscal, uncovertebral,
and/or posterior spinal facet joint osteophy-
tosis. These alterations in turn may lead to
narrowing of the central spinal canal, lateral
recesses of the central spinal canal, and spinal
(intervertebral) neural foramen. The protocols
in this unit will allow an accurate assessment
of these phenomena.
Literature Cited
Resnick, D. 1985. Degenerative diseases of the ver-
tebral column. Radiology 156:3-14.
*
Figure A8.3.2 Degenerative spinal musculoligamentous alterations. (A) Sagittal T2-weighted (TR = 4000
msec, TE = 100 msec) fat suppressed image shows interspinous ligament hyperintensity (arrows) indicating
degeneration/rupture. (B) Transverse T1-weighted (TR = 500 msec, TE = 10 msec) image shows the multifidus
muscle (arrow) to be swollen and relatively hyperintense on the right side. (C) Transverse T2-weighted (TR
= 4000 msec, TE = 100 msec) fat suppressed image shows hyperintense degeneration of the multifidus
muscle (asterisk) on the right side.
Schneck, C. 1985. The anatomy of lumbar spondy- Yousem, D.M., Atlas, S.W., Goldberg, H.I., and
losis. Clin. Orthop. 193:20-37. Grossman, R.I. 1991. Degenerative narrowing of
Tsuruda, J.S., Norman, D., Dillon, W., Newton, the cervical spine neural foramina: Evaluation
with high-resolution 3DFT gradient-echo imag-
T.H., and Mills, D.G. 1990. Three-dimensional
ing. Am. J. Neuroradiol. 12:229-230.
gradient recalled MR imaging as a screening tool
for the diagnosis of cervical radiculopathy. Am.
J. Neurol. 1263-1271. Contributed by J. Randy Jinkins and
Van Dyke, C., Ross, J.S., Tkach, L., Masaryk, T.J., David D. Stark
and Modic, M.T. 1989. Gradient-echo MR im- Downstate Medical Center
aging of the cervical spine: Evaluation of ex- State University of New York
tradural disease. Am. J. Neuroradiol. 10:627- Brooklyn, New York
Spondylosis 632.
Extradural Spine Deformans

A8.3.9 A8.3.10
Current Protocols in Magnetic Resonance Imaging Supplement 1 Supplement 1 Current Protocols in Magnetic Resonance Imaging
Spinal Inflammation
In part because of the epidemic of acquired immunodeficiency syndrome (AIDS), the
frequency of central nervous system (CNS) infections has been on the rise (Post et al.,
1986; Provenzale and Jinkins, 1997; Thurnher et al., 1997). The use of diagnostic imaging
in the evaluation of infections involving the CNS has become an essential component of
patient care. Patients with inflammatory disease of the CNS usually present with nonspe-
cific signs and symptoms that often require a more detailed investigation. Early deline-
ation of the location, extent, and type of infection may be an important determinant of
outcome. While the modern imaging of spinal inflammatory conditions can be carried
out using various methods ranging from conventional radiography through the newer
computed imaging techniques, MRI has proven to be the most sensitive, if not specific,
imaging technique. This unit will outline the various techniques realizing this potential
(Modic et al., 1985; Chang et al., 1989; Sharif et al., 1990; Sklar et al., 1993; Jinkins et
al., 1995; Reddy et al., 1995).
CONVENTIONAL AND FAST SPIN ECHO
Improvements in technology have made MR the investigation of choice in the diagnostic
evaluation of spinal infections. MR provides anatomic localization and characterization
of spinal infections that is far superior to any other imaging technique. MR imaging has
the advantage of multiplanar direct visualization of the spinal cord, subarachnoid space,
extradural soft tissues, and spinal column noninvasively without necessitating the in-
trathecal injection of a contrast agent. In addition, pathologic changes in the spine are
seen earlier on MR than on any other imaging modality. Another advantage of MR is its
easy sequential repeatability and reproducibility in patients undergoing treatment for
spinal infection.
The following sequences comprise the preferred protocol for high field MR machines. In
addition, in all cases intravenous (i.v.) contrast enhanced protocols are mandatory to
evaluate the integrity of the blood-nerve and blood cord barrier in certain cases of
suspected CNS infection, or in cases of epidural and/or spinal column/paraspinal infection
(e.g., spondyloclodiscitis and perispinal abscess/phlegmon).
Table A8.4.1 lists the hardware necessary to perform the procedure, along with appropri-
ate parameters. The available gradient strength will depend on the scanner, and the echo
times given in other tables will be varied accordingly (i.e., the smaller the gradient
strength, the longer the echo time for a particular scan).
This entire protocol should take 40 to 60 min to complete.
Table A8.4.1 Equipment Parameters for Spine Imaging in Cases of
Inflammation
Coil type Cervical, thoracic, lumbar: phase array
surface coil (or other depending upon
machine compatibility and availability)
Gradient coil strength 25 mT/m (or whatever the system permits)
Flow compensation Optional (any level, if available)
Peripheral gating Optional (thoracic spine only)
Respiratory gating Optional (thoracic spine only)
Use of contrast agents Yes: intravenous gadolinium
Contributed by J. Randy Jinkins and David D. Stark
Current Protocols in Magnetic Resonance Imaging (2001) A8.4.1-A8.4.10
Copyright © 2001 by John Wiley & Sons, Inc.
UNIT A8.4 NOTE: Be sure that technicians and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.
Materials
Normal saline (0.9% NaCl), sterile
Gadolinium-based MR contrast agent (e.g., Magnevist, Omniscan, or Prohance)
Set up patient and equipment
1. Interview (screen) the patient to ensure that he or she has no contraindications such
as cardiac pacemakers or other implants containing ferromagnetic materials. Also be
sure to find out if the patient has any health conditions that may require the presence
of special emergency equipment during the scanning procedure, or necessitate any
other precautions.
Generally, standard screening forms are used for all patients scanned in a magnetic
resonance system.
BASIC The presence of any ferromagnetic materials may be a health hazard to the patient when
PROTOCOL he or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any implants; see Shellock (1996)
for discussion of what implants may be safely scanned using magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
2. If the procedure is a research protocol, have the patient sign any necessary consent
form.
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
4. Inform the patient about what will occur during the procedure, what he or she will
experience while in the magnet, and how to behave, including the following.
a. If earphones or headphones are used to protect the ears from the loud sounds
produced by the gradients, the patient will be asked to wear these, but will be able
to communicate with you at any time during the imaging.
b. The patient will be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this works).
c. For good results, the patient should not talk, and should avoid or minimize
swallowing or other movement, during each scan—i.e., as long as the banging
sounds continue. Between scans, talking and swallowing are allowed in most
cases, but should be avoided when comparative positional studies are being
performed; the patient will be informed when this is the case.
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
5. Have the patient mount onto the table. Either before or right after the patient lies
down, set up any triggering devices or other monitoring equipment that is to be used.
6. Center the coil over the region where the key information is desired.
Make sure that the body is constrained to prevent motion, especially if high-resolution scans
are to be run.
7. If needed, place a pillow or other support under the knees to make the patient more
Spinal comfortable.
Extradural Spine Inflammation
A8.4.1 A8.4.2
Supplement 1 Supplement 1 Current Protocols in Magnetic Resonance Imaging
 Table A8.4.2 Primary Clinical Imaging Parameters for Sequence 1 (Pilot Scan) Table A8.4.3 Primary Clinical Imaging Parameters for Sequence 2 (T1-Weighted
Image)
Patient position Supine
Scan type Gradient echo Patient position Supine
Imaging plane (orientation) Transverse, sagittal, coronal Scan type Conventional spin echo
Central slice or volume center Laser light centered at: Imaging plane (orientation) Sagittal
Cervical spine: thyroid cartilage Central slice or volume center Centered on:
Thoracic spine: nipple line Cervical: 3rd cervical vertebra
Lumbar spine: iliac crests Thoracic: 6th thoracic vertebra
Echo time (TE) As short as possible Lumbar: third lumbar vertebra
Repeat time (TR) As short as possible Echo time (TE) 10 msec
Flip angle (FA) 15° Repeat time (TR) 500 msec
Fields of view (FOVx, FOVy) Cervical: 240 mm, 240 mm Flip angle (FA) 90°
Thoracic: 320 mm, 320 mm Fields of view (FOVx, FOVy) Cervical: 240 mm, 240 mm
Lumbosacral: 280 mm, 280 mm Thoracic: 320 mm, 320 mm
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm Lumbosacral: 280 mm, 280 mm (may
Thoracic: 1.25 mm, 1.25 mm use rectangular field of view, e.g., half
or three-quarter field, if available, or
Lumbrosacral: 1.09 mm, 1.09 mm
tailor to region of interest)
Number of data points collected (Nx, Ny) 256, 256
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm
Display matrix (Dx, Dy) 256, 256
Thoracic: 1.25 mm, 1.25 mm
Slice thickness (Δz) 5 mm
Lumbosacral: 1.09 mm, 1.09 mm
Number of slices 3
Number of data points collected (Nx, Ny) 256, 256
Slice gap Not applicable
Display matrix (Dx, Dy) 256, 256
Number of acquisitions (Nacq) 1
Slice thickness (Δz) Cervical: 3 mm
Scan time ∼10 sec
Thoracic: 3 mm
Lumbar: 1 mm
Number of slices As many as needed to cover the region
8. Use the centering light to position the patient (cervical spine: thyroid cartilage; of interest
thoracic spine: nipple line; lumbar spine: iliac crests) and put him or her into the Slice gap Cervical: 0.5 mm
center of the magnet. Thoracic: 1 mm
Once this step has been performed, so long as the patient does not move on the table, the Lumbar: 1 mm
table itself can be moved and then replaced in the same position as before without Number of acquisitions (Nacq) 2
jeopardizing the positioning of one scan relative to another. Flow compensation Yes (if available)
Saturation pulses Yes; anterior cervical/thoracic/lumbar
9. If the patient is unable to hold still, provide an appropriate sedative. slab to saturate larynx/vessels
Slice series Left to right or the reverse depending
Sequence 1: Rapid positioning pilot on preference
10. To validate the patient’s position, run the system’s pilot (or scout) scan (sequence 1) Scan time ∼4 min
to ensure correct location of the neck in three dimensions, using the imaging sequence
given in Table A8.4.2 or similar parameters.
This sequence usually consists of three orthogonal planes to allow subsequent localization. Sequence 3: Sagittal T2-weighted fast spin echo, fat suppressed
The images are often also used later to determine where to place the saturation pulses and 14. Review the pilot scans and ensure that the saturation pulse is correctly placed anterior
to set up total coverage of the volume of interest. to above the slab of interest.

Sequence 2: Sagittal T1-weighted conventional spin echo 15. Run sequence 3 according to Table A8.4.4.
11. Set the imaging parameters as shown in Table A8.4.3.
Sequence 4: Transverse T1-weighted conventional spin echo
12. Use the pilot image to locate the spine in three dimensions to ensure coverage of the 16. Using the midline sagittal T1-weighted image acquired in Sequence 2, set the
region of interest (e.g., cervical, thoracic, lumbosacral spine). transverse acquisition parameters as follows:
13. Let the patient know you are ready and begin the scan. a. Cervical spine: stacked images from C1 through C1-T1
b. Thoracic spine: stacked images through levels of interest
Spinal
Extradural Spine Inflammation

A8.4.3 A8.4.4
Current Protocols in Magnetic Resonance Imaging Supplement 1 Supplement 1 Current Protocols in Magnetic Resonance Imaging
 Table A8.4.4 Primary Clinical Imaging Parameters for Sequence 3 (T2-Weighted Table A8.4.5 Primary Clinical Imaging Parameters for Sequence 4 (T1-Weighted
Image, FSE)a Image)

Patient position Supine Patient position Supine

Scan type Fast spin echo Scan type Conventional spin echo
Imaging plane (orientation) Sagittal Imaging plane (orientation) Transverse
Central slice or volume center Centered on area of interest (as in Central slice or volume center Centered on the area of interest (as in
Sequence 2, Table A8.4.3) Sequence 2, Table A8.4.3)
Echo time (TE) 100 msec Echo time (TE) 10 msec
Echo train length (ETL) 8 Repeat time (TR) 500 msec
Repeat time (TR) 4000 msec Flip angle (FA) 90°
Flip angle (FA) 90° Fields of view (FOVx, FOVy) As in Sequence 2, Table A8.4.3
Fields of view (FOVx, FOVy) As in Sequence 2, Table A8.4.3 Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm
Resolution (Δx, Δy) Cervical: 0.47 mm, 0.47 mm Thoracic: 1.25 mm, 1.25 mm
Thoracic: 0.63 mm, 0.63 mm Lumbosacral: 1.09 mm, 1.09 mm
Lumbosacral: 0.55 mm, 0.55 mm Number of data points collected (Nx, Ny) 256, 256
Number of data points collected (Nx, Ny) 512, 512 Display matrix (Dx, Dy) 256, 256
Display matrix (Dx, Dy) 512, 512 Slice thickness (Δz) Cervical: 3 mm
Slice thickness (Δz) Cervical: 3 mm Thoracic: 3–8 mm
Thoracic: 3 mm Lumbar: 4 mm
Lumbar: 1 mm Number of slices Varies with spinal level
Number of slices Varies with spinal level Slice gap Cervical: 1 mm
Slice gap Cervical: 0.5 mm Thoracic: 1–2 mm
Thoracic: 1 mm Lumbar: 1 mm
Lumbar: 1 mm Number of acquisitions (Nacq) 2
Number of acquisitions (Nacq) 1 Slice locations See text (Basic Protocol, step 16)
Flow compensation Yes (if available) Flow compensation Yes (if available)
Saturation pulses Yes; anterior cervical/thoracic/lumbar Scan time ∼4 min
slabs to saturate larynx/vessels/heart
Fat suppression Yes: chemical saturation or STIR
(short tau inversion recovery)
22. Establish an i.v. line from which the contrast agent can be injected, and attach this
Slice series Left to right or the reverse depending
line securely to the patient so that movement into or out of the magnet will not pull
on preference
Scan time
at the patient’s arm.
∼4 min
aFSE: fast spin echo. It is preferable to insert the line prior to imaging and to leave the patient in the magnet,
with no intervening motion, between the scans that are run before contrast agent injection
and those run after injection.
c. Lumbosacral spine: 5 slices each, angled to the plane of the intervertebral disc at
L3-4, L4-5, and L5-S1; one or more slices each, angled to the intervertebral disc Run pilot scan
at L1-2 and L2-3. 23. Run a rapid three-plane positioning pilot scan (see sequence 1).

17. Supplement additional slices according to visible disease present or to clinical query. Scan sequences
24. Leaving the patient in the magnet, inject the contrast agent, flush the i.v. line with 10
18. Run sequence 4 according to Table A8.4.5. cc saline, and then immediately run sagittal (sequence 6) and transverse (sequence
7) T1-weighted image sequences (see sequences 2 and 4).
Sequence 5: Transverse T1-weighted fast spin echo
19. Using the midline T1-weighted image acquired in sequence 2, repeat the setup as in A dose of 0.1 mmol/kg of contrast agent is usually given i.v.
Table A8.4.6. This is performed to demonstrate solid or cavitary infectious disease (Post et al., 1990).

20. Run sequence 5 according to Table A8.4.6. Fat suppression techniques may be very useful in order to suppress osseous vertebral
marrow fat and perispinal soft tissue fat, thereby clearly distinguishing enhancement from
normal fatty tissue. Fat saturation or STIR (short tau inversion recovery) techniques may
Sequences 6 and 7: Sagittal (Sequence 6) and transverse (Sequence 7) i.v. contrast
be used to suppress fat.
enhanced T1-weighted spin echo
21. Remove patient from the magnet. Using the same equipment, set up equipment and
Spinal
patient as in steps 5 to 9. Extradural Spine Inflammation

A8.4.5 A8.4.6
Current Protocols in Magnetic Resonance Imaging Supplement 1 Supplement 1 Current Protocols in Magnetic Resonance Imaging
 Table A8.4.6 Primary Clinical Imaging Parameters for Sequence 5 (T2-Weighted Table A8.4.7 Primary Clinical Imaging Parameters for Sequence 8 (T1-Weighted
Image, FSE)a Image)

Patient position Supine Patient position Supine

Scan type Fast spin echo Scan type Conventional spin echo
Imaging plane (orientation) Transverse Imaging plane (orientation) Coronal
Central slice or volume center Centered on the region of interest (as Central slice or volume center Centered on the area of interest (as in
in Sequence 2, Table A8.4.3) sequence 2, Table A8.4.3)
Echo time (TE) 100 msec Echo time (TE) 10 msec
Echo train length (ETL) 6 Repeat time (TR) 500 msec
Repeat time (TR) 4000 msec Flip angle (FA) 90°
Flip angle (FA) 90° Fields of view (FOVx, FOVy) As in Sequence 2, Table A8.4.3
Fields of view (FOVx, FOVy) As in Sequence 2, Table A8.4.3 Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm Thoracic: 1.25 mm, 1.25 mm
Thoracic: 1.25 mm, 1.25 mm Lumbosacral: 1.09 mm, 1.09 mm
Lumbosacral: 1.09 mm, 1.09 mm Number of data points collected (Nx, Ny) 256, 256
Number of data points collected (Nx, Ny) 256, 256 Display matrix (Dx, Dy) 256, 256
Display matrix (Dx, Dy) 256, 256 Slice thickness (Δz) 3–4 mm
Slice thickness (Δz) Cervical: 3 mm Number of slices Varies with spinal level
Thoracic: 3–8 mm Slice gap 1–2 mm
Lumbar: 4 mm Number of acquisitions (Nacq) 2
Number of slices Varies with spinal level Slice locations From front to back of spinal column
Slice gap Cervical: 1 mm Saturation pulses No
Thoracic: 1–2 mm Fat suppression Yes: fat suppression or STIR (short tau
Lumbar 1 mm inversion recovery)
Number of acquisitions (Nacq) 2 Scan time ∼4 min
Slice locations See text (Basic Protocol, step 16)
Flow compensation Yes (if available)
Saturation pulses No COMMENTARY
Scan time ∼6 min
Background Information myelitis, granuloma, or abscess. Infection
aFSE: fast spin echo.
The most common symptoms of spinal in- within the vertebral body usually starts in the
fection include malaise, back pain, and focal anterior (ventral) subchondral region. It can
CORONAL CONVENTIONAL SPIN ECHO-ENHANCED ALTERNATE tenderness. Spinal infections may present with spread to the adjacent disk and/or directly to
PROTOCOL signs that vary in degree of severity from a mild adjacent vertebrae, sparing the disk by tracking
In some instances, a coronal acquisition may be helpful to analyse the perispinal tissues fever to a life-threatening septicemia and para- beneath the longitudinal ligaments in the po-
for abscess formation. plegia. Compromised hosts such as diabetics, tential subligamentous space. Alternatively, the
intravenous drug abusers, and chronically ill infection can be disseminated directly to the
Run pilot scan patients are particularly susceptible to spinal leptomeninges or the parenchyma of the spinal
1. Run a rapid three-plane positioning pilot scan (see Basic Protocol, sequence 1). infections. In patients who are not abusing cord. This may result in arachnoiditis, myelitis,
intravenous medications, the most common or meningeal/parenchymal granuloma/abscess
Sequence 8: Contrast enhanced coronal T1-weighted sequence route of spread to the spine is still a hemato- formation.
2. Leaving patient in the magnet, inject the contrast agent, flush i.v. line with 10 ml genous one, originating from the genitourinary It has been reported that MR has a sensitivity
saline, and then immediately run sagittal (sequence 6) and transverse (sequence 7) tract, gastrointestinary tract, gastrointestinal of 96%, a specificity of 92%, and an accuracy
T1-weighted image sequences (see Basic Protocol, sequences 2 and 4). tract, skin, or respiratory tract. Infections of the of 94% in the diagnosis of vertebral osteomyeli-
spine also can spread via direct extension from tis. This is roughly equivalent to the combined
A dose of 0.1 mmol/kg of contrast agent is usually given. infections involving adjacent paraspinal tis- results of technetium radionuclide bone scans
sues, or by inoculation occurring from pene- and gallium radionuclide studies.
Fat suppression techniques may be very useful in order to suppress osseous vertebral
marrow fat and perispinal soft tissue fat, thereby clearly distinguishing enhancement from trating trauma, therapeutic procedure, or diag- MR is also the most sensitive imaging mo-
normal fatty tissue. nostic studies. The usual pathologic manifesta- dality for the detection of diskitis (Modic et al.,
tions are the disk space infection, osteomyelitis 1985). Sagittal sections are especially useful
3. Run the sequence according to Table A8.4.7. of the vertebral body (or bodies), epidural for demonstrating infection of the disk space
phlegmon (solid soft tissue inflammation), and adjacent vertebral body. The transverse
As stated above, fat suppression may be useful (see sequences 6 and 7).
epidural abscess, meningitis, paraspinal phleg- sections are excellent for assessing paraverte-
Spinal mon/abscess, and intramedullary spinal cord bral soft tissue extension. The infected disk
Extradural Spine Inflammation

A8.4.7 A8.4.8
Current Protocols in Magnetic Resonance Imaging Supplement 1 Supplement 1 Current Protocols in Magnetic Resonance Imaging
 condition. However, in the early stages of
osteomyelitis, where only the vertebral body is
A B involved, this differentiation becomes difficult.
Gadolinium (Gd) is valuable in cases of
suspected spinal infection because it demon-
strates the focus of spinal involvement, directs
needle biopsy if necessary, and provides a sen-
sitive means of following the resolving infec-
tion on medical therapy (Post et al., 1990). In
one study, although the diagnosis of discitis and
osteomyelitis could not be made in all patients
on the basis of unenhanced MR alone, enhance-
ment with intravenous Gd resulted in the cor-
rect diagnosis in all cases.
Gd-enhanced studies also provide good de-
marcation of the epidural or paraspinous soft
tissue spread of spinal infection. The majority
of epidural inflammatory processes homogene-
ously enhance, indicating that they are pre-
C dominantly phlegmonous (Fig. A8.4.1). On the
other hand, extradural abscesses have a sharp
margin of enhancement surrounding the central
nonenhancing fluid pus.
Gd-enhanced MR is the investigation of
choice in cases of intramedullary infection of
the spinal cord. The enhanced MR may show a
ring- or solid-enhancing lesion within the sub-
stance of an enlarged cord. There is invariably
a surrounding area of nonenhancing cord
edema. In acute/subacute infectious arach-
noiditis, Gd-enhanced MR shows matting of
nerve roots and diffuse leptomeningeal and
nerve root enhancement.
Critical Parameters and
Troubleshooting
Cerebrospinal fluid (CSF) flow, as well as
Figure A8.4.1 Thoracic spondylodiscitis (Staphylococcus aureus). (A) Sagittal T1-weighted (TR
= 500 msec, TE = 10 msec) image shows poor definition of the intervertebral disc at T5-T6 and an cardiac, laryngeal, body wall, and other sources
anterior and epidural spinal mass. Also the vertebral body marrow of T5 and T6 is hypointense. (B) can produce artifacts that can on occasion sig-
Sagittal T2-weighted (TR = 4000 msec, TE = 100 msec) MR shows abnormal hyperintensity in the nificantly degrade the images. Proper spatial
marrow of T5 and T6 with early segmental collapse of the spinal column. Anterior epidural and saturation pulses, cardiac/respiratory gating,
prevertebral mass for motion (arrows) is also noted. (C) Intravenous gadolinium enhanced T1- and flow compensation gradients can reduce
weighted (TR = 500 msec, TE = 10 msec) image shows abnormal enhancement of the T5 and T6
these artifacts significantly. In spite of this, in
vertebral body marrow, as well as the enhancing prespinal and anterior epidural inflammatory spinal
some instances these artifacts may be difficult
mass (arrows).
or impossible to overcome in any given case.
the focus and extent of the inflammatory dis-
ease. On occasion spinal inflammation can be
confused with other types of pathology such as
neoplasia. Although such specificity is a rela-
tive limitation of MRI, this is easily overcome
in most cases by imaging directed needle bi-
opsy or aspiration.
Literature Cited
Chang, K.H., Han, M.H., Choi, Y.W, et al. 1989.
Tuberculous arachnoiditis of the spine: Findings
on myelography, CT, and MR imaging. Am. J.
Neuroradiol. 10:1255-1262.
Jinkins, J.R., Gupta, R., Chang, K.H., and Ro-
driguez-Carbajal, J. 1995. MR imaging of cen-
tral nervous system tuberculosis. Radiol. Clin.
North Am. 33:771-786.
Modic, M.T., Feiglin, D.H., Piraino, D.W., et al.
1985. Vertebral osteomyelitis: Assessment using
MR. Radiology 157:157-166.
Post, M.J.D., Sheldon, J.J., Hensley, G.T., et al.
1986. Central nervous system disease in ac-
quired immunodeficiency syndrome: Prospec-
tive correlation using CT, MR imaging, and
pathologic studies. Radiology 158:141-148.
Post, M.J.D., Sze, G., Quencer, R.M., et al. 1990.
Gadolinium enhanced MR in spinal infection. J.
Comput. Assist. Tomogr. 15:721-729.
Provenzale, J.M. and Jinkins, J.R. 1997. Brain and
spine imaging findings in AIDS patients. Radiol.
Clin. North Am. 35:1127-1166.
Reddy, S., Leite, C.C., and Jinkins, J.R. 1995. Im-
aging of infectious disease of the spine. In Spine:
State of the Art Reviews, Vol. 9 (R.R. Lee, ed.)
pp. 119-140. Hanley & Belfus, Philadelphia.
Sharif, H.S., Clark, D.C., Aabed, M.Y., et al. 1990.
Granulomatous spinal infections: MR imaging.
Radiology 177:101-107.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Sklar, E.M., Post, M.J.D., and Lebwohl, N.H. 1993.
Imaging of infection of the lumbosacral spine.
Neuroimag. Clin. North Am. 3:577-590.
Thurnher, M.M., Jinkins, J.R., and Post, M.J.D.
1997. Diagnostic imaging of infections and neo-
plasms affecting the spine in patients with AIDS.
Neuroimag. Clin. North Am. 7:341-357.

space and portions of the vertebral bodies ad-
jacent to the disk are relatively hypointense on
T1-weighted imaging following gadolinium ad-
ministration, including thin or thick enhance-
ment of the periphery of the disk or diffuse
enhancement of the majority of the disk space.
A loss of disk height and paravertebral soft
tissue alterations are also typically observed.
With increasing involvement of the adjacent
vertebral bodies in the infectious process, the
clear distinction between intervertebral disk Anticipated Results
and vertebral body becomes increasingly diffi- The goal of studying the spine in cases of Contributed by J. Randy Jinkins and
cult on MR imaging. The associated increase clinically suspected inflammation involving David D. Stark
in water content of the disk and adjacent bone the spinal column and peri-intraspinal tissues Downstate Medical Center
is reflected in the signal changes on T1- and is to confirm this diagnosis, exclude other eti- State University of New York
T2-weighted imaging. These MR signal altera- ologies of disease if possible, and to determine Brooklyn, New York
tions usually precede lesions visible on conven-
tional radiographs of the spine. Malignant le-
sions can usually be differentiated from infec-
tion by the sparing of disk spaces in the former Spinal
Extradural Spine Inflammation

A8.4.9 A8.4.10
Current Protocols in Magnetic Resonance Imaging Supplement 1 Supplement 1 Current Protocols in Magnetic Resonance Imaging
Spinal Primary Neoplasia/Metastasis UNIT A8.5 Table A8.5.2 Primary Clinical Imaging Parameters for Sequence 1 (Pilot Scan)

Patient position Supine

Primary and metastatic neoplasia may involve the spine in patients in all age categories. Scan type Gradient echo
Sometimes presenting with acute but potentially reversible neurological deficit, such Imaging plane (orientation) Transverse
patients may benefit from emergency MRI, used to visualize the spinal compartment to Central slice or volume center Centered on:
assess the extent and degree of spinal cord or cauda equina compression frequently Cervical spine: thyroid cartilage
associated with such disease (Daffner et al., 1986; Beltran et al., 1987; Colman et al., Thoracic spine: nipple line
1988; Avrahami et al., 1989; Carmody et al., 1989; Algra et al., 1991; Coletti et al. 1991). Lumbar spine: iliac crests
Echo time (TE) As short as possible
CONVENTIONAL AND FAST SPIN ECHO BASIC Repeat time (TR) As short as possible
PROTOCOL Flip angle (FA) 15°
Although findings of destructive, mass-forming, and marrow-replacing neoplastic disease
Fields of view (FOVx, FOVy) Cervical: 240 mm, 240 mm
can often be used to diagnose advanced involvement of the spinal column, contrast agents
Thoracic: 320 mm, 320 mm
are almost always utilized in this subset of patients to determine clearly the focus and
Lumbosacral: 280 mm, 280 mm
extent of the enhancing, solid portions of epidural neoplasia as an indication of the degree
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm
and regional expression of the neural tissue compression prior to decompressive surgery
Thoracic: 1.25 mm, 1.25 mm
and/or chemo-radiation therapy (Jinkins et al., 1998).
Lumbosacral: 1.09 mm, 1.09 mm
The following sequences comprise the preferred protocol for high-field MR machines. In Number of data points collected (Nx, Ny) 256, 256
all cases intravenous (i.v.) contrast-enhanced protocols are mandatory to evaluate the Display matrix (Dx, Dy) 256, 256
integrity of the blood-nerve and blood-cord barrier in certain cases of suspected central Slice thickness (Δz) 5 mm
nervous system (CNS) infection, or in cases of epidural and/or spinal column/paraspinal Number of slices Variable
infection (e.g., spondyloclodiscitis and perispinal abscess/phlegmon; Sze et al., 1988; Slice gap Not applicable
Sugimura et al., 1991; Ragland et al., 1996). Number of acquisitions (Nacq) 1
Table A8.5.1 lists the hardware necessary to perform the procedure, along with appropri- Scan time ∼10 sec
ate parameters. The available gradient strength will depend on the scanner, and the echo
times (i.e., TE) given in other tables will be varied accordingly (i.e., the smaller the gradient of special emergency equipment during the scanning procedure, or necessitate any
strength, the longer the echo time for a particular scan). other precautions.
The entire protocol should take 45 to 50 min to complete. Generally standard screening forms are used for all patients scanned in a magnetic
NOTE: Be sure that technicians and nurses have immediate access to any emergency resonance system.
equipment that may be relevant to a given study, or that may be needed for a particular The presence of any ferromagnetic metals may be a health hazard to the patient when he
patient, such as crash carts or oxygen. or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
Materials (1996) for discussion of what implants may be safely scanned using magnetic resonance.
Normal saline (0.9% NaCl), sterile Patients may be accompanied into the magnet room by a friend or family member, who can
Extravascular contrast agent (e.g., Magnevist, Omniscan, or Prohance) sit in the room during the scan and comfort the patient as needed. This companion must
Set up patient and equipment be screened as well to ensure the absence of loose metal objects on the body or clothing.
1. Interview (screen) the patient to ensure that he or she has no contraindications such 2. If the procedure is a research protocol, have the patient sign any necessary consent
as cardiac pacemakers or other implants containing ferromagnetic materials. Also be form.
sure to find out if the patient has any health conditions that may require the presence
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
Table A8.5.1 Equipment Parameters for Spine Imaging in Cases of that might be found in clothing.
Suspected Primary or Metastatic Neoplasia
4. Inform the patient about what will occur during the procedure, what he or she will
Coil type Cervical, thoracic, lumbar: phase array experience while in the magnet, and how to behave, including the following.
surface coil (or other depending upon a. If earphones or headphones are used to protect the ears from the loud sounds
machine compatibility and availability)
produced by the gradients, the patient will be asked to wear these, but will be able
Gradient coil strength 25 mT/m (or whatever the system permits)
to communicate with you at any time during the imaging.
Flow compensation pulse Optional (any level, if available)
Peripheral gating Optional (thoracic spine only) b. The patient will be given a safety squeeze-bulb or similar equipment to request
Respiratory gating Optional (thoracic spine only) assistance at any time (demonstrate how this works).
Spinal Primary
Use of contrast agents Yes Neoplasia/
Extradural Spine Metastasis
Contributed by J. Randy Jinkins and David D. Stark A8.5.1 A8.5.2
Current Protocols in Magnetic Resonance Imaging (2001) A8.5.1-A8.5.10
Copyright © 2001 by John Wiley & Sons, Inc. Supplement 2 Supplement 2 Current Protocols in Magnetic Resonance Imaging
 Table A8.5.3 Primary Clinical Imaging Parameters for Sequence 2 (T1-Weighted Table A8.5.4 Primary Clinical Imaging Parameters for Sequence 3 (T2-Weighted
Image) Image, FSE)a

Patient position Supine Patient position Supine

Scan type Conventional spin echo Scan type Fast spin echo
Imaging plane (orientation) Sagittal Imaging plane (orientation) Sagittal
Central slice or volume center Centered on: Central slice or volume center Centered on area of interest (as in
Cervical: 3rd cervical vertebra sequence 2, Table A8.5.3)
Thoracic: 6th thoracic vertebra Echo time (TE) 100 msec
Lumbar: 3rd lumbar vertebra Echo train length (ETL) 8
Echo time (TE) 10 msec Repeat time (TR) 4000 msec
Repeat time (TR) 500 msec Flip angle (FA) 90°
Flip angle (FA) 90° Fields of view (FOVx, FOVy) As in sequence 2, Table A8.5.3
Fields of view (FOVx, FOVy) Cervical: 240 mm, 240 mm Resolution (Δx, Δy) Cervical: 0.47 mm, 0.47 mm
Thoracic: 320 mm, 320 mm Thoracic: 0.63 mm, 0.63 mm
Lumbosacral: 280 mm, 280 mm (may Lumbosacral: 0.55 mm, 0.55 mm
use rectangular field of view, e.g., half Number of data points collected (Nx, Ny) 512, 512
or three-quarter field, if available, or Display matrix (Dx, Dy) 512, 512
tailor to region of interest) Slice thickness (Δz) Cervical: 3 mm
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm Thoracic: 3 mm
Thoracic: 1.25 mm, 1.25 mm Lumbar: 4 mm
Lumbosacral: 1.09 mm, 1.09 mm Number of slices Varies with spinal level
Number of data points collected (Nx, Ny) 256, 256 Slice gap Cervical: 0.5 mm
Display matrix (Dx, Dy) 256, 256 Thoracic: 1 mm
Slice thickness (Δz) Cervical: 3 mm Lumbar: 1 mm
Thoracic: 3 mm Number of acquisitions (Nacq) 1
Lumbar: 4 mm Flow compensation Yes (if available)
Number of slices 10, or more, as needed to cover the Saturation pulses Yes; anterior cervical/thoracic/lumbar
region of interest slabs to saturate larynx/vessels/heart
Slice gap Cervical: 0.5 mm Fat suppression Yes
Thoracic: 1 mm Slice series Left to right or the reverse depending
Lumbar: 1 mm on preference
Number of acquisitions (Nacq) 2 Scan time ∼4 min
Flow compensation Yes (if available) aFSE, fast spin echo.
Saturation pulses Yes; anterior cervical/thoracic/lumbar
slab to saturate larynx/vessels
7. If needed, place a pillow or other support under the knees to make the patient more
Slice series Left to right or the reverse depending
on preference
comfortable.
Scan time ∼4 min 8. Use the centering light to position the patient (cervical spine: thryoid cartilage;
thoracic spine: nipple line; lumbar spine: iliac crests) and put him or her into the
center of the magnet.
c. For good results the patient should not talk, and should avoid or minimize
swallowing or other movement, during each scan—i.e., as long as the banging Once this step has been performed, so long as the patient does not move on the table, the
sounds continue. Between scans, talking and swallowing are allowed in most table itself can be moved and then replaced in the same position as before without
cases, but should be avoided when comparative positional studies are being jeopardizing the positioning of one scan relative to another.
performed; the patient will be informed when this is the case. 9. If the patient is unable to hold still, provide an appropriate sedative.
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
5. Have the patient mount onto the table. Either before or right after the patient lies Sequence 1: Rapid positioning pilot
down, set up any triggering devices or other monitoring equipment that is to be used. 10. To validate the patient’s position, run the system’s pilot (or scout) scan (sequence 1)
to ensure correct location of the neck in three dimensions, using the imaging sequence
6. Center the coil over the region where the key information is desired. given in Table A8.5.2 or similar parameters.
Make sure that the body is constrained to prevent motion, especially if high-resolution scans This sequence usually consists of three orthogonal planes to allow subsequent localization.
are to be run. Spinal Primary The images are often also used later to determine where to place the saturation pulses and
Neoplasia/ to set up total coverage of the volume of interest.
Extradural Spine Metastasis

A8.5.3 A8.5.4
Current Protocols in Magnetic Resonance Imaging Supplement 2 Supplement 2 Current Protocols in Magnetic Resonance Imaging
 Table A8.5.5 Primary Clinical Imaging Parameters for Sequence 4 (T1-Weighted Table A8.5.6 Primary Clinical Imaging Parameters for Sequence 5 (T2-Weighted
Image) Image, FSE)a

Patient position Supine Patient position Supine

Scan type Conventional spin echo Scan type Fast spin echo
Imaging plane (orientation) Transverse Imaging plane (orientation) Transverse
Central slice or volume center Centered on the area of interest (as in Central slice or volume center Centered on the region of interest (as
sequence 2, Table A8.5.3) in sequence 2, Table A8.5.3)
Echo time (TE) 10 msec Echo time (TE) 100 msec
Repeat time (TR) 500 msec Echo train length (ETL) 8
Flip angle (FA) 90° Repeat time (TR) 4000 msec
Fields of view (FOVx, FOVy) As in sequence 2, Table A8.5.3 Flip angle (FA) 90°
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm Fields of view (FOVx, FOVy) As in sequence 2, Table A8.5.3
Thoracic: 1.25 mm, 1.25 mm Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm
Lumbosacral: 1.09 mm, 1.09 mm Thoracic: 1.25 mm, 1.25 mm
Number of data points collected (Nx, Ny) 256, 256 Lumbosacral: 1.09 mm, 1.09 mm
Display matrix (Dx, Dy) 256, 256 Number of data points collected (Nx, Ny) 256, 256
Slice thickness (Δz) Cervical: 3 mm Display matrix (Dx, Dy) 256, 256
Thoracic: 3–8 mm Slice thickness (Δz) Cervical: 3 mm
Lumbar: 4 mm Thoracic: 3–8 mm
Number of slices Varies with spinal level Lumbar: 4 mm
Slice gap Cervical: 1 mm Number of slices Varies with spinal level
Thoracic: 1–2 mm Slice gap Cervical: 1 mm
Lumbar: 1 mm Thoracic: 1–2 mm
Number of acquisitions (Nacq) 2 Lumbar: 1 mm
Slice locations See text (Basic Protocol, step 16) Number of acquisitions (Nacq) 2
Saturation pulses No Slice locations See text (Basic Protocol, step 16)
Scan time ∼4 min Saturation pulses No
Scan time ∼4 min
aFSE, fast spin echo.
Sequence 2: Sagittal T1-weighted conventional spin echo
11. Set the imaging parameters as shown in Table A8.5.3.
12. Use the pilot image to locate the spine in three dimensions to ensure coverage of the 18. Run the sequence according to Table A8.5.5.
region of interest (e.g., cervical, thoracic, lumbosacral spine).
Sequence 5: Transverse T2-weighted fast spin echo
13. Let the patient know you are ready and begin the scan. 19. Using the midline T1-weighted image acquired in sequence 2, repeat the setup as in
Table A8.5.6.
Sequence 3: Sagittal T2-weighted fast spin echo, fat suppressed
14. Review the pilot scans and ensure that the saturation pulse is correctly placed anterior 20. Run sequence 5 according to Table A8.5.6.
to above the slab of interest.
Sequences 6 and 7: Sagittal and transverse i.v. contrast enhanced T1-weighted spin
15. Run sequence 3 according to Table A8.5.4. echo
21. Remove patient from the magnet. Use the same equipment and perform equipment
Sequence 4: Transverse T1-weighted conventional spin echo and patient setup as in steps 5 to 9.
16. Using the midline sagittal T1-weighted image acquired in sequence 2, set the
transverse acquisition parameters as follows: 22. Establish an i.v. line from which the contrast agent can be injected, and attach this
line securely to the patient so that movement into or out of the magnet will not pull
a. Cervical spine: stacked images from C1 through C7-T1. at the patient’s arm. Put the patient back into the MR machine. The contrast agent
b. Thoracic spine: stacked images through levels of interest. will be used to assess the extent of the neoplasm.
c. Lumbosacral spine: 5 slices each, angled to the plane of the intervertebral disc at It is preferable to insert the line prior to imaging and to leave the patient in the magnet,
L3-4, L4-5, and L5-S1; one slice each, angled to the intervertebral disc at L1-2 with no intervening motion, between the scans run before contrast agent injection and those
and L2-3. run after injection.
17. Supplement additional slices according to visible disease present or to clinical query.
Spinal Primary Run pilot scan
Neoplasia/
Extradural Spine Metastasis 23. Run a rapid three-plane positioning pilot scan (see sequence 1).

A8.5.5 A8.5.6
Current Protocols in Magnetic Resonance Imaging Supplement 2 Supplement 2 Current Protocols in Magnetic Resonance Imaging
Scan sequences COMMENTARY
24. Leaving the patient in the magnet, inject the contrast agent, flush the i.v. line with 10 deflected laterally by the midline septum com-
Background Information
ml saline, and then immediately run sagittal (sequence 6) and transverse (sequence Primary bony tumors affecting the spinal plex, a fibrous structure attached to the poste-
7) T1-weighted image sequences (see sequences 2 and 4). column (e.g., osteosarcoma, chondrosarcoma, rior longitudinal ligament and the posterior
A dose of 0.1 mmol/kg of contrast agent is usually given. Ewing’s sarcoma, myeloma, chordoma, giant surface of the vertebral body. It has been shown
cell tumor, aneurysmal bone cyst, osteoblas- repeatedly that radionuclide bone scanning is
Fat suppression techniques may be very useful in order to suppress osseous vertebral toma, etc.) are not common, but may present less sensitive than MR for the detection of bony
marrow fat and perispinal soft tissue fat, thereby clearly distinguishing enhancement from with epidural extension and clinically relevant metastatic disease of the spine. On MR, the
normal fatty tissue. compression of the spinal cord, spinal nerves, bony metastases are seen to replace hyperin-
and/or cauda equina. The primary bony com- tense marrow on T1-weighted imaging with
CORONAL, CONTRAST ENHANCED ACQUISITION ALTERNATE ponent may produce a pathologic vertebral hypointense tumor/peritumoral edema; mar-
PROTOCOL fracture. These changes will be evident on MRI row tumor/edema becomes hyperintense on T2-
In some instances (e.g., neurofibromatosis with multiple, bilateral neoplasms extending without the use of an intravenous contrast weighted acquisitions, and can be especially
through the neural foramina), a coronal acquisition may be helpful to analyze the agent. However, intravenous gadolinium (Gd) well visualized on fat-suppressed studies (Me-
perispinal tissues for tumor extension either inward to or outward from the central spinal can be very helpful in delineating the extent and hta et al., 1995). After i.v. contrast administra-
canal. degree of neural tissue compression, which aids tion, the degree of abnormal contrast enhance-
in preoperative decompression surgical plan- ment of the involved vertebral marrow varies
Sequence 8: Contrast enhanced coronal T1-weighted image ning. from patient to patient and from location to
Run the sequence according to sequence 8, Table A8.5.7. As stated above, fat suppression Malignant perispinal neoplasms (e.g., soft location in the same patient. Epidural tumor on
may be useful. tissue sarcomas, renal cell carcinoma) and me- MR tends to be generally isointense to muscle
tastatic nodal diseases (including lymphoma) tissue on T1-weighted images and isointense or
in the perispinal regions may secondarily in- hyperintense on T2-weighted acquisitions. En-
Table A8.5.7 Primary Clinical Imaging Parameters for Sequence 8 (T1-Weighted
Image) volve the spinal column and spinal canal. Bony hancement of the epidural tumor component
destruction and neural compromise may pro- after i.v. gadolinium administration is typically
Patient position Supine duce pain and neurologic deficits. MR clearly moderate to intense and usually homogeneous
Scan type Conventional spin echo demonstrates any extension of the neoplasm (Fig. A8.5.1); this enhancement may also vary
into the spinal canal and the resulting effects on in degree and pattern from location to location.
Imaging plane (orientation) Coronal
the spinal cord and/or cauda equina. The spinal cord and nerve roots are not infre-
Central slice or volume center Centered on the area of interest (as in
Following the lung and liver, the skeletal quently seen to be compressed by the metastatic
sequence 2, Table A8.5.3)
system is the most common site of metastatic epidural tumor mass.
Echo time (TE) 10 msec
neoplastic disease, and in the skeletal system,
Repeat time (TR) 500 msec the spine is the most common area affected. Critical Parameters and
Flip angle (FA) 90° Hematogenously disseminated bony spinal Troubleshooting
Fields of view (FOVx, FOVy) As in sequence 2, Table A8.5.3 neoplastic metastases are reported in up to 10% Cerebrospinal fluid (CSF) flow, cardiac, la-
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm of patients with malignancies. The frequency ryngeal, body wall and other sources of motion
Thoracic: 1.25 mm, 1.25 mm of neoplastic metastases to the spine is greatest can produce artifacts that can on occasion sig-
Lumbosacral: 1.09 mm, 1.09 mm in the thoracic region, and least frequent in the nificantly degrade the images. Proper spatial
Number of data points collected (Nx, Ny) 256, 256 cervical area. The involvement of the spine is (e.g., prevertebral) saturation pulses and some-
Display matrix (Dx, Dy) 256, 256 most often multisegmental. The vertebral body times flow compensation pulses and/or car-
Slice thickness (Δz) 3–4 mm tends to be involved first, with secondary ex- diac/respiratory gating can reduce these arti-
Number of slices Varies with spinal level tension into the pedicles. However, hemato- facts significantly. In many instances these ar-
Slice gap 1–2 mm genously disseminated metastases may occur tifacts may be difficult or impossible to easily
directly to any spinal structure, including the overcome from patient to patient.
Number of acquisitions (Nacq) 2
posterior bony neural arch. At least 5% of Another problem that may surface during
Slice locations From front to back of spinal column
patients eventually will reveal subsequent the MRI evaluation of patients with primary or
Saturation pulses No epidural extension of the metastatic disease. metastatic disease of the spinal column is pain
Fat suppression Yes: fat saturation or STIR (short tau Neoplasms commonly encountered involving severe enough to prevent patient cooperation
inversion recovery) the vertebrae and adjacent epidural space of the for the duration of a sometimes holospine (e.g.,
Scan time ∼4 min spine include carcinomas (e.g., lung, breast, or cervical, thoracic, lumbar spine) evaluation.
aFSE, fast spin echo. prostate carcinoma), lymphoma, and myeloma. For this reason, patients should be on call from
Epidural tumor may compress the intraspinal the hospital floor with an immediately preced-
neural structures (e.g., spinal cord, nerve roots) ing dose of pain medication, and the physician
and cause consonant neurological deficits. Ex- in charge should be reachable by phone or in
tension of tumor directly from the vertebral person in order to supplement the pain medica-
Spinal Primary body into the anterior epidural space of the tion as needed in the MRI suite. In addition, all
Neoplasia/ spinal cord may be divided in the midline and
Extradural Spine Metastasis

A8.5.7 A8.5.8
Current Protocols in Magnetic Resonance Imaging Supplement 2 Supplement 2 Current Protocols in Magnetic Resonance Imaging

A B
Figure A8.5.1 Spinal lymphoma with epidural extension. (A) Sagittal T1-weighted (TR = 500 msec, TE
= 10 msec) image shows hypointensity within the marrow of multiple thoracic vertebral bodies. Note that
the spinal cord is displaced posteriorly (arrows). (B) Intravenous gadolinium enhanced T1-weighted (TR
= 500 msec, TE = 10 msec) image shows poor enhancement of the abnormal thoracic bodies and a
multilevel enhancing epidural lymphomatous mass (arrowheads).
efforts should be undertaken by the technical
staff to insure a speedy examination.
In some critical settings, simply T1-
weighted images in two planes (e.g., sagittal
and transverse) before and after intravenous
gadolinium administration may be all that is
required to visualize clearly the extent of
epidural tumor extension. Deleting the T2-
weighted acquisitions will speed up the overall
examination time in cases when the entire spine
must be examined in an emergency situation.
This decision to curtail the examination must
be made at the discretion of the imaging phy-
sician, taking each case into consideration in-
dividually.
Anticipated Results
The goal of studying the spine in cases of
known or suspected neoplastic disease is to
reveal the area or areas of involvement, the
Current Protocols in Magnetic Resonance Imaging
presence of secondary pathologic changes
(e.g., pathologic fractures), and the degree and
extent of epidural extension of tumor with neu-
ral tissue (e.g., spinal cord, cauda equina) com-
pression. This information will allow the refer-
ring physician to accurately plan the type of
therapy as well as the specific approach.
Literature Cited
Algra, P.R., Bloem, J.L., Tissing, H., Falke, T.H.,
Arndt, J.W., and Verboom, L.J. 1991. Detection
of vertebral metastases: Comparison between
MR imaging and bone scintigraphy. Radiog-
raphics 11:219-232.
Avrahami, E., Tadmor, R., Dally, O., and Hadar, H.
1989. Early MR demonstration of spinal metas-
tases in patients with normal radiographs and CT
and radionuclide bone scans. J. Comput. Assist.
Tomogr. 13(4):598-602.
Beltran, J., Noto, A.M., Chakeres, D.W., and Chris-
Spinal Primary
toforidis, A.J. 1987. Tumors of the osseous
Extradural Spine
A8.5.9
Supplement 2
spine: Staging with MR imaging versus CT. Ra-
diology 162:565-569.
Carmody, R.F., Yang, P.J., Seeley, G.W., Seeger, J.F.,
Unger, E.C., and Johnson, J.E. 1989. Spinal cord
compression due to metastatic disease: Diagno-
sis with MR imaging versus myelography. Radi-
ology 173:225-229.
Coletti, P.M., Dang, H.T., Deseran, M.W., Kerr,
R.M., Boswell, W.D., and Ralls, P.W. 1991. Spi-
nal MR imaging in suspected metastases: Corre-
lation with skeletal scintigraphy. Magn. Res.
Imag. 9:335-349.
Colman, L.K., Porter, B.A., Redmond, J., III, Olson,
D.O., Stimac, G.K., Dunning, D.M., and Friedl,
K.E. 1988. Early diagnosis of spinal metastases
by CT and MR studies. J. Comput. Assist. To-
mogr. 12:423-426.
Daffner, R.H., Lupetin, A.R., Dash, N., Deeb, Z.L.,
Sefczek, R.J., and Schapiro, R.L. 1986. MRI in
the detection of malignant infiltration of bone
marrow. Am. J. Roentgenol. 146:353-358.
Jinkins, J.R. and Leite da Costa, C. 1998. Differen-
tial diagnosis of primary or secondary paraspinal
abnormalities originating in the spine, perispi-
nous tissues, or remote sources. In Neurodiag-
nostic Imaging Pattern Analysis and Differential
Diagnosis (J. R. Jinkins and C. Da Costa Leite,
ed.) pp. 773-779. Lippincott-Raven, Philadel-
phia.
Neoplasia/
Metastasis
A8.5.10
Supplement 2
Mehta, R.C., Marks, M.P., Hinks, R.S., Glover,
G.H., and Enzmann, D.R. 1995. MR evaluation
of vertebral metastases: T1-weighted, short-in-
version-time inversion recovery, fast spin-echo,
and inversion-recovery fast spin-echo se-
quences. Am. J. Neuroradiol. 16:281-288.
Ragland, R.L., Knorr, J.R., Kamath, S.V., Landis,
E.S., Tenreiro-Picon, O.R., and Weyreuther, M.
1996. Magnetic resonance patterns of epidural
impression from spinal metastases: review of
200 cases. Int. J. Neuroradiol. 1:69-72.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Sugimura, K., Kajitani, A., Okizuka, H., Sugihara,
M., Mizutani, M., and Ishida, T. 1991. Assessing
response to therapy of spinal metastases with
gadolinium-enhanced MR imaging. J. Magn.
Reson. Imag. 1:481-484.
Sze, G., Krol, G., Zimmerman, R.D., and Deck,
M.D. 1988. Malignant extradural spinal tumors:
MR im ag in g with Gd-DTPA. Radiology
167:217-223.
Contributed by J. Randy Jinkins and
David D. Stark
Downstate Medical Center
State University of New York
Brooklyn, New York
Current Protocols in Magnetic Resonance Imaging
Spinal Trauma
While at the present time the cranium is not being studied routinely by MR in cases of
trauma, the spinal column presents an ideal situation for the study of trauma from the
acute through the chronic phases. The reason for this is that computed tomography of the
cranium often adequately delineates the traumatic pathology affecting the intracranial
neural tissue. In the spinal column however, although computed tomography is excellent
for the evaluation of bony trauma, the spinal neural tissue is usually not adequately
visualized. On the other hand, MR demonstrates traumatic change quite well within the
spinal cord and epidural tissues (Mayer and Kulkarni, 1987; Takahashi et al., 1987; Mathis
et al., 1988; Mirvis et al., 1988; Pan et al., 1988; Emery et al., 1989; Flanders et al., 1990;
Mendolsohn et al., 1990; Kerslake et al., 1991; Curati et al., 1992; Silberstein et al., 1992;
Bashir et al., 1993; Davis et al., 1993; Kliewer et al., 1993; Post et al., 1994; Leite et al.,
1997; Jinkins, 1998a,b; Jinkins et al. 1998).
CONVENTIONAL FAST SPIN ECHO AND GRADIENT-RECALLED ECHO
ACQUISITIONS
Although computed tomography (CT) and conventional radiography can be used to show
gross traumatic alterations affecting the spinal column itself and the perispinal soft tissues,
the contents of the spinal canal are not accurately delineated using these techniques.
Magnetic resonance (MR), on the other hand, excellently images the spinal cord and
subarachnoid space, and often the individual spinal nerve roots. In addition, MR is also
able, in many respects, to accurately and sensitively evaluate the elements of the spinal
column and perispinal soft tissues, and coupled with an intravenously administered
contrast agent (e.g., gadolinium, Gd), it can assess the integrity of the blood–central
nervous system (CNS) barrier (i.e., the combined blood-cord, the blood-nerve, and the
relative blood-meningeal barriers; Jinkins, 1993a,b; 1999).
The following sequences comprise the preferred protocol for high-field MR machines. In
most or all cases of acute trauma, gradient-recalled echo acquisitions may be desirable in
order to search for acute blood products (i.e., deoxyhemoglobin). In addition, occasionally
intravenous (i.v.) contrast enhanced alternate protocols may be useful to evaluate the
integrity of the blood-nerve barrier in certain cases of suspected compressive radicu-
lopathy.
Table A8.6.1 lists the hardware necessary to perform the procedure, along with appropri-
ate parameters. The available gradient strength will depend on the scanner, and the echo
times (e.g., TE) given in other tables will be varied accordingly (the smaller the gradient
strength, the longer the echo time for a particular scan).
This entire protocol should take 45 to 50 min to complete.
Table A8.6.1 Equipment Parameters for Spine Imaging in Cases of
Spinal Trauma
Coil type Cervical, thoracic, lumbar: phase array
surface coil (or other depending upon
machine compatibility and availability)
Peripheral gating Thoracic spine only (optional)
Flow compensation Any level (optional, if available)
Respiratory gating Thoracic spine only (optional)
Use of contrast agent Yes
Contributed by J. Randy Jinkins and David D. Stark
Current Protocols in Magnetic Resonance Imaging (2002) A8.6.1-A8.6.12
Copyright © 2002 by John Wiley & Sons, Inc.
UNIT A8.6 NOTE: Be sure that technicians and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.
Set up patient and equipment
1. Interview (screen) the patient to ensure that he or she has no contraindications such
as cardiac pacemakers or other implants containing ferromagnetic materials. Also be
sure to find out if the patient has any health conditions that may require the presence
of special emergency equipment during the scanning procedure, or necessitate any
other precautions.
Generally standard screening forms are used for all patients scanned in a magnetic
resonance system.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(1996) for discussion of what implants may be safely scanned using magnetic resonance.
BASIC
PROTOCOL Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
2. If the procedure is a research protocol, have the patient sign any necessary consent
form.
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
4. Inform the patient about what will occur during the procedure, what he or she will
experience while in the magnet, and how to behave, including the following.
a. If earphones or headphones are used to protect the ears from the loud sounds
produced by the gradients, the patient will be asked to wear these, but will be able
to communicate with you at any time during the imaging.
b. The patient will be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this works).
c. For good results the patient should not talk, and should avoid or minimize
swallowing or other movement, during each scan—i.e., as long as the banging
sounds continue. Between scans, talking and swallowing are allowed in most
cases, but should be avoided when comparative positional studies are being
performed; the patient will be informed when this is the case.
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
5. Have the patient mount onto the table. Either before or right after the patient lies
down, set up any triggering devices or other monitoring equipment that is to be used.
6. Center the coil over the region where the key information is desired.
Make sure that the body is constrained to prevent motion, especially if high-resolution scans
are to be run.
7. If needed, place a pillow or other support under the knees to make the patient more
comfortable.
8. Use the centering light to position the patient (cervical spine: thyroid cartilage;
thoracic spine: nipple line; lumbar spine: iliac crests) and put him or her into the
Extradural Spine
center of the magnet.
Spinal Trauma
A8.6.1 A8.6.2
Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging
 Table A8.6.2 Primary Clinical Imaging Parameters for Sequence 1 (Pilot Scan) Table A8.6.3 Primary Clinical Imaging Parameters for Sequence 2 (T1-Weighted
Image)
Patient position Supine
Scan type Gradient echo Patient position Supine
Imaging plane (orientation) Transverse Scan type Conventional spin echo
Central slice or volume center Centered on: Imaging plane (orientation) Sagittal
Cervical spine: thyroid cartilage Central slice or volume center Centered on:
Thoracic spine: nipple line Cervical: 3rd cervical vertebra
Lumbar spine: iliac crests Thoracic: 6th thoracic vertebra
Echo time (TE) As short as possible Lumbar: 3rd lumbar vertebra
Repeat time (TR) As short as possible Echo time (TE) 10 msec
Flip angle (FA) 15° Repeat time (TR) 500 msec
Fields of view (FOVx, FOVy) Cervical: 240 mm, 240 mm Flip angle (FA) 90°
Thoracic: 320 mm, 320 mm Fields of view (FOVx, FOVy) Cervical: 240 mm, 240 mm
Lumbosacral: 280 mm, 280 mm Thoracic: 320 mm, 320 mm
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm Lumbosacral: 280 mm, 280 mm (may
Thoracic: 1.25 mm, 1.25 mm use rectangular field of view, e.g., half
Lumbosacral: 1.09 mm, 1.09 mm or three-quarter field, if available, or
Number of data points collected (Nx, Ny) 256, 256 tailor to region of interest)
Display matrix (Dx, Dy) 256, 256 Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm
Thoracic: 1.25 mm, 1.25 mm
Slice thickness (Δz) 5 mm
Lumbosacral: 1.09 mm, 1.09 mm
Slice gap Not applicable
Number of data points collected (Nx, Ny) 256, 256
Number of acquisitions (Nacq) 1
Display matrix (Dx, Dy) 256, 256
Scan time ∼10 sec
Slice thickness (Δz) Cervical: 3 mm
Thoracic: 3 mm
Lumbar: 4 mm
Once this step has been performed, so long as the patient does not move on the table, the Number of slices 10, or more, as needed to cover the
table itself can be moved and then replaced in the same position as before without region of interest
jeopardizing the positioning of one scan relative to another. Slice gap Cervical: 0.5 mm
9. If the patient is unable to hold still, provide an appropriate sedative. Thoracic: 1 mm
Lumbar: 1 mm
Sequence 1: Rapid positioning pilot Number of acquisitions (Nacq) 2
10. To validate the patient’s position, run the system’s pilot (or scout) scan (sequence 1) Flow compensation Yes (if available)
to ensure correct location of the neck in three dimensions, using the imaging sequence Saturation pulses Yes; anterior cervical/thoracic/lumbar
slab to saturate larynx/vessels
given in Table A8.6.2 or similar parameters.
Slice series Left to right or the reverse depending
This sequence usually consists of three orthogonal planes to allow subsequent localization. on preference
The images are often also used later to determine where to place the saturation pulses and Scan time ∼4 min
to set up total coverage of the volume of interest.

Sequence 2: Sagittal T1-weighted conventional spin echo 15. Run sequence 3 according to Table A8.6.4.
11. Set the imaging parameters as shown in Table A8.6.3.
Sequence 4: Transverse T1-weighted conventional spin echo
12. Use the pilot image to locate the spine in three dimensions to ensure coverage of the 16. Using the midline sagittal T1-weighted image acquired in sequence 2, set the trans-
region of interest (e.g., cervical, thoracic, lumbosacral spine). verse acquisition parameters as follows:
13. Let the patient know you are ready and begin the scan. a. Cervical spine: stacked images from C1 through C7-T1.
b. Thoracic spine: stacked images through levels of interest.
Sequence 3: Sagittal T2-weighted fast spin echo, fat suppressed
Fat suppression is necessary to determine vertebral marrow edema in cases of frank or c. Lumbosacral spine: 5 slices each, angled to the plane of the intervertebral disc at
occult bony fracture and in visualizing spinal ligament injury/rupture (Emery et al., 1989; L3-4, L4-5, and L5-S1; one slice each, angled to the intervertebral disc at L1-2
Kliewer et al., 1993). and L2-3.
17. Supplement additional slices according to visible disease present or to clinical query.
14. Review the pilot scans and ensure that the saturation pulse is correctly placed anterior
to above the slab of interest. 18. Run the sequence according to Table A8.6.5.
Extradural Spine Spinal Trauma

A8.6.3 A8.6.4
Current Protocols in Magnetic Resonance Imaging Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging
 Table A8.6.4 Primary Clinical Imaging Parameters for Sequence 3 (T2-Weighted Table A8.6.5 Primary Clinical Imaging Parameters for Sequence 4 (T1-Weighted
Image, Fast Spin Echo) Image)

Patient position Supine Patient position Supine

Scan type Fast spin echo Scan type Conventional spin echo
Imaging plane (orientation) Sagittal Imaging plane (orientation) Transverse
Central slice or volume center Centered on area of interest (as in Central slice or volume center Centered on the area of interest (as in
sequence 2, Table A8.6.3) sequence 2, Table A8.6.3)
Echo time (TE) 10 msec
Echo time (TE) 100 msec
Repeat time (TR) 500 msec
Echo train length (ETL) 8
Flip angle (FA) 90°
Repeat time (TR) 4000 msec Fields of view (FOVx, FOVy) As in sequence 2, Table A8.6.3
Flip angle (FA) 90° Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm
Fields of view (FOVx, FOVy) As in Sequence 2, Table A8.6.3 Thoracic: 1.25 mm, 1.25 mm
Resolution (Δx, Δy) Cervical: 0.47 mm, 0.47 mm Lumbosacral: 1.09 mm, 1.09 mm
Thoracic: 0.63 mm, 0.63 mm Number of data points collected (Nx, Ny) 256, 256
Lumbosacral: 0.55 mm, 0.55 mm Display matrix (Dx, Dy) 256, 256
Number of data points collected (Nx, Ny) 512, 512 Slice thickness (Δz) Cervical: 3 mm
Display matrix (Dx, Dy) 512, 512 Thoracic: 3–8 mm
Slice thickness (Δz) Cervical: 3 mm Lumbar: 4 mm
Thoracic: 3 mm Number of slices Varies with spinal level
Lumbar: 4 mm Slice gap Cervical: 1 mm
Number of slices Varies with spinal level Thoracic: 1–2 mm
Slice gap Cervical: 0.5 mm Lumbar: 1 mm
Thoracic: 1 mm Number of acquisitions (Nacq) 2
Lumbar: 1 mm Slice locations See text (Basic Protocol, step 16)
Number of acquisitions (Nacq) 1 Saturation pulses No
Scan time ∼4 min
Flow compensation Yes (if available)
Saturation pulses Yes; anterior cervical/thoracic/lumbar
slabs to saturate larynx/vessels/heart Table A8.6.6 Primary Clinical Imaging Parameters for Sequence 5 (T2-Weighted
Fat suppression Yes; chemical saturation or STIR Image, Fast Spin Echo)
(short tau inversion recovery)
Slice series Left to right or the reverse depending Patient position Supine
on preference Scan type Fast spin echo
Scan time ∼4 min Imaging plane (orientation) Transverse
Central slice or volume center Centered on the region of interest (as
in sequence 2, Table A8.6.3)
Sequence 5: Transverse T2-weighted fast spin echo Echo time (TE) 100 msec
19. Using the midline T1-weighted image acquired in sequence 2, repeat the setup as in Echo train length (ETL) 8
Table A8.6.6. Repeat time (TR) 4000 msec
Flip angle (FA) 90°
20. Run sequence 5 according to Table A8.6.6. Fields of view (FOVx, FOVy) As in sequence 2, Table A8.6.3
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm
Sequence 6: Sagittal gradient-recalled echo Thoracic: 1.25 mm, 1.25 mm
Gradient-recalled echo acquisitions may be used in the sagittal and/or transverse planes Lumbosacral: 1.09 mm, 1.09 mm
to clearly distinguish between various types of hemorrhage and to clearly identify acute Number of data points collected (Nx, Ny) 256, 256
blood products (e.g., deoxhemoglobin). Fast spin echo sequences will overlook this type Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) Cervical: 3 mm
of hemorrhage.
Thoracic: 3–8 mm
Lumbar: 4 mm
21. Run the sequence according to Table A8.6.7.
Number of slices Varies with spinal level
Slice gap Cervical: 1 mm
Sequence 7: Transverse gradient-recalled echo Thoracic: 1–2 mm
22. Run the sequence according to Table A8.6.8. Lumbar: 1 mm
Number of acquisitions (Nacq) 2
Slice locations See text (Basic Protocol, step 16)
Saturation pulses No
Extradural Spine Spinal Trauma Scan time ∼4 min
A8.6.5 A8.6.6
Current Protocols in Magnetic Resonance Imaging Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging
 Table A8.6.7 Primary Clinical Imaging Parameters for Sequence 6 (T2* ALTERNATE CONTRAST ENHANCED ACQUISITIONS
Gradient-Recalled Echo) PROTOCOL
For some situations, such as an MRI scan that has findings that do not match the clinical
Patient position Supine findings, an i.v. paramagnetic contrast material may be indicated.
Scan type 2-D gradient echo
Materials
Imaging plane (orientation) Sagittal
Central slice or volume center Centered on the region of interest (as Normal saline (0.9% NaCl), sterile
in sequence 2, Table A8.6.3) Extravascular contrast agent (e.g., Magnevist, Omniscan, or Prohance)
Echo time (TE) 15 msec
Repeat time (TR) 500 msec Set up patient and equipment
Flip angle (FA) 10° to 20° 1. Use the same equipment and perform equipment and patient setup as in the Basic
Fields of view (FOVx, FOVy) As in sequence 2 Protocol (steps 1 to 4).
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm 2. Establish an intravenous line from which the contrast agent can be injected, and attach
Thoracic: 1.25 mm, 1.25 mm this line securely to the patient so that movement into or out of the magnet will not
Lumbosacral: 1.09 mm, 1.09 mm
pull at the patient’s arm. Resume steps 5 to 9 in the Basic Protocol.
Number of data points collected (Nx, Ny) 256, 256
Display matrix (Dx, Dy) 256, 256 It is preferable to insert the line prior to imaging and to leave the patient in the magnet,
Slice thickness (Δz) 3 mm with no intervening motion, between the scans run before contrast agent injection and those
Number of slices Varies with spinal level run after injection.
Slice gap ≤1 mm
Number of acquisitions (Nacq) 3
Scan pilot
Flow compensation Yes (if available)
3. Run a rapid three-plane positioning pilot scan (see Basic Protocol, sequence 1).
Saturation pulses Yes
Sequences 8 and 9: Contrast enhanced scans
Scan time ∼6 min
4. Leaving the patient in the magnet, inject the contrast agent, flush the i.v. line with 10
ml saline, and then immediately run sagittal (sequence 8) and transverse (sequence
Table A8.6.8 Primary Clinical Imaging Parameters for Sequence 7 (T2* 9) T1-weighted image sequences (see Basic Protocol, sequences 2 and 4; sequences
Gradient-Recalled Echo) 8 and 9 are the same as sequences 2 and 4, respectively, but with the contrast agent
injected into the patient).
Patient position Supine
Scan type 2-D gradient echo A dose of 0.1 mmol/kg of contrast agent is usually given.
Imaging plane (orientation) Transverse
Central slice or volume center Centered on the region of interest (as COMMENTARY
in sequence 2, Table A8.6.3)
Background Information 1998a,b). In medical imaging, there is a con-
Echo time (TE) 15 msec
The special structures that are involved in tinuing dilemma between deciding what can be
Repeat time (TR) 500 msec the process of trauma include the spinal col- done and what should be done in any given
Flip angle (FA) 10° to 20° umn, related ligaments and attached muscles, patient. In the case of trauma to the spine, the
Fields of view (FOVx, FOVy) As in sequence 2, Table A8.6.3 the dura, arachnoid, and pia, the spinal cord, benefits of MR imaging include the advantage
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm nerve roots or spinal nerves, and the regional of superior spatial and contrast resolution in
Thoracic: 1.25 mm, 1.25 mm blood vessels. It is important to note that what multiple planes, the possibility of specific tis-
Lumbosacral: 1.09 mm, 1.09 mm is observed on the imaging will depend upon sue characterization (e.g., hemorrhage), and the
Number of data points collected (Nx, Ny) 256, 256 not only the structures involved, but also the relative ability to gain potentially prognostic
Display matrix (Dx, Dy) 256, 256 severity of the injury and the timing of the information noninvasively (Fig. A8.6.2). The
Slice thickness (Δz) 3-4 mm imaging study in relation to the traumatic inci- drawbacks of MR imaging include the need to
Number of slices Varies with spinal level dent. As is true of all traumas to tissues, there work with unstable patients, perhaps in traction
Slice gap ≤1 mm will be a temporal evolution of the pathophysi- and on life support devices, the difficulty of
Number of acquisitions (Nacq) 1 ology, and this will in part be reflected in the performing examinations on a 24-hr emergency
Slice locations See text (Basic Protocol, step 16) imaging findings. basis, the overall expense of the examination,
Saturation pulses Yes Direct nonpenetrating and penetrating and the inability of MR scan to detect nonde-
trauma are two of the most common forms of forming spinal fractures. In addition, it remains
Scan time ∼2 min
trauma encountered in clinical practice. Many true that many patients have a poor prognosis
of the injuries to the spinal column, neural in spite of undergoing either medical imaging
tissue, and vascular tissue involved are excel- procedures or extended treatment regimens. In
lently evaluated both in the acute and chronic the future, progressive new therapies may
phases by MR imaging (Fig. A8.6.1; Jinkins, change this situation. Ideally, MR imaging
Extradural Spine Spinal Trauma

A8.6.7 A8.6.8
Current Protocols in Magnetic Resonance Imaging Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging
 


*

Figure A8.6.1 Spinal cord contusion. Sagittal T2-weighted (TR = 4000 msec, TE = 100 msec) MR
shows multiple areas of spinal column marrow hyperintensity indicating bony injury and hyperin-
tensity within the cervical spinal cord (asterisk) indicating contusion.

* *

* *
Figure A8.6.3 Hemorrhagic spinal cord contusion. (A) Sagittal T1-weighted (TR = 500 msec, TE
= 10 msec) image shows cervical spinal cord swelling (asterisk). (B) Transverse T2* gradient recalled
echo (TR = 500 msec, TE = 15 msec, flip angle = 15°) image shows an area of focal hypointensity
(arrow) indicating acute hemorrhage (deoxyhemoglobin).

Figure A8.6.2 Traumatic epidural intraspinal hematoma. Sagittal T1-weighted (TR = 500 msec,
TE = 10 msec) image shows hyperintense epidural hematoma formation throughout the thoraco-
lumbar area anteriorly and posteriorly (asterisks). Extradural Spine Spinal Trauma

A8.6.9 A8.6.10
Current Protocols in Magnetic Resonance Imaging Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging

Figure A8.6.4 Traumatic spinal ligamentous injury. Sagittal T2-weighted (TR = 4000 msec, TE =
100 msec) fat suppressed image shows a focal segmental dehiscence of the ligamenta flava (arrow)
indicating traumatic rupture of these ligaments.
should be performed in either the acute or
chronic stage of trauma in order to identify and
direct treatment in patients who may benefit
from intervention aimed at halting or reversing
the pathological process, thereby preserving or
regaining function.
Clinical Parameters and
Troubleshooting
The gradient-recalled echo sequences are
important to include as a part of an acute spine
injury evaluation. The reason for this is that
acute hemorrhage (e.g., deoxyhemoglobin)
will be missed on routine fast spin echo scans
(Fig. A8.6.3; Jinkins, 1998a). Similarly, spinal
ligament tears will be overlooked on routine
fast spin echo T2-weighted studies, if simulta-
neous fat suppression is not used (Fig. A8.6.4;
Jinkins, 1998b). It is important not to overlook
these factors in the MRI analysis of spinal
trauma.
If the patient is in an emergency situation,
only run sequence 1, 2, 3, and 6.
Current Protocols in Magnetic Resonance Imaging
Anticipated Results
In general, the aims of imaging in the patient
with trauma include the detection of the specific
structures or tissues involved, the determina-
tion of the extent of the injury, the assessment
of general patient prognosis, and the enhance-
ment of therapeutic planning. These goals are
all aimed at a potentially positive affect on a
patient’s outcome. The results of these and
other outcomes will largely define how medical
care, including imaging studies, is dispensed in
the future in patients with trauma; these deci-
sions in turn will affect all those involved in the
health care of these patients.
Literature Cited
Bashir, E.F., Cybulski, G.R., Chaudhri, K., and
Choudhury, A.R. 1993. Magnetic resonance im-
aging and computed tomography in the evalu-
ation of penetrating gunshot injury of the spine.
Spine 18:772-795.
Curati, W.L., Kingsley, D.P.E., Kendall, B.E., and
Moseley, I.F. 1992. MRI in chronic spinal cord
trauma. Neuroradiology 35:30-35.
Davis, P.C., Reisner, A., Hudgins, P.A., Davis, W.E.,
and O’Brien, M.S. 1993. Spinal injuries in chil-
dren: Role of MR. Am. J. Neuroradiol. 14:607-
717.
Emery, S.E., Pathria, M.N., Wilber, R.G., Masaryk,
T., and Bohlman, H.H. 1989. MR imaging of
post-traumatic spinal ligament injury. J. Spinal
Disorders 2:229-233.
Flanders, A.E., Schaefer, D.M., Doan, H.T.,
Mishkin, M.M., Gonzalez, C.F., and Northrup,
B.E. 1990. Acute cervical spine trauma: Corre-
lation of MRI findings with degree of neurologic
deficit. Radiology 177:25-33.
Jinkins, J.R. 1993a. MR of enhancing nerve roots in
the unoperated lumbosacral spine. Am. J.
Neuroradiol. 14:193-202.
Jinkins, J.R. 1993b. Magnetic resonance imaging of
benign nerve root enhancement in the unoper-
ated and postoperative lumbosacral spine.
Neuroimaging Clin. North Am. 3:525-541.
Jinkins, J.R. 1998a. Magnetic resonance evaluation
of the symptomatic patient with acute or remote
spinal trauma. Part I. Direct nonpenetrat-
ing/blunt and penetrating trauma. J. Hong Kong
Coll. Radiol. 1:17-26.
Jinkins, J.R. 1998b. Magnetic resonance evaluation
of the symptomatic patient with acute or remote
spinal trauma. Part II. Distraction trauma. J.
Hong Kong Coll. Radiol. 1:106-112.
Jinkins, J.R. 1999. Magnetic resonance evaluation
of the symptomatic patient with acute or remote
spinal trauma. Part III. Indeterminate/ degenera-
tive/acquired traumatic spinal effects. Blunt and
penetrating trauma. J. Hong Kong Coll. Radiol.
2:14-20.
Jinkins, J.R., Reddy, S., Leite, C.C., Bazan, C., III,
and Xiong, L. 1998. MR of parenchymal spinal
cord signal change as a sign of active advance-
ment in clinically progressive posttraumatic
syringomyelia. Am. J. Neuroradiol. 19:177-182.
Kerslake, R.W., Jaspan, T., and Worthington, B.S.
1991. Magnetic resonance imaging of spinal
trauma. Br. J. Radiol. 64:386-402.
Kliewer, M.A., Gray, L., Paver, J., Richardson,
W.D., Vogler, J.B., McElhaney, J.H., and Myers,
B.S. 1993. Acute spinal ligament disruption: MR
Extradural Spine Spinal Trauma
A8.6.11 A8.6.12
Supplement 5 Supplement 5
imaging with anatomic correlation. J. Magn.
Reson. Imaging 3:855-861.
Leite, C.C., Escobar, B.E., Bazan, C., III, and Jink-
ins, J.R. 1997. MR imaging of cervical facet
dislocation. Neuroradiology 39:583-588.
Mathis, J.M., Wilson, J.T., Barnard, J.W., and
Zelenik, M.E. 1988. MR imaging of spinal cord
avulsion. Am. J. Neuroradiol. 9:1232-1233.
Mayer, J.S. and Kulkarni, M.V. 1987. MR imaging
of incisional spinal cord injury. Am. J. Neurora-
diol. 8:925-927.
Mendolsohn, D.B., Zollars, L., Weatherall, P.T., and
Girson, M. 1990. MR of cord transection. J.
Comput. Assist. Tomogr. 14:909-911.
Mirvis, S.E., Geisler, F.H., Jelinek, J.J., Joslyn, J.N.,
and Gellad, F. 1988. Acute cervical spine trauma:
Evaluation with 1.5 T MRI. Radiology 166:807-
816.
Pan, G., Kulkarni, M., MacDougall, D.J., and Miner,
M.E. 1988. Traumatic epidural hematoma of the
cervical spine. Diagnosis with magnetic reso-
nance imaging. J. Neurosurg. 68:798-801.
Post, M.J., Becerra, J.L., Madsen, P.W., Puckett, W.,
Quencer, R.M., Bunge, R.P., and Sklar, E.M.
1994. Acute spinal subdural hematoma: MR im-
aging and CT findings with pathologic corre-
lates. Am. J. Neuroradiol. 15:1895-1905.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Silberstein, M., Tress, B.M., and Hennessy, O. 1992.
Delayed neurologic deterioration in the patient
with spinal trauma: Role of MR imaging. Am. J.
Nuroradiol. 13:1373-1381.
Takahashi, M., Sakamoto, Y., Miyawaki, M., and
Bussaka, H. 1987. Increased MR signal intensity
secondary to chronic cervical cord compression.
Neuroradiology 29:550-556.
Contributed by J. Randy Jinkins and
David D. Stark
Downstate Medical Center
State University of New York
Brooklyn, New York
Current Protocols in Magnetic Resonance Imaging
Extradural Spinal Cord/Cauda Equina UNIT A8.7 Table A8.7.2 Primary Clinical Imaging Parameters for Sequence 1 (Pilot Scan)

Compression Patient position Supine

Scan type Gradient echo
Clinical signs and symptoms suggesting acute or subacute compression of the spinal cord Imaging plane (orientation) Transverse
or cauda equina constitute a medical emergency requiring urgent diagnosis in order to Slice or volume center Centered on:
Cervical spine: thyroid cartilage
effect appropriate therapy aimed at alleviating the pathologic process responsible for the
Thoracic spine: nipple line
compressive phenomenon. The etiology of the compressive pathology is nonspecific, and Lumbar spine: iliac crests
runs the gamut of disease-process categories from benign to malignant. The purpose of Echo time (TE) As short as possible
MR imaging in such cases is to determine the level(s), degree, and, if possible, the type Repeat time (TR) As short as possible
of disease process in order to assist in therapeutic planning gauged toward relieving the Flip angle (FA) 15°
neurologic compression (Schnebel et al., 1989; Kent et al., 1992; Jinkins, 1999). Fields of view (FOVx, FOVy) Cervical: 240 mm, 240 mm
Thoracic: 340 mm, 340 mm
CONVENTIONAL AND FAST SPIN ECHO ACQUISITIONS BASIC Lumbosacral: 280 mm, 280 mm
PROTOCOL Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm
The following sequences comprise the preferred protocol for high-field MR machines. Thoracic: 1.25 mm, 1.25 mm
On some machines, alternate gradient-recalled echo acquisitions may be more desirable. Lumbosacral: 1.09 mm, 1.09 mm
In addition, occasionally, intravenous (i.v.) contrast enhanced alternate protocols may be Number of data points collected (Nx, Ny) 256, 256
useful to evaluate the integrity of the blood-nerve barrier in certain cases of suspected Display matrix (Dx, Dy) 256, 256
compressive radiculopathy, or in cases of epidural masses of ambiguous etiology (e.g., Slice thickness (Δz) 5 mm
epidural neoplasia/infection versus intervertebral disc fragment). Number of slices Variable
Slice gap Not applicable
Table A8.7.1 lists the hardware necessary to perform the procedure, along with appropri-
Number of acquisitions (Nacq) 1
ate parameters. The available gradient strength will depend on the scanner, and the echo
Scan time ∼10 sec
times (e.g., TE) given in other tables will be varied accordingly (the smaller the gradient
strength, the longer the echo time for a particular scan).
The entire protocol should take 35 to 45 min to complete. The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
NOTE: Be sure that technicians and nurses have immediate access to any emergency composition of the items, it is best to exclude patients with any metal implants; see Shellock
(1996) for discussion of what implants may be safely scanned using magnetic resonance.
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen. Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
Set up patient and equipment be screened as well to ensure the absence of loose metal objects on the body or clothing.
1. Interview (screen) the patient to ensure that he or she has no contraindications such 2. If the procedure is a research protocol, have the patient sign any necessary consent
as cardiac pacemakers or other implants containing ferromagnetic materials. Also be form.
sure to find out if the patient has any health conditions that may require the presence
of special emergency equipment during the scanning procedure, or necessitate any 3. Have the patient remove all jewelry and change into a gown to eliminate any metal
other precautions. that might be found in clothing.
Generally standard screening forms are used for all patients scanned in a magnetic 4. Inform the patient about what will occur during the procedure, what he or she will
resonance system. experience while in the magnet, and how to behave, including the following.
a. If earphones or headphones are used to protect the ears from the loud sounds
Table A8.7.1 Equipment Parameters for Spine Imaging in Cases of produced by the gradients, the patient will be asked to wear these, but will be able
Suspected Extradural Spinal Cord/Cauda Equina Compression to communicate with you at any time during the imaging.
b. The patient will be given a safety squeeze-bulb or similar equipment to request
Coil type Cervical, thoracic, lumbar: phase array
surface coil (or other depending upon assistance at any time (demonstrate how this works).
machine compatibility and availability) c. For good results the patient should not talk, and should avoid or minimize
Flow compensation Any level (optional, if available) swallowing or other movement during each scan—i.e., as long as the banging
Peripheral gating Thoracic spine only (optional) sounds continue. Between scans, talking and swallowing are allowed in most
Respiratory gating Thoracic spine only (optional) cases, but should be avoided when comparative positional studies are being
Extradural Spinal
Use of contrast agents See Alternate Protocol (optional) Cord/Cauda performed; the patient will be informed when this is the case.
Equina
Extradural Spine Compression d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
Contributed by J. Randy Jinkins and David D. Stark A8.7.1 A8.7.2
Current Protocols in Magnetic Resonance Imaging (2002) A8.7.1-A8.7.9
Copyright © 2002 by John Wiley & Sons, Inc. Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging
 Table A8.7.3 Primary Clinical Imaging Parameters for Sequence 2 (T1-Weighted Table A8.7.4 Primary Clinical Imaging Parameters for Sequence 3 (T2-Weighted
Image) Image, Fast Spin Echo)

Patient position Supine Patient position Supine

Scan type Conventional spin echo Scan type Fast spin echo
Imaging plane (orientation) Sagittal Imaging plane (orientation) Sagittal
Central slice or volume center Centered on: Slice or volume center Centered on area of interest (as in
Cervical: 3rd cervical vertebra sequence 2, Table A8.7.3)
Thoracic: 6th thoracic vertebra Echo time (TE) 100 msec
Lumbar: 3rd lumbar vertebra Echo train length (ETL) 8
Echo time (TE) 10 msec Repeat time (TR) 4000 msec
Repeat time (TR) 500 msec Flip angle (FA) 90°
Flip angle (FA) 90° Fields of view (FOVx, FOVy) As in sequence 2, Table A8.7.3
Fields of view (FOVx, FOVy) Cervical: 240 mm, 240 mm Resolution (Δx, Δy) Cervical: 0.47 mm, 0.47 mm
Thoracic: 320 mm, 320 mm Thoracic: 0.63 mm, 0.63 mm
Lumbosacral: 280 mm, 280 mm (may Lumbosacral: 0.55 mm, 0.55 mm
use rectangular field of view [e.g., half Number of data points collected (Nx, Ny) 512, 512
or three-quarter field] if available, or
Display matrix (Dx, Dy) 512, 512
tailor to region of interest)
Slice thickness (Δz) Cervical: 3 mm
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm
Thoracic: 3 mm
Thoracic: 1.25 mm, 1.25 mm
Lumbar: 4 mm
Lumbosacral: 1.09 mm, 1.09 mm
Number of slices Varies with spinal level
Number of data points collected (Nx, Ny) 256, 256
Slice gap Cervical: 0.5 mm
Display matrix (Dx, Dy) 256, 256
Thoracic: 1 mm
Slice thickness (Δz) Cervical: 3 mm Lumbar: 1 mm
Thoracic: 3 mm
Number of acquisitions (Nacq) 1
Lumbar: 4 mm
Flow compensation Yes (if available)
Number of slices As many as needed to cover the region
of interest Saturation pulses Yes; anterior cervical/thoracic/lumbar
slabs to saturate larynx/vessels/heart
Slice gap Cervical: 0.5 mm
Thoracic: 1 mm Fat suppression Yes
Lumbar: 1 mm Slice series Left to right or the reverse depending
Number of acquisitions (Nacq) 4 on preference
Flow compensation Yes (if available) Scan time ∼4 min
Saturation pulses Yes; anterior cervical/thoracic/lumbar
slab to saturate larynx/vessels/heart
Slice series Left to right or the reverse depending
9. If the patient is unable to hold still, provide an appropriate sedative.
on preference
Scan time ∼8 min Sequence 1: Rapid positioning pilot
10. To validate the patient’s position, run the system’s pilot (or scout) scan (sequence 1)
to ensure correct location of the neck in three dimensions, using the imaging sequence
5. Have the patient mount onto the table. Either before or right after the patient lies given in Table A8.7.2 or similar parameters.
down, set up any triggering devices or other monitoring equipment that is to be used. This sequence usually consists of three orthogonal planes to allow subsequent localization.
6. Center the coil over the region where the key information is desired. The images are often also used later to determine where to place the saturation pulses and
to set up total coverage of the volume of interest.
Make sure that the body is constrained to prevent motion, especially if high-resolution scans
are to be run.
Sequence 2: Sagittal T1-weighted conventional spin echo
7. If needed, place a pillow or other support under the knees to make the patient more 11. Set the imaging parameters as shown in Table A8.7.3.
comfortable.
12. Use the pilot image to locate the spine in three dimensions to ensure coverage of the
8. Use the centering light to position the patient (cervical spine: thyroid cartilage; region of interest (e.g., cervical, thoracic, or lumbosacral spine).
thoracic spine: nipple line; lumbar spine: iliac crests) and put him or her into the
center of the magnet. 13. Let the patient know you are ready and begin the scan.
Once this step has been performed, so long as the patient does not move on the table, the Extradural Spinal
table itself can be moved and then replaced in the same position as before without Cord/Cauda
Equina
jeopardizing the positioning of one scan relative to another. Extradural Spine Compression

A8.7.3 A8.7.4
Current Protocols in Magnetic Resonance Imaging Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging
 Table A8.7.5 Primary Clinical Imaging Parameters for Sequence 4 (T1-Weighted Table A8.7.6 Primary Clinical Imaging Parameters for Sequence 5 (T2-Weighted
Images) Image, Fast Spin Echo)

Patient position Supine Patient position Supine

Scan type Conventional spin echo Scan type Fast spin echo
Imaging plane (orientation) Transverse Imaging plane (orientation) Transverse
Central slice or volume center Centered on the area of interest (as in Central slice or volume center Centered on the region of interest (as
sequence 2, Table A8.7.3) in sequence 2, Table A8.7.3)
Echo time (TE) 10 msec Echo time (TE) 100 msec
Repeat time (TR) 500 msec Echo train length (ETL) 8
Flip angle (FA) 90° Repeat time (TR) 4000 msec
Fields of view (FOVx, FOVy) As in sequence 2, Table A8.7.3 Flip angle (FA) 90°
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm Fields of view (FOVx, FOVy) As in sequence 2, Table A8.7.3
Thoracic: 1.25 mm, 1.25 mm Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm
Lumbosacral: 1.09 mm, 1.09 mm Thoracic: 1.25 mm, 1.25 mm
Number of data points collected (Nx, Ny) 256, 256 Lumbosacral: 1.09 mm, 1.09 mm
Display matrix (Dx, Dy) 256, 256 Number of data points collected (Nx, Ny) 256, 256
Slice thickness (Δz) Cervical: 3 mm Display matrix (Dx, Dy) 256, 256
Thoracic: 3-8 mm Slice thickness (Δz) Cervical: 3 mm
Lumbar: 4 mm Thoracic: 3-8 mm
Number of slices Varies with spinal level Lumbar: 4 mm
Slice gap Cervical: 1 mm Number of slices Varies with spinal level
Thoracic: 1–2 mm Slice gap Cervical: 1 mm
Lumbar: 1 mm Thoracic: 1–2 mm
Number of acquisitions (Nacq) 2 Lumbar: 1 mm
Slice location See text (Basic Protocol, step 16) Number of acquisitions (Nacq) 2
Saturation pulses No Slice locations See Basic Protocol, step 16
Scan time ∼4 min Saturation pulses No
Scan time ∼4 min

Sequence 3: Sagittal T2-weighted fast spin echo, fat suppressed ALTERNATE GRADIENT-RECALLED ECHO ACQUISITIONS
14. Review the pilot scans and ensure that the saturation pulse is correctly placed anterior PROTOCOL 1
With some machines, or according to preferences, gradient-recalled echo acquisitions
to above the slab of interest. may be used in the sagittal and/or transverse planes to clearly distinguish between the
15. Run sequence 3 according to Table A8.7.4. spinal cord and the extradural tissue comprising this structure. This is especially true in
cases where the epidural mass consists of acute blood products (e.g., deoxyhemoglobin)
Sequence 4: Transverse T1-weighted conventional spin echo that are paramagnetic and are better characterized by T2* gradient-recalled echo images
16. Using the midline sagittal T1-weighted image acquired in sequence 2, set the than by T2 fast spin echo acquisitions.
transverse acquisition parameters as follows:
Sequence 6: Sagittal gradient-recalled echo
a. Cervical spine: stacked images from C1 through C7-T1. 1. Run the sequence according to Table A8.7.7.
b. Thoracic spine: stacked images through levels of interest.
c. Lumbosacral spine: 5 slices each, angled to the plane of the intervertebral disc at Sequence 7: Transverse gradient-recalled echo
L3-4, L4-5, and L5-S1; one slice each, angled to the intervertebral disc at L1-2 2. Run the sequence according to Table A8.7.8.
and L2-3.
ALTERNATE CONTRAST ENHANCED ACQUISITIONS
17. Supplement additional slices according to visible disease present or to clinical query. PROTOCOL 2
In some situations (e.g., infection, neoplasia) i.v. paramagnetic contrast material may be
18. Run the sequence according to Table A8.7.5. indicated.

Sequence 5: Transverse T2-weighted fast spin echo Materials

19. Using the midline T1-weighted image acquired in sequence 2, repeat the setup as in Normal saline (0.9% NaCl), sterile
Table A8.7.6. Extradural Spinal Intravenous MRI contrast agent (e.g., Magnevist, Omniscan, or Prohance)
Cord/Cauda
20. Run sequence 5 according to Table A8.7.6. Equina
Extradural Spine Compression

A8.7.5 A8.7.6
Current Protocols in Magnetic Resonance Imaging Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging
 Table A8.7.7 Primary Clinical Imaging Parameters for Sequence 6 (T2* Gradient Set up patient and equipment
Recalled Echo) 1. Use the same equipment and perform patient setup as in Basic Protocol, steps 1 to 4.
Patient position Supine 2. Establish an i.v. line from which the contrast agent can be injected, and attach this
Scan type 2-D gradient echo line securely to the patient so that movement into or out of the magnet will not pull
Imaging plane (orientation) Sagittal at the patient’s arm. Resume steps 5 to 9 in the Basic Protocol.
Central slice or volume center Centered on the region of interest (as It is preferable to insert the line prior to imaging and to leave the patient in the magnet,
in sequence 2, Table A8.7.3) with no intervening motion, between the scans run before contrast agent injection and those
Echo time (TE) 15 msec run after injection.
Repeat time (TR) 500 msec Scan pilot
Flip angle (FA) 10° to 20° 3. Run a rapid three-plane positioning pilot scan (see Basic Protocol, sequence 1).
Fields of view (FOVx, FOVy) As in sequence 2, Table A8.7.3
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm Sequences 8 and 9: i.v. contrast enhanced T1-weighted spin echo
Thoracic: 1.25 mm, 1.25 mm 4. Leaving the patient in the magnet, inject the contrast agent, flush the line with 10 ml
Lumbosacral: 1.09 mm, 1.09 mm saline, and then immediately run sagittal (sequence 8) and transverse (sequence 9)
Number of data points collected (Nx, Ny) 256, 256 T1-weighted image sequences (see Basic Protocol, sequences 2 and 4, Tables A8.7.3
Display matrix (Dx, Dy) 256, 256 and A8.7.5; sequences 8 and 9 are the same sequences 2 and 4, respectively, but with
Slice thickness (Δz) 3 mm the contrast agent injected into the patient).
Number of slices Varies with spinal level
An i.v. dose of 0.1 mmol/kg of contrast agent is usually given.
Slice gap ≤1 mm
Number of acquisitions (Nacq) 3
Slice location See Basic Protocol, step 16 COMMENTARY
Flow compensation Yes (if available) Background Information order to minimize permanent neural deficit.
Saturation pulses Yes Acute or subacute compression of the spinal The causes of such compression are nonspe-
Scan time 6 min cord in the cervical or thoracic spinal cord, or cific and can be degenerative, traumatic, in-
of the cauda equina in the thoracolumbar re- flammatory, or neoplastic, or may constitute
gion, is a major neurologic emergency almost acquired conditions of a different nature than
always requiring urgent medical attention in the pathologic entities mentioned (Fig. A8.7.1).
Table A8.7.8 Primary Clinical Imaging Parameters for Sequence 7 (T2* Gradient
Recalled Echo)

Patient position Supine

Scan type Gradient echo  
Imaging plane (orientation) Transverse
Central slice or volume center Centered on the region of interest (as
in sequence 2, Table A8.7.3)
Echo time (TE) 15 msec
Repeat time (TR) 500 msec
Flip angle (FA) 10° to 20°
Fields of view (FOVx, FOVy) As in sequence 2, Table A8.7.3
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm
Thoracic: 1.25 mm, 1.25 mm
Lumbosacral: 1.09 mm, 1.09 mm
Number of data points collected (Nx, Ny) 256, 256
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 3–4 mm
Number of slices Varies with spinal level
Slice gap ≤1 mm
Number of acquisitions (Nacq) 1
Slice location See Basic Protocol, step 16
Flow compensation Yes (if available)
Figure A8.7.1 Benign multilevel extradural spinal cord/cauda equina compression. (A) Sagittal
Saturation pulses Yes T2-weighted (TR = 4000 msec, TE = 100 msec) fat suppressed image shows multilevel spinal cord
Extradural Spinal
Scan time ∼2 min Cord/Cauda compression in the cervical region due to generalized spondylosis. (B) Sagittal T2-weighted (TR =
Equina 4000 msec, TE = 100 msec) fat suppressed image shows multilevel degenerative cauda equina
Extradural Spine Compression compression in the lumbosacral region also secondary to generalized spondylosis.
A8.7.7 A8.7.8
Current Protocols in Magnetic Resonance Imaging Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging
Regardless of etiology, MR imaging, an almost
imperative element, is the acute or even emer-
gent evaluation of these patients in order to
analyze the degree, level(s), and extent of spinal
cord/cauda equina compression. In addition,
MR imaging may enable a listing of the nar-
rowest possible differential diagnostic possibil-
ity(ies). This will materially aid, in the most
timely fashion, the therapeutic planning of such
acutely neurologically debilitated patients.
Clinical Parameters and
Troubleshooting
Cerbrospinal fluid (CSF) flow, cardiac, la-
ryngeal, body wall, and other sources of motion
can produce artifacts that can on occasion sig-
nificantly degrade the images. Proper spatial
(e.g., prevertebral) saturation pulses and some-
times flow compensation pulses and/or car-
diac/respiratory gating can reduce these arti-
facts significantly. In many instances these ar-
tifacts may be difficult or impossible to
overcome from patient to patient.
Anticipated Results
The goal of starting the patient with clini-
cally suspected spinal cord or cauda equina
compression is to pinpoint the level(s), degrees,
Current Protocols in Magnetic Resonance Imaging
and extent of disease in order to significantly
aid in the urgent treatment planning in these
patients in an effort to minimize permanent
neurologic deficit and enhance timely clinical
patient outcome.
Literature Cited
Jinkins, J.R. 1999. MR evaluation of stenosis in-
volving the neural foramina lateral recesses and
central canal of the lumbosacral spine. Neuroi-
mag. Clin. North Am. 7:493-511.
Kent, D.L., Haynor, D.R., Larson, E.B., and Deyo,
R.A. 1992. Diagnosis of lumbar spinal stenosis
in adults: A metaanalysis of the accuracy of CT,
MR, and myelgraphy. Am. J. Roentgenol.
158:1135-1144.
Schnebel, B., Kingston, S., Watkins, R., and Dillon,
W. 1989. Comparison of MRI to contrast CT in
the diagnosis of spinal stenosis. Spine 14:332-
337.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Contributed by J. Randy Jinkins and
David D. Stark
Downstate Medical Center
State University of New York
Brooklyn, New York
Extradural Spine
A8.7.9
Supplement 5
Post Surgical Spinal Evaluation UNIT A8.8
One of the most challenging areas of diagnosis is to be found in acquiring and interpreting
medical images in a patient who has undergone surgery for spinal degenerative disease.
This may have involved partial or complete laminectomy, discectomy (surgical removal
of the native and/or herniated portions of the intervertebral disc), intervertebral bony
fusion and/or spinal instrumentation. The imaging findings may be either of an expected
or unexpected nature. In some instances, the treated tissues may be left with benign
scarring; in other cases, there may be a recurrence of disease or the appearance of a
different type of disease process engendered by the surgical procedure. In still other
situations, the observation may represent a true acute/subacute complication of the
operation. All of these possibilities complicate medical image analysis (Ghazi et al., 1992;
Jinkins, 1993; Jinkins et al., 1993; Grane et al., 1998; Shafaiae et al., 1997; Van Goethem
et al., 1997).
In order to critically evaluate the posttherapeutic patient, it is imperative to understand
several factors in reasonably specific detail. These factors include (a) the primary clinical
syndrome and diagnosis, (b) the surgical treatment(s) undergone by the patient, (c) the
elapsed time since the operative procedure(s), and (d) the current clinical syndrome. The
answers to these questions will in large part determine which imaging modality or
modalities are chosen for the patient evaluation, how the images are acquired, and whether
or not an enhancing agent is used.
CONVENTIONAL AND FAST SPIN ECHO ACQUISITIONS BASIC
PROTOCOL
Progressive improvements in magnetic resonance (MR) imaging have substantially
improved the ability of the medical imaging physician to critically analyze the postop-
erative lumbosacral spine following surgery for degenerative spinal disease. Compared
to other imaging methods, MR images acquired in multiple planes have superior diag-
nostic potential, in part because of their generally greater spatial and contrast resolution
characteristics for many tissues. Nevertheless, the similarity of some postsurgical
pathologic processes in regard to MR signal intensity often makes even this imaging
technique somewhat difficult to interpret. The development of intravenously (i.v.) admin-
istered paramagnetic MR contrast agents (e.g., gadolinium) has materially assisted the
diagnostic sensitivity and specificity of MR imaging in the evaluation of the failed back
surgery syndrome (FBSS) because of the improvement in differential contrast enhance-
ment afforded by these agents.
In general, T2-weighted fast spin-echo images are superior to conventional spin-echo
images of the lumbosacral spine in part because of improved image quality resulting from
superior spatial resolution and reduced motion artifact. Sagittal (with fat suppression) and
transverse (without fat suppression) fast spin-echo T2-weighted images are helpful in
assessing neural foramen narrowing, central spinal canal and lateral recess spinal stenosis,
hydration status of the intervertebral disc, abnormal signal intensity of the disc and
cancellous bone and signal intensity of abnormal intra- or perispinal soft tissue masses
(e.g., disc herniation, epidural scar, epidural phlegmon, epidural abscess, posterior facet
joint synovial cyst). Sagittal and transverse T1-weighted spin-echo images obtained before
and immediately after the bolus i.v. injection of a gadolinium product are almost
imperative in the evaluation of the postoperative lumbar spine, in order to evaluate the
integrity of the blood-nerve barrier in certain cases of suspected compressive radicu-
lopathy, or in cases of epidural masses of ambiguous etiology (e.g., epidural infection
versus intervertebral disc fragment versus epidural fibrosis).
Extradural Spine
Contributed by J. Randy Jinkins and David D. Stark A8.8.1
Current Protocols in Magnetic Resonance Imaging (2003) A8.8.1-A8.8.10
Copyright © 2003 by John Wiley & Sons, Inc. Supplement 8
 Table A8.8.1 Equipment Parameters for Spine Imaging in Cases of Table A8.8.2 Primary Clinical Imaging Parameters for Sequence 1 (Pilot Scan)
Postsurgical Evaluation
Patient position Supine
Coil type Cervical, thoracic, lumbar: phased array, Scan type Gradient echo
surface coil (or other depending upon Imaging plane (orientation) Transverse
machine compatibility and availability)
Central slice or volume center Centered on:
Peripheral gating Thoracic spine only (optional) Cervical spine: thyroid cartilage
Respiratory gating Thoracic spine only (optional) Thoracic spine: nipple line
Flow compensation pulse Any level (if available) Lumbar spine: iliac crest
Use of contrast agents Yesa Echo time (TE) As short as possible
aIn emergency situations, contrast agents may not be necessary. Repeat time (TR) As short as possible
Flip angle (FA) 15°
Fields of view (FOVx, FOVy) Cervical: 240 mm, 240 mm
Fat-suppression techniques can also often be used after i.v. gadolinium administration Thoracic: 320 mm, 320 mm
because they improve the relative intensity and homogeneity of contrast enhancement Lumbosacral: 280 mm, 280 mm
when it exists. However, it may make the critical interpretation of possible abnormal Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm
postoperative intrathecal nerve root enhancement (i.e., sterile radiculitis) difficult or Thoracic: 1.25 mm, 1.25 mm
impossible because small degrees of apparently normal nerve root enhancement are not Lumbosacral: 1.09 mm, 1.09 mm
infrequently observed when fat suppression MR imaging techniques are used with i.v. Number of data points collected (Nx, Ny) 256, 256
gadolinium enhancement. Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 5 mm
The following sequences comprise the preferred protocol for high-field MR machines. Number of slices Variable
On some machines alternate gradient recalled echo acquisitions may be more desirable. Slice gap Not applicable
This entire protocol should take 45 to 50 min to complete. Number of acquisitions (Nacq) 1
Scan time ∼10 sec
Table A8.8.1 lists the hardware necessary to perform the procedure, along with appropri-
ate parameters. The available gradient strength will depend on the scanner, and the echo
times given in other tables will be varied accordingly (the smaller the gradient strength,
2. If the procedure is a research protocol, have the patient sign any necessary consent
the longer the echo time for a particular scan).
form.
NOTE: Be sure that technicians and nurses have immediate access to any emergency 3. Have the patient remove all jewelry and change into a gown to eliminate any metal
equipment that may be relevant to a given study, or that may be needed for a particular that might be found in clothing.
patient, such as crash carts or oxygen.
4. Inform the patient about what will occur during the procedure, what he or she will
Materials experience while in the magnet, and how to behave, including the following:
Normal saline (0.9% NaCl) sterile
a. If earphones or headphones are used to protect the ears from the loud sounds
Extravascular contrast agent (e.g., Mangevist, Omniscan, or Prohance)
produced by the gradients, the patient will be asked to wear these, but will be able
Set up patient and equipment to communicate with you at any time during the imaging.
1. Interview (screen) the patient to ensure that he or she has no contraindications such b. The patient will be given a safety squeeze-bulb or similar equipment to request
as cardiac pacemakers or other implants containing ferromagnetic materials. Also be assistance at any time (demonstrate how this works).
sure to find out if the patient has any health conditions that may require the presence c. For good results the patient should not talk, and should avoid or minimize
of special emergency equipment during the scanning procedure, or necessitate any swallowing or other movement during each scan—i.e., as long as the banging
other precautions. sounds continue. Between scans, talking and swallowing are allowed in most
Generally standard screening forms are used for all patients scanned in a magnetic cases, but should be avoided when comparative positional studies are being
resonance system. performed; the patient will be informed when this is the case.
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact 5. Have the patient mount onto the table. Either before or right after the patient lies
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(1996) for discussion of what implants may be safely scanned using magnetic resonance.
down, set up any triggering devices or other monitoring equipment that is to be used.

Patients may be accompanied into the magnet room by a friend or family member, who can 6. Center the coil over the region where the key information is desired.
sit in the room during the scan and comfort the patient as needed. This companion must Make sure that the body is constrained to prevent motion, especially if high-resolution scans
be screened as well to ensure the absence of loose metal objects on the body or clothing. are to be run.
Post Surgical
Spinal Evaluation Extradural Spine

A8.8.2 A8.8.3
Supplement 8 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 8
 Table A8.8.3 Primary Clinical Imaging Parameters for Sequence 2 (T1-Weighted Table A8.8.4 Primary Clinical Imaging Parameters for Sequence 3 (T2-Weighted
Image) Image, FSE)a

Patient position Supine Patient position Supine

Scan type Conventional spin echo Scan type Fast spin echo
Imaging plane (orientation) Sagittal Imaging plane (orientation) Sagittal
Central slice or volume center Centered on: Central slice or volume center Centered on area of interest (as in
Cervical: 3rd cervical vertebra sequence 2, Table A8.8.3)
Thoracic: 6th thoracic vertebra Echo time (TE) 100 msec
Lumbar: 3rd lumbar vertebra Echo train length (ETL) 8
Echo time (TE) 10 msec Repeat time (TR) 4000 msec
Repeat time (TR) 500 msec Flip angle (FA) 90°
Flip angle (FA) 90° Fields of view (FOVx, FOVy) As in sequence 2, Table A8.8.3
Fields of view (FOVx, FOVy) Cervical: 240 mm, 240 mm Resolution (Δx, Δy) Cervical: 0.47 mm, 0.47 mm
Thoracic: 320 mm, 320 mm Thoracic: 0.63 mm, 0.63 mm
Lumbosacral: 280 mm, 280 mm (may Lumbosacral: 0.55 mm, 0.55 mm
use rectangular field of view [e.g., half Number of data points collected (Nx, Ny) 512, 512
or three-quarter field] if available, or Display matrix (Dx, Dy) 512, 512
tailor to region of interest)
Slice thickness (Δz) Cervical: 3 mm
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm Thoracic: 3 mm
Thoracic: 1.25 mm, 1.25 mm Lumbar: 4 mm
Lumbosacral: 1.09 mm, 1.09 mm
Number of slices Varies with spinal level
Number of data points collected (Nx, Ny) 256, 256
Slice gap Cervical: 0.5 mm
Display matrix (Dx, Dy) 256, 256 Thoracic: 1 mm
Slice thickness (Δz) Cervical: 3 mm Lumbar: 1 mm
Thoracic: 3 mm Number of acquisitions (Nacq) 1
Lumbar: 4 mm
Flow compensation Yes (if available)
Number of slices As many as needed to cover the region
Saturation pulses Yes; anterior cervical/thoracic/lumbar
of interest
slabs to saturate larynx/vessels/heart
Slice gap Cervical: 0.5 mm
Fat suppression Yes
Thoracic: 1 mm
Lumbar: 1 mm Slice series Left to right or the reverse depending
on preference
Number of acquisitions (Nacq) 2
Scan time ∼4 min
Flow compensation Yes (if available)
aFSE: fast spin echo.
Saturation pulses Yes; anterior cervical/thoracic/lumbar
slab to saturate larynx/vessels
Slice series Left to right or the reverse depending Sequence 1: Rapid positioning pilot
on preference 10. To validate the patient’s position, run the system’s pilot (or scout) scan (sequence 1)
Scan time ∼4 min to ensure correct location of the neck in three dimensions, using the imaging sequence
given in Table A8.8.2 or similar parameters.
This sequence usually consists of three orthogonal planes to allow subsequent localization.
7. If needed, place a pillow or other support under the knees to make the patient more The images are often also used later to determine where to place the saturation pulses and
comfortable. to set up total coverage of the volume of interest.
8. Use the centering light to position the patient (cervical spine: thyroid cartilage; NOTE: In an emergency situation (e.g., postoperative acute hemorrhage), only sequences
thoracic spine: nipple line; spine: iliac crests) and put him or her into the center of 2 to 5 may be indicated.
the magnet.
Once this step has been performed, so long as the patient does not move on the table, the Sequence 2: Sagittal T1-weighted conventional spin echo
table itself can be moved and then replaced in the same position as before without 11. Set the imaging parameters as shown in Table A8.8.3.
jeopardizing the positioning of one scan relative to another.
12. Use the pilot image to locate the spine in three dimensions to ensure coverage of the
9. If the patient is unable to hold still, provide an appropriate sedative. region of interest (cervical, thoracic, lumbosacral) spine.
13. Let the patient know you are ready and begin the scan.
Post Surgical
Spinal Evaluation Extradural Spine

A8.8.4 A8.8.5
Supplement 8 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 8
 Table A8.8.5 Primary Clinical Imaging Parameters for Sequence 4 (T1-Weighted Table A8.8.6 Primary Clinical Imaging Parameters for Sequence 5 (T2-Weighted
Image) Image, FSE)a

Patient position Supine Patient position Supine

Scan type Conventional spin echo Scan type Fast spin echo
Imaging plane (orientation) Transverse Imaging plane (orientation) Transverse
Central slice or volume center Centered on the area of interest (as in Central slice or volume center Centered on the region of interest (as
sequence 2, Table A8.8.3) in sequence 2, Table A8.8.3)
Echo time (TE) 10 msec Echo time (TE) 100 msec
Repeat time (TR) 500 msec Echo train length (ETL) 8
Flip angle (FA) 90° Repeat time (TR) 4000 msec
Fields of view (FOVx, FOVy) As in sequence 2, Table A8.8.3 Flip angle (FA) 90°
Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm Fields of view (FOVx, FOVy) As in sequence 2, Table A8.8.3
Thoracic: 1.25 mm, 1.25 mm Resolution (Δx, Δy) Cervical: 0.94 mm, 0.94 mm
Lumbosacral: 1.09 mm, 1.09 mm Thoracic: 1.25 mm, 1.25 mm
Number of data points collected (Nx, Ny) 256, 256 Lumbosacral: 1.09 mm, 1.09 mm
Display matrix (Dx, Dy) 256, 256 Number of data points collected (Nx, Ny) 256, 256
Slice thickness (Δz) Cervical: 3 mm Display matrix (Dx, Dy) 256, 256
Thoracic: 3–8 mm Slice thickness (Δz) Cervical: 3 mm
Lumbar: 4 mm Thoracic: 3–8 mm
Number of slices Varies with spinal level Lumbar: 4 mm
Slice gap Cervical: 1 mm Number of slices Varies with spinal level
Thoracic: 1 mm Slice gap Cervical: 1 mm
Lumbar: 1 mm Thoracic: 1–2 mm
Number of acquisitions (Nacq) 2 Lumbar: 1 mm
Flow compensation Yes (if available) Number of acquisitions (Nacq) 2
Slice locations See text (Basic Protocol, step 16) Slice locations See text (Basic Protocol, step 16)
Saturation pulses No Flow compensation Yes (if available)
Scan time ∼4 min Saturation pulses No
Scan time ∼4 min
aFSE: fast spin echo.
Sequence 3: Sagittal T2-weighted fast spin echo
14. Review the pilot scans and ensure that the saturation pulse is correctly placed anterior
to above the slab of interest. Sequence 6 and 7: Sagittal (sequence 6) and transverse (sequence 7) i.v. contrast
enhanced T1-weighted spin echo
15. Run sequence 3 according to Table A8.8.4. 21. Remove patient from the magnet.
22. Establish an i.v. line from which the contrast agent can be injected, and attach this
Sequence 4: Transverse T1-weighted conventional spin echo
line securely to the patient so that movement into or out of the magnet will not pull
16. Using the midline sagittal T1-weighted image acquired in sequence 2, set the trans-
at the patient’s arm. Push the patient back to the magnet.
verse acquisition parameters as follows:
It is preferable to insert the line prior to imaging and to leave the patient in the magnet,
a. Cervical spine: stacked images from C1 through C7-T1. with no intervening motion, between the scans run before contrast agent injection and those
b. Thoracic spine: Stacked images through levels of interest run after injection.
c. Lumbosacral spine: 5 slices each, angled to the plane of the intervertebral disc at Scan pilot
L3-4, L4-5, and L5-S1; one slice each, angled to the intervertebral disc at L1-2 23. Run a rapid three-plane positioning pilot scan (see sequence 1).
and L2-3.
Scan sequences
17. Supplement additional slices according to visible disease present or to clinical query. 24. Leaving the patient in the magnet, inject the contrast agent, flush the i.v. line with 10
cc saline, and then immediately run sagittal (sequence 6) and transverse (sequence
18. Run the sequence according to Table A8.8.5.
7) T1-weighted image sequences (see sequences 2 and 4; sequences 6 and 7 are the
same as sequences 2 and 4, respectively).
Sequence 5: Transverse T2-weighted fast spin echo
19. Using the midline T1-weighted image acquired in sequence 2, repeat the setup as in A dose of 0.1 mmol/kg of contrast agent is usually given i.v.
Table A8.8.6. Fat-suppression techniques may be very useful in order to suppress osseous vertebral
Post Surgical marrow fat and perispinal soft tissue fat, thereby clearly distinguishing enhancement from
Spinal Evaluation 20. Run sequences according to Table A8.8.6. normal fatty tissue. Extradural Spine

A8.8.6 A8.8.7
Supplement 8 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 8
 ALTERNATE CORONAL CONTRAST ENHANCED ACQUISITION
PROTOCOL
In some instances, a coronal acquisition may be helpful to analyze the perispinal tissues A B
for infectious phlegmon or frank abscess formation.

Sequence 8: Contrast enhanced coronal T1-weighted image

Run the sequence according to sequence 8, Table A8.8.7. Fat suppression may be useful.
*
Table A8.8.7 Primary Clinical Imaging Parameters for Sequence 8 (T1-Weighted
Image)

Patient position Supine *

Scan type
Imaging plane (orientation)
Central slice or volume center
Echo time (TE)
Repeat time (TR)
Flip angle (FA)
Fields of view (FOVx, FOVy)
Resolution (Δx, Δy)
Number of data points collected (Nx, Ny)
Display matrix (Dx, Dy)
Slice thickness (Δz)
Number of slices
Slice gap
Number of acquisitions (Nacq)
Slice location
Saturation pulses
Fat suppression
Scan time
COMMENTARY
Background Information
Despite the relatively loose application of
criteria for judging operative success, lum-
bosacral spinal surgery has been so often un-
successful (10% to 40%) that failed back sur-
gery is now labeled as a syndrome—the failed
back surgery syndrome (FBSS). In general,
surgery for lumbar disc herniation relieves pain
in most patients, and produces good long-time
outcome in almost 90% of patients. Repeat
surgery, however, is less successful, and ac-
cording to several studies only 60% to 82% of
patients with recurrent disc herniation improve
after surgery. In patients who only have
epidural scar tissue, the success rate of reinter-
vention is as low as 17% to 38%. Therefore, the
Post Surgical obvious solution to this problem is to attempt
Spinal Evaluation to avoid the first operation that statistically is
Conventional spin echo
Coronal
Centered on the region of interest (as
in sequence 2, Table A8.8.3)
10 msec
500 msec
90°
As in sequence 2, Table A8.8.3
Cervical: 0.94 mm, 0.94 mm
Thoracic: 1.25 mm, 1.25 mm
Lumbosacral: 1.09 mm, 1.09 mm
256, 256
256, 256
3–4 mm
Varies with spinal level
1–2 mm
2 tempted surgical fusion failure. Still further
From front to back of spinal column
No
Yes: fat saturation or STIR (short tall
inversion recovery)
∼4 min
likely to lead to a less than satisfactory result,
and thus not create a clinical condition that
requires repeat spinal surgery.
The FBSS is characterized by postsurgical
intractable pain in the low back and lower
extremity or extremities, combined with vary-
ing degrees of functional incapacitation. The
major identifiable causes of the FBSS include
“clinically relevant” epidural fibrosis (i.e.,
fibrosis causing perineural constriction, teth-
ering of the nerve root), recurrent or residual
disc herniation, postoperative spinal infec-
tion, sterile arachnoiditis (Fig. A8.8.1), post-
surgical pseudomeningocele formation and
lateral recess, and foraminal or central spinal
stenosis that may preexist or follow the spinal
surgery. Other less common causes of the
FBSS include postoperative hemorrhage, sur-
Figure A8.8.1 Postoperative chronic adhesive arachnoiditis. (A) Sagittal T2-weighted (TR = 4000
msec, TE = 100 msec) fat suppressed image shows matting of the nerve roots superiorly (arrows)
and are “empty thecal sac” inferiorly (asterisk) indicating adhesive fibrosis of the contents of the
thecal sac. (B) Transverse T2-weighted (TR = 4000 msec, TE = 100 msec) image at L5 shows again
the “empty thecal sac” (asterisk) in this region due to adhesion of the intrathecal nerve roots to the
peripheral walls of the thecal sac.
gery inadvertently performed on the wrong side because some disc herniations can be some-
or at the incorrect segmental level, direct nerve what vascularized and therefore may enhance
injury at the time of surgery, spinal pain of relatively early. The practical basis for this
mechanical origin, e.g., posterior spinal facet imaging strategy is that the vessels in the scar
(i.e., zygapophyseal) joint disease, and at- tissue are relatively homogeneously distrib-
uted, whereas in disc herniation the vessels are
causes of the FBSS are recurrent or residual quite heterogeneous or are centrally absent.
clinical symptoms related to anterior spinal disc Therefore a centrally nonenhancing epidural
protrusion, spinal nerve root sterile radiculitis mass would be labeled a partially vascularized
(i.e., neuritis), disc herniation at a spinal level disc herniation (Fig. A8.8.2), but homogeneous
other than that operated on, posterior facet joint enhancement of an epidural process would be
fracture, and progressive spinal instability with termed epidural fibrosis. It should be noted that
attendant spondylolisthesis. These conditions imaging after 20 to 30 min of the i.v. contrast
should be distinguished from acceptable and agent administration is not helpful because
expected postoperative findings found after many recurrent disc herniations enhance more
successful lumbar surgery (i.e., that surgery or less homogeneously within this delayed time
associated with clinical sign and symptom re- frame as the gadolinium progressively seeps
lief). into the extruded disc material.
Intravenously administered gadolinium
Critical Parameters and compounds are also an important adjunct to the
Troubleshooting MR evaluation of the postoperative lumbosac-
Cerebrospinal fluid (CSF) flow, cardiac, la- ral spine in the elucidation and differentiation
ryngeal, body wall, and other sources of motion of spinal, leptomeningeal, and/or neural in-
can produce artifacts that can on occasion sig- flammation (infectious or aseptic). The proper
nificantly degrade the images. Proper spatial differentiation of these varied pathologic phe-
(e.g., prevertebral) saturation pulses and some- nomena on MR imaging should allow im-
times flow compensation pulses and/or car- proved patient triage toward appropriate medi-
diac/respiratory gating can reduce these arti- cal-surgical therapy aimed at treating the spe-
facts significantly. In many instances these ar- cific pathologic change.
tifacts may be difficult or impossible to easily
overcome from patient to patient. Anticipated Results
The initiation of MR imaging within ∼2 min It should be clear to those who perform and
of the i.v. gadolinium injection is important interpret medical images of the spine following Extradural Spine

A8.8.8 A8.8.9
Supplement 8 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 8

A B
Figure A8.8.2 Postoperative recurrent disc herniation. (A) Transverse T1-weighted (TR = 500
msec, TE = 10 msec) image shows mildly hyperintense mass (arrow) at L4-5 on the right side of
the spinal canal anteriorly. (B) i.v. gadolinium enhanced T1-weighted (TR= 500 msec, TE = 10 msec)
image shows rim enhancement surrounding a recurrent disc herniation (arrow).
one or more forms of surgical therapy that the
images are often difficult to interpret, in part
because of the superimposition of the original
disease process, alteration engendered by the
surgery and/or because of a complication of the
surgical procedure. Although long-term expe-
rience in this area is without a doubt helpful in
regard to improving interpretive skills, never-
theless, certain sequellae can be predicted re-
gardless of the interpreter’s background.
Once the normal or expected post-surgical
findings are understood, the subtle and gross
changes that depart from these observations can
be better analyzed. The importance of a high
level of competence in the domain of post-
therapeutic neurodiagnostic imaging is in the
knowledge that the patient returning for restudy
may be acutely in distress or even in medical
danger (e.g., postoperative spondylitis). In fact,
the clinical presentation post-therapeutically
may well be more severe or dire than was
observed pretherapeutically. Thus, an in-depth
appreciation of the broad range of clinicoradi-
ologic possibilities, as presented here, should
place the medical imaging physician in an ex-
cellent position to provide an experienced di-
agnostic evaluation in the patient presenting
with recurrent or new signs and symptoms
following any one of the spectrum of possible
spinal surgical procedures.
Post Surgical
Spinal Evaluation
A8.8.10
Supplement 8
Literature Cited
Ghazi, J., Golimbu, C.N., and Engler, G.L. 1992.
MRI of spinal fusion pseudoarthrosis. J. Com-
put. Assist. Tomogr. 16:324-326.
Grane, P., Josephsson, A., Seferlis, A., and Tullberg,
T. 1998. Septic and aseptic post-operative disci-
tis in the lumbar spine—evaluation by MR im-
aging. Acta Radiol. 39:108-115.
Jinkins, J.R. 1993. Magnetic resonance imaging of
benign nerve root enhancement in the unoper-
ated and postoperative lumbosacral spine.
Neuroimaging Clin. North Am. 3:525-541.
Jinkins, J.R., Osborn, A.G., Garrett, D., Jr., and
Hunt, S. 1993. Spinal nerve enhancement with
Gd-DTPA: MR correlation with the postopera-
tive lumbosacral spine. Am. J. Neuroradiol.
14:383-394.
Shafaiae, F., Bunscuh, C., and Jinkins, J.R. 1997.
The posttherapeutic lumbosacral spine. In Post-
therapeutic Neurodiagnostic Imaging (J.R. Jink-
ins, ed.) pp. 223-243. Lippincott-Raven. Phila-
delphia.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Van Goethem, J.W., Parizel, P.M., van den Hauwe,
L., and DeSchepper, A.M. 1997. Imaging find-
ings in patients with failed back surgery syn-
drome. J. Belge Radiol. 80(2):81-84.
Contributed by J. Randy Jinkins and
David D. Stark
Downstate Medical Center
State University of New York
Brooklyn, New York
Current Protocols in Magnetic Resonance Imaging
Intramedullary Spine Disease UNIT A9.1
One of the most significant impacts of magnetic resonance (MR) has been its ability to
exquisitely depict normal and pathologic anatomy of the spine. Direct acquisitions
acquired in multiple planes coupled with the ability to study the spine with different T1-
and T2-weighted images have enabled critical assessment of the spinal cord and its
surroundings not previously available to the medical imaging specialist. The development
of contrast media has further extended the capability of MR imaging of the spinal cord
by improving its sensitivity and has allowed the use of the method in certain additional
types of pathology. The first protocol (see Basic Protocol) deals with intramedullary
disease, i.e., that involving the cord, the next one deals with extramedullary-intradural
disease. Both can cause myelopathy.
MYELOPATHY BASIC
Our basic protocol consists of a sagittal T1-weighted conventional spin echo, a sagittal PROTOCOL
T2-weighted fast spin echo, and a transverse T2*-weighted gradient echo. Most of the time
we will also add T1-weighted images in the sagittal and transverse planes before and after
administration of gadolinium. It is important that at least one pair of pre- and post-contrast
images be identical to facilitate detection of subtle enhancement. Thus, if fat saturation
is used following administration of contrast, it should also be used at least in one plane
prior to injection of contrast agent. At 1.5 T, we use a dose of 0.1 mmol/kg; on our low
field open magnet, we use 0.2 mmol/kg to get comparable enhancement. The reason for
this is that T1 increases with field strength so that, for a given TR of 500 msec, the low
field acquisition is actually less T1-weighted than a high field acquisition. Coupled to this
is the need for a longer TE to accommodate the longer echo sampling time required for a
lower bandwidth acquisition. The resultant increased T2-weighting competes with one’s
already marginal T1-weighted contrast—hence the need for double dose at lower field.
When the sensory level is fairly specific, we will limit the examination to the cervical or
thoracic spine; for most myelopathy workups; however, we do the complete spinal cord
from foramen magnum to conus. This can usually be accomplished in one acquisition
using phased array coils without loss of signal-to-noise. The cervical and thoracic portions
are then magnified and filmed (or viewed) separately. While we always acquire transverse
images through the cervical spine, we generally only acquire transverse images at the
level of a suspected abnormality in the thoracic cord.
Table A9.1.1 lists the hardware necessary to perform the procedure. Subsequent tables
list imaging parameters appropriate for high field—i.e., 1.0 to 1.5 T. For lower fields, the
echo times (i.e., TE) are generally increased to accommodate lower bandwidths and the
number of acquisitions (Nacq) is generally doubled.
This protocol takes about 30 min from start to finish.
Table A9.1.1 Equipment Parameters
Coil type Circularly polarized (quadrature) neck coil and/or
torso phased array coil
Cardiac gating No
Peripheral gating For safety only
Respiratory gating No
Respirator If required by patient
Oxygen If required by patient
Motion cushions Useful
Use of contrast agents Yes
Intradural Spine
Contributed by William G. Bradley A9.1.1
Current Protocols in Magnetic Resonance Imaging (2001) A9.1.1-A9.1.12
Copyright © 2001 by John Wiley & Sons, Inc.
 Materials 7. If intravenous gadolinium or i.v. sedation will be given, start an i.v. line and attach it
Normal saline (0.9% NaCl; 500-ml bag) securely to the patient so that movement into or out of the magnet will not pull at the
K-50 tubing patient’s arm. Hang a bag of saline and adjust drip to keep it open.
23- to 25-G butterfly needle It is preferable to insert the line prior to imaging and to leave the patient in the magnet,
Intravenous MRI contrast agent (e.g., Magnevist, Omniscan, Prohance, or with no intervening motion, between the scans run before contrast agent injection and those
OptiMark at a dose of 0.1 mmol/kg for high field or 0.2 mmol/kg for low field). run after injection.

NOTE: Be sure that technicians and nurses always have immediate access to any 8. If needed, place a pillow or other support under the head and knees to make the patient
emergency equipment that may be relevant to a given study, or that may be needed for a more comfortable.
particular patient—i.e., crash carts and oxygen. 9. Use the centering light to position the patient (Table A9.1.2) at the threshold of the
magnet and then put him or her into the center of the magnet.
Set up equipment and patient
1. Interview (screen) the patient to ensure that he or she has no contraindications such Once this step has been performed, so long as the patient does not move on the table, the
as cardiac pacemakers or other implants containing ferromagnetic materials. Also be table itself can be moved and then replaced in the same position as before without
sure to find out if the patient has any health conditions that may require the presence jeopardizing the positioning of one scan relative to another.
of special emergency equipment during the scanning procedure, or necessitate any 10. If the patient is unable to hold still, provide an appropriate sedative; if the patient is
other precautions. in pain, provide an appropriate analgesic.
Generally standard screening forms are used for all patients scanned in a magnetic Sequence 1: Rapid positioning pilot (Fig. A9.1.1)
resonance system.
11. To verify patient position, run the pilot (or scout) scan to ensure correct location of
The presence of any ferromagnetic metals may be a health hazard to the patient when he the neck or back in three dimensions, using the imaging sequence given in Table
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact A9.1.2 or similar parameters.
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(1996) for discussion of what implants may be safely scanned using magnetic resonance. This sequence usually consists of three orthogonal planes to allow subsequent localization.
The images are often also used later to determine where to place the saturation pulses and
Patients may be accompanied into the magnet room by a friend or family member, who can to set up coverage of the volume of interest.
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing. Sequence 2: Sagittal T1-weighted conventional spin echo (Fig. A9.1.2)
12. Set the imaging parameters as shown in Table A9.1.3.
2. If the procedure is a research protocol, have the patient sign any necessary consent
form. 13. Use the pilot image to locate the spine in three dimensions to ensure coverage of the
region of interest (e.g., cervical, thoracic, lumbosacral spine).
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
14. Tell the patient (through the speaker system) that you are ready to begin the scan.
that might be found in clothing.
4. Inform the patient about what will occur during the procedure, what he or she will Table A9.1.2 Primary Clinical Imaging Parameters for Pilot Scan (Sequence 1)
experience while in the magnet and how to behave, including the following.
Patient position Supine
a. If earphones or headphones are used to protect the ears from the loud sounds Scan type Gradient Echo
produced by the gradients, the patient will be asked to wear these, but will be able Imaging plane (orientation) Transverse, sagittal, coronal
to communicate with you at any time during the imaging. Central slice or volume center Cervical: on thyroid cartilage
b. The patient will be given a safety squeeze-bulb or similar equipment to request Thoracic: on sternal notch
assistance at any time (demonstrate how this works). Echo time (TE) As short as possible
c. For good results the patient should not talk, and should avoid or minimize Repeat time (TR) As short as possible
swallowing or other movement, during each scan—i.e., as long as the banging Flip angle (FA) 15°
sounds continue. Between scans, talking and swallowing are allowed in most Fields of view (FOVx, FOVy) Cervical: 280 mm, 280 mm
cases, but other motion should be avoided when comparative positional studies Thoracic: 350 mm, 350 mm
are being performed; the patient will be informed when this is the case. Resolution (Δx, Δy) Cervical 1.09 mm, 2.19 mm
Thoracic: 1.37 mm, 2.73 mm
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
Number of data points collected (Nx, Ny) 256, 128
5. Help the patient mount onto the table. Either before or right after the patient lies down, Display matrix (Dx, Dy) 256, 256
set up any triggering devices or other monitoring equipment that is to be used. Slice thickness (Δz) 5 mm
Number of slices 5 in each plane
6. Center the coil over the region of the spine where the key information is desired. Slice gap 1 mm
Make sure that the body is constrained to prevent motion, especially if high resolution Number of acquisitions (Nacq) 1
Intramedullary
Spine Disease scans are to be run. Scan time 30 sec (all 3 planes) Intradural Spine

A9.1.2 A9.1.3
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 A B

C
Figure A9.1.2 Sagittal T1-weighted spin echo image.

Table A9.1.3 Primary Clinical Imaging Parameters for T1-Weighted Spin Echo
(Sequence 2)

Patient position Supine

Scan type Conventional spin echo
Imaging plane (orientation) Sagittal
Central slice or volume center Slice centered on:
Cervical: the third cervical
vertebra
Thoracic: the 6th thoracic vertebra
Echo time (TE) 14 msec
Repeat time (TR) 500 msec
Flip angle (FA) 90°
Fields of view (FOVx, FOVy) Cervical: 260 mm, 260 mm
Figure A9.1.1 Scout images (A) transverse, (B) coronal, and (C) sagittal. Thoracic: 320 mm, 320 mm (may
use rectangular field of view [e.g.,
half or three-quarter field] if
Sequence 3: Sagittal T2-weighted fast spin echo (FSE) (Fig. A9.1.3) available, or tailor to region of
15. Review the pilot scans and ensure the saturation pulses are correctly placed anterior interest)
to or above the slab of interest. Resolution (Δx, Δy) Cervical: 1.02 mm, 1.02 mm
16. Run sequence 3 according to Table A9.1.4. Thoracic: 1.25 mm, 1.25 mm
Number of data points collected (Nx, Ny) 256, 256
Sequence 4: Transverse T2*-weighted gradient echo (Fig. A9.1.4) Display matrix (Dx, Dy) 256, 256
17. Using the midline sagittal T1-weighted image acquired in sequence 2 as a localizer: Slice thickness (Δz) 3 mm
Number of slices 12
a. Cervical spine: acquired stacked images from C1 through T1. Slice gap 1 mm
b. Thoracic spine: acquire single or stacked images through the levels of interest. Number of acquisitions (Nacq) 3
c. Supplement with additional slices according to visible disease present or to clinical Flow compensation Yes (if available)
query. Saturation pulses Superior, inferior, anterior
Scan time 3 min, 20 sec
Intramedullary 18. Run sequence 4 according to Table A9.1.5.
Spine Disease Intradural Spine

A9.1.4 A9.1.5
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Figure A9.1.3 Sagittal T2-weighted fast spin echo image.
Figure A9.1.4 Transverse T2*-weighted gradient echo image.

Table A9.1.4 Primary Clinical Imaging Parameters for T2-Weighted FSE

(Sequence 3) Table A9.1.5 Primary Clinical Imaging Parameters for T2*-Weighted Gradient
Echo (Sequence 4)
Patient position Supine
Patient position Supine
Scan type Fast spin echo
Scan type 3-D gradient echo
Imaging plane (orientation) Sagittal
Imaging plane (orientation) Transverse
Central slice or volume center Centered on area of interest (as in
Central slice or volume center Volume centered on the area of
sequence 2, Table A9.1.3)
interest (as in sequence 2, Table
Echo time (TE) 102 msec A9.1.3)
Echo train length (ETL) 8 Echo time (TE) 16 msec
Repeat time (TR) 4000 msec Repeat time (TR) 35 msec
Flip angle (FA) 90° Flip angle (FA) 5°
Fields of view (FOVx, FOVy) As in sequence 2, Table A9.1.3 Fields of view (FOVx, FOVy) 220 mm, 220 mm
Resolution (Δx, Δy) As in sequence 2, Table A9.1.3 Resolution (Δx, Δy) 0.86 mm, 0.86 mm
Number of data points collected (Nx, Ny) 256, 256 Number of data points collected (Nx, Ny) 256, 256
Display matrix (Dx, Dy) 256, 256 Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 3 mm Slice thickness (Δz) 2-3 mma
Number of slices 12 Number of slices 36
Slice gap 1 mm Slice gap 0
Number of acquisitions (Nacq) 3 Number of acquisitions (Nacq) 1
Flow compensation Yes (if available) Flow compensation Yes (if available)
Saturation pulses Superior, inferior, anterior Saturation pulses No
Scan time 3 min, 12 sec Scan time 5 min, 26 sec
aIf 3 mm, zero interpolate (ZIP) to 1.5 mm.

Intramedullary
Spine Disease Intradural Spine

A9.1.6 A9.1.7
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Sequence 5: Transverse T1-weighted conventional spin echo (Fig. A9.1.5)
19. Using the midline T1-weighted image acquired in sequence 2 as a localizer, position
slices as in step 17 and run sequence 5 according to Table A9.1.6. A B
Sequences 6 and 7: Post contrast sequences (Fig. A9.1.6)
20. Leaving the patient unchanged in position in the magnet, inject the contrast agent,
flush the line with 10 ml saline, and then immediately run “post contrast” sagittal

Figure A9.1.6 Post-contrast T1-weighted spin echo images: (A) sagittal and (B) transverse.

(sequence 6) and transverse (sequence 7) T1-weighted image sequences (using the

parameters in Tables A9.1.3 and A9.1.6, respectively).

TRAUMA ALTERNATE
For trauma cases, the most important reason for the examination is to determine if there PROTOCOL
has been cord hemorrhage. This is best detected using T2*-weighted gradient echo images.
For the trauma protocol, a sagittal T2*-weighted sequence is added (Table A9.1.7) and the
Figure A9.1.5 Transverse T1-weighted spin echo image. enhanced sequences 6 and 7 are dropped.

Table A9.1.6 Primary Clinical Imaging Parameters for T1-Weighted Spin Echo Table A9.1.7 Primary Clinical Imaging Parameters for T2*-Weighted Gradient
(Sequence 5) Echo (Sequence 8)

Patient position Supine Patient position Supine

Scan type Conventional spin echo Scan type 2-D gradient echo
Imaging plane (orientation) Transverse Imaging plane (orientation) Sagittal
Central slice or volume center Centered on the area of interest Central slice or volume center Slice centered on area of interest
(as in sequence 2, Table A9.1.3) (as in sequence 2, Table A9.1.3)
Echo time (TE) 16 msec Echo time (TE) 20 msec
Repeat time (TR) 500 msec Repeat time (TR) 450 msec
Flip angle (FA) 90° Flip angle (FA) 15°
Fields of view (FOVx, FOVy) 200 mm, 200 mm Fields of view (FOVx, FOVy) 260 mm, 260 mm
Resolution (Δx, Δy) 0.78 mm, 1.04 mm Resolution (Δx, Δy) 1.02 mm, 1.35 mm
Number of data points collected (Nx, Ny) 256, 192 Number of data points collected (Nx, Ny) 256, 192
Display matrix (Dx, Dy) 256, 256 Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 4 mm Slice thickness (Δz) 4 mm
Number of slices 23 Number of slices 12
Slice gap 1 mm Slice gap 1 mm
Number of acquisitions (Nacq) 3 Number of acquisitions (Nacq) 3
Flow compensation Yes (if available) Flow compensation Yes (if available)
Saturation pulses Superior, inferior Saturation pulses Superior, inferior, anterior
Intramedullary Scan time 5 min, 25 sec Scan time 4 min, 32 sec
Spine Disease Intradural Spine

A9.1.8 A9.1.9
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Set up equipment and patient
1. Repeat steps 1 to 19 in the Basic Protocol.
Sequence 8: Sagittal T2*-weighted gradient echo
2. Run sequence 8 according to Table A9.1.7.
COMMENTARY
Background Information
Just as the gadolinium chelates have proven
useful in MR imaging of the brain, the same is
true of the spinal cord (Haughton et al., 1999;
Najem et al., 1999). Intramedullary tumors and
inflammatory processes which are perfused
and which have blood-cord barrier breakdown
enhance with gadolinium. The most common
cord tumor is the glioma of which there are two
primary types: astrocytoma and ependymoma
(Haughton et al., 1999; Najem et al., 1999).
Astrocytomas tend to occur higher in the cord
than ependymomas, often involving the cervi-
cal region or even extensive portions of the
cervical and thoracic cord. They are infiltrative
lesions which are difficult to resect surgically
and respond poorly to radiation therapy. Epen-
dymomas tend to occur lower in the cord, near
the conus medullaris, and are generally better
behaved than astrocytomas. Specifically, they
are more easily removed surgically and tend to
have a more favorable response to radiotherapy.
As with other processes in the cord, it is impor-
tant to define the upper and lower extent of
disease. In the case of ependymomas, it is
particularly important to image the entire neu-
raxis as the tumor can seed through the cere-
brospinal fluid (CSF) spaces throughout the
spinal and intracranial subarachnoid spaces
(Haughton et al., 1999; Najem et al., 1999).
Acute tumefactive multiple sclerosis (MS)
can simulate a cord tumor. Often, the key to the
diagnosis is not the additional administration
of gadolinium to the cord lesion but rather an
additional MRI of the brain to search for
periventricular lesions. We have found thin
slice, sagittal, fast FLAIR (fluid attenuated in-
version recovery) to be particularly useful for
this assessment (Hashemi et al., 1995; Palmer
et al., 1999). In the absence of intracranial
findings, tumefactive MS may not be distin-
guishable from cord tumor, either on the basis
of enhancement features or lack of same. In
such cases, it is useful to rescan the patient in
6 or 12 weeks, as acute tumefactive MS tends
to resolve over this period of time while a tumor
would tend to remain unchanged in size or to
Intramedullary grow.
Spine Disease
A9.1.10
While MS is, by definition, a relapsing dis-
ease, it may be difficult to make the diagnosis
definitively on the basis of the initial presenta-
tion. Similarly, it may be difficult to distinguish
the initial presentation of MS from another
demyelinating process such as ADEM (acute
disseminated encephalomyelitis). ADEM is a
monophasic demyelinating process which is
typically found in children and young adults
following an exanthematous viral infection or
vaccination. It is an autoimmune reaction to the
patient’s own white matter and can involve
either the brain or the spinal cord. It tends to
produce somewhat larger lesions than those
seen with multiple sclerosis. Since it is im-
mune-mediated, ADEM responds to steroids.
Syringohydromyelia is a difficult diagnosis
to make by myelography as the overall cord
contour may not be enlarged (Haughton et al.,
1999; Najem et al., 1999). Delayed post-myelo-
gram CT (computed tomography) has been
useful in demonstrating diffusion of contrast
material into the central cavity; however, it is
much less sensitive than MRI. When a syrinx
is identified on an MR study, its upper and lower
margins should be identified. The foramen
magnum should be evaluated for the presence
of low-positioned cerebellar tonsils, as a Chiari
I malformation is often associated with a syr-
inx. Actually, since this form of syrinx is really
enlargement of the ependyma-lined central ca-
nal, the more proper term is “hydromyelia”. If
a syrinx is found without a Chiari I malforma-
tion, then gadolinium should be given, as syr-
inges are often found in association with cord
tumors. These tumors can be cranial or caudad
to the syrinx and can be intra- or extramedullary
(Haughton et al., 1999; Najem et al., 1999).
They are thought to arise from the obstruction
of flow of CSF between the central canal and
the subarachnoid space via the perivascular
spaces of the cord (Fischbein et al., 1999).
Hemangioblastomas are highly vascular tu-
mors which can not only produce a syrinx but
also form a tumor cyst within the cord. When
a hemangioblastoma is found within the cord,
the cerebellum should also be evaluated as
asymptomatic lesions may be harbored there as
Current Protocols in Magnetic Resonance Imaging
well. When found in both locations, the patient
may have von Hippel-Lindau disease.
Trauma can injure the spinal cord through
compression from posterior buckling of the
posterior longitudinal ligament or frank
retropulsion of a fractured vertebral body into
the canal (Davis et al., 1991). The key MR
finding in spinal trauma is cord hemorrhage.
Since hemorrhage may either be identified on
the basis of a short T1 or a short T2, both
T1-weighted spin echo and T2*-weighted gradi-
ent echo images should be performed. (T2-
weighted spin echo images tend to be prone to
motion artifact and the commonly used T2-
weighted fast spin echo sequences tend to mini-
mize susceptibility effects and the detection of
short T2 hemorrhage.) Cord edema without
hemorrhage (i.e., a “bland contusion”) tends to
resolve with minimal (if any) neurologic defi-
cits, while hemorrhagic contusions tend to be
associated with more serious disability. In the
acute or subacute setting, it is also important to
exclude persistent causes of cord compression
(i.e., extruded fragments or retropulsed bony
fragments), which may have been overlooked
on plain films or CT. The late sequellae of
trauma include myelomalacia (gliosis of the
cord), cystic myelomalacia, and frank syrinx
formation. In the extreme, cord transection can
be easily diagnosed by MR.
MR is useful for diagnosing vascular mal-
formations of the cord of which there are three
types: intramedullary arteriovenous malforma-
tions (AVMs), extramedullary (radiculomen-
ingeal or dural) AVMs, and cavernous
angiomas (Haughton et al., 1999; Najem et al.,
1999). As in the brain, cavernous angiomas
generally appear as areas of low signal on
T2-weighted images (due to hemosiderin). In
institutions where the bright CSF, low flip an-
gle, gradient echo technique has been replaced
by a T2-weighted fast spin echo, these magnetic
susceptibility effects may be less obvious (as
noted above). True AVMs of the cord (i.e., nidus
within cord parenchyma) are unusual. Radicu-
lomedullary vascular malformations of the
dura and radicular vessels are more common.
These may produce scalloping of the protein
cord margins. The flow void produced by these
enlarged vessels in the subarachnoid space
must be distinguished from the normal CSF
flow voids noted particularly posterior to the
cord. High cord signal on T2-weighted images
distal to the AVM may represent ischemia sec-
ondary to a vascular steal phenomenon.
Current Protocols in Magnetic Resonance Imaging
Critical Parameters and
Troubleshooting
The most common problem with spine im-
aging is CSF motion artifacts. These can simu-
late vascular flow voids, giving the appearance
of arteriovenous malformations, particularly
on bright CSF gradient echo acquisitions (e.g.,
Table A9.1.5 and Table A9.1.7). In such cases,
T2-weighted fast spin echo imaging has been
shown to have particular utility due to the natu-
ral flow compensation inherent in the multiple
180° pulses. Increasing the bandwidth to lower
the echo spacing and decreasing the time avail-
able for motion dephasing is another trick to
decrease these motion artifacts.
Swapping phase encoding and read direc-
tions such that the phase encoding direction is
craniocaudad rather than anterior-posterior
also serves to minimize CSF motion artifacts
overlying the cord which could potentially
simulate syringohydromyelia.
Patients with severe back pain should be
appropriately medicated with morphine or
Demerol (rather than merely given sedation) to
be able to lie motionless for their study.
Anticipated Results
MRI of the cord should demonstrate abnor-
mal signal intensity corresponding to clinical
symptoms. These symptoms should correlate
with the location of the tracts with the cord.
Literature Cited
Davis, S.J., Teresi, L.M., Bradley, W.G., Ziemba,
M.A., and Blaze, A.E. 1991. Cervical spine hy-
perextension injuries: MR findings. Radiology
180:245-251.
Fischbein, N.J., Dillon, W.P., Cobbs, C., and Wein-
stein, P.R. 1999. The “presyrinx” state: A revers-
ible myelopathic condition that may precede
syringomyelia. A.J.N.R. 20:7-20.
Hashemi, R.H., Bradley, W.G., Chen, D.-Y., Jordan,
J.E., Queralt, J.A., Cheng, A.E., and Henrie, J.N.
1995. Suspected multiple sclerosis: MR imaging
with a thin-section fast-FLAIR pulse sequence.
Radiology 196:505-510.
Haughton, V.M., Daniels, D.L., Czervionke, L.F.,
Williams, A.L., and Rand, S.D. 1999. Cervical
spine. In Magnetic Resonance Imaging 3rd edi-
tion (D.D. Stark and W.G. Bradley, eds.)
pp.1833-1850. Mosby, St. Louis.
Najem, E.S., Bazan, C. III, and Jinkins, J.R. 1999.
Thoracic Spine. In Magnetic Resonance Imag-
ing 3rd edition (D.D. Stark and W.G. Bradley,
eds.) pp. 1851-1882. Mosby, St. Louis.
Palmer, S., Bradley, W.G., Chen, D.-Y., and Patel, S.
1999. Subcallosal striations: An early finding of
MS on sagittal, thin slice, fast FLAIR images.
Radiology 210:149-153.
Intradural Spine
A9.1.11
 Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Key References
Shellock, 1996. See above.
Covers a number of important patient management
issues related to MR imaging, including recom-
mended safety procedures, a list of metallic implants
that have been tested for MR compatibility, and a
list of other sources on MR safety.
Contributed by William G. Bradley
Long Beach Memorial Medical Center
Long Beach, California
Congenital Heart Disease UNIT A10.1
One of the primary roles of cardiac MR imaging has been the assessment of congenital
heart disease. Initially used to define cardiac anatomy as an adjunct to echocardiography,
cardiac MR now plays a pivotal role in both anatomical and functional assessment of
shunts, admixture lesions, transpositions, and the surgical correction of these lesions.
Unlike X-ray angiography, cardiac MR is noninvasive, does not use iodinated contrast
agents or ionizing radiation, and is, therefore, suitable for both children and adults. This
unit presents the basic techniques for the evaluation of congenital heart disease. While
sequence parameters described are meant to be as generic as possible, parameters are
most appropriate for the Siemens 1.5 T Vision or Symphony and may need to be altered
for magnets of different field strengths and manufacturers. Basic Protocol 1 (imaging
congenital heart disease) will take 30 to 45 min to complete, depending upon the
complexity of the heart disease. Alternate Protocol 1 (intracardiac shunt assessment) and
Alternate Protocol 2 (valve assessment), each take an additional ∼15 min to perform when
run with Basic Protocol 1. Basic Protocol 2 (great vessel assessment) will take 30 to 40
min, but can be performed following Basic Protocol 1, for a total examination time of 1
to 1.5 hr.

IMAGING CONGENITAL HEART DISEASE BASIC

PROTOCOL 1
The basic components of the congenital heart MR examinations are (1) a black-blood
technique to assess anatomy, and (2) a bright-blood gradient recalled cine technique to
assess for the presence of intracardiac shunts, to assess for valvular abnormalities, and to
assess ventricular function.

For young children, nonbreath-hold black-blood methods that employ double inversion
recovery half-Fourier single-shot turbo spin echo (HASTE) or fast spin echo (FSE)
techniques are the most optimal. These are not only nonbreath-hold, but are faster than
traditional spin echo (SE) imaging. Segmented k-space cine gradient echo images (fast
low angle shot—FLASH; fast cardiac gated gradient echo—FASTCARD), especially
those that are fully balanced by refocused gradients (true fast imaging with steady state
free precession—true FISP), give better results in a short examination time. Most of these
require breath-hold, but some, such as true FISP, can give adequate images with shallow
respiration.
While black-blood imaging provides an anatomical overview of the heart, cine images
are vital for the diagnosis of many entities including small intracardiac shunts, which may
otherwise be missed, and for the assessment of surgically created shunts and baffles.
Bright-blood cine images can provide qualitative assessment of congenital valvular
disease (aortic stenosis and bicuspid valve, Ebstein’s anomaly) and qualitative assessment
of shunt size by visual assessment of the flow jet.
Additional phase-contrast imaging can help to quantify shunts and can provide velocity
information of the blood flow that can be used to calculate the pressure gradient across a
stenotic valve or aortic coarctation.
Finally, contrast-enhanced techniques are the best means of assessing pulmonic stenoses
in patients with tetralogy of Fallot or congenital rubella, congenital arteriovenous mal-
formations, pulmonary slings, or anomalous pulmonary venous return.
Table A10.1.1 lists the hardware necessary to perform an MR examination for congenital
Intramedullary
heart disease assessment. An MR system with higher gradient strengths (>20 mT/meter) Congenital Heart
Spine Disease Disease

A9.1.12 Contributed by Pamela K. Woodard and Jie Zheng A10.1.1

Current Protocols in Magnetic Resonance Imaging (2001) A10.1.1-A10.1.19
Current Protocols in Magnetic Resonance Imaging Copyright © 2001 by John Wiley & Sons, Inc.
 Table A10.1.1 Equipment Parameters for Cardiac Imaging, Basic Protocol 1 2. Have the patient change into a gown and remove all jewelry. If the procedure is a
research protocol, have the patient sign any necessary consent form.
Coil type Torso phased array coil (or dedicated cardiac
coil, if available) 3. Tell the patient what will occur during the procedure. In particular, instruct the patient
Gradient coil strength At least ≥20 mT/m (or whatever the system how to hold his or her breath.
permits)
a. Headphones or earphones are used to protect the ears from the noise induced by
Cardiac gating Yes, preferably fiberoptic cables
the gradients. Instruct the patient that he or she will be able to hear you via these
Peripheral gating No
headphones and that microphones in the magnet will permit them to talk to you
Respiratory gating No
at any time.
Respirator If required by patient
Oxygen Yes, 2 liters nasal cannula for most patients (to b. Provide adults and older children with a safety squeeze-bulb alarm (or similar
ensure breath-holding >15 sec) device) and demonstrate how this works. (Infants and very young children will be
Motion cushions Under feet, can be used for patient comfort monitored by the nurse anesthetist.)
Use of contrast agents No c. The patient may call out at any time if he or she feels it necessary.
Power injector No At the authors’ institution, patients are instructed to take a breath in, blow it out, take a
Monitoring equipment Heart rate, oxygen saturation, and blood pressure breath in again, relax, and hold it. This allows images to be obtained in mid-respiratory
can be monitored with MRI-compatible cycle (∼15 to 20 sec will be required for each breath-hold for each cine gradient echo
equipment. This is required if i.v. contrast, acquisition). Some of the newer sequences, such as true FISP, may require shorter
anesthesia, or sedation is administered. breath-holds. Assess the need for supplementary oxygen to improve the breath-holding
capacity, and if necessary, administer 2 liters oxygen via nasal cannula. Advise the patient
and faster rise times (≤300 μsec) that allow gated turbo spin echo and segmented cine of the importance of not moving during the acquisition periods and of not taking deep
breaths during the nonbreath-hold acquisitions.
gradient echo imaging is preferable.
4. Have the patient mount onto the table. For some systems, it will be important that the
NOTE: Be sure that technologists and nurses have immediate access to any emergency patient lie within the center of the spine coil, which may be the posterior elements
equipment that may be relevant to a given study, or that may be needed for a particular for the phased array system.
patient, such as crash carts or oxygen.
5. Once the patient is on the table, place ECG (electrocardiogram) leads on the patient
Set up patient and equipment according to manufacturer’s guidelines.
1. Interview and screen the patient to ensure that he or she has no contraindications to
an MR examination. Contraindications include a cardiac pacemaker or defibrillator,
intracerebral aneurysm clip, or ferromagnetic materials in or near vital structures, Table A10.1.2 Primary Clinical Imaging Parameters for Sequence 1 (Turbo
including the eyes. FLASH Scout Scan)

A number of references have been published listing the MR compatibility of specific Patient position Supine
implanted devices (Shellock, 1999). Contacting the manufacturer of a device may also be Scan type Gradient echo
helpful if MR compatibility remains a question. Imaging plane (orientation) 3 transverse, 1 coronal, and 1
In addition, question the patient about a history of claustrophobia, as this is often a reason sagittal (5 images)
for terminating an examination. Central slice or volume center Center of chest
Echo time (TE) As short as possible
If contrast administration is planned, obtain an allergy history. Especially, ask about a Repeat time (TR) As short as possible
history of allergy to Gd-DTPA if contrast agent is used.
Flip angle (FA) 15°
Also, question the patient about health conditions, including those related to his or her Fields of view (FOVx, FOVy) 400 mm, 400 mm
heart disease, that could require emergency equipment during the scanning procedure. Resolution (Δx, Δy) 1.56 mm, 4.17 mm
Determine whether blood oxygenation, pulse or blood pressure monitoring is required in Number of data points collected (Nx, Ny) 256, 96
addition to cardiac gating.
Display matrix (Dx, Dy) 256, 256
If the patient is a young child, arrange to have the child sedated under the auspices of the Slice thickness (Δz) 6–10 mm
hospital’s anesthesia services. Number of slices 5
Generally standard screening forms are used for all patients scanned in a magnetic Slice gap Variable
resonance system. Number of acquisitions (Nacq) 1
Swap read and phase encoding No
Patients may be accompanied into the magnet room by a friend or family member, who can Slice location Variable
sit in the room during the scan and comfort the patient as needed. This companion must
Saturation pulses Not applicable
be screened as well to ensure the absence of loose metal objects on the body or clothing.
ECG gating Yes
Congenital Heart Scan time 5 sec Congenital Heart
Disease Disease

A10.1.2 A10.1.3
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 The types of electrodes available include individual electrodes in a packet of 3 or 4, or 4 Table A10.1.4 Primary Clinical Imaging Parameters for Sequence 3
together on a single patch (Quatrode, InVivo). Electrodes with graphite tips cause the least (T1-Weighted TSE)
artifact. Lead placement may be either on the chest or back. Leads placed on the chest are
less likely to come off if the patient moves and may be more comfortable to the patient. Patient position Supine
Alternatively, placement of leads on the back may reduce motion artifacts related to Scan type Fast spin echo
breathing. This is less of a problem with fiberoptic cables. Make sure that the ECG tracing Imaging plane (orientation) Transverse
shows high (positive) R-waves. If it does not, it is imperative that the leads either be Central slice or volume center Cardiac region to be assessed
repositioned or that the lead polarity be adjusted to alternative options.
Echo time (TE) 12 msec (or minimum)
6. Place a pillow or other support under the knees to make the patient more comfortable. Echo train length (ETL) 9
Repeat time (TR) <900 msec (and <90% of the
7. No i.v. cannula is necessary for a basic congenital heart examination. R-to-R interval)
8. Center the patient’s chest in the phased array coil and steady it with the Velcro or Delay time (TD) after R-wave 0 msec
buckle straps provided by the manufacturer. Make sure that the phased array body Flip angle (FA) 180°a
coil (and spine coil, if necessary) are plugged in. Fields of view (FOVx, FOVy) 300–350 mm, 300r–350r mm,
with r = rectangular field of view,
9. Center the heart to the “0” location with the laser light marker device available on depending on body habitus
most systems and bring the patient into the magnet. Resolution (Δx, Δy) 1.17–1.37 mm, 1.17r–2.73r mm
(depends upon FOV, with r =
10. Once the patient has been centered in the magnet, check again to be sure that the ECG rectangular field of view)
tracing demonstrates sharp positive R-waves for suitable triggering. If not, then with Number of data points collected (Nx, Ny) 256, 128–256, depending on body
the patient still within the magnet, exchange lead polarity until a suitable tracing is habitus
obtained. Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 6–8 mm
Number of slices 1
Table A10.1.3 Primary Clinical Imaging Parameters for Sequence 2 (HASTE) Slice gap Not applicable
Number of acquisitions (Nacq) 1
Patient position Supine Swap read and phase encoding No
Scan type Single shot fast spin echo Saturation pulses No
Imaging plane (orientation) Transverse (may also be run ECG gating Yes
sagittal or coronal)
Scan time 15–20 sec per image
Central slice or volume center Center of heart
aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this
Echo time (TE) 43–63 msec (effective)
sequence is 90°.
Repeat time (TR) Infinity
Delay time (TD) R-to-R interval, use to blank
R-wave for every other beat trigger
Flip angle (FA) 180°
Fields of view (FOVx, FOVy) 300–350 mm, 300–350 mm
Resolution (Δx, Δy) 1.17–1.37 mm, 1.41–1.65 mm A B
(will depend upon FOV)
Number of data points collected (Nx, Ny) 256, 212
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 5–8 mm (children especially
benefit from thinner slice
thickness)
Number of slices 21
Slice gap None
Number of acquisitions (Nacq) 1
Swap read and phase encoding No
Saturation pulses No
Slice series Ascending
ECG gating Yes (can run without gating, but Figure A10.1.1 Transaxial black blood HASTE (A) and cine GRE (B) images at the same cardiac
images are better with gating) level. Image from the cine sequence demonstrates the jet from a small membranous ventricular
Scan time Depends upon volume to be septal defects (VSD; arrow). The VSD cannot be identified on the static black-blood images (RV,
covered, (2 × (number of slices) × right ventricle; LV, left ventricle).
Congenital Heart Congenital Heart
Disease (R-to-R inverval)) Disease

A10.1.4 A10.1.5
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 A B

Figure A10.1.3 Image from a bright-blood cine GRE sequence

C shows a small patent ductus arteriosus (PDA) connecting the descend-
ing aorta and the main pulmonary artery. An arrow points to a small
jet. This PDA was barely visible on black-blood anatomic MR imaging.

RV
LV
A B

RV
LV

Figure A10.1.2 Multiple orthogonal views can be prescribed from initial images. Here a transaxial
black blood image (A) was used to set up a long-axis two-chamber view (B). The long-axis view
was then used to set up a horizontal long-axis view (C) of the heart. (RV, right ventricle; LV, left
ventricle).
Figure A10.1.4 Horizontal long-axis black-blood image can be used to set up multiple short-axis
cine sequences that can be used to provide information about left ventricular function. (RV, right
ventricle; LV, left ventricle).

If still not satisfactory, then bring the patient table out of the magnet, check the lead Sequence 2: 2-D Transverse half-Fourier turbo spin echo (HASTE or double
connections, and reposition the leads until a satisfactory tracing is obtained. inversion recovery FSE)
At the authors’ institution either a single shot HASTE or dual-segment HASTE sequence
Sequence 1: Rapid multi-plane scout is obtained at least transaxially, and often in multiple planes (sagittal and coronal) to cover
11. To assist with placement of subsequent acquisitions, run a gradient echo multi-plane the thorax. The authors have found that the dual-segment sequence provides better
scout scan according to Table A10.1.2. signal-to-noise ratio (SNR). This sequence can be run acquiring ≤5 slices in a single
Some systems will allow the imager to modify the sequence to add groups of images, or to breath-hold, or 21 slices can be performed with shallow breathing. This is an ideal
change image orientation. The authors often run a 5-image turbo FLASH scout. Those with sequence to run for anatomic imaging on pediatric patients. For best image results, slice
faster MR systems can run a true FISP sequence. series (excitation order) should be ascending. The dual segment sequence images often
This is a nonbreath-hold sequence. have an Ny/2 artifact (signal from the anterior chest wall runs through the image). This
can be displaced posteriorly and out of the body by not using a rectangular field of view.
Dual-segment HASTE is a segmented HASTE imaging sequence with two separated
Congenital Heart
acquisitions for one slice. Unlike conventional single-shot HASTE sequence, each Congenital Heart
Disease segmented acquisition acquires half of required k-space data in one cardiac cycle with a Disease

A10.1.6 A10.1.7
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A10.1.5 Primary Clinical Imaging Parameters for Sequence 4 (Cine Table A10.1.6 Clinical Imaging Parameters for Cine GRE, Gradient Refocused
Gradient Echo) (True FISP)

Patient position Supine Patient position Supine

Scan type Segmented k-space 2-D cine Scan type Segmented k-space 2-D cine
gradient echo gradient echo
Imaging plane (orientation) As determined by imager Imaging plane (orientation) As determined by imager
Central slice or volume center Cardiac region to be assessed Central slice or volume center Cardiac region to be assessed
Echo time (TE) 4.8 msec (or minimum) Echo time (TE) 1.6 msec (or minimum)
Number of lines per segment 9 Number of lines per segment 10
Repeat time (TR) 80–100 msec (temporal resolution) Repeat time (TR) 64 msec (temporal resolution)
Delay time (TD) 0 msec Delay time (TD) 0 msec
Flip angle (FA) 20° Flip angle (FA) 60°
Fields of view (FOVx, FOVy) 300–350 mm, 300r–350r mm, Fields of view (FOVx, FOVy) 300–350 mm, 300r–350r mm,
with r = rectangular field of view, with r = rectangular field of view,
depending on body habitus depending on body habitus
Resolution (Δx, Δy) 1.17–1.37 mm, 1.17r–2.73r mm Resolution (Δx, Δy) 1.17–1.37 mm, 1.17r–2.92r mm
(depends on FOV, with r = (depends on FOV, with r =
rectangular FOV) rectangular field of view)
Number of data points collected (Nx, Ny) 256, 128–256, depending on body Number of data points collected (Nx, Ny) 256, 120–256, depending on body
habitus habitus
Display matrix (Dx, Dy) 256, 256 Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 8 mm Slice thickness (Δz) 5–8 mm
Number of slices 1 Number of slices 1
Slice gap Not applicable Slice gap Not applicable
Number of acquisitions (Nacq) 1 Number of acquisitions (Nacq) 1
Swap read and phase encoding No Swap read and phase encoding No
Saturation pulses No Saturation pulses No
Number of cardiac phasesa (R-to-R interval) × 85%/TR Number of cardiac phases (R-to-R interval) × 85%/TR
(~8–15, multiple frames per slice (multiple frames per slice position)
position) ECG gating Yes
ECG gating Yes Scan time 5–10 sec
Scan time 15–20 sec
aSee annotation under Basic Protocol 1, step 14.
Sequence 4: 2-D fast cine gradient echo (GRE) in multiple planes
Cine gradient echo images can be performed at 10-mm intervals through the heart
sequential phase-encoding order. This segmentation reduces the data acquisition window transversely, or selectively. At the authors’ institution, routine cine imaging through the
in each cardiac cycle and thus shortens the TE, which increases signal-to-noise ratio. heart transaxially has proved invaluable in demonstrating small atrioseptal defects or
12. Run sequence 2 according to Table A10.1.3. ventriculoseptal defects that otherwise would not be identified (Fig. A10.1.1). These
transaxial images are also useful in the visual assessment of the mitral and tricuspid
Sequence 3: 2-D T1-weighted turbo spin echo (depending on quality of sequence 2) valves, along with 4-chambered long axis views (Fig. A10.1.2). In addition, selected
13. If the quality of the HASTE images is poor, instruct the patient to hold their breath images can be performed through surgically created shunts (e.g., Blalock-Taussig or
and run a T1-weighted turbo spin echo (TSE) sequence according to the parameters Glenn Shunts) to assess patency or through areas of suspected pathology (e.g., in a region
in Table A10.1.4 transaxially through the heart, or in other imaging planes to further of suspected patent ductus; Fig. A10.1.3). Double oblique sagittal and coronal images are
assess specific regions. obtained through the aorta and pulmonary trunk. Long axis images can be set up off of
the coronal image to show both the aortic and mitral valves in a single plane. Two to three
The HASTE images can then be used to locate specific areas in question and position short axis images (Fig. A10.1.4) can be obtained if ventricular function is a question.
T1-weighted TSE images. The effective (or total) TR should be ∼85% to 90% of (or 100
msec less than) the patient’s R-to-R interval (time between R waves). For T1-weighting, TR 14. Set the imaging parameters as shown in Table A10.1.5. Instruct the patient to hold
should be <900 msec. It should be noted that shorter TR times, however, will limit the their breath and run sequence 4.
number of slices that can be obtained in each acquisition. If one wishes to cover the entire
heart in adults, rather than using the sequence to assess a limited region, it may be This sequence will provide multiple cine frames throughout the cardiac cycle in a single
necessary to run this sequence multiple times. In addition, usually only one image can be slice position per breath-hold. The parameters may be adjusted depending on the patient’s
acquired in a single breath-hold. Either one image can be acquired at a time (this eliminates breath-holding capability and heart rate. For patients with slower heart rates, sequences
Congenital Heart respiratory motion artifact), or the patient can be instructed to perform shallow breathing. that provide a greater number of lines per segment can help to shorten the required Congenital Heart
Disease Disease

A10.1.8 A10.1.9
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 breath-hold. The number of cardiac phases should be set according to one of the following Table A10.1.7 Primary Clinical Imaging Parameters for Sequence 5
formulas: (Through-Plane Phase Contrast)

Number of cardiac phases = (R-to-R interval) × 85%/TR Patient position Supine

or Scan type Cine phase contrast, 2-D gradient
echo
Number of cardiac phases = (R-to-R interval − 100 msec)/TR. Imaging plane (orientation) Single slice through plane of
For imagers with faster magnets, gradient refocused, true FISP-type sequences (Table ascending aorta and pulmonary
A10.1.6) can provide sharp, detailed images with minimal breath-holding. trunk (3–4 cm above valve). Run
the sequence two times, once
through each vessel.
ALTERNATE INTRACARDIAC SHUNT ASSESSMENT Central slice or volume center Aorta or pulmonary trunk
PROTOCOL 1
Gradient recalled echo cine MRI has been demonstrated to be very useful in determining Echo time (TE) 6.5 msec (per rf-pulse)
the presence of atrial septal defect (ASD), ventricular septal defect (VSD) or patent Number of lines per segment 1
foramen ovale (PFO). Cine sequences are vital in diagnosing the presence of intracardiac Repeat time (TR) 28 msec (temporal resolution)
shunts and in performing assessment of the lesion. These should be performed at least Delay time (TD) 0 msec
transaxially throughout the region of interest (ROI). Without cine MR, it is quite possible Flip angle (FA) 30°
to miss a small ASD, VSD, or PFO on black-blood anatomic imaging alone. Bright-blood Fields of view (FOVx, FOVy) 300–320 mm, 300–320 mm,
cine MR not only demonstrates the presence of the lesion, but also demonstrates the extent depending on body habitus
of the lesion and direction of the shunting. Further assessment can be made via additional Resolution (Δx, Δy) 1.17–1.25 mm, 1.17–2.21 mm
cine imaging orthogonal to the jet. Cine imaging in long axis and short axis orientations Number of data points collected (Nx, Ny) 256, 145–256, depending on body
can be used to demonstrate sequelae of the lesion, for instance right ventricular size and habitus
function in the presence of a VSD. Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 6 mm
In addition to gradient recalled echo cine imaging alone, phase contrast sequence can Number of slices 1
confirm shunt direction and provide quantitative information (Didier and Higgins, 1986; Slice gap Not applicable
Schectem et al., 1987). In-plane imaging can provide qualitative assessment. Number of acquisitions (Nacq) 1
Swap read and phase encoding No
Through-plane phase contrast imaging through the ascending aorta and pulmonic outflow Saturation pulses No
tract can be performed to estimate right and left ventricular stroke volumes in order to Number of cardiac phases (R-to-R interval) × 85%/TR
assess the extent of a shunt (shunt ratio). In normal individuals, the left and right multiple phases
ventricular stroke volumes should be equivalent. Stroke volumes can be estimated using Vascular options venc = 250 cm/sec
phase contrast sequences, as described in this unit, or can be assessed using multiple short ECG gating Yes
axis cine sequences through the heart to calculate an exact stroke volume using software Scan time Depends upon heart rate (Ny ×
packages provided by a vendor. Using either method, assessing the difference in left and (R-to-R inverval))
right ventricular stroke volumes can accurately assess the extent of a shunt if no valve
disease is present to introduce additional factors that could alter ventricular stroke volume
5. Use the software provided by most systems for quantitative analysis to convert the
(Didier and Higgins, 1986).
signal obtained into velocity measurements.
Set up patient and equipment 6. Once the images are acquired, use the software provided by the system to place a
1. Repeat Basic Protocol 1. circular ROI to match the diameter of the aorta on each phase image (or, if the system
allows, each magnitude image corresponding to each phase image). The position of
Sequence 5: Through plane cine phase contrast for shunt assessment the ROI should be hand-adjusted on each image to insure accurate placement.
2. To ensure accuracy of velocity measurement and to avoid aliasing, make sure the venc
(velocity encoding anti-aliasing limit) of the sequence is above the assumed velocity 7. Use the software to generate velocity curves and tabulate the area within the specific
of aortic or ventricular flow. Unless a stenosis is present, a venc of 250 cm/sec should ROI.
be adequate. 8. Repeat steps 6 and 7 for the pulmonary trunk.
3. Use adequate temporal resolution—as many phases as possible. 9. For each vessel, calculate the mean velocity in systole. Multipy the average velocity
4. Acquire the sequence described in Table A10.1.7 two times, once through the (systole) by cross-sectional area to estimate the stroke volume. The ratio of stroke
ascending aorta and once through the pulmonary trunk, each 3 to 4 cm above the volumes (left to right) is the shunt ratio.
valve plane. In normal individuals, the left and right ventricular stroke volumes should be equal.
Congenital Heart It is not necessary to ask the patient to hold his or her breath. Note that the ratio may be affected by concomitant valve disease. Congenital Heart
Disease Disease

A10.1.10 A10.1.11
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 ALTERNATE VALVE ASSESSMENT
PROTOCOL 2
Cine images (sequence 4, Table A10.1.5) can be obtained to visually assess the jet from
a stenosis or turbulence caused by regurgitant flow. Aortic and pulmonic valves can be
visually assessed in a double-oblique (oblique sagittal and coronal) orientation (Fig.
A10.1.5). A long-axis view, set up off of the coronal image, can be used to assess the
aortic valve and mitral valve simultaneously. Transaxial view may also help in the
assessment of the mitral and tricuspid valves.
Qualitative evaluations can be made to assess valve stenoses, or aortic and pulmonic
insufficiency using phase contrast sequences.
For stenotic valve pressure gradients, the authors use a through-plane phase-contrast
sequence with a high venc (in order to obtain accurate velocity measurement and avoid
A B
C
Figure A10.1.5 Cine GRE images demonstrate post-stenotic dephasing. (A) A coronal image
shows a domed, bicuspid aortic valve. A jet (arrow) indicates stenosis. (B) In the same patient, the
black area of post-stenotic dephasing or jet (arrow) can be identified on a transaxial cine image.
Congenital Heart (C) A double-oblique image through the pulmonic valve in a patient with tetralogy of Fallot shows
Disease a jet (arrow) distal to subpulmonic soft tissue, indicative of infundibular subpulmonic stenosis.
A10.1.12
Current Protocols in Magnetic Resonance Imaging
aliasing). This is set up through and perpendicular to the “jet” created by the dephased
flow. This usually turns out to be a plane ∼2 cm above and parallel to the valve. Once
images are obtained, the manufacturer’s quantification package usually can be used to
obtain both mean and peak velocity measurements. Follow the manufacturer’s instruc-
tions for use of this package. However, the software will usually allow placement of an
ROI to encompass the jet as seen on either the magnitude or phase images. One will need
to manually adjust the position of the ROI for each frame because of the constant motion
of the heart. One can estimate the mean pressure gradient and peak pressure gradient
across a stenotic valve by using a modified Bernoulli’s equation ΔP = 4v2 (the difference
between pressure, P, is proportional to the square of velocity, v; P is in units of mmHg;
v is peak velocity in units of meters/sec; Didier et al., 2000). A hemodynamically
significant stenosis is usually considered to have a pressure gradient >25 mmHg. Ber-
noulli’s estimation of the pressure gradient from velocity can also be used to assess the
hemodynamic significant of an aortic coarctation.
Regurgitant volume can be estimated, but is affected by other regurgitant or stenotic
valves, therefore, assessment is less reliable. Aortic or pulmonic valve regurgitant vol-
umes can be assessed by directly calculating the difference between the right and left
ventricular stroke volumes using multiple short axis cine sequences or, more simply, by
estimating right and left ventricular stroke volumes using phase contrast sequences. This
is done in a fashion similar to evaluation for shunt assessment. Run the cine phase contrast
sequences through the plane above and parallel to the aortic and pulmonic valve. Calculate
the mean velocity in systole for each using the manufacturer supplied quantification
package, making sure the area of the ROI matches the cross sectional area of the vessel
(i.e., aorta or pulmonary trunk) being measured. Average velocity (systole) multiplied by
cross-sectional area estimates the stroke volume. The difference between right and left
ventricular stroke volume is the regurgitant volume (Chatzimavroudis et al., 1997).
If there is another diseased valve, regurgitant flow can be estimated by calculating the
area under the retrograde volumetric flow curve (not velocity curve) in diastole (Higgins,
2000).
Set up patient and equipment
1. Repeat Basic Protocol 1.
Sequence 6: Through-plane cine phase contrast for valve assessment
2. To ensure accuracy of the velocity measurement and to avoid aliasing, make sure the
venc of the sequence is above the assumed velocity of the flow to be measured. For
insufficiency calculations, a venc of 250 cm/sec should be adequate. A venc of 500
cm/sec should be used for assessment of stenoses.
3. Use adequate temporal resolution to have as many cardiac phases as possible.
4. For insufficiency assessment, acquire the sequence described in Table A10.1.8 two
times, once through the ascending aorta and once through the pulmonary trunk, each
3 to 4 cm above the valve plane. For stenosis assessment, acquire the sequence
perpendicular to the stenotic jet.
This is not a breath-hold sequence.
5. Use the software provided by the manufacturer for quantitative analysis to convert
the signal obtained into velocity measurements as in steps 6 through 8 below.
6. For insufficiency assessment (to determine right and left ventricular stroke volumes),
place a circular ROI to match the diameter of the aorta on each phase image (or if the Congenital Heart
system allows, each magnitude image corresponding to each phase image). The Disease
A10.1.13
Current Protocols in Magnetic Resonance Imaging
 Table A10.1.8 Primary Clinical Imaging Parameters for Sequence 6 volume, for both the aortic and pulmonary artery measurements, calculate the mean
(Through-Plane Phase Contrast) velocity in systole. Average velocity (systole) multiplied by cross-sectional area
estimates the stroke volume. The ratio of stroke volume (left to right) is the shunt
Patient position Supine
ratio.
Scan type Cine phase contrast, 2-D gradient
echo In normal individuals, the left and right ventricular stroke volumes should be equal.
Imaging plane (orientation) For aortic or pulmonic
Note that the calculated regurgitant volume may be affected by disease in other valves, or
insufficiency: through plane of
by the presence of a shunt.
ascending aorta and pulmonary
trunk (3–4 cm above valve). Run
the sequence two times, once ASSOCIATED GREAT VESSEL ASSESSMENT BASIC
through each vessel. PROTOCOL 2
For aortic, mitral or pulmonic
Contrast-enhanced gradient recalled echo sequences can be used for assessment of both
stenosis: perpendicular to jet congenital aortic and pulmonary artery anomalies. This includes assessment of pulmonary
Central slice or volume center Above aortic and pulmonic valve artery stenoses or atresia as seen in tetralogy of Fallot, and aortic arch anomalies such as
plane for insufficiency; through seen in coarctation, William’s syndrome (supravalvular aortic obstruction), or Shone’s
jet for valve stenosis syndrome (congenital hypoplasic left heart syndrome). Assessment of pulmonary embo-
Echo time (TE) 6.5 msec (per rf-pulse) for venc = lism is covered in UNIT A13.1. Performing two scans, one timed to pulmonary artery contrast
250 cm/sec; 5.5 msec for venc = enhancement and the second timed to aortic enhancement, can allow assessment of
500 msec/sec (needed for stenotic complex congenital heart disease using a single contrast bolus.
valves as jet velocity is anticipated
to be high) Table A10.1.9 lists the hardware necessary to perform an MR examination for associated
Number of lines per segment 1 great vessel assessment.
Repeat time (TR) 23 msec (temporal resolution)
Delay time (TD) 0 msec Materials
Flip angle (FA) 30° Normal saline (0.9% NaCl), sterile
Fields of view (FOVx, FOVy) 300–320 mm, 300–320 mm, Gadolinium-based MR contrast agent (e.g., Magnevist, Omniscan, Prohance, or
depending on body habitus OptiMARK)
Resolution (Δx, Δy) 1.17–1.25 mm, 1.17–2.21 mm
Number of data points collected (Nx, Ny) 256, 145–256, depending on body Set up patient and equipment
habitus 1. Set up patient as in Basic Protocol 1, steps 1 to 6; skip step 5.
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 6 mm 2. Question the patient about allergies, especially allergic reaction to prior administra-
Number of slices 1 tion of MR contrast agents.
Slice gap Not applicable
Number of acquisitions (Nacq) 1 Table A10.1.9 Equipment Parameters for Cardiac Imaging, Basic Protocol 2
Swap read and phase encoding No
Saturation pulses No Coil type Torso phased array coil (or
Number of cardiac phases (R-to-R interval) × 85%/TR dedicated cardiac coil, if available)
Vascular options venc = 250 cm/sec for Gradient coil strength ≥20 mT/m (or whatever the
insufficiency; venc = 500 cm/sec system permits)
for stenotic valve assessment Cardiac gating No
ECG gating Yes Peripheral gating No
Scan time Depends upon heart rate (Ny × Respiratory gating No
(R-to-R inverval)) Oxygen Yes, 2 liters nasal cannula for
most patients (to ensure
breath-holding >15 sec)
position of the ROI should be hand-adjusted on each image to insure accurate Motion cushions Under feet, can be used for patient
placement. For stenosis assessment, place a circular ROI to encompass the through- comfort.
plane measurement of the jet. Use of contrast agents Yes
7. Use the software to generate velocity curves and, for stroke volume measurements Power injector Yes, mandatory for dynamic
contrast enhancement of vessels
to calculate regurgitant volumes to tabulate the area within the ROI.
Monitoring equipment Heart rate, oxygen saturation, and
8. For stenosis assessment, determine the mean and peak velocity within the jet. Use blood pressure should be
Congenital Heart the modified Bernoulli’s equation to calculate pressure gradients. For regurgitant monitored with MRI-compatible Congenital Heart
Disease equipment Disease

A10.1.14 A10.1.15
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 12
 3. Instruct the patient in breath-holding techniques. This is especially vital for protocols test bolus, use 2 ml of gadolinium-based i.v. contrast agent, followed by 15 ml sterile
during which i.v. contrast will be administered, as there are limited chances to acquire normal saline. Inject at 2 ml/sec.
a respiratory artifact-free study. Some manufacturers provide an automatic bolus-tracking program and will inform the
4. Inform the patient that you will be administering a contrast agent, and that the patient imager to start the scan when the signal in the desired vessel reaches a prescribed signal
may feel a cold sensation in their arm at the time of infusion. intensity. (The imager usually indicates the desired vessel with ROI placement.) If the
manufacturer has this bolus-tracking program in place, it is not necessary to perform a
The total amount of contrast agent is 40 ml. test bolus. Rather, follow the instructions provided by the manufacturer for contrast
bolus-tracking program.
5. Start a 20- to 22-G antecubital i.v. Attach this line securely to the patient so that
movement into or out of the magnet will not pull at the patient’s arm. This is not a breath-hold sequence.
10. Using images from the test-bolus, determine the time to peak enhancement (each
It is preferable to insert the i.v. line prior to imaging and to leave the patient in the magnet,
with no intervening motion between the scans run before contrast agent injection and those
image takes 1 sec to acquire).
run after injection.
Sequence 8: 3-D contrast enhanced gradient recalled echo MR angiography
6. Repeat Basic Protocol 1, steps 8 and 9. 11. Instruct the patient to hold their breath and run the 3-D gradient recalled MR
7. Perform a multi-image turbo FLASH scout sequence according to the parameters in angiography sequence according to the parameters in Table A10.1.11 without contrast
Table A10.1.2, if not already obtained. in the coronal plane, using a slab volume large enough to cover the desired vessels,
yet small enough to be obtained in a single breath-hold.
8. Perform a transaxial black-blood HASTE or double inversion recovery TSE/FSE
12. Prepare to run the 3-D gradient recalled MR angiography sequence according to the
sequence according to the parameters in Table A10.1.3 through the heart, if not
parameters in Table A10.1.11 with contrast agent. The dose should be ∼0.2 mmol/kg,
already obtained.
at an injection rate of 2 ml/sec. As 2 ml was used for the test dose, 38 ml can be used
for the MR angiogram (this is the remainder of two bottles of contrast agent).
Sequence 7: Test bolus
Calculate the scan delay using the following equation.
9. Perform the test bolus sequence according to the parameters in Table A10.1.10. This
can be performed in any plane that demonstrates the desired vessel(s) well. For the
Table A10.1.11 Primary Clinical Imaging Parameters for Sequence 8 (MR
Angiography)
Table A10.1.10 Primary Clinical Imaging Parameters for Sequence 7 (Test
Bolus) Patient position Supine
Scan type 3-D gradient recalled echo
Patient position Supine Imaging plane (orientation) Coronal
Scan type T1-weighted gradient echo (Turbo Central slice or volume center Great vessels
FLASH), inversion or saturation Echo time (TE) 0.77 msec
recovery preparation Repeat time (TR) 2.2 msec
Imaging plane (orientation) Transverse or sagittal Flip angle (FA) 25°
Central slice or volume center Through vessel to be imaged Fields of view (FOVx, FOVy) 300–350 mm, 300–350 mm,
Echo time (TE) 1.8–3.2 msec (or minimum) depending on body habitus
Repeat time (TR) 2.4–6 msec (effective 1000 msec) Resolution (Δx, Δy) 0.59–0.68 mm, 0.79–0.92 mm
Inversion time (TI) 20 msec Number of data points collected (Nx, Ny) 512, 380, depending on body
Delay time (TD) Not applicable habitus
Flip angle (FA) 8° Display matrix(Dx, Dy) 512, 512
Fields of view (FOVx, FOVy) 300–350 mm, 300–350 mm Slice thickness (Δz) 1.5 mm (interpolated from 3 mm)
Resolution (Δx, Δy) 1.17–1.37 mm, 2.34–2.73 mm Number of slices 64 (interpolated from 32)
Number of data points collected (Nx, Ny) 256, 128 Slab thickness 96 mm (can be less, for a shorter
Display matrix (Dx, Dy) 256, 256 breath-hold)
Slice thickness (Δz) 10 mm Slice gap 0 mm
Number of slices 1 Number of acquisitions (Nacq) 1
Slice gap Not applicable Number of repetitions 1 (or 2)
Number of acquisitions (Nacq) 1 Swap read and phase encoding No
Number of repetitions 50 ZIP 2a Yes (by factor of 2)
Swap read and phase encoding No Saturation pulses No
Saturation pulses No ECG gating No
ECG gating No Scan time 20–27 sec (or 40–54 sec),
Congenital Heart Scan time 50 sec depending on volume covered Congenital Heart
Disease aSlice interpolation. Disease

A10.1.16 A10.1.17
Supplement 12 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 scan delay = (time to peak enhancement) + (injection duration)/2 − ally, if the TR (or TR plus TD) is 100 msec less
(scan time)/2 than the R-to-R interval, imaging will fall in
mid-diastole, during the least cardiac motion.
Time to peak enhancement is obtained from the test bolus; scan time is determined from A cine gradient echo sequence images multiple
the precontrast scan time; and injection duration is the amount of contrast agent to be phases of the cardiac cycle at a single slice
injected divided by the injection rate (e.g., if the contrast agent is to be injected at 2 ml/sec,
position. For optimal use of cine gradient echo
and the amount of contrast agent is 38 ml, the injection duration is 19 sec).
sequences, the number of cardiac phases times
13. Instruct the patient to hold their breath and begin the scan. Inject the contrast agent TR should be ∼85% to 90% (or 100 msec less
at the scan delay calculated in step 12 after the start of the scan. than) of the R-to-R interval.
to surgical repair. Faster and stronger gradients
permit the acquisition of most of this informa-
tion, if not all, in a single MR examination.
Acknowledgments
The authors would like to thank Mehdi
Poustchi-Amin, M.D., Instructor, Mallinckrodt
Institute of Radiology, for providing Figures
A10.1.2 and A10.1.4.

If two vascular phases are to be imaged (e.g., pulmonary arterial and pulmonary venous),
or if the patient has a shunt, two measurements can be acquired, one right after the other,
with a delay interval for a second breath-hold programmed between the two measurements.
Images will be constructed after both measurements are acquired.
14. Use the noncontrast-enhanced images as a mask and subtract them from the contrast-
enhanced images at the console or a satellite system, using vendor-provided software.
COMMENTARY
Background Information
Cardiac MR is ideally suited for congenital
heart disease assessment, providing informa-
tion on both anatomy and function. Often, it
can provide answers to questions that otherwise
could not be answered noninvasively. Echo-
cardiography has long been the principal
method of noninvasive imaging of congenital
heart disease, however, with advances in func-
tional imaging, cardiac MR can provide both
functional assessment of valves, shunts, and
ventricular motion, as well as high resolution
anatomic assessment. Furthermore, MR assess-
ment is not limited to the heart, but can provide
information regarding the aorta and pulmonary
arteries that often cannot be obtained with
transthoracic echocardiography. Information
provided by MR may include assessment of
pulmonary artery stenoses in patients with
tetralogy of Fallot, congenital rubella, or pul-
monic atresia, assessment of the aorta for right-
sidedness versus left-sidedness, or aneurysmal
dilatation, as well as assessment for the pres-
ence of bronchial or chest wall collaterals.
Moreover, 3-D breath-held MR angiography
methods can be used to assess for the presence
of anomalous coronary arteries or congenital
aneurysms, as found in patients with Kawasa-
ki’s disease.
Critical Parameters and
Troubleshooting
Cardiac gating
An adequate ECG tracing with sharp, up-
going R-waves and smaller T-waves is required
Congenital Heart for most cardiac imaging. MR compatible ECG
Disease
Motion and breathing artifacts Literature Cited
Sequences that can be acquired relatively Chatzimavroudis, G.P., Walker, P.G., Oshinski, J.N.,
rapidly, such as multislice HASTE and true Franch, R.H., Pettigrew, R.I., and Yoganathan,
A.P. 1997. Slice location dependence of aortic
FISP, can be obtained during shallow breathing
regurgitation measurements with MR phase ve-
with adequate results. Thus, they are particu- locity mapping. Magn. Reson. Med. 37(4):545-
larly useful in imaging the pediatric patient. 551.
Likewise, cine gradient echo images can be Didier, D. and Higgins, C.B. 1986. Identification
performed with multiple signal averages with- and localization of ventricular septal defect by
out breath-hold. As described in UNIT A11.2, an gated magnetic resonance imaging. Am. J.
anterior saturation band can be placed over the Cardiol. 57:1363-1368.
leads with conductive gel should be used. Leads
subcutaneous fat of the anterior chest wall in Didier, D., Ratib, O., Lerch., and Friedli, B. 2000.
with graphite tips may reduce susceptibility Detection and quantification of valvular heart
order to help reduce breathing-related motion
artifact in the regions of the leads. To reduce disease with dynamic cardiac MR imaging. Ra-
artifact.
unnecessary noise, good electrical contact be- diographics 20:1279-1299.
tween the electrodes and skin is essential. For Higgins, C.B. 2000. Cardiac imaging. Radiology
good skin contact, chest hair may need to be
Anticipated Results 217:4-10.
The goal in studying congenital heart dis-
shaved at the points of electrode contact. If Sechtem, U., Pflugfelder, P., Cassidy, M.C., Holt,
ease with MR is to define the anatomy of the
difficulty persists, the skin may be cleaned with W., Wolfe, C., and Higgins, C.B. 1987. Ventricu-
disease. If the patient has been operated upon, lar septal defect: Visualization of shunt flow and
NuPrep. In patients with congenital heart dis-
the radiologist should look for complications determination of shunt size by cine MR imaging.
ease, the position of the heart and axes of the
known to occur subsequent to the particular AJR Am. J. Roentgenol. 149:689-692.
conduction pathways across patients will vary.
type of surgery performed. In patients who may Shellock, F.G. 1999. Pocket Guide to MR Proce-
This will be especially true in patients who have
have an intracardiac shunt, cine sequences can dures and Metallic Objects. Lippincott-Raven,
enlarged hearts with ventricular apex displace- Philadelphia.
be performed to identify its location. In some
ment or abnormal situs. Situs and other anoma-
instances, the shunt can be quantified. The
lous cardiac anatomy may necessitate right- Key References
valves should be assessed to identify the pres-
sided lead placement or placement of the leads Chen, J.T.T. 1997. Essentials of Cardiac Imaging,
ence of stenosis or insufficiency. The aortic
on the patient’s back rather than chest. Check 2nd ed. Lippincott-Raven, Philadelphia.
valve can be interrogated to determine whether
to see that the leads are positioned correctly This textbook is an excellent reference for the inter-
it is tricuspid (normal) or bicuspid. Pressure pretation of all congenital heart imaging, going over
and, if necessary, vary the lead polarity options
gradients across stenotic valves can be deter- the anatomy, physiology and embryology of each
(I, II, AVR, etc). If gating is still inadequate,
mined. In some instances, valvular insuffi- congenital heart defect in detail.
reposition the ECG leads, using new adhesive
ciency can be quantified. Short axis cine images
pads. If T-waves are larger than the R-waves, Higgins, 2000. See above.
can provide information regarding ventricular
the electrode patch or individual electrodes Provides an overview of multiple modalities and
size and function. Finally, contrast-enhanced
may need to be moved more laterally. If ECG their uses in cardiac imaging, including cardiac
non-ECG gated MR angiography can provide MR.
tracings are no longer visible, check the patient
information about the pulmonary arteries and
to ensure that leads and cable have remained
aorta, assessing pulmonary artery size, the pres-
connected and that there is no break in the cable
ence of pulmonary artery stenoses, aortic dis-
or cable casing. If patient motion is continually Contributed by Pamela K. Woodard and
section, or aortic aneurysm. Three-dimensional
a problem, consider fiberoptic lead use. In the Jie Zheng
(3-D) coronary MR angiography can also be Mallinckrodt Institute of Radiology
authors experience, while fiber optic leads
performed to determine the presence of Washington University Medical Center
eliminate patient motion artifact and reduce
congenitally anomalous coronary arteries of St. Louis, Missouri
MR interference, interference induced by the
importance, especially, if the patient is to go on
rf-pulse is not entirely eliminated.
MR imaging parameters
For most cardiac imaging, the effective TR
must be shorter than the R-to-R interval. Ide- Congenital Heart
Disease

A10.1.18 A10.1.19
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
Assessment of Right Ventricular Dysplasia
Magnetic resonance imaging is currently considered a noninvasive modality of choice for
evaluation of patients with suspected right ventricular dysplasia. It has the unique ability
to provide tissue characterization in addition to providing functional information. This
unit presents the basic techniques for the evaluation of right ventricular dysplasia. The
sequence parameters described are most appropriate for the GE 1.5T Signa, and may need
to be altered for magnets of different field strengths and manufacturers. The Basic
Protocol will take 45 min to 1 hr to complete. Including the optional sequence 8 to the
Basic Protocol will take an additional ∼10 min to perform.
The MR imaging protocol for right ventricular dysplasia is aimed at recognizing two
important aspects of the disease: (a) fibro-fatty infiltration of the right ventricle and (b)
global and regional right ventricular dysfunction. Hence, the protocol includes (a)
black-blood imaging to identify intramyocardial fatty infiltration and (b) bright-blood
cine imaging to visualize right ventricular global and regional dysfunction.
For black-blood techniques, breath-hold imaging with double inversion recovery fast-spin
echo (FSE) techniques are preferred over traditional spin echo (SE) imaging. These
techniques substantially shorten the imaging time and are devoid of respiratory motion
artifacts. Black blood inversion in combination with half-Fourier single-shot turbo-spin
echo (HASTE) imaging has not been systematically evaluated, but is currently not
recommended due to the blurring of detail by this sequence caused by the long echo train
length. For bright-blood imaging, steady-state free precession imaging is the most preferred
technique (fast imaging employing steady state acquisition–FIESTA, true fast imaging
with steady-state-free precession—true FISP, balanced fast field echo—balanced FFE). If
those cine sequences are not available, segmented k-space cine gradient echo images (e.g.,
fast, low-angle shot, FLASH; fast, cardiac-gated gradient echo, FASTCARD) can be used.
All images are optimally performed during end expiratory breath-hold.
Table A10.2.1 lists the hardware necessary to perform the procedure, along with the
appropriate parameters. The available gradient strength will depend on the scanner and
the echo times given in other tables below may be varied accordingly (the smaller the
gradient strength, the longer the echo time for a particular scan).
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.
Table A10.2.1 Equipment Parameters
Coil type Cardiac phased array coil preferred (or
torso phased array coil)
Gradient coil strength 25 mT/m (or whatever the system permits)
Cardiac gating Yes, preferably fiber optic cables
Peripheral gating No
Respiratory gating Yes, to monitor breath hold
Respirator If required by patient
Oxygen If required by patient
Motion cushions NA
Use of contrast agents No
Contributed by Harikrishna Tandri and David A. Bluemke
Current Protocols in Magnetic Resonance Imaging (2003) A10.2.1-A10.2.17
Copyright © 2003 by John Wiley & Sons, Inc.
UNIT A10.2 Materials
Nu-Prep ECG gel
BASIC Vermed patches or other non-metallic MR-compatible ECG pad
PROTOCOL 50 mg Metoprolol (optional)
Set up patient and equipment
1. Interview (screen) the patient to ensure that he or she has no contraindications such
as cardiac pacemakers or other implants containing ferromagnetic materials that may
be problematic for patient safety or good image acquisition. Also be sure to find out
if the patient has any health conditions that may require the presence of special
emergency equipment during the scanning procedure, or necessitate any other pre-
cautions.
Generally, standard screening forms (see APPENDIX 1) are used for all patients scanned in
a magnetic resonance system. The MRI centers should allow 20 min before the scheduled
MRI exam time for patient preparation.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the image quality. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(2001) for a discussion of what implants may be safely scanned using magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing,
as well as other items as described above.
2. If the procedure is a research protocol, have the patient sign any necessary consent
forms.
3. Have the patient remove any metallic objects, including jewelry, dentures, hearing
aids, hairpins, and other objects and secure all items in a safe location. Have the
patient change into a gown to eliminate any metal that might be found in clothing.
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating
and image artifacts.
5. Inform the patient about what will occur during the procedure, what he or she will
experience while in the magnet, and how to behave, including the following:
a. If earplugs, earphones, or headphones are used to protect the ears from the loud
sounds produced by the gradients, the patient will be asked to wear one, but will
be able to communicate with you at any time during the imaging.
b. The patient may be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this works).
c. For good results, the patient should not talk, and should avoid or minimize other
movement during each scan—i.e., as long as the banging sounds continue.
Between scans, talking is allowed in most cases, but should be avoided when
comparative positional studies are being performed; the patient will be informed
when this is the case.
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
e. The procedure will take ∼45 min.
f. The MRI technologist should answer any questions that the patient may have.
6. Have the patient mount onto the table. Either before or right after the patient lies
Assessment of
Congenital Heart Right Ventricular down, set up any triggering devices or other monitoring equipment that is to be used.
Disease Dysplasia
A10.2.1 A10.2.2
Supplement 10 Supplement 10 Current Protocols in Magnetic Resonance Imaging
 Explain to the patient that a soft pad, which has a receiving coil, will be placed on Table A10.2.2 Primary Clinical Imaging Parameters for Sequence 1:
the chest and that it will aid in creating the images. Sagittal Scout

The head or feet first position will depend on the manufacturer (feet first is preferred if Patient position Supine
possible). Scan type 2-D gradient echo
7. Practice breath-holds with the patient while preparing for ECG lead placement. Imaging plane (orientation) Sagittal
Practice a 20-sec breath-hold with the patient until satisfied that the patient under- Pulse sequence database (PSD) 2-D fast gradient echo
stands the importance of breath-holding at resting lung volume. Ask the patient to Central slice or volume center Laser light centered on mid point
between the sternal notch and the
take two deep breaths in and out, take another breath in, and hold it.
xiphoid process
8. Set up ECG leads as follows: Echo time (TE) Minimum (e.g., 1.5 msec)
Receiver bandwidth (RBW) ±31.25 kHz
a. Use an anterior lead placement if possible. Number of lines per segment 32
b. Rub the area where electrodes go with Nu-Prep ECG (electrocardiogram) gel. Repeat time (TR) 1 R-to-R interval
c. Apply gel to gauze and rub skin gently. Flip angle (FA) 15°
d. Remove excess gel and apply the ECG patches. Fields of view (FOVx, FOVy) 400 mm, 400 mm
Resolution (Δx, Δy) 1.56 mm, 2.50 mm
9. Place the cardiac phased array coil over the patient’s chest just below the clavicle so Number of data points collected (Nx, Ny) 256, 160
that the coil lies in the center of the chest and over the patient’s heart, which is the Display matrix (Dx, Dy) 256, 256
region of interest. Secure it with the Velcro or buckle. Slice thickness (Δz) 10 mm
Number of slices 4
10. If needed, place a pillow or other support under the knees to make the patient more Slice gap 8 mm
comfortable. Number of acquisitions (Nacq) 1
11. Align the centering light to the center mark on the anterior coil. Move the patient to Swap read and phase encoding No
the center of the magnet. Number of cardiac phases 1
ECG gating Yes
Once this step has been performed, so long as the patient does not move on the table, the Scan time 20 heart beats (5 heart beats per slice)
table itself can be moved and then returned to the same position as before without
jeopardizing the positioning of one scan relative to another.
12. If the patient is very anxious, give a mild sedative 30 min prior to the procedure (this 17. Repeat step 16 until all slices are covered.
may help reduce the patient’s anxiety). Figure A10.2.1 shows the prescription for transverse black blood sequence.
If the patient is known to have frequent ventricular ectopy, the authors recommend the use The transverse imaging plane provides the best view of the right ventricular anterior wall
of 50 mg oral Metoprolol 1 hr prior to the procedure provided that the patient has no and the right ventricular outflow tract. This imaging plane is useful to demonstrate
contraindications. If ventricular arrhythmias are frequent and will substantially impact intramyocardial hyperintense T1 signals and to evaluate for outflow tract enlargement.
image quality, the exam should be terminated at this point. Transverse black blood images are acquired at the authors’ institution using a double
inversion recovery fast-spin echo sequence with blood suppression. If the center has a
Sequence 1: Sagittal scout dedicated cardiac coil, best results are obtained using the anterior coil elements only
13. Set up the scan parameters for sequence 1 as shown in Table A10.2.2. Upon (posterior coil switched off) to prevent wrap-around artifacts. Also, the authors recommend
completion of the setup, instruct the patient to take in a deep breath and exhale, take using an anterior saturation band as shown in Figure A10.2.2. An example of a good quality
transverse black blood image sequence is shown in Figure A10.2.3.
in another deep breath, and hold it. Run sequence 1 according to Table A10.2.2.
Sequence 3: Transverse bright blood cine images
Sequence 2: Transverse black blood images
18. Set up sequence 3 as shown in Table A10.2.4. Prescribe similarly for the transverse
14. Set up parameters for sequence 2 as shown in Table A10.2.3. Review the sagittal scout
black blood images as in sequence 2, i.e., starting from the diaphragm to the
images and select an image, which shows the mid left ventricle cavity. Place the
pulmonary artery. Inform the patient that a new scan is about to begin.
localizer line along the orientation of the transverse plane and apply grid lines
beginning from the diaphragm to the pulmonary artery. 19. Instruct the patient to take in a deep breath and exhale, ask the patient to repeat, take
in another deep breath, and hold it. Run sequence 3 according to Table A10.2.4 with
15. Tell the patient that the next scan involves a series of ∼20-sec breath-holds.
the number of slices set to one slice.
16. Instruct the patient to take in a deep breath and exhale, repeat, take in another deep
20. Repeat step 19 until the entire region of interest is covered, i.e., all slices are covered.
breath, and hold it. Run sequence 2 according to Table A10.2.3 with the number of
slices set to one. Cine imaging in the transverse plane is optimal to assess right ventricular global and
Assessment of
regional function visually. One can also assess the right ventricular and atrial chamber
Congenital Heart Right Ventricular enlargement in this plane. As previously mentioned, the most preferred method of cine
Disease Dysplasia imaging is using steady-state free precession technique. This allows better endocardial
A10.2.3 A10.2.4
Current Protocols in Magnetic Resonance Imaging Supplement 10 Supplement 10 Current Protocols in Magnetic Resonance Imaging
 Table A10.2.3 Primary Clinical Imaging Parameters for Sequence 2:
Transverse Black Blood Images

Patient position Supine

Scan type Double inversion recovery fast-spin
echo
Imaging plane (orientation) Transverse
Pulse sequence database (PSD) FSE-xl
Central slice or volume center Center of heart
Echo time (TE) Minimum-full (e.g., 5–10 msec)
Receiver bandwidth (RBW) ±62.5 kHz
Echo train length (ETL) 24 (preferred) to 32
Repeat time (TR) 2 R-to-R intervals
Inversion time (TI) ∼625 msec
Flip angle (FA) NAa
Fields of view (FOVx, FOVy) 260 mm, 260 mm (depends on patient
size and available signal)
Resolution (Δx, Δy) 1.02 mm, 1.02 mm
Number of data points collected (Nx, Ny) 256, 256
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 5 mm
Number of slices Variable (usually 8–9 in total)
Slice gap 5 mm
Number of acquisitions (Nacq) 1
Swap read and phase encoding No Figure A10.2.1 Prescription of transverse black blood images on sagittal scout image.
Slice locations From pulmonary bifurcation to dome
of diaphragm
ZIP 512 Yes
Saturation pulses Anterior saturation band over
subcutaneous fat
Fat suppression No
ECG gating Yes
Scan time ∼2–3 min (13–18 sec per slice)
aThe flip angle of the inversion pulse is 180°.

definition when compared to fast gradient echo. For best results while using steady-state
free precession in GE scanners, one should manually inspect resonance peaks to insure
that the center frequency is water.
An example of good quality image obtained from the transverse bright blood sequence is
shown in Figure A10.2.4.

Sequence 4: Vertical long axis scout

21. Review the transverse black blood images and select an image that shows both the
left ventricular cavity and the left atrium, in order to place the localizer in the long
axis of the left ventricle parallel to the interventricular septum. Draw the grid line
from the left ventricular apex to the mid mitral valve plane, and parallel to the Figure A10.2.2 Sagittal scout image showing the positioning of an anterior saturation band for
interventricular septum as shown in Figure A10.2.5. Instruct the patient to take in a the transverse black blood images.
deep breath and exhale, repeat, take in another deep breath, and hold it. Run sequence
4 according to the parameters shown in Table A10.2.5.

Sequence 5: Four-chamber cine scout

22. On the diastolic vertical long axis cine image obtained from sequence 4, place the Assessment of
Congenital Heart Right Ventricular
localizer along the long axis of the left ventricle. Draw the grid line from the left Disease Dysplasia

A10.2.5 A10.2.6
Current Protocols in Magnetic Resonance Imaging Supplement 10 Supplement 10 Current Protocols in Magnetic Resonance Imaging
Figure A10.2.3 An example of a good quality transverse black blood sequence image. Note the
use of anterior coil elements only and the use of an anterior saturation band

Table A10.2.4 Primary Clinical Imaging Parameters for Sequence 3:

Transverse Bright Blood Cine Images
Figure A10.2.4 An example of a transverse bright blood cine sequence image.
Patient position Supine
Scan type Cardiac cine gradient echo
Imaging plane (orientation) Transverse
Pulse sequence database (PSD) 2-D FIESTA/true-FISP/balanced FFE
Central slice or volume center Center of heart
Echo time (TE) Minimum (e.g., 1.3 msec)
Receiver bandwidth (RBW) ±125 kHz
Number of lines per segment 12–16
Repeat time (TR) ∼30–40 msec (temporal resolution)
Delay time (TD) Minimum (i.e, 10 msec)
Flip angle (FA) 45°
Fields of view (FOVx, FOVy) 360 mm, 360 mm
Resolution (Δx, Δy) 1.41 mm, 2.25 mm
Number of data points collected (Nx, Ny) 256, 160
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 8 mm
Number of slices Variable (usually 5–7 in total)
Slice gap 4 mm
Number of acqusitions (Nacq) 1
Swap read and phase encoding No
Slice locations From pulmonary bifurcation to dome
of diaphragm
Saturation pulses None
Fat suppression No
Number of cardiac phases 20–30 (heart rate dependent)
Figure A10.2.5 Prescription of a vertical long axis view of the left ventricle from a transverse
ECG gating Yes image.
Scan time ∼1 min (10–14 R-to-R intervals per Assessment of
Congenital Heart Right Ventricular
slice) Disease Dysplasia

A10.2.7 A10.2.8
Current Protocols in Magnetic Resonance Imaging Supplement 10 Supplement 10 Current Protocols in Magnetic Resonance Imaging
 Table A10.2.5 Primary Clinical Imaging Parameters for Sequence 4: Vertical Long Table A10.2.6 Primary Clinical Imaging Parameters for Sequence 5:
Axis Scout Four-Chamber Cine Scout

Patient position Supine Patient position Supine

Scan type Gradient echo Scan type Cardiac cine gradient echo
Imaging plane (orientation) Oblique sagittal Imaging plane (orientation) Oblique coronal
Pulse sequence database (PSD) Fast gradient echo Pulse sequence database (PSD) 2-D FIESTA/true-FISP/balanced FFE
Central slice or volume center Center of left ventricle Central slice or volume center Center of left ventricle
Echo time (TE) Minimum (e.g., 1.3 msec) Echo time (TE) Minimum (e.g., 1.3 msec)
Receiver bandwidth (RBW) ±32 kHz Receiver bandwidth (RBW) ±125 kHz
Number of lines per segment 32 Number of lines per segment 12–16
Repeat time (TR) Minimum (∼128–320 msec, temporal Repeat time (TR) ∼30−40 msec (temporal resolution)
resolution) Delay time (TD) Minimum (e.g., 10 msec)
Delay time (TD) Minimum (e.g., 10 msec) Flip angle (FA) 45°
Flip angle (FA) 15° Fields of view (FOVx, FOVy) 360 mm, 360 mm
Fields of view (FOVx, FOVy) 360 mm, 360 mm Resolution (Δx, Δy) 1.41 mm, 2.25 mm
Resolution (Δx, Δy) 1.41 mm, 2.81 mm Number of data points collected (Nx, Ny) 256, 160
Number of data points collected (Nx, Ny) 256, 128 Display matrix (Dx, Dy) 256, 256
Display matrix (Dx, Dy) 256, 256 Slice thickness (Δz) 8 mm
Slice thickness (Δz) 8 mm Number of slices 1
Number of slices 1 Slice gap NA
Slice gap NA Number of acquisitions (Nacq) 1
Number of acquisitions (Nacq) 1 Swap read and phase encoding No
Swap read and phase encoding No Number of cardiac phases 20–30 (heart rate dependent)
Number of cardiac phases 1 ECG gating Yes
ECG gating Yes Scan time 8–11 sec (10–14 R-to-R intervals)
Scan time 4 heart beats

Assessment of
Figure A10.2.6 Prescription of a four-chamber view from the vertical long axis image. Congenital Heart Right Ventricular Figure A10.2.7 Prescription of short-axis cine sequence on the four-chamber diastolic image.
Disease Dysplasia Note that the coverage starts from the mitral valve plane to the apex of the ventricle.

A10.2.9 A10.2.10
Current Protocols in Magnetic Resonance Imaging Supplement 10 Supplement 10 Current Protocols in Magnetic Resonance Imaging
 Table A10.2.7 Primary Clinical Imaging Parameters for Sequence 6: Table A10.2.8 Primary Clinical Imaging Parameters for Sequence 7:
Short-Axis Cine Short-Axis Black Blood Images

Patient position Supine Patient position Supine

Scan type Cardiac cine gradient echo Scan type Double inversion recovery fast-spin
Imaging plane (orientation) Oblique transverse echo
Pulse sequence database (PSD) 2-D FIESTA/true-FISP/balanced FFE Imaging plane (orientation) Oblique transverse
Central slice or volume center Center of heart Pulse sequence database (PSD) FSE-xl
Echo time (TE) Minimum (e.g., 1.3 msec) Central slice or volume center Center of heart
Receiver bandwidth (RBW) ±125 kHz Echo time (TE) Minimum-full (e.g., 5–10 msec)
Number of lines per segment 12–16 Receiver bandwidth (RBW) ±62.5 kHz
Repeat time (TR) ∼30−40 msec (temporal resolution) Echo train length (ETL) 24 (preferred) to 32
Delay time (TD) Minimum (e.g., 10 msec) Repeat time (TR) 1 or 2 R-to-R intervals
Flip angle (FA) 45° Inversion time (TI) TI set by system (e.g., for TR of 2
Fields of view (FOVx, FOVy) 360 mm, 360 mm R-to-R intervals, TI is 600 msec; for 1
Resolution (Δx, Δy) 1.41 mm, 2.25 mm R-to-R interval, TI is 200 msec)
Number of data points collected (Nx, Ny) 256, 160 Flip angle (FA) NAa
Display matrix (Dx, Dy) 256, 256 Fields of view (FOVx, FOVy) 260 mm, 260 mm (depends on patient
size and available signal)
Slice thickness (Δz) 8 mm
Resolution (Δx, Δy) 1.02 mm, 1.02 mm
Number of slices Variable (usually 6–8 in total)
Number of data points collected (Nx, Ny) 256, 256
Slice gap 4 mm
Display matrix (Dx, Dy) 256, 256
Number of acquisitions (Nacq) 1
Slice thickness (Δz) 5 mm
Swap read and phase encoding No
Number of slices 5 (total)
Slice locations From the mitral plane to the apex of
the heart Slice gap 5 mm
Number of cardiac phases 20–30 (heart rate dependent) Number of acquisitions (Nacq) 1
ECG gating Yes Swap read and phase encoding No
Scan time ∼1–2 min (10–14 R-to-R intervals per Slice locations From the mitral plane to the apex of
slice) the heart
ZIP 512 Yes
Saturation pulses Anterior saturation band over
subcutaneous fat
ECG gating Yes
Scan time ∼1 min (13–18 sec per slice when TR =
2 R-to-R interval); ∼40 sec (6–9 sec
per slice when TR = 1 R-to-R interval)
aThe flip angle of the inversion pulse is 180°.

ventricular apex to the mid mitral valve plane as shown in Figure A10.2.6. Set up
parameters of sequence 5 as shown in Table A10.2.6. Instruct the patient to take in a deep
breath and exhale, repeat, take in another deep breath, and hold it. Run sequence 5.
Other than the scan plane, slice location, and number of slices, the parameters are identical
to sequence 3.
On the long axis scout image prescribe four-chamber scout cine image as shown in Figure
A10.2.6. This view provides a four-chamber view of the heart, similar to the echocardiog-
raphic four-chamber view. In this view one can assess the relative chamber sizes of the left
and right ventricle and ventricular function visually.
Sequence 6: Short-axis cine
23. On the diastolic four-chamber scout image obtained from sequence 5, place the
localizer line parallel to the mitral valve plane and perpendicular to the long axis of
Figure A10.2.8 An example of short-axis bright blood cine images. Note that the coverage starts the left ventricle. Draw the grid lines at the mitral valve plane and extend them up to
Assessment of
from the mitral valve plane to the apex of the left ventricle. Congenital Heart Right Ventricular the left ventricular apex as shown in Figure A10.2.7. Set up parameters as shown in
Disease Dysplasia Table A10.2.7 with number of slices set to one.
A10.2.11 A10.2.12
Current Protocols in Magnetic Resonance Imaging Supplement 10 Supplement 10 Current Protocols in Magnetic Resonance Imaging
 Table A10.2.9 Primary Clinical Imaging Parameters for Sequence 8:
Transverse Black Blood Images with Fat Suppression

Patient position Supine

Scan type Double inversion recovery fast-spin
echo
Imaging plane (orientation) Transverse
Pulse sequence database (PSD) FSE-xl
Central slice or volume center Center of heart
Echo time (TE) Minimum-full (e.g., 3.1 msec)
Receiver bandwidth (RBW) ±62.5 kHz
Echo train length (ETL) 24 (preferred) to 32
Repeat time (TR) 2 R-to-R intervals
Inversion time (TI) ∼625 msec
Flip angle (FA) NAa
Fields of view (FOVx, FOVy) 260 mm, 260 mm (depends on patient
size and available signal)
Resolution (Δx, Δy) 1.02 mm, 1.02 mm
Number of data points collected (Nx, Ny) 256, 256
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 5 mm
Number of slices Variable (usually 8–9 in total)
Slice gap 5 mm
Number of acquisitions (Nacq) 1
Figure A10.2.9 A short-axis black blood image. Swap read and phase encoding No
Slice locations From pulmonary bifurcation to dome
24. Instruct the patient to take in a deep breath and exhale, repeat, take in another deep of diaphragm
breath, and hold it. Run sequence 6 according to Table A10.2.7. ZIP 512 Yes
Saturation pulses Chemical shift fat suppression
25. Repeat step 24 until the entire region of interest is covered. Fat suppression Yes
The parameters in Table A10.2.7 are similar to those in Table A10.2.4. An example of short ECG gating Yes
axis cine images is shown in Figure A10.2.8. Coverage for this sequence should start from Scan time 2–3 min (13–18 sec per slice)
the mitral valve plane to the apex of the heart. This is important as this view is often used aThe flip angle of the inversion pulse is 180°.
for quantification of ventricular volumes. Manual prescan should be performed in GE
scanners to optimize image quality.
29. Instruct the patient to take in a deep breath and exhale, repeat, take in another deep
Sequence 7: Short-axis black blood images breath, and hold it. Run sequence 8 according to Table A10.2.9 with number of slices
26. Instruct the patient to take in a deep breath and exhale, repeat, take in another deep set to one.
breath, and hold it. Run sequence 7 according to Table A10.2.8 with number of slices 30. Repeat step 29 until the entire region of interest is covered.
set to one.
The prescription for this sequence is exactly the same as in sequence 2.
27. Repeat step 26 until the entire region of interest is covered.
An example of good quality transverse black blood fat suppressed image is shown in Figure
The parameters for this sequence are similar to those used in sequence 2. A10.2.10.
An example of good quality short axis black blood images is shown in Figure A10.2.9.
This sequence is optional and usually requires an additional 10 min to the total imaging
This view is useful to assess the inferior (diaphragmatic) border of the right ventricle. Care time. The authors recommend performing this sequence if time permits in all patients. This
should be taken to adjust the field of view appropriately to minimize wrap artifacts. In sequence is particularly useful in patients who have abundant epicardial fat, which makes
scanners with a surface coil, using anterior coil elements only and switching off the it difficult to distinguish the true myocardial border. Also in patients who have had
posterior coil to reduce wrap artifacts is recommended. sub-optimal image quality in the non-fat-suppressed transverse black blood images,
repeating the sequence with fat suppression may improve confidence in diagnosis.
Sequence 8: Transverse black blood images with fat suppression (optional)
Fat-suppressed transverse black blood images are acquired at the authors’ institution using
28. Set up imaging parameters for this sequence as shown in Table A10.2.9. Copy the
double inversion recovery fast-spin echo sequence. If the center has a dedicated cardiac
prescription from sequence 2 on to this sequence. Make sure the option of fat coil, best results are obtained using the anterior coil elements only (posterior coil switched
suppression is chosen. Manually align the imaging center frequency to water peak, Assessment of off) to prevent wrap around artifacts. Also, the authors recommend using an anterior
while maximizing fat suppression. Congenital Heart Right Ventricular
Disease Dysplasia
saturation band as shown in Figure A10.2.2.

A10.2.13 A10.2.14
Current Protocols in Magnetic Resonance Imaging Supplement 10 Supplement 10 Current Protocols in Magnetic Resonance Imaging

Figure A10.2.10 Transverse black blood images with fat suppression. Note the use of anterior
coil elements only.
COMMENTARY
Background Information
Arrhythmogenic right ventricular dysplasia
(ARVD) is a heritable cardiomyopathy charac-
terized by fibro-fatty infiltration of the right
ventricular (RV) myocardium (Marcus et al.,
1982). Affected individuals have increased in-
cidence of sudden death due to malignant ven-
tricular arrhythmias. The disease is more com-
mon in men and is a substantial cause of sudden
deaths in athletes <35 years old (Thiene et al.,
1988). Fibro-fatty infiltration of the RV leads to
progressive RV failure and, in the late stages of
the disease process, the left ventricle may also
be involved (Marcus et al., 1982; Thiene et al.,
1988). The diagnosis of ARVD is based on a set
of major and minor criteria encompassing struc-
tural, electrocardiographic and histological cri-
teria proposed by the Task Force of the Working
Group on Cardiomyopathies in 1994 (McKenna
et al., 1994). As such, the accurate evaluation of
right ventricular anatomy and function is crucial
to diagnosis of this disease.
Conventional non-invasive imaging modali-
ties like the echocardiogram and radionuclide
ventriculography have several limitations, es-
pecially in patients with abnormal RV geome-
Current Protocols in Magnetic Resonance Imaging
try. By contrast, magnetic resonance (MR) im-
aging provides excellent anatomic details of the
RV with high spatial and temporal resolutions.
Also, it alone has the ability to depict tissue
characterization non-invasively. For the above
reasons, MR imaging is considered the non-in-
vasive imaging modality of choice to evaluate
ARVD (Pennell and Casolo, 1997). The im-
proved contrast between the blood pool and the
myocardium allows accurate and highly repro-
ducible quantitation to be performed making it
ideal for follow-up evaluation and to study the
disease progression.
Critical Parameters and
Troubleshooting
Cardiac gating
An adequate ECG tracing is crucial for car-
diovascular MR imaging. For best results the
authors recommend the use of the NuPrep gel
to clean the skin prior to the application of the
MR-compatible ECG leads. Excess hair needs
to be shaved in order to ensure adequate skin
contact. The use of fiber-optic leads decreases
Congenital Heart
patient motion artifacts and yields better results.
Disease
A10.2.15
Supplement 10
Cardiac arrhythmias
A significant proportion of patients being
evaluated for ARVD have frequent ventricular
ectopy, which can be problematic. The follow-
ing is recommended to improve image quality
in these patients.
Ideally, if the patient is known to have fre-
quent ectopy, a beta blocker (50 mg Meto-
prolol) given orally 1 to 2 hr prior to the scan
minimizes ectopy. Often this information be-
comes apparent while the patient is on the
scanner table. If the ectopy are frequent, then
the resulting scan quality will be poor and the
test becomes uninterpretable. In this case, it is
prudent to reschedule the scan at a later time on
the same day if the schedule permits and to give
the patient a beta blocker while waiting. If it is
not possible to reschedule the scan on the same
day, then do not perform the scan and resched-
ule it for another day with adequate beta blocker
pre-medication.
Breathing artifacts
All images should be obtained during the
breath-hold to minimize motion artifacts. Ade-
quate patient education to make sure that the
patient understands the importance of breath
holding is invaluable. The patient should be told
that movement and/or speaking during the
exam will cause the images to be less than
optimal, and if at all possible to refrain from
such activity during the scan. For pediatric
patients performing exams without breath
holding, shallow breathing is recommended. In
spite of all this, a proportion of patients fail to
hold their breath. In pediatric patients, an alter-
native is to insure the number of signal aver-
ages, e.g., to 4, and perform the exam without
breath-holding. Inversion prepared HASTE se-
quences are an alternative. In both cases, the
diagnostic value of black blood images is sub-
stantially reduced. The use of an anterior satu-
ration band over the subcutaneous fat of the
anterior chest wall will help reduce breathing-
related motion artifacts.
Anticipated Results
The diagnosis of ARVD is made in the pres-
ence of major and minor criteria proposed by
the Task Force of cardiomyopathies (McKenna
et al., 1994). These criteria encompass electri-
cal and morphologic abnormalities of the right
ventricle. The goal of MR imaging in ARVD is
to accurately depict RV structure and function.
MR abnormalities that are currently included
Assessment of
Right Ventricular in the Task Force criteria are RV chamber en-
Dysplasia largement and RV global and/or regional dys-
A10.2.16
Supplement 10
function. Intramyocardial fat on MR imaging
is currently not recognized as a criteria for
ARVD by the Task Force, as the reliability of
this finding is unknown. When evaluating RV
structure, special attention should be paid to the
RV inflow, RV apex, and the outflow tract, as
these are the commonly involved sites in
ARVD. Fat infiltration can often be difficult to
visualize, as epicardial fat is normally present
abundantly near the atrio-ventricular groove
and near the RV apex. A clear line of demarca-
tion often exists between the epicardial fat and
the RV myocardium in normal subjects and a
disruption of this line is often seen in patients
with ARVD. There is also increased trabecula-
tion of the RV with a prominent moderator band
in patients with ARVD. Assessing RV function
visually can be very challenging as the RV has
a complex morphology and the contraction
pattern differs from the left ventricle. The
authors recommend assessing RV function
from the transverse bright blood cine images.
The RV on transverse cine images starts off as
a large triangle in diastole, which becomes a
smaller triangle in systole. Most of the contrac-
tion occurs in the long-axis of the RV from the
movement of the tricuspid valve towards the
RV apex. The RV anterior wall should be ex-
amined for bulges and aneurysms, as these are
morphologic abnormalities associated with
ARVD. Whenever feasible, quantification of
RV volumes and ejection fraction should be
performed. Quantification of RV volumes can
be performed on the short-axis bright blood
images using MASS software (Medis). Endo-
cardial margins of the end diastolic and end
systolic images are manually contoured and
volumes calculated using the “Simpson’s” rule.
These are more reproducible than qualitative
estimates and may increase sensitivity of MR
imaging.
Acknowledgements
This work was supported in part by the Johns
Hopkins ARVD center (funded by a private
grant from the Bogle Foundation) and National
Institutes of Health Research Grant 1 UO1
HL65594-01A1 (D. Bluemke, PI, MRI ARVD
reading center). The authors would like to thank
Scott Pride, R.T., and Donna Green, R.T. for
providing helpful comments.
Literature Cited
Marcus, F.I., Fontaine, G.H., Guiraudon, G., Frank,
R., Laurenceau, J.L., Malergue, C., and Grosgo-
geat, Y. 1982. Right ventricular dysplasia: A
report of 24 adult cases. Circulation 65:384-
398.
Current Protocols in Magnetic Resonance Imaging
McKenna, W.J., Thiene, G., Nava, A., Fontaliran, F.,
Blomstrom-Lundqvist, C., Fontaine, G., and
Camerini, F. 1994. Diagnosis of arrhythmogenic
right ventricular dysplasia/cardiomyopathy. Br.
Heart J. 71:215-218.
Pennell, D. and Casolo, G. 1997. Right ventricular
arrhythmia: Emergence of magnetic resonance
imaging as an investigative tool. Eur. Heart J.
18:1843-1845.
Thiene, G., Nava, A., Corrado, D., Rossi, L., and
Pe nnelli, N. 1988. Right ventricular
cardiomyopathy and sudden death in young peo-
ple. N. Engl. J. Med. 318:129-133.
Current Protocols in Magnetic Resonance Imaging
Internet Resources Pericardial Disease UNIT A11.1
http://www.arvd.com
An Internet site providing an overview of clinical
features of ARVD along with a current MRI protocol Although echocardiography is considered the primary technique for evaluating the
at the authors’ institution. pericardium, magnetic resonance imaging (MRI) may be indicated in patients who have
poor acoustic windows or who have complex pericardial disease. This unit presents the
basic technique for evaluating nonneoplastic disease of the pericardium with optional
Contributed by Harikrishna Tandri and contrast-enhanced sequences for evaluating suspected neoplastic disease. The parameters
David A. Bluemke are based on experience on a Siemens 1.5 T Vision or Symphony and should be altered
The Johns Hopkins University School of
Medicine
accordingly for different field strengths and manufacturers.
Baltimore, Maryland
IMAGING OF NONNEOPLASTIC PERICARDIUM BASIC
PROTOCOL
To image the pericardium, cardiac imaging packages with gated echo-train fast (or turbo)
spin-echo and segmented k-space cine gradient-echo images give better results in a
significantly shorter examination time.
The core components of the imaging protocol are gated T1-weighted and T2-weighted
turbo spin-echo images of the heart and pericardium, supplemented by a series of cine
gradient echo images. The spin-echo images are used to provide high-resolution imaging
of the pericardium and pericardial space and to help characterize effusions and masses.
The electrocardiogram (ECG)–gated cine gradient-echo images generally provide views
of a single two-dimensional (2-D) slice at multiple phases of the cardiac cycle. When
viewed in the cine mode, these can be used to evaluate the motion of the heart dynamically.
Cine MR images can help detect compression or diastolic collapse of the right side of the
heart in cardiac tamponade and abnormal septal motion in patients with constrictive
pericarditis. With fast breath-hold cine acquisitions, the total examination time for the
Basic Protocol should be <45 min.
For patients with pericardial masses or possible neoplastic involvement of the pericar-
dium, additional contrast-enhanced imaging of the pericardium can be performed by
repeating T1-weighted imaging following intravenous gadolinium chelate.
Table A11.1.1 lists the hardware necessary to perform MRI of the pericardium. Higher
gradient strengths with imaging sequences that allow gated turbo spin-echo and seg-
mented cine gradient-echo imaging are preferable.
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as a crash cart or oxygen. Also ensure only magnetic field compatible oxygen
tanks are utilized if wall-source oxygen is unavailable.
Table A11.1.1 Equipment Parameters for Pericardial Imaging
Coil type Torso phased-array coil (or dedicated cardiac
coil, if available)
Gradient coil strength 25–40 mT/m (or whatever the system permits)
Cardiac gating Yes, preferably fiber-optic cables
Peripheral gating No
Oxygen Yes, 2 liters via nasal cannula for most patients
(to ensure breath-holding >15 sec)
Use of contrast agents Possible in patients with suspected neoplastic
disease (see Alternate Protocol)
Congenital Heart Acquired Heart
Disease Disease
A10.2.17 Contributed by Vivian S. Lee A11.1.1
Current Protocols in Magnetic Resonance Imaging (2001) A11.1.1-A11.1.14
Supplement 10 Copyright © 2001 by John Wiley & Sons, Inc.
 Set up equipment and patient
1. Interview and screen the patient to ensure that he or she has no contraindications,
such as cardiac pacemaker or defibrillator or other ferromagnetic materials near vital
structures. Also, ensure that the patient does not have health conditions, including
those related to cardiac or pericardial disease, that could require emergency equip-
ment during the scanning procedure. Assess the need for the monitoring of blood
pressure, pulse, and blood oxygenation during the examination.
9. Center the patient’s chest in the phased-array coil.
10. Use the centering light to position the patient and put him or her into the center of
the magnet.
Once this step has been performed, so long as the patient does not move on the table, the
table itself can be moved and then replaced in the same position as before without
jeopardizing the positioning of one scan relative to another.

Generally standard screening forms are used for all patients scanned in a magnetic 11. If the patient is unable to hold still, provide an appropriate sedative.
resonance system. 12. Once the patient has been centered in the magnet, check again to be sure that the ECG
The presence of any ferromagnetic metals may be a health hazard to the patient when he tracing demonstrates sharp positive R waves for suitable triggering. If not, then with
or she is inside the magnet and will also affect the imaging. If in doubt as to the exact patient still positioned within the magnet, toggle lead polarity until a suitable tracing
composition of the items, it is best to exclude patients with any metal implants. is obtained. If still not satisfactory, then bring the patient out of the magnet, check
Patients may be accompanied into the magnet room by a friend or family member, who can lead connections, and if necessary, reposition leads until a satisfactory tracing is
sit in the room during the scan and comfort the patient as needed. This companion must obtained.
be screened as well to ensure the absence of loose metal objects on the body or clothing.
Sequence 1: Rapid three-plane positioning scout
2. If the procedure is a research protocol, have the patient sign any necessary consent 13. For localization of subsequent acquisitions, run the system’s three-plane scout scan
form. (see Table A11.1.2). If possible, have the patient breath-hold at end expiration for the
3. Have the patient remove all jewelry and change into a gown to eliminate any metal scout for better positioning of subsequent acquisitions.
that might be found in clothing.
Sequence 2: 2-D transverse double inversion recovery half-Fourier single-shot turbo
4. Inform the patient about what will occur during the procedure, what he or she will spin echo (HASTE) (optional)
experience while in the magnet, and how to behave, including the following: 14. Use this optional ultrafast turbo spin-echo sequence (Table A11.1.3) for an overview
a. If earphones or headphones are used to protect the ears from the loud sounds of the chest and pericardium to localize lesions for subsequent spin-echo imaging.
produced by the gradients, the patient will be asked to wear these, but will be able Position a series of transverse slices off the scout images to ensure coverage of the
to communicate with you at any time during the imaging. pericardium to the superior recess (to the level of the inferior aspect of the aortic
b. The patient will be given a safety squeeze-bulb or similar equipment to request arch). Instruct the patient to perform shallow breathing.
assistance at any time (demonstrate how this works).
c. For good results the patient should not talk, and should avoid or minimize
swallowing or other movement, during each scan—i.e., as long as the banging
sounds continue. Between scans, talking and swallowing are allowed in most Table A11.1.2 Primary Clinical Imaging Parameters for Sequence 1 (Pilot
cases, but should be avoided when comparative positional studies are being Scan)
performed; the patient will be informed when this is the case.
Patient position Supine
d. Nevertheless, the patient may call out at any time if he or she feels it necessary. Scan type Gradient echo
5. Inform the patient of the need to hold his or her breath (preferably at end expiration) Imaging plane (orientation) Three-plane (transverse, coronal,
sagittal)
for ∼15 to 20 sec for each cine gradient-echo acquisition. Assess the need for
Central slice or volume center Center of chest
supplementary oxygen to improve breath-holding capacity, and if necessary, admin-
Echo time (TE) As short as possible
ister 2 liters oxygen via nasal cannula. Advise the patient of the importance of not
Repeat time (TR) As short as possible
moving during the acquisition periods and of not taking deep breaths during the
Flip angle (FA) 15°
non-breath-hold acquisitions.
Fields of view (FOVx, FOVy) 500 mm, 500 mm
6. When the patient is on the table, place MRI-compatible electrodes (Quatrode from Resolution (Δx, Δy) 1.95 mm, 3.91 mm
In Vivo) and ECG leads according to manufacturer’s guidelines. Make sure that the Number of data points collected (Nx, Ny) 256, 128
ECG tracing shows in high (positive) R waves. If it does not, either reposition the Display matrix (Dx, Dy) 256, 256
leads or toggle the lead polarity to alternative options, if available on the system. Slice thickness (Δz) 10 mm
Lead placement on the back is useful and may reduce motion artifacts related to breathing. Number of slices 5–10
Slice gap 0–10 mm
7. Place a pillow or other support under the knees to make the patient more comfortable.
Number of acquisitions (Nacq) 1
8. Place a 22-G intravenous catheter in the antecubital fossa if the gadolinium contrast Swap read and phase encoding No
is deemed necessary. Slice location Variable
A power injector can be used to administer contrast, but is not necessary. Saturation pulses Not applicable
Pericardial Scan time 5–15 sec Acquired Heart
Disease Disease

A11.1.2 A11.1.3
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 6
 The sequence can be performed with breath-holding at end expiration, but this is not 18. Set the imaging parameters as shown in Table A11.1.5 to obtain a single slice per
absolutely necessary. ECG gating is not necessary for this sequence but can be performed breath-hold. Ensure that the TR plus delay time is less than 85% to 90% of two R-to-R
for better images. intervals. Instruct the patient to breath-hold at end expiration, if possible, for best
reproducibility.
Sequence 3: 2-D transverse T1-weighted turbo spin echo (TSE)
15. Use the pilot scan and HASTE images to locate the pericardial mass in question or 19. Determine if the inversion time needs to be adjusted.
to define the extent of the pericardium. The inversion time determines the degree of fat suppression. Delay times are selected to
For T1-weighting, TR should be ∼85% to 90% of the patient’s R-to-R interval (time between allow imaging in diastole.
R waves) and less than 900 msec. Short TR times will limit the number of slices that can be
obtained in each acquisition. For greater coverage, this sequence can be run twice using Sequence 5: 2-D transverse and coronal fast cine gradient echo
the parameters shown in Table A11.1.4. If the area of coverage can be localized (e.g., for 20. For global pericardial disease, such as pericarditis, perform transverse fast cine
evaluation of a mass), the sequence need only be performed once with minimum slice gap. gradient-echo images through the entire pericardium, acquiring additional selected
ECG gating is necessary in this sequence.
sagittal images for evaluation of the inferior pericardium.
16. Acquisition times are too long to allow breath-hold imaging, and therefore with
multiple acquisitions (signal averaging), instruct the patient to perform shallow
breathing. Table A11.1.4 Primary Clinical Imaging Parameters for Sequence 3
(T1-Weighted TSE)
Sequence 4: 2-D transverse T2-weighted inversion-recovery turbo spin echo
17. If necessary, perform limited transverse T2-weighted turbo spin-echo images through Patient position Supine
the pericardium. Scan type Fast spin echo
For example, using the HASTE and T1-weighted TSE images, transverse T2-weighted Imaging plane (orientation) Transverse
images can be performed through the mass in question, or through the pericardium at the Central slice or volume center Center of heart
level of an effusion, to help characterize the abnormality. An inversion-recovery technique Echo time (TE) 12 msec (or minimum)
will provide useful fat suppression and improve contrast between pathology and normal Ech train length (ETL) 3
pericardial and epicardial fat. ECG gating is necessary in this sequence.
Repeat time (TR) <900 msec (and <90% of the
R-to-R interval)
Delay time after R wave 0 msec
Table A11.1.3 Primary Clinical Imaging Parameters for Sequence 2 (HASTE,
Optional) Flip angle (FA) 180°a
Fields of view (FOVx, FOVy) 300–350 mm, 300r–350r mm,
Patient position Supine with r = 3/4 (rectangular field of
Scan type Single-shot fast spin echo view) depending on body habitus
Imaging plane (orientation) Transverse Resolution (Δx, Δy) 1.17–1.37 mm, 1.43–1.67 mm
Central slice or volume center Center of heart Number of data points collected (Nx, Ny) 256, 210r, with r = 3/4
(rectangular field of view)
Echo time (TE) 43–63 msec (effective)
depending on body habitus
Repeat time (TR) Infinity (echo spacing is ∼4 msec)
Display matrix (Dx, Dy) 256, 256
Delay time (TD) 700 msec
Slice thickness (Δz) 4–8 mm
Flip angle (FA) 180°a
Number of slices 10–12
Fields of view (FOVx, FOVy) 300–350 mm, 300r–350r mm,
Slice gap 0.4–0.8 mm or 4–8 mm if
with r = 3/4 (rectangular field of
concatenatedb
view) depending on body habitus
Number of acquisitions (Nacq) 3–5
Resolution (Δx, Δy) 1.17–1.37 mm, 1.70–1.99 mm
Swap read and phase encoding No
Number of data points collected (Nx, Ny) 256, 176r, with r = 3/4
(rectangular field of view) Saturation pulses No
depending on body habitus ECG gating Yes
Display matrix (Dx, Dy) 256, 256 Scan time 3–8 min
aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this
Slice thickness (Δz) 5–8 mm
Number of slices 20 sequence is 90°.
bIf the z-axis coverage needed is on the order of 60 to 80 mm, then a single acquisition with about 8
Slice gap 0.8–2 mm
to 10 slices at 6 to 8 mm thickness and minimal gap (0.6 to 0.8 mm) can be performed. For complete
Number of acquisitions (Nacq) 1 coverage of the chest, two separate acquisitions of 8 to 10 slices can be performed, each with a gap
Swap read and phase encoding No equal to the slice thickness (e.g., 8 mm slice thickness with 8 mm gap). In this example, the second
Saturation pulses No acquisition should be performed with the slice positions shifted by 8 mm and with the same thickness
ECG gating Not necessary and gap, so that with the two acquisitions together, complete coverage of the chest will be obtained
Scan time 20–30 sec with 8-mm-thick slices and effectively no gap.
Pericardial Acquired Heart
Disease aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this Disease
sequence is 90°. A11.1.5
A11.1.4
Supplement 6 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A11.1.5 Primary Clinical Imaging Parameters for Sequence 4 (Short Tau Table A11.1.6 Primary Clinical Imaging Parameters for Sequence 5 (Cine
Inversion Recovery) Gradient Echo)

Patient position Supine Patient position Supine

Scan type Inversion recovery fast spin echo Scan type Segmented k-space cine gradient
Imaging plane (orientation) Transverse echo
Central slice or volume center Center of heart Imaging plane (orientation) Transverse (coronal is optional)
Echo time (TE) 57–76 msec Central slice or volume center Center of heart
Echo train length (ETL) 33 Echo time (TE) 4.8 msec (or minimum)
Repeat time (TR) <900 msec (TR + TD should be <2 Number of lines per segment 9 (see step 21)
× (85%–90% of the R-to-R Repeat time (TR) 80–100 msec (temporal resolution)
interval)) Delay time (TD) 0 msec
Inversion time (TI) 170 msec Flip angle (FA) 20°
Delay time (TD) 500 msec Fields of view (FOVx, FOVy) 300–350 mm, 300 r–350r mm,
Flip angle (FA) 180° with r = 3/4 (rectangular field of
Fields of view (FOVx, FOVy) 300–350 mm, 300r–350r mm, view) depending on body habitus
with r = 3/4 (rectangular field of Resolution (Δx, Δy) 1.17–1.37 mm, 1.70–1.99 mm
view) depending on body habitus Number of data points collected (Nx, Ny) 256, 176r, with r = 3/4
Resolution (Δx, Δy) 1.17–1.37 mm, 1.67–1.94 mm (rectangular field of view)
Number of data points collected (Nx, Ny) 256, 180r, with r = 3/4 depending on body habitus
(rectangular field of view) Display matrix (Dx, Dy) 256, 256
depending on body habitus Slice thickness (Δz) 5–7 mm
Display matrix (Dx, Dy) 256, 256 Number of slices 1
Slice thickness (Δz) 8 mm Slice gap Not applicable
Number of slices 1 Number of acquisitions (Nacq) 1
Slice gap Not applicable Swap read and phase encoding No
Number of acquisitions (Nacq) 1 Saturation pulses No
Swap read and phase encoding No Number of cardiac phases (R-to-R interval) × 85%/TR
Saturation pulses No ECG gating Yes
ECG gating Yes Scan time 15–20 sec
Scan time 12–15 sec

IMAGING OF NEOPLASTIC PERICARDIUM ALTERNATE

PROTOCOL
Cine gradient echo images can be performed selectively. For evaluation of the effect of For patients with suspected neoplastic involvement of the pericardium, contrast-enhanced
pericardial masses on hemodynamics, acquisitions can be positioned selectively. ECG images may also be needed in addition to the Basic Protocol. Intravenous gadolinium
gating is necessary in this sequence. contrast (typically, 0.1 mmol/kg) is administered, and turbo T1 spin-echo images are
21. Set the imaging parameters as shown in Table A11.1.6 to obtain a single slice per obtained. If necessary, nonenhanced images can be subtracted from contrast-enhanced
breath-hold (parameters shown are for a segmented k space approach with nine lines images using postprocessing software to depict enhancing lesions.
of k space per segment and four echoes shared). Adjust the parameters depending on Materials
the patient’s breath-holding capability and heart rate. For patients with slower heart
rates, use sequences that provide a greater number of lines per segment to help reduce Normal saline (0.9% NaCl), sterile
acquisition times. Set the number of phases of the cardiac study according to the Gadolinium-based MR contrast agent (e.g., Magnevist, Omniscan, or Prohance)
formula (for view sharing off):
Set up equipment and patient
Number of cardiac phases = (R-to-R interval) × 85%/TR 1. Use the same equipment and perform equipment and patient setup as for the previous
method (see Basic Protocol).
22. Instruct the patient to breath-hold at end expiration, if possible, for best reproduci-
bility. If patient is unable to breath-hold, perform the sequence with multiple 2. Run sequences 1 to 5.
acquisitions (e.g., three signal averages) with the patient breathing normally or 3. Leaving the patient in the magnet, inject the contrast agent, and flush the line with
shallowly. 20 ml saline.
A dose of 0.1 mmol/kg of contrast agent is usually given.
Pericardial Acquired Heart
Disease Disease

A11.1.6 A11.1.7
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Sequence 6: 2-D Transverse T1-weighted turbo spin echo
4. Repeat the sequence described in Table A11.1.4 ∼3 to 5 min after intravenous
injection of gadolinium contrast material. Ensure that all imaging parameters are kept
the same to facilitate postprocessing.
COMMENTARY
Background Information
The pericardium is a fibroserous sac that
contains the heart and proximal portions of the
ascending aorta and main pulmonary artery. By
MRI, it appears as a thin low-signal-intensity
line that should measure no greater than 4 mm
in thickness (Fig. A11.1.1). It is typically well
delineated by MRI because it is surrounded by
epicardial and mediastinal fat. There is nor-
mally a small amount of fluid, typically 15 to
50 ml, that lubricates the pericardial sac.
Transthoracic echocardiography is consid-
ered the first-line diagnostic modality for pa-
tients with suspected pericardial disease (Mar-
tin et al., 1998). However, pericardial thicken-
ing in the absence of pericardial fluid is difficult
to diagnose, such as in the setting of constrictive
pericarditis. Certain patients have poor acoustic
windows, and the imaging technique is ex-
tremely dependent on the skill of the echo-
cardiographer. Computed tomography (CT)
imaging offers wide field-of-view imaging. Its
sensitivity for the detection of pericardial cal-
cifications, as seen in constrictive pericarditis,
is advantageous. However, tissue contrast is
otherwise inferior to MRI, and motion artifacts
limit visualization of the noncalcified pericar-
dium. However, with the advent of multidetec-
tor ECG-gated CT, dynamic acquisitions may
become possible.
Indications for MRI of the pericardium in-
clude the diagnosis of constrictive pericarditis,
evaluation of pericardial effusions, including
differentiation of hemorrhagic from nonhem-
orrhagic effusions, detection of primary or sec-
ondary pericardial neoplastic disease, and
evaluation of congenital anomalies of the peri-
cardium.
The distinction between constrictive peri-
carditis and restrictive cardiomyopathy can be
a difficult clinical problem since clinical pres-
entation and hemodynamic patterns are similar
(Masui et al., 1992). Although unable to visu-
alize calcifications, MRI is still an ideal tech-
nique for diagnosing constrictive pericarditis.
Typical findings include pericardial thickening
>4 mm, dilated inferior vena cava, dilated atria,
and compressed ventricles (Fig. A11.1.2).
Pericardial
Disease
A11.1.8
A
Thickening can be confined to the right side of
the heart, and abnormal motion of the interven-
tricular septum may be seen during early dias-
tole. Patients with restrictive cardiomyopathy
do not usually have thickened pericardium.
Pericardial effusions can develop from dis-
eases such as uremia, collagen vascular dis-
eases, postpericardiotomy, and postinfarction
syndromes, as well as viral, tuberculous, or
idiopathic acute pericarditis and neoplasia.
Magnetic resonance imaging can be used to
help characterize fluid collections and to detect
the presence of adhesions. Transudative effu-
sions are typically low in signal intensity on
T1-weighted images and high on T2-weighted
images. Proteinaceous or hemorrhagic effu-
sions may be high in signal intensity on T1-
weighted images.
The most common pericardial mass is a
pericardial cyst, which classically lies along the B
right cardiophrenic angle and demonstrates
uniform high signal intensity on T2-weighted
images (Fig. A11.1.3). Neoplastic involvement
is more commonly due to secondary involve-
ment such as from breast and lung carcinoma,
lymphoma, and melanoma. When direct exten-
sion of tumor occurs, the pericardial effusion is
typically hemorrhagic. Mesothelioma is the
most common malignant pericardial tumor
(Hancock, 1990). In cases of suspected neo-
plastic disease, contrast-enhanced imaging
may be necessary to identify and delineate
tumor that is associated with complex pericar-
dial effusions. MRI can also help to differenti-
ate pericardial disease from adjacent para-
cardiac pathology (Barakos et al., 1989; Fig.
A11.1.4).
Congenital absence of the pericardium, typi-
cally partial and left sided, can be evident by
chest radiography and is readily confirmed by
unenhanced MRI.
Critical Parameters and
Troubleshooting
ECG tracings Figure A11.1.1 Normal pericardium. (A, B) Transverse HASTE images at two levels demonstrat-
Since all but the HASTE sequences used for ing normal low signal intensity line measuring <4 mm, corresponding to normal pericardium
(arrows).
imaging the pericardium must be ECG gated,
Acquired Heart
Disease
A11.1.9
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 A
A

B
B

Figure A11.1.3 Pericardial cyst. (A) Transverse and (B) coronal HASTE images show a well-de-
fined oval mass (arrows) with high signal intensity on T2-weighted images in the right cardiophrenic
angle, a characteristic location for a pericardial cyst. Contrast-enhanced images (not shown)
Figure A11.1.2 Constrictive pericarditis. (A) Transverse HASTE and (B) transverse cine gradient- confirmed a lack of enhancement in this lesion. Acquired Heart
Pericardial
Disease echo image show thickening of the pericardium (arrowheads), enlargement of the atria, and Disease
associated compression of the right ventricle (RV).
A11.1.10 A11.1.11
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 it is vital that an adequate ECG tracing be
obtained while the patient is in the bore of the
A magnet. For an ideal ECG tracing, sharp up-go-
ing R waves with significantly smaller (or in-
verted) T waves are needed, as triggering typi-
cally depends on the ECG value crossing a
particular threshold. Magnetic resonance–
compatible ECG leads with conductive gel
should be used. For good skin contact, the
patient’s skin may need to be shaved in some
cases. The position of the heart and axes of the
conduction pathways may vary considerably
across patients. The manufacturer’s recommen-
dations for lead positioning should be used as
a starting point. If polarity is not as desired (e.g.,
inverted R waves), check that leads are posi-
tioned correctly (RA to the right side, LA to the
left side, etc.), and then if needed, toggle
through the lead polarity options (I, II, AVR,
etc.). Finally, as a last resort, reposition ECG
leads, using new adhesive pads to ensure ade-
quate adhesion and conduction gel. If ECG
tracings are lost during positioning of the pa-
tient or scanning, check the patient to ensure
that leads have remained properly connected.
Customization of gated MRI parameters
Magnetic resonance imaging of the heart
B and pericardium requires careful customization
of imaging parameters according to patient
heart rate and breath-holding capability. For a
multislice turbo spin-echo sequence, the num-
ber of slices that can be imaged in one acquisi-
tion depends on TR, which in turn depends on
the patient’s heart rate. The TR must be defined
as shorter than the R-to-R interval (which is
inversely related to patient heart rate) to ensure
an adequate trigger window. Typically, trigger
windows of 10% to 15% are needed to allow
for variability of heart rates. For patients who
have extremely variable heart rates (e.g., fre-
quent premature beats), then the shorter R-to-R
intervals should be used to set TR
For single-slice cine gradient-echo images,
multiple phases of the heart cycle are imaged
at a single slice position. The lines of k space
(number of phase-encoding lines) collected for
each cardiac phase during each R-to-R interval
defines the number of lines per segment. The
more lines per segment, the longer the TR and
the fewer the cardiac phases that can be ac-
quired (the TR times the number of cardiac
phases should equal ∼90% of the R-to-R inter-
val). The total number of heartbeats needed to
acquire sufficient data to generate all images
Pericardial Figure A11.1.4 Pericardial and cardiac lymphoma. (A) Transverse HASTE image and (B) coronal depends on the total number of phase encoding
Disease cine gradient-echo image of the heart demonstrate a large mass (arrows) that invades through the
pericardial space into the right heart in a patient with lymphoma.
A11.1.12
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
steps needed and the number of lines per seg-
ment (number of heart beats should equal ap-
proximately the number of phase encoding
lines per image divided by the number of lines
per segment). While sequences could be cus-
tomized for every patient, to simplify the proc-
ess, system protocols can be set up for specific
categories of patients, such as “slow heart rate”
and “fast heart rate,” with the aim of limiting
acquisition times to <20 sec.
On some systems, a view-sharing option
will be available for cine gradient-echo se-
quences that enables almost double or greater
temporal resolution (number of cardiac phases)
in the same acquisition time.
Motion or breathing artifacts
Occasionally, despite a patient’s effort to
breath-hold, substantial motion artifact is seen
on the MRI. Cine gradient-echo images can be
performed with multiple signal averages with
the patient breathing normally but shallowly.
An anterior saturation band can be placed over
the anterior chest wall subcutaneous fat to help
reduce motion-related artifacts from breathing.
Anticipated Results
The goals of MRI of the pericardium are to
visualize the pericardium, characterize abnor-
malities of the pericardium itself or the pericar-
dial space, and identify and characterize peri-
cardial masses. With ECG gating, MRI can
provide clear visualization of anatomic abnor-
malities, such as pericardial thickening or peri-
cardial effusions, as well as allow an assess-
ment of consequent hemodynamic abnormali-
ties. For example, in patients with constrictive
pericarditis, the effect of constriction on the
right ventricle can be seen during dynamic cine
gradient-echo imaging. Similarly, in patients
with large pericardial effusions, cine gradient-
echo images depict the compression of right-
sided heart chambers, which indicates physi-
ologically significant changes of tamponade.
The distinction between pericardial disease
and paracardial masses can also be performed
using MRI. With pericardial disease, involve-
ment of the pericardium can be visualized di-
rectly. The use of contrast material can supple-
ment cine gradient-echo imaging in helping to
depict the extent of enhancing pericardial neo-
plasm in the setting of a complex pericardial
effusion. However, in many cases, contrast ma-
terial is not needed to diagnose and characterize
pericardial pathology.
Acquired Heart
Disease
A11.1.13
 Literature Cited
Barakos, J.A., Brown, J.J., and Higgins, C.B. 1989.
MR imaging of secondary cardiac and para-
cardiac lesions. Am. J. Roentgenol. 153:47-50.
Hancock, E.W. 1990. Neoplastic pericardial disease.
Cardiol. Clin. 8:673-682.
Martin, E.T., Fuisz, A.R., and Pohost, G.M. 1998.
Imaging cardiac structure and pump function.
Cardiol. Clin. 16:135-160
Masui, T., Finck, S., and Higgins, C.B. 1992. Con-
strictive pericarditis and restrictive cardiomyo-
pathy: Evaluation with MR imaging. Radiology
182:369-373.
Key References
Dupuis, K., Thangaraj, V., and Edelman, R.R. 1996.
Practical MRI for the technologist and imaging
specialist. In Clinical Magnetic Resonance Im-
aging (R.R. Edelman, J.R. Hesselink, and M.B.
Zlatkin, eds.) pp. 52-87. W.B. Saunders, Phila-
delphia.
Pericardial
Disease
A11.1.14
Covers a wide range of practical information for
setting patients up for MRI and also includes a
useful description of cardiac imaging sequence pa-
rameters.
Higgins, C.B. 1992. Essentials of Cardiac Radiol-
ogy and Imaging. J.B. Lippincott, New York.
Contains a detailed description of the spectrum of
modalities that can be used to image the pericar-
dium, including plain film radiography and MRI.
White, C.S. 1996. MR evaluation of the pericardium
and cardiac malignancies. MRI Clin. North Am.
4:237-251.
Reviews the anatomy and pathology of the pericar-
dium with an emphasis on MRI findings.
Contributed by Vivian S. Lee
New York University Medical Center
New York, New York
Current Protocols in Magnetic Resonance Imaging
Cardiac Masses UNIT A11.2
Although echocardiography is considered the primary technique for evaluating cardiac
masses, magnetic resonance imaging (MRI) may be indicated in patients who have poor
acoustic windows or in whom lesions are incompletely characterized by echocardiogra-
phy. This unit presents the basic technique for evaluating cardiac masses, with optional
contrast-enhanced sequences for specific indications. The parameters are based on
experience on a Siemens 1.5 T Vision or Symphony and should be altered accordingly
for different field strengths and manufacturers.
IMAGING OF CARDIAC MASSES BASIC
PROTOCOL
To image the heart, cardiac imaging packages that allow gated fast imaging of the heart
using echo-train fast (or turbo) spin-echo and segmented k-space cine gradient-echo
images are necessary for better imaging with examination times that are practical in the
clinical setting.
Similar to that described for pericardial disease in UNIT A11.1, the core components of the
imaging protocol are gated T1-weighted and T2-weighted turbo spin-echo images and cine
gradient-echo images through the heart. The spin-echo images are used to provide
high-resolution multiplanar imaging of the heart and to help to characterize masses. A
fat-suppressed imaging sequence is optional for suspected germ cell tumors, lipomas, or
lipomatous hypertrophy of the interatrial septum. The two-dimensional electrocardio-
gram (2-D ECG)–gated cine gradient-echo images provide views of a single slice at
multiple phases of the cardiac cycle. When viewed in the cine mode, images can be used
to evaluate the motion of the heart dynamically. Cine MRI can help distinguish a mass in
a cardiac chamber from flowing blood and also is useful for evaluating physiologic effects
of masses, such as mitral valvular obstruction by a left atrial myxoma. With fast
breath-hold cine acquisitions, the total examination time for the Basic Protocol should be
<45 min.
For patients with certain entities, such as thrombus versus mass or suspected intramural
tumors, additional contrast-enhanced imaging of the heart can be performed by repeating
T1-weighted imaging following intravenous gadolinium contrast material.
Table A11.2.1 lists the hardware necessary to perform MRI of the heart. Higher gradient
strengths with imaging sequences that allow gated turbo spin-echo and segmented cine
gradient-echo imaging are preferable.
Materials
Normal saline (0.9% NaCl), sterile
Gadolinium-based MR contrast agent (e.g., Magnevist, Omniscan, or Prohance)
Table A11.2.1 Equipment Parameters for Pericardial Imaging
Coil type Torso phased-array coil (or dedicated cardiac
coil, if available)
Gradient coil strength 25–40 mT/m (or whatever the system permits)
Cardiac gating Yes, preferably fiber-optic cables
Peripheral gating No
Oxygen Yes, 2 liters nasal cannula for most patients (to
ensure breath-holding >15 sec)
Use of contrast agents Possible in patients with certain diseases (see
Alternate Protocol) Acquired Heart
Disease
Contributed by Vivian S. Lee A11.2.1
Current Protocols in Magnetic Resonance Imaging (2001) A11.2.1-A11.2.12
Copyright © 2001 by John Wiley & Sons, Inc.
 Set up equipment and patient
1. As discussed in UNIT A11.1 on pericardial disease, it is important to interview and screen
the patient to ensure that he or she has no contraindications such as cardiac pacemaker
or defibrillator or other ferromagnetic materials near vital structures. Also, ensure
that the patient does not have health conditions, including those related to cardiac or
pericardial disease, that may require emergency equipment during the scanning
procedure.
Generally standard screening forms are used for all patients scanned in a magnetic
resonance system.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants.
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
2. If the procedure is a research protocol, have the patient sign any necessary consent
form.
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
4. Inform the patient about what will occur during the procedure, what he or she will
experience while in the magnet, and how to behave, including the following:
a. If earphones or headphones are used to protect the ears from the loud sounds
produced by the gradients, ask the patient to wear these, but inform the patient
that he or she will be able to communicate with you at any time during the imaging.
b. Give the patient a safety squeeze-bulb or similar equipment to request assistance
at any time (demonstrate how this works).
c. For good results inform the patient not to talk and to avoid or minimize swallowing
or other movement, during each scan—i.e., as long as the banging sounds continue.
Between scans, talking and swallowing are allowed in most cases but should be avoided
when comparative positional studies are being performed; the patient will be informed
when this is the case.
d. Inform the patient that, if necessary, he or she may call out at any time.
5. Inform the patient of the need to hold his or her breath (preferably at end expiration)
for ∼15 to 20 sec for each cine gradient-echo acquisition. Assess need for supplemen-
tary oxygen to improve breath-holding capacity, and if necessary, administer 2 liters
oxygen via nasal cannula. Advise the patient of the importance of not moving during
the acquisition periods and of not taking deep breaths during the non-breath-hold
acquisitions.
6. When the patient is on the table, place MRI-compatible ECG leads (Quatrode from
In Vivo) according to manufacturer’s guidelines. When possible, place leads on the
back as this may reduce motion artifacts related to breathing. Make sure that the ECG
tracing resulting from the lead placement results in high (positive) R waves. If it does
not, either reposition the leads or toggle the lead polarity to alternative options, if
available on the system.
7. Place a pillow or other support under the knees to make the patient more comfortable.
Cardiac Masses
A11.2.2
Current Protocols in Magnetic Resonance Imaging
8. Place a 22-G intravenous catheter in the antecubital fossa if the gadolinium contrast
is deemed likely to be necessary.
A power injector can be used to administer contrast, but it is not necessary.
It is preferable to insert the line prior to imaging and to leave the patient in the magnet,
with no intervening motion, between the scans run before contrast agent injection and those
run after injection.
9. Center the patient’s chest in the phased-array coil.
10. Use the centering light to position the patient and put him or her into the center of
the magnet.
Once this step has been performed, so long as the patient does not move on the table, the
table itself can be moved and then replaced in the same position as before without
jeopardizing the positioning of one scan relative to another.
11. If the patient is unable to hold still, provide an appropriate sedative.
12. Once the patient has been centered in the magnet, check again to be sure that the ECG
tracing demonstrates sharp positive R waves for suitable triggering. If not, then, with
patient still positioned within the magnet, toggle lead polarity until a suitable tracing
is obtained. If this is still not satisfactory, then bring the patient out of the magnet,
check lead connections, and if necessary, reposition leads until a satisfactory tracing
is obtained.
Sequence 1: Rapid three-plane positioning scout
13. For localization of subsequent acquisitions, run the system’s three-plane scout scan
(see Table A11.2.2). If possible, have the patient breath-hold at end expiration for the
scout to better positioning of subsequent acquisitions.
Sequence 2: 2-D transverse double inversion recovery half-Fourier single-shot turbo
spin echo (db-HASTE)
14. Use this ultrafast turbo spin-echo sequence (Table A11.2.3) with blood-nulling
preparatory pulse for an overview of the heart to localize lesions for subsequent
spin-echo imaging. Position a series of transverse slices off the scout images to ensure
coverage of the entire heart.
This is a T2-weighted sequence, and most tumors will appear high in signal intensity on
HASTE.
The sequence can be performed with breath-holding at end expiration, but this is not
absolutely necessary.
ECG-gating, also not necessary, can be used for improved images.
Sequence 3: 2-D transverse T1-weighted turbo spin-echo (TSE)
15. Use the pilot scan and HASTE images to locate the mass in question or to define the
extent of the cardiac abnormality. For greater coverage, run this sequence twice using
the parameters shown in Table A11.2.4. If the area of coverage can be localized (e.g.,
for evaluation of a mass), only perform the sequence once with minimum slice gap
(Fig. A11.2.1).
For T1 weighting, TR should be ∼85% to 90% of the patient’s R-to-R interval (time between
R waves) and <900 msec. Short TR times will limit the number of slices that can be obtained
in each acquisition.
16. Because acquisition times are too long to allow breath-hold imaging, with multiple
Acquired Heart
acquisitions, instruct the patient to perform shallow breathing. Disease
A11.2.3
Current Protocols in Magnetic Resonance Imaging
 Table A11.2.2 Primary Clinical Imaging Parameters for Sequence 1 (Pilot Table A11.2.4 Primary Clinical Imaging Parameters for Sequence 3
Scan) (T1-Weighted TSE)

Patient position Supine Patient position Supine

Scan type Gradient echo Scan type Turbo spin echo
Imaging plane (orientation) Three-plane (transverse, coronal, Imaging plane (orientation) Transverse
sagittal) Central slice or volume center Center of heart
Central slice or volume center Center of chest Echo time (TE) 12 msec (or minimum)
Echo time (TE) As short as possible Echo train length (ETL) 3
Repeat time (TR) As short as possible Repeat time (TR) <900 msec (and <90% of the
Flip angle (FA) 15° R-to-R interval)
Fields of view (FOVx, FOVy) 500 mm, 500 mm Delay time after R wave 0 msec
Resolution (Δx, Δy) 1.95 mm, 3.91 mm Flip angle (FA) 180°a
Number of data points collected (Nx, Ny) 256, 128 Fields of view (FOVx, FOVy) 300–350 mm, 300r–350r mm,
Display matrix (Dx, Dy) 256, 256 with r = 3/4 (rectangular field of
Slice thickness (Δz) 10 mm view) depending on body habitus
Number of slices 5–10 Resolution (Δx, Δy) 1.17–1.37 mm, 1.43–1.67 mm
Slice gap 0–10 Number of data points collected (Nx, Ny) 256, 210r, with r = 3/4
Number of acquisitions (Nacq) 1 (rectangular field of view)
depending on body habitus
Swap read and phase encoding No
Display matrix (Dx, Dy) 256, 256
Saturation pulses Not applicable
Slice thickness (Δz) 4–8 mm
Scan time 5–15 sec
Number of slices 10–12
Slice gap 0.4 to 0.8 mm or 4–8 mm if
concatenatedb
Table A11.2.3 Primary Clinical Imaging Parameters for Sequence 2 (HASTE) Number of acquisitions (Nacq) 3–5
Swap read and phase encoding No
Patient position Supine Saturation pulses No
Scan type Single-shot fast spin echo ECG gating Yes
Imaging plane (orientation) Transverse Scan time 3–8 min
Central slice or volume center Center of heart aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this
Echo time (TE) 43–63 msec (effective) sequence is 90°.
bIf the z-axis coverage needed is on the order of 60 to 80 mm, then a single acquisition with about 8
Repeat time (TR) Infinity (echo spacing is ∼4 msec)
Delay time (TD) 700 msec to 10 slices at 6 to 8 mm thickness and minimal gap (0.6 to 0.8 mm) can be performed. For complete
coverage of the chest, two separate acquisitions of 8 to 10 slices can be performed, each with a gap
Flip angle (FA) 180°a
equal to the slice thickness (e.g., 8 mm slice thickness with 8 mm gap). In this example, the second
Fields of view (FOVx, FOVy) 300–350 mm, 300r–350r mm, acquisition should be performed with the slice positions shifted by 8 mm and with the same thickness
with r = 3/4 (rectangular field of and gap, so that with the two acquisitions together, complete coverage of the chest will be obtained
view) depending on body habitus with 8-mm-thick slices and effectively no gap.
Resolution (Δx, Δy) 1.17–1.37 mm, 1.70–1.99 mm
Number of data points collected (Nx, Ny) 256, 176r, with r = 3/4
Sequence 4: 2-D fat-suppressed transverse T2-weighted inversion-recovery turbo
(rectangular field of view)
depending on body habitus spin echo (optional)
Display matrix (Dx, Dy) 256, 256 17. Perform limited transverse T2-weighted turbo spin-echo images through the heart as
Slice thickness (Δz) 5–8 mm needed.
Number of slices 20 An inversion-recovery technique will provide useful fat suppression and allow charac-
Slice gap 0.8–2 mm terization of fatty masses, such as lipomas of the heart (Fig. A11.2.2).
Number of acquisitions (Nacq) 1 18. Set the imaging parameters as shown in Table A11.2.5 to obtain a single slice per
Swap read and phase encoding No breath-hold. Instruct the patient to breath-hold at end expiration, if possible, for best
Saturation pulses No reproducibility.
ECG gating Not necessary
The TR plus delay time should be <85% to 90% of two R-to-R intervals.
Scan time 20–30 sec
aThe system displays the flip angle of the refocusing pulse. 19. Adjust the inversion time depending on the particular manufacturer. Select delay
times to allow imaging in diastole.
Acquired Heart
Cardiac Masses The inversion time determines the degree of fat suppression obtained. Disease

A11.2.4 A11.2.5
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 A A

B
B

Figure A11.2.1 Left atrial myxoma. (A) Transverse T1-weighted turbo spin echo and (B) trans-
Cardiac Masses verse cine gradient echo show a large mass in the left atrium (M).
Figure A11.2.2 Atrial lipoma. (A) Transverse HASTE image and (B) transverse short TI inversion-
recovery image through the right atrium demonstrate a large mass in the right atrial wall (curved
arrows) and also involving the interatrial septum (shorter arrowhead in A). The lesion is high in
signal intensity on the HASTE image. The loss of signal intensity in the mass on the inversion-re-
covery sequence parallels that of subcutaneous fat and supports the diagnosis of right atrial lipoma
Acquired Heart
with involvement of the internal septum. Disease

A11.2.6 A11.2.7
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A11.2.5 Primary Clinical Imaging Parameters for Sequence 4 (Short Tau Table A11.2.6 Primary Clinical Imaging Parameters for Sequence 5 (Cine
Inversion Recovery, Optional) Gradient Echo)

Patient position Supine Patient position Supine

Scan type Inversion recovery fast spin echo Scan type Segmented k-space cine gradient
Imaging plane (orientation) Transverse echo
Central slice or volume center Center of heart Imaging plane (orientation) Transverse (coronal is optional)
Echo time (TE) 57–76 msec Central slice or volume center Center of heart
Echo train length (ETL) 33 Echo time (TE) 4.8 msec (or minimum)
Repeat time (TR) <900 msec [TR + TD should be <2 Number of lines per segment 9 (see step 21)
× (85% to 90%) × (R-to-R Repeat time (TR) 80–100 msec (temporal resolution)
interval)] Delay time (TD) 0 msec
Inversion time (TI) 170 msec Flip angle (FA) 20°
Delay time (TD) 500 msec Fields of view (FOVx, FOVy) 300–350 mm, 300r–350r mm,
Flip angle (FA) 180° with r = 3/4 (rectangular field of
Fields of view (FOVx, FOVy) 300–350 mm, 300r–350r mm, view) depending on body habitus
with r = 3/4 (rectangular field of Resolution (Δx, Δy) 1.17–1.37 mm, 1.70–1.99 mm
view) depending on body habitus Number of data points collected (Nx, Ny) 256, 176r, with r = 3/4
Resolution (Δx, Δy) 1.17–1.37 mm, 1.67–1.94 mm (rectangular field of view)
Number of data points collected (Nx, Ny) 256, 180r, with r = 3/4 depending on body habitus
(rectangular field of view) Display matrix (Dx, Dy) 256, 256
depending on body habitus Slice thickness 5–7 mm
Display matrix (Dx, Dy) 256, 256 Number of slices 1
Slice thickness (Δz) 8 mm Slice gap Not applicable
Number of slices 1 Number of acquisitions (Nacq) 1
Slice gap Not applicable Swap read and phase encoding No
Number of acquisitions (Nacq) 1 Saturation pulses No
Swap read and phase encoding No Number of cardiac phases (R-to-R interval) × 85%/TR
Saturation pulses No ECG gating Yes
ECG gating Yes Scan time 15–20 sec
Scan time 12–15

Sequence 5: 2-D transverse and coronal fast cine gradient echo
20. Generally, perform transverse cine gradient-echo imaging, slice by slice, through the
entire heart. Obtain additional coronal or sagittal oblique views for evaluation of
specific regions, for example, through the inferior vena cava for evaluation of vascular
extension of tumor thrombus from hepatoma or renal cell carcinoma.
Cine gradient-echo images can be performed selectively. A four-chamber view or two-
chamber view of the heart may be useful to determine whether an atrial mass interferes
with mitral or tricuspid valvular function.
21. Set the imaging parameters as shown in Table A11.2.6 to obtain a single slice per
breath-hold. Adjust the parameters depending on the patient’s breath-holding capa-
bility and heart rate. For patients with slower heart rates, use sequences that provide
a greater number of lines per segment to help reduce acquisition times. Set the number
of phases of the cardiac study according to the formula:
Number of cardiac phases = (R-to-R interval) × 85%/TR
acquisitions (three-signal averages, for example) with the patient breathing normally
or shallowly.
Sequence 6: 2-D transverse T1-weighted turbo spin echo (optional)
For patients with specific clinical indications, such as the differentiation between cardiac
tumor and thrombus or suspected intramural masses, contrast-enhanced images may also
be needed.
23. Leaving the patient in the magnet, inject the contrast agent, flush the line with 20 ml
saline.
Intravenous gadolinium contrast material at a dose of 0.1 mmol/kg is administered and
turbo T1-weighted spin-echo images repeated. Non-enhanced images can be subtracted
from contrast-enhanced images using post-processing software to depict enhancing le-
sions.
24. Repeat the sequence described in Table A11.2.4, ensuring that all imaging parameters
are kept the same to facilitate post-processing.

Parameters shown in Table A11.2.6 are for a segmented k-space approach with nine lines
of k-space per segment and four echoes shared. COMMENTARY
22. Instruct the patient to breath-hold at end expiration, if possible, for best reproduci- Background Information with this technique. The limited field of view,
bility. If patient is unable to breath-hold, perform the sequence with multiple The primary screening test for suspected dependence on adequate acoustic window and
intracardiac masses is transthoracic echo- skill of the echocardiographer, and limited Acquired Heart
Cardiac Masses cardiography. However, there are limitations visualization of certain areas of the heart such Disease

A11.2.8 A11.2.9
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 as the apex or left atrial appendage require
additional performance of transesophageal
echocardiography in many cases, which is in-
vasive and also limited in field of view (Martin
et al., 1998). Magnetic resonance imaging can
provide a noninvasive yet comprehensive
anatomic and physiologic assessment of car-
diac masses by relying on spin-echo imaging
for anatomy and cine gradient-echo imaging
for flow and function.
The most common cardiac mass is throm-
bus, which typically lies in a dilated left atrium
or ventricle. Thrombus can usually be distin-
guished from most neoplastic masses by mor-
phology and the lack of contrast enhancement.
Benign tumors include myxomas, typically
arising in the left atrium with a pedicle attached
to the atrial septum, and lipomas, usually in the
right atrium as a discrete mass or more com-
monly a uniform fatty infiltration of the intera-
trial septum (Fig. A11.2.2). Lipomas can be
characterized by using fat suppression se-
quences to identify the presence of fat within
lesions. In children, rhabdomyomas can be
found, typically within the ventricles, and are
associated with tuberous sclerosis. Malignant
tumors are most commonly secondary, such as
direct extension from adjacent lung or medi-
astinal malignancies, vascular extension from
renal cell carcinoma or hepatoma, or hemato-
genous metastases, such as melanoma or sar-
coma. Primary cardiac malignancies are ex-
tremely rare, the most common being angiosar-
coma (Higgins, 1992). Unlike benign tumors,
malignancies of the heart typically are less
well-circumscribed and demonstrate aggres-
sive infiltrative appearance, best seen on con-
trast-enhanced images. Magnetic resonance
imaging can also help to differentiate cardiac
masses from adjacent paracardiac pathology
(Barakos et al., 1989).
Critical Parameters and
Troubleshooting
ECG tracings
As discussed in UNIT A11.1 on pericardial im-
aging, ECG tracings are vital for all imaging of
the heart. Sharp up-going R waves with signifi-
cantly smaller (or inverted) T waves are needed
since triggering typically depends on the ECG
value crossing a particular threshold. Magnetic
resonance–compatible ECG leads with con-
ductive gel should be used. For good skin con-
tact, the patient’s skin may need to be shaved.
The position of the heart and axes of the con-
Cardiac Masses duction pathways may vary considerably
A11.2.10
across patients. The manufacturer’s recommen-
dations of lead positioning should be used as a
starting point. If polarity is not as desired (e.g.,
inverted R waves), check that leads are posi-
tioned correctly (RA to the right side, LA to the
left side, etc), and then, if needed, toggle
through the lead polarity options (I, II, AVR,
etc.). Finally, as a last resort, reposition ECG
leads, using new adhesive pads to ensure ade-
quate adhesion and conduction gel. If ECG
tracings are lost during positioning of the pa-
tient or scanning, check the patient to ensure
that leads have remained properly connected.
Customization of gated MRI parameters
Magnetic resonance imaging of the heart
and pericardium requires careful customization
of imaging parameters according to patient
heart rate and breath-holding capability (see
UNIT A11.1 on pericardial imaging). For a multi-
slice turbo spin-echo sequence, the number of
slices that can be imaged in one acquisition
depends on the TR, which in turn depends on
the patient’s heart rate. For single-slice cine
gradient-echo images, multiple phases of the
heart cycle are imaged at a single slice position.
The more lines of k space (number of phase-en-
Figure A11.2.3 Right atrial pseudotumor. Transverse cine gradient-echo image shows a
coding lines) that are collected for each cardiac typical nodular appearance of the right atrial pseudotumor, seen along the posterior wall of
phase during each TR interval, the faster the the right atrium (arrowheads). This is a normal structure and should not be mistaken for tumor
acquisition time, but fewer cardiac phases can or thrombus.
be acquired. While sequences could be custom-
ized for every patient, to simplify the process,
Appropriate sedation should be administered (Fig. A11.2.2). Vascular or adjacent extension
system protocols should be set up for specific
when necessary. of tumor into the heart can be visualized be-
categories of patients, such as “slow heart rate”
cause of the large field of view available with
and “fast heart rate” with the aim of limiting Pseudotumors MRI.
acquisition times <20 sec.
In ∼59% to 90% of patients, a nodular or Contrast material can supplement cine gra-
On some systems, a view-sharing option
linear density can be seen along the posterior dient-echo imaging in helping to differentiate
will be available for cine gradient-echo se-
wall of the right atrium (Mirowitz and Gutier- nonenhancing masses such as thrombus from
quences that enables almost double or greater
rez, 1992; Meier and Hartnell, 1994). This neoplastic disease. However, in many cases,
temporal resolution (number of cardiac phases)
characteristic lesion can be misinterpreted as contrast material is not needed to diagnose and
in the same acquisition time.
an intracardiac mass and cause potential diag- characterize cardiac masses.
Motion or breathing artifacts nostic difficulties if unrecognized. The mass is
caused by normal fibromuscular elements of
Cine gradient-echo images can be per-
Literature Cited
the right atrium, such as the crista terminalis Barakos, J.A., Brown, J.J., and Higgins, C.B. 1989.
formed with multiple signal averages when MR imaging of secondary cardiac and para-
and Chiari network (Fig. A11.2.3).
patients are unable to breath-hold adequately. cardiac lesions. Am. J. Roentgenol. 153:47-50.
Supplemental oxygen via nasal cannula can Higgins, C.B. 1992. Essentials of Cardiac Radiol-
improve a patient’s breath-holding capability
Anticipated Results
ogy and Imaging. J.B. Lippincott, New York.
The goal of MRI of the cardiac masses is to
substantially and should be considered in all Martin, E.T., Fuisz, A.R., and Pohost, G.M. 1998.
localize and characterize as best possible
patients. Imaging cardiac structure and pump function.
masses within the heart. With adequate ECG
Cardiol. Clin. 16:135-160
Imaging infants and children gating, MRI can provide clear visualization of
Meier, R.A. and Hartnell, G.G. 1994. MRI of right
anatomic abnormalities (Fig. A11.2.1) as well
MR imaging techniques must be modified atrial pseudomass: Is it really a diagnostic prob-
as an assessment of consequent hemodynamic lem? J. Comput. Assist. Tomogr. 18:398-401.
for the evaluation of suspected masses in infants
abnormalities. Magnetic resonance imaging
and children. A head coil should be used for Mirowitz, S.A. and Gutierrez, F.R. 1992. Fibromus-
can be used to characterize lipomas and lipo- cular elements of the right atrium: Pseudomass
small infants. Non-breath-hold modifications Acquired Heart
matous hypertrophy of the interatrial septum at MR imaging. Radiology 182:231-233.
such as multiple signal averages should be used. Disease
A11.2.11
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Key References
Dupuis, K., Thangaraj, V., and Edelman, R.R. 1996.
Practical MRI for the technologist and imaging
specialist. In Clinical Magnetic Resonance Im-
aging (R.R. Edelman, J.R. Hesselink, and M.B.
Zlatkin, eds.) pp. 52-87. W.B. Saunders, Phila-
delphia.
Covers a wide range of practical information for
setting patients up for MRI and also includes a
useful description of cardiac imaging sequence pa-
rameters.
Higgins, 1992. See above.
Cardiac Masses
A11.2.12
Contains a detailed description of methods to evalu-
ate cardiac masses, including plain film diagnoses
and MRI.
White, C.S. 1996. MR evaluation of the pericardium
and cardiac malignancies. MRI Clin. North Am.
4:237-251.
Reviews MRI findings of cardiac masses.
Contributed by Vivian S. Lee
New York University Medical Center
New York, New York
Current Protocols in Magnetic Resonance Imaging
Myocardial Perfusion and Viability UNIT A11.3
Rapid magnetic resonance imaging of the heart during the first pass of a bolus-injected
contrast agent can be used to assess myocardial perfusion (Manning et al., 1991; Rossum
et al., 1990; Wilke et al., 1994; Wilke et al., 1993). In this context, the term perfusion
refers to transport of the contrast agent into myocardial tissue by blood flow, distribution
in the tissue, and washout. Following the course of enhancement after injection of an
extracellular MR contrast agent such as Gd-DTPA (gadolinium diethylenetriamine pen-
taacetic acid) provides a means to determine myocardial blood flow. The rate of contrast
enhancement depends on the rate of transport of the contrast agent through a tissue region,
indicating sensitivity to the level of tissue blood flow. Assessment of perfusion over the
entire heart with MRI is generally accomplished by 2-dimensional imaging of multiple
slices. Cardiac motion appears frozen if the acquisition time for each image is short on
the time scale of a heart beat, and if the image acquisition is synchronized to the heart
rhythm.
Compared with other organs, the heart has a relatively high resting blood flow and blood
volume on the order of ∼1 ml/min per gram of tissue and 0.1 ml per gram of tissue,
respectively. The wash-in of an injected contrast agent takes, under normal conditions,
only a couple of heart beats. This determines the required temporal resolution for dynamic
first pass studies. Multislice coverage of the heart and sequential imaging with a temporal
resolution of ∼1 image per heart beat for each slice location can be met with rapid
T1-weighted, gradient echo imaging techniques. T1-weighted imaging is the method of
choice for imaging of perfusion in the heart. Contrast agent bolus dosages typically used
for T1-weighted myocardial perfusion imaging of the heart give rise to a several-fold signal
intensity increase over the precontrast level of signal intensity. T2*-weighted imaging is
by definition sensitive to magnetic susceptibility changes produced by the introduction
of paramagnetic contrast agents. T2*-weighted imaging is more prone to susceptibility-
induced artifacts–the arrival of the contrast agent in the ventricular cavity gives rise to a
marked drop of T2*-weighted signal intensity that extends into myocardial tissue.
Extracellular contrast agents such as Gd-DTPA, are commonly used for MRI perfusion
studies. Extracellular contrast agents cross the cell membrane only after severe myocar-
dial injury and the loss of myocardial viability has occurred (Judd et al., 1995; Kim et al.,
1996; Lima et al., 1995; Tong et al., 1993a,b). The distribution volume of an extracellular
contrast agent is larger in injured tissue than in normal tissue. Given sufficient time, an
extracellular contrast agent reaches an approximate equilibrium distribution where the
contrast enhancement of tissue relative to the contrast enhancement in the ventricular
blood pool is proportional to the distribution volume (Pereira et al., 1996; 1999). Loss of
viability and leakage of the contrast agent into the cell results in T1-weighted signal
hyperenhancement.
Both first pass perfusion studies and imaging of delayed hyperenhancement can be
combined in one patient exam. The MRI protocols for myocardial perfusion and viability
assessment are presented together in this unit due to the complimentary information that
is obtained with both protocols. Also, the imaging techniques used for both protocols are
closely related. Basic Protocol 1 should be used in patients with symptoms of coronary
artery disease or ischemic cardiomyopathies. Basic Protocol 1 assesses the functional
severity of coronary artery lesions. Basic Protocol 1 takes ∼30 min, including the time
for patient preparation and study set-up. If imaging of perfusion during pharmacological
stress is included, the duration is increased to ∼45 min. Basic Protocol 2 should be used,
in addition to Basic Protocol 1, when a patient presents with symptoms of myocardial
Acquired Heart
Disease
Contributed by Michael Jerosch-Herold and Arthur E. Stillman A11.3.1
Current Protocols in Magnetic Resonance Imaging (2001) A11.3.1-A11.3.21
Copyright © 2001 by John Wiley & Sons, Inc.
 infarction or post-coronary revascularization to determine the presence and extent of Table A11.3.1 Equipment Parameters for First Pass Perfusion Imaging
nonviable myocardium. Basic Protocol 2 adds an additional 20 min for the completion of
the examination. Coil type Quadrature phase-array torso coil.
Gradient coil strength 24 mT/m (or higher if the system permits)
Cardiac gating Yes, fiber-optic systems are preferred for studies
BASIC IMAGING MYOCARDIAL PERFUSION DURING FIRST-PASS performed during pharmacologic stress or
PROTOCOL 1 CONTRAST-ENHANCEMENT vasodilation
By its very nature, the short repetition times and high bandwidths of fast imaging Peripheral gating For safety only
techniques used for myocardial perfusion imaging can create a challenge for achieving Respiratory gating No
an acceptable signal-to-noise ratio (SNR) that allows reliable interpretation of the images. Respirator If required by patient
It is, therefore, generally only possible to obtain acceptable SNRs at field strengths of Oxygen If required by patient
≥1.0 T. The only exception occurs with echo-planar imaging sequences, which may work Breath-holding Optional
better at lower field strengths due to the inherent reduction of T2* artifacts at lower field Motion cushions Can be used for patient comfort
strengths. Single-shot echo-planar imaging (EPI) techniques applied to the heart are more Use of contrast agents Extracellular or intravascular contrast agent
prone to image artifacts than multi-shot EPI techniques or segmented gradient-echo/EPI administered with power injector through
sequences. The latter can be used for cardiac perfusion imaging to improve the speed of intravenous (i.v.) needle in antecubital vein
image acquisition, without incurring a significant penalty in image quality. Power injector Important for rapid and reproducible
administration of contrast agent
Fast T1-weighted imaging sequences, such as spoiled gradient-echo imaging with TR’s as Infusion pump Required for pharmacological stress or
short as 2 msec and a magnetization preparation for T1-weighting, are applied to image vasodilation
the contrast-enhancement during the first pass of the injected contrast agent. Once the Monitoring equipment Heart rate, oxygen saturation, and blood pressure
should be monitored with MRI-compatible
limits of gradient system performance have been reached with gradient-echo imaging,
equipment. EKG rhythm strip monitoring should
multi-shots between conventional gradient echo and echo-planar imaging provide an be available if pharmacologic stress or
additional boost in image acquisition speed. Instead of reading out a single gradient echo vasodilation is planned.
after each slice-selective radio-frequency excitation pulse, segmented pulse sequences
create a train of typically 3 to 5 gradient echoes after each radio-frequency excitation
pulse. Reading out the closely spaced gradient echoes without the need to (re)apply a slice contrast enhancement should ideally be proportional to the contrast agent concentration.
selective radio-frequency pulse before each gradient echo boosts the rate of image Such an approximate linear relationship between regional signal intensity and contrast
acquisition. agent concentration is only observed at lower contrast agent dosages–typically, <0.05
mmol/kg of Gd-DTPA for fast, IR-prepared gradient-echo sequences (TR ≤3 msec; TE ≤2
The image intensity obtained with an inversion-recovery (IR) or saturation-recovery-pre- msec; Jerosch-Herold et al., 1998). A more marked contrast enhancement can be obtained
pared fast gradient echo sequence is strongly T1-weighted. With an IR preparation the with higher contrast agent dosages, but then the kinetics of the contrast agent, and the
signal intensity depends not only on the inversion time after the preparation but also on correlation of contrast enhancement with tissue blood flow can not be assessed in a strict
the state of the magnetization before the inversion pulse is applied. During rapid serial quantitative manner. The choice of contrast agent dosage is therefore in part dictated by
imaging with electrocardiograph-triggered (ECG-triggered) IR-prepared gradient echo the quantitative requirements for a perfusion study.
sequences, therefore, the signal intensity depends on the heart rate. A saturation recovery
magnetization preparation consists of a non-slice selective 90° radio frequency pulse, Table A11.3.1 lists the required equipment for first-pass perfusion imaging.
followed by a gradient crusher pulse to dephase the transverse magnetization (Tsekos et
al., 1995). Driving the magnetization into a well-defined state, with all magnetization NOTE: Adenosine or dipyridamole is not FDA (Food and Drug Administration)-approved
components nulled or dephased, allows one to obtain a T1-weighted gradient-echo signal for use with MRI. Gd-DTPA is not FDA-approved for cardiac imaging. While any
that is independent of the duration of an any previous relaxation recovery delay. The physician can use any drug already FDA-approved for another reason off-label, the
saturation recovery preparation prevents modulation of the image intensity, when the protocol described herein suggests the off-label use of two drugs simultaneously. This
ECG-triggered image acquisition rate varies due to fluctuations in heart rate. Saturation- protocol has been developed for research only.
recovery prepared gradient echo imaging is therefore preferred over the inversion-recov-
Materials
ery preparation when heart rate variations, or an unreliable ECG trigger are of concern
(Tsekos et al., 1995). Furthermore, multi-slice imaging is more easily implemented with Normal saline (0.9% NaCl), sterile
a saturation-recovery preparation than with an IR magnetization preparation (Wilke et al., Extravascular GD-DTPA contrast agent (e.g., Magnevist or Omniscan)
1997). 16-G i.v. needle and injection line
Disposable syringes for power injector
Rapid contrast agent administration is crucial for assessing myocardial perfusion with
contrast agents as this improves the sensitivity for detecting changes of myocardial blood Set up patient and equipment
flow (Kroll et al., 1996). The goal is to assure that the primary bottleneck to the rate of 1. The same precautions as with any other MRI exam should be followed when
Myocardial preparing a patient for an MRI examination of myocardial perfusion. Check for
Perfusion and contrast enhancement is the rate of transport of contrast agent through myocardial tissue, Acquired Heart
Viability and not the rate at which the contrast agent is injected. The regional image intensity possible counter-indications and for ferromagnetic implants. For patients with heart Disease

A11.3.2 A11.3.3
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 disease, check heart rate, blood pressure, and other vital signs while the patient
undergoes the MRI examination. Emergency equipment such as a defibrillator, and
emergency medications for treating cardiac arrests (e.g., epinephrine) should be
nearby. If measurement of perfusion during vasodilation is being contemplated, a
cardiac nurse or a cardiologist should be present to monitor the patient during the
examination, and be prepared to initiate emergency procedures if an adverse event
occurs.
Generally, standard screening forms are used for all patients scanned in a magnetic
resonance system.
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
2. If the procedure is a research protocol, have the patient sign any necessary consent
form.
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating
and image artifacts.
5. Inform the patient about what will occur during the procedure, what he or she will
experience while in the magnet, and how to behave, including the following:
a. If earphones or headphones are used to protect the ears from the loud sounds
produced by the gradients, the patient will be asked to wear these, but will be able
to communicate with you at any time during the imaging.
b. The patient will be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this works).
c. For good results, the patient should not talk, and should avoid or minimize
swallowing or other movement, during each scan–i.e., as long as the banging
sounds continue. Between scans, talking and swallowing are allowed in most
cases, but should be avoided when comparative positional studies are being
performed; the patient will be informed when this is the case.
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
6. An intravenous (i.v.) line should be started before the examination for administration
of the contrast agent. Myocardial perfusion studies require either a central or
peripheral injection of contrast agent, preferably with a power injector at a rate of 5
to 10 ml/sec. A venous catheter line advanced through the antecubital vein to the
superior vena cava, or a 16-G needle in an antecubital vein, will allow injection of
the contrast agent as a compact bolus. Use a dual-port connector with the needle to
allow injection of contrast agent and infusion of pharmacological agent, such as
adenosine.
7. If a power injector is used, the contrast agent and saline should be loaded under sterile
conditions into the proper (disposable) syringes of the power injector. Great care
should be taken to expel all air bubbles from the reservoirs and catheter lines. Luer
fittings should be examined to ensure that they are secure and tight. Injection dosage
and rate should be set appropriately. Newer power injector systems have touch-sen-
sitive liquid crystal display (LCD) screens to set the injection parameters. Contrast
Myocardial agent injections should generally be followed without delay by an injection of
Perfusion and physiologic saline solution to assure that the entire contrast agent dosage is injected
Viability
A11.3.4
Current Protocols in Magnetic Resonance Imaging
into the vein. A constant saline drip before and after bolus administration is recom-
mended to keep i.v. injection lines open.
For older power injector models that cannot be brought near the magnet, the contrast agent
is preloaded through an i.v. catheter and a long i.v. line is connected to a remotely
positioned power injector.
8. Position the patient supinely on the scanning table. Provide him/her with a cushion
under the legs for comfort.
9. Place the ECG electrodes around the heart with ≤4 in. of separation between them.
This results in an improved quality of the ECG signal and more reliable triggering
of the pulse sequence. Align the ECG wires parallel to the magnetic-flux lines of the
static magnetic field. Conductive cables should not cross inside the bore, and they
should also be kept away from the sides of the bore. These precautions are necessary
to minimize the risk of burns due to radio-frequency power deposition.
ECG triggering is often a source of problems in cardiac MRI examinations due to
degradation of the ECG signal by superimposed voltages that are generated by blood flow
through the great vessels while the patient is positioned in a strong magnetic field. Special
MR-compatible physiologic recorders, which can be interfaced with the MRI scanner, often
use customized ECG electrodes and proprietary filtering techniques to reduce artifacts in
the ECG.
The ECG electrodes may be placed on the chest or back. Placing them on the chest is best
for monitoring, but in this position the ECG signal is more susceptible to artifacts from
breathing motion, as well as image artifacts from the leads affecting the right ventricle.
The dry rub method improves electrode contact to the skin and reduces contact resistance.
If the baseline of the ECG moves up and down with respiration, then the ECG leads need
to be moved up higher on the chest. Asking the patient to avoid strong breathing motion
can help. Reducing the distance between the ECG leads can reduce interference from
switching of the gradient coils.
The quality of the ECG signal is normally just sufficient for counting the heart rate and
triggering the MR image acquisition with the R wave of the QRS complex, but not good
enough to diagnose the on-set of ischemia from changes in the ST segment. (Q, R, S, and
T are the standard nomenclature for distinctive features on the ECG trace that correspond
to well defined events during the cardiac cycle. The reader is referred to introductory
physiology textbooks for further details.) It is therefore important to supplement the ECG
signal with measurements of the arterial oxygen saturation and blood pressure, especially
if the patient will be undergoing a stress protocol or pharmacological vasodilation in the
magnet. If pharmacologic stress or vasodilation is planned, fiber optic leads that reduce
the signal artifact caused by patient motion and radiofrequency interference is preferred.
A rhythm strip should be run between each sequence acquisition during adenosine,
dipyridamole, or dobutamine administration. While ischemic changes cannot be moni-
tored, the strips can be monitored for heart block or arrhythmias.
10. Wrap blood pressure cuff around the arm to monitor blood pressure. MRI compatible
monitoring systems use a pump that inflates and deflates the blood pressure cuff at
predefined time intervals or when the technologist manually initiates a blood pressure
measurement. Place a fiber optic sensor on the patient’s finger to measure the arterial
oxygen saturation. Take note of the baseline heart rate, systemic blood pressure, and
oxygen saturation.
11. Use a dedicated cardiac coil for these studies to maximize the achievable SNR. Center
the coil at the level of the heart, or over the patient’s heart, if a relatively small surface
coil is used.
In many cases phase-array coils with anterior and posterior elements fitted under the
patient and over the patient’s chest work best for this purpose. Acquired Heart
Disease
A11.3.5
Current Protocols in Magnetic Resonance Imaging
 Table A11.3.2 Primary Clinical Imaging Parameters for Sequence 1 (Pilot
Scan)

Patient position Supine

Scan type Fast gradient echo with dark
blood preparation
Imaging plane (orientation) Transverse, pseudo-long axis,
double oblique short axis
Central slice or volume center Laser light centered on coil center
(at heart)
Echo time (TE) 2.3 msec
Repeat time (TR) 5 msec
Figure A11.3.1 Scout images for transverse (step 16), pseudo-vertical long axis (step 17), and pseudo-horizontal long
Flip angle (FA) 10°
axis (step 18) in left-to-right order. The images were acquired with a fast steady-state free-precession (SSFP) scout imaging
Fields of view (FOVx, FOVy) 400 mm, 400 mm technique (“true FISP”), which was used because of the excellent definition of anatomical structures that is achieved with
Resolution (Δx, Δy) 1.56 mm, 3.13 mm this technique. The use of SSFP imaging is currently limited to MRI systems with high-performance gradient systems. The
Number of data points collected (Nx, Ny ) 256, 128 thick white lines denote the image plane orientation for the next scout image to the right.
Display matrix (Dx, Dy ) 256, 256
Slice thickness (Δz) 10 mm 19. Acquire one mid-ventricular short axis view as second localizer. This second localizer
Number of slices 1 is used in setting up sequences 2 and 3 to determine the smallest possible field of
Slice gap Not applicable view dimensions that do not cause wrap-around that obscures the view of the heart.
Number of acquisitions (Nacq) 1
Swap read and phase encoding No Sequence 2: Magnetization-prepared, T1-weighted FLASH (fast low angle shot)
Slice location Not applicable perfusion imaging
Saturation pulses Not applicable 20. Call up the rapid, multi-slice T1-weighted spoiled gradient-echo imaging as described
Scan time ∼600 msec in Table A11.3.3. Either a saturation-recovery or an inversion-recovery preparation
can be used, although the former appears to be more advantageous for multi-slice,
arrhythmia-insensitive imaging (Tsekos et al., 1995; Wilke et al., 1997). Table
12. Instruct the patient on the procedure for breath-holding during the MRI examination. A11.3.4 gives the parameters for an inversion-recovery (IR) prepared echo-planar
First pass studies can be performed with the patient breathing normally, although the sequence, that can be used as an alternative to the sequence in Table A11.3.3 for
absence of breathing motion facilitates interpretation of the images. myocardial perfusion imaging. Figure A11.3.2 shows a diagram of the pulse sequence
protocol.
13. Provide the patient with headphones or earplugs. The intercom equipment should be
used during the MRI study to guide the patient and assure his/her cooperation. Alert IR-prepared segmented echo-planar pulse sequences represent a useful alternative on some
the patient before an i.v. injection or infusion, particularly in cases where the infusion scanners for first pass perfusion imaging (Ding et al., 1998; Fischer and Lorenz, 1997;
of a pharmacological agent induces hemodynamic changes. Tell the patient to signal Reeder et al., 1999). Instead of reading out all image data after a single radio-frequency
(RF) pulse as is done in single-shot EPI, in multi-shot sequences RF pulses are applied
any discomfort or the onset of pain. Communication with the patient is important for
repeatedly to produce short echo trains. The echo train length after each radio-frequency
the success and safety of the study! pulse is 4 to 6 echoes, and the optimal TR is on the order of 10 to 15 msec at 1.5 T (Epstein
14. Position the patient in the magnet such that the heart is in the center of the magnet. and Arai, 2000). The pulse sequence is optimized for high temporal resolution (∼6 to 7
slices per 2 R-to-R intervals). A slice-selective inversion-recovery magnetization prepara-
15. Check that a strong ECG R-wave is received before proceeding. Otherwise the ECG tion is applied for each slice. For multi-slice acquisitions, the slice-selective IR preparation
leads have to be reconfigured to produce a more prominent R-wave peak. for a given slice can be applied right before the EPI readout for a preceding slice to
maximize the efficiency in image acquisition while maintaining a sufficiently long inversion
Sequence 1: Scout imaging of the heart time to optimize the T1-weighting. For example, the IR preparation pulse for slice 2 is
applied just before the EPI readout of slice 1. This construction constrains the choice of TI
16. Acquire a transverse scout with a fast gradient echo sequence, preferably with dark
but in practice the duration of the EPI read-out matches well the optimal TI values. Call
blood preparation for better visualization of ventricular cavities. Typical sequence up the rapid, multi-slice, T1-weighted, multi-shot echo-planar imaging sequence as de-
parameters for such a pulse sequence are listed in Table A11.3.2. scribed in Table A11.3.3.
17. Oblique pseudo-long axis scout: Prescribe on the previous image a slice through the 21. Preset the contrast agent injection dosage at ∼0.04 to 0.1 mmol/kg times patient body
middle of the left ventricle and acquire an image for this view as localizer. weight in kilograms divided by the concentration of Gd-DTPA when TR/TE/FA of the
18. Horizontal long axis scout: Prescribe on previous scout, a slice parallel to the septum, saturation-recovery prepared spoiled gradient echo sequence are on the order of 2.4
centered in the LV cavity, positioned from the LV apex, through the mid portion of msec/1.2 msec/18°, respectively. Injection rate depends on the size of the i.v. needle
the mitral valve. Acquire image. An example of the images obtained by following or catheter but should be on the order of 5 ml/sec for a compact bolus.
Myocardial
Perfusion and steps 16 to 18 is shown in Figure A11.3.1. It is useful to remember that the injection rate specified on the control panel of the injector Acquired Heart
Viability is often a nominal injection rate that is not reached if the catheter size is too small. Disease

A11.3.6 A11.3.7
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A11.3.3 Primary Clinical Imaging Parameters for Sequence 2 (First Table A11.3.4 Primary Clinical Imaging Parameters for Sequence 2 (First
Pass Perfusion, T1-weighted FLASH Variant) Pass Perfusion, Multishot EPI variant)

Patient position Supine Patient position Supine

Scan type Rapid serial, multislice 2-D Scan type Rapid, IR-prepared, segmented
gradient echo imaging with (multi-shot) EPI sequence
magnetization preparation for Imaging plane (orientation) Double oblique short axis view
T1-weighting Central slice or volume center Centered over the heart
Imaging plane (orientation) Double oblique short axis view Echo time (TE) As short as possible (usually <2
Central slice or volume center Centered over the heart msec)
Echo time (TE) 1.2 msec, or as short as possible Echo train length (ETL) 4–6
(usually <2 msec) Repeat time (TR) As short as possible (usually <15
Repeat time (TR) 2.4 msec or short as possible msec)
(usually <3 msec) Inversion time (TI) 100–200 msec
Inversion time (TI) 10 msec for saturation recovery ECG trigger delay Minimum
preparation and 100–150 msec for Flip angle (FA) 20°
inversion recovery preparation
Fields of view (FOVx, FOVy ) 360 mm, 360r mm, with r = 3/4
Flip angle (FA) 15°–18° (rectangular field of view)
Fields of view (FOVx, FOVy) 320 mm, 320r mm, with r = 6/8 Resolution (Δx, Δy) 2.80 mm, 2.80 mm
(rectangular field of view)
Number of data points collected (Nx, Ny ) 128, 128r, with r = 3/4
Resolution (Δx, Δy) 2.50 mm, 2.76 mm (rectangular field of view)
Number of data points collected (Nx, Ny) 128, 118r, with r = 6/8 Display matrix (Dx, Dy ) 128, 128
(rectangular field of view)
Slice thickness (Δz) 8 mm
Display matrix (Dx, Dy) 128, 128
Number of slices 3–6, or as many as needed to
Slice thickness (Δz) 10 mm cover the heart from base to apex
Number of slices 2–6, or as many as needed to Slice gap 0.25 of slice thickness
cover the heart from base to apex
Number of acquisitions (Nacq) 1
Slice gap 0.3–0.5 of slice thickness
Number of repetitions ∼40
Number of acquisitions (Nacq) 1
Swap read and phase encoding Only if this reduces aliasing
Number of repetitions 40 (number of serial images for artifacts and allows reduction of
first pass study) FOV
Swap read and phase encoding Only if this reduces aliasing Slice location User define short axis locations
artifacts and allows reduction of between base and apex
FOV
Saturation pulses No
Slice location User define short axis locations
Slice series Interleaved
between base and apex
ECG gating Yes
Saturation pulses No
Scan time ∼40 R-to-R intervals
Slice series Interleaved
ECG gating Yes
Scan time ∼40 R-to-R intervals
23. Use the scout images acquired earlier to prescribe the slices for the perfusion study.
Except for the visualization of apical perfusion defects, double oblique slices that
give a short axis view of the heart are recommended for this study. For multi-slice
22. Adjust the number of phase encoding steps (Ny) to a value between 80 and 120 such acquisitions, set the gap between slices to 30% to -50% of the chosen slice thickness.
that the combination of phase encoding steps and the number of slices result in an An arbitrary gap thickness may be chosen if maintenance of adequate temporal resolution
image acquisition time that matches the heart rate. Maintain an in-plane spatial (1 image per 1 to 2 heart beats) allows only imaging of 2 to 3 slices. The choice of slice
resolution of <3 mm. locations is often guided by findings from wall-motion studies that are performed before-
hand. Slice positions are then chosen based on the location of wall-motion defects.
For a heart rate of 60 beats/min (R-to-R interval of 1000 msec) the repetition time (TR)
times the number of phase encoding steps (Ny) plus the time for the magnetization 24. Minimize the fields of view without causing aliasing (“wrap-around”) artifacts. Use
preparation (TI) determine the time required to image one slice (TACQ): TACQ = Ny*TR + TI. the double oblique short-axis localizer (see sequence 1) for this purpose. Choose the
This time (TACQ) multiplied by the number of slices should be less than the R-to-R interval read-out direction parallel to the chest wall as this reduces the possibility of aliasing
to maintain a temporal resolution of 1 image per heart beat for each slice location. The and other artifacts. See Figure A11.3.3 showing perfusion images for three slice
idle time at the end of the R-to-R interval should be ≥40 msec to anticipate possible small
Myocardial increases in heart rate during the image acquisition that could cause intermittent changes
locations.
Perfusion and in rate of image acquisition. Acquired Heart
Viability Disease

A11.3.8 A11.3.9
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Multislice SR-Prepared FLASH
R R

T
ECG

FLASH FLASH FLASH FLASH FLASH

Slice 1 Slice 2 Slice 3 Slice 1 Slice 2
RF

SR-pulse SR-pulse

Figure A11.3.2 Schematic sequence diagram of multislice perfusion imaging sequence with
nonslice-selective saturation-recovery (SR) preparation before each FLASH readout. The delay
between the SR preparation and the FLASH readout can be kept as short as 10 msec and still
provides good T1-weighting of the signal intensity. By comparison, an inversion recovery preparation
requires a relaxation delay (TI) of ∼100 msec. A slice-selective magnetization-preparation would
lead to modulation of the signal intensity from spins flowing into the magnetization-prepared volume.
A nonslice-selective preparation pulse eliminates this flow-dependent modulation of the signal
intensity.

Figure A11.3.4 Magnified multislice perfusion images in a patient with an inferior perfusion defect.
The perfusion defect results in a reduced and slowed contrast enhancement after the first pass
through the left ventricle. Patient had shown a fixed defect in the inferior wall segment on single plot
on emission computed tomography (SPECT).

25. Start the pharmacological stress protocol before the contrast injection if this is part
Figure A11.3.3 Selected frames from a multislice first pass perfusion study acquired in a 53 year
old male patient with a scarred infarct in the inferior wall. The images are arranged in separate rows
for each slice position, and in each row the images show the sequential appearance of the contrast
agent bolus in the right ventricle, the left ventricle, and enhancement of myocardial tissue. Images
were acquired while the patient held his breath. The images in this example have not been cropped
and they illustrate how the rectangular field of view is adjusted to avoid wrap-around in the phase
encoding direction, which can otherwise obstruct the view of the heart.
Myocardial
Perfusion and
Viability
of the study. Note that adenosine requires 3 min to reach steady state. Slowly
increment the dosage of the pharmacological agent, such as dipyridamole or adeno-
sine if induction of hyperemia is desired. Monitor the patient’s blood pressure and
heart rate. Run a rhythm strip between each dose increase and between each sequence
acquisition. Observe for an atrial block or any other abnormalities in heart-rhythm.
Discontinue the infusion immediately in case problems develop. Have aminophylline
ready in case dipyridamole was used, and it becomes necessary to relieve or reverse
its effects.
The first pass measurement during pharmacological stress can be repeated at 10 to 15 min
after a previous first pass study, for example, a baseline “rest” study. The first pass studies
can be repeated up to 3 times with a dosage of 0.03 to 0.05 mmol/kg of Gd-DTPA and with
the sequence parameters shown in Table A11.3.2.
The effect of potent coronary vasodilators such as adenosine or dipyridamole is adequate
for perfusion imaging when the heart rate increases by 10 bpm and diastolic blood pressure
decreases by 10 mm Hg. The patients often feel some chest pressure and slight discomfort
during vasodilation with adenosine or dipyridamole. Patients on beta-blockers will not Acquired Heart
show a significant rise in heart rate during i.v. adenosine or dipyridamole. Disease

A11.3.10 A11.3.11
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 26. The power-injector should be armed, i.e., readied for injection.
27. If breath-holding is used, the patient should be asked to take a deep breath and hold
his/her breath.
This pulse sequence does not necessarily require breath-holding to avoid breathing motion
artifacts in the image. It is nevertheless recommended that the patient holds his/her breath
as long as possible so that the position of the heart appears fixed in consecutive images.
28. Begin the scan as soon as the patient starts to hold his/her breath. Listen for regular
sounds from the gradients that indicate regular triggering of the pulse sequence by
the ECG. The contrast agent injection should be started after acquisition of 3 to 4
“baseline” images.
Process data and view for sequence 2
29. View the acquired series of sequential images in cine mode to determine that the
contrast agent bolus was successfully injected. It is important to ascertain that a
sufficient number of images were acquired to track the passage of the contrast agent
through the first pass at least to the beginning of recirculation in the left ventricular
cavity. This initial visual assessment should be performed immediately after image
acquisition. Absence of any contrast enhancement due to a problem with the contrast
agent injection or other malfunctions should be corrected and the study repeated no
earlier than 5 min after the previous contrast injection.
The initial visual assessment already provides an impression about any spatial heteroge-
neities and temporal delays in myocardial contrast enhancement. This may indicate
regional impairments in myocardial blood flow, but potential pitfalls in interpretation like
a drop-off in signal intensity due to an inhomogeneous surface coil profile need to be kept
in mind. Figure A11.3.4 shows an example from a “rest” study in a patient with a fixed
perfusion defect in the inferior wall. Pharmacological vasodilation is necessary to detect
mild to moderate perfusion defects.
BASIC IMAGING MYOCARDIAL VIABILITY
PROTOCOL 2
With imaging of hyperenhancement, the requirements for very short image acquisition
times can be somewhat relaxed to improve the signal-to-noise ratio (SNR) and the spatial
resolution. Nevertheless the avoidance of image artifacts from cardiac motion still
requires that the image acquisition be performed only during a relatively small fraction
of the cardiac cycle, and preferably during the most quiescent diastolic period of the
cardiac cycle. The most successful techniques employed for imaging of delayed (hyper-)
enhancement in injured and/or infarcted myocardium have resorted to the segmented
acquisition of the k-space data, but are otherwise very similar to the imaging techniques
employed for first pass imaging. The goal is to maximize the T1-weighted contrast, with
minimum interference from T2* effects and cardiac motion.
T1-weighted imaging is applied ∼10 to 20 min after injection of the contrast agent to
identify severely injured, nonviable myocardial segments (Tong et al., 1993a). For delayed
enhancement studies, one is interested in the imaging differences of distribution volume
of the contrast agent. By the time the contrast agent distribution is in a semi-equilibrium,
the duration of the contrast agent injection has no noticeable effect, unless prolonged over
minutes. Hand injections of the contrast agent are therefore acceptable. Image sequences
that employ segmented acquisitions of k-space data require breath-holding for the
duration of each image acquisition to achieve optimal image quality.
Myocardial Timing for imaging of hyperenhancement is an important aspect of this study. The time
Perfusion and to reach 90% of equilibrium concentration depends on the distribution volume, but
Viability
A11.3.12
Current Protocols in Magnetic Resonance Imaging
generally does not >15 min (Tong et al., 1993a). Larger infarcts can show a core zone that
initially lacks enhancement even at ≥5 min after contrast agent injection. This phenome-
non, linked to microvascular obstruction, was shown to carry a graver prognosis for the
patient, than if the core no-enhancement zone was absent (Rochitte et al., 1998; Wu et al.,
1998b). Microvascular obstruction creates a severe bottleneck for the delivery of contrast
into the affected zone, but given sufficient time a central no-enhancement zone will also
display hyperenhancement compared to remote zones (Wu et al., 1998a). The observation
of hypoenhancement and hyperenhancement will depend on the delay between contrast
agent injection and acquisition of the images. A hypoenhancement zone with a surround-
ing hyperenhanced rim indicates that contrast agent has not yet reached the core zone. A
dark core zone can not simply be interpreted as a zone with reduced distribution volume.
The measurement should be repeated after an additional delay.
In cases where relatively low dosages of contrast agent are used for the first pass study,
one would follow the first pass study with a second and higher dosage injection of contrast
agent to image the late-enhancement. Typical dosages used for imaging of viability are
on the order of 0.1 to 0.2 mmol/kg of Gd-DTPA. For example, with a bolus dosage of
0.05 mmol/kg of Gd-DTPA for a first pass study an additional dosage of 0.15 mmol/kg
of Gd-DTPA contrast agent should be injected for imaging of viability. The time delay
after injection of the contrast agent and before imaging of the delayed hyperenhancement
should be used for optimizing the T1-contrast through adjustment of the image parameters,
primarily the inversion time TI.
Set up patient and equipment
1. The same precautions as with any other MRI exam should be followed when
preparing a patient for an MRI examination of myocardial viability (see Basic
Protocol 1). This includes checking for possible counter-indications and for ferro-
magnetic implants. For patients with suspected acute myocardial infarction, heart
rate, blood pressure, and other vital signs should be checked while the patient
undergoes the MRI examination. Emergency equipment such as a defribillator, and
emergency medications for treating cardiac arrests (e.g., epinephrine) should be
nearby.
2. An i.v. injection line for administration of contrast agent should be started before the
examination.
3. Position the patient supinely on the scanning table. Provide him/her with a cushion
under the legs for comfort.
4. Place the ECG electrodes around the heart with ≤4 in. of separation between each
one. This results in an improved quality of the ECG signal and more reliable
triggering of the pulse sequence. Align the ECG wires parallel to the magnetic flux
lines of the static magnetic field. Do not cross the conductive cables inside the bore,
and also keep away from the sides of the bore.
These precautions are necessary to minimize the risk of burns due to radio-frequency power
deposition.
ECG triggering is often a source of problems in cardiac MRI examinations due to
degradation of the ECG signal by superimposed voltages that are generated by blood flow
through the great vessels while the patient is positioned in a strong magnetic field. Special
MR-compatible physiologic recorders, which can be interfaced with the MRI scanner, often
use customized ECG electrodes and proprietary filtering techniques to reduce artifacts in
the ECG.
The ECG electrodes may be placed on the chest or back. Placing them on the chest is best Acquired Heart
for monitoring, but in this position the ECG signal is more susceptible to artifacts from Disease
A11.3.13
Current Protocols in Magnetic Resonance Imaging
 breathing motion, as well as artifacts from the leads affecting the right ventricle. The dry Table A11.3.5 Primary Clinical Imaging Parameters for Sequence 3
rub method improves electrode contact to the skin and reduces contact resistance. If the (Distribution Volume Assessment)
baseline of the ECG moves up and down with respiration, then the ECG leads need to be
moved up higher on the chest. Asking the patient to avoid strong breathing motion might Patient position Supine
help. Reducing the distance between the ECG leads might reduce interference from Scan type Rapid segmented 2-D gradient
switching of the gradient coils. echo with magnetization
preparation for T1-weighting
5. Use a dedicated cardiac coil for these studies to maximize the achievable signal-to-
Imaging plane (orientation) Double oblique short axis view
noise ratio. Center the coil at the level of the heart, or over the patient’s heart, if a
Central slice or volume center Centered over the heart
relatively small surface coil is used.
Echo time (TE) <3.5 msec, or as short as possible
In many cases phase-array coils with anterior and posterior elements fitted under the Number of lines per segment 23–30
patient and over the patient’s chest work best for this purpose. Repeat time (TR) <8 msec, or as short as possible
Inversion time (TI) ∼150–250 msec
6. Instruct the patient on the procedure for breath-holding during the MRI examination.
Delay time (TD) Adjust to TD to obtain diastolic
Breath-holding durations are relatively short (i.e., ∼10 sec) for T1-weighted imaging
images
of viability with segmented k-space acquisition, as described below.
Flip angle (FA) 15°
7. Provide the patient with headphones or earplugs. Tell the patient how to communicate Fields of view (FOVx, FOVy) 350 mm, 350r mm, with r = 3/4
to the persons performing the exam while he/she lies in the magnet. The intercom (rectangular field of view)
equipment should be used during the MRI study to guide the patient and assure his/her Resolution (Δx, Δy) 1.37 mm, 1.59 mm
cooperation. Number of data points collected (Nx, Ny) 256, 220r, with r = 3/4
(rectangular field of view)
8. Position the patient in the magnet such that the heart is in the center of the magnet. (segmented acquisition of phase
encoding steps if possible)
9. Check that you receive a strong R-wave before proceeding. Otherwise the ECG leads Display matrix (Dx, Dy) 256, 256
have to be reconfigured to produce a more prominent R-wave peak. Slice thickness (Δz) 6 mm
10. Run the scout sequence as listed in sequence 1 of Basic Protocol 1. Number of slices 1; repeated breath-holds for
acquisition of contiguous slices
Sequence 3: T1-weighted gradient-echo imaging with segmented acquisition from base to apex
11. Call up the protocol with a pulse sequence for rapid, T1-weighted imaging as Slice gap 0.0–0.5 of slice thickness
described in Table A11.3.5. The parameters in Table A11.3.5 are for an ECG-triggered Number of acquisitions (Nacq) 1
FLASH sequence with segmented acquisition of k-space lines (33 lines per segment) Swap read and phase encoding Only if this reduces aliasing
artifacts and allows reduction of
and a non-slice selective inversion recovery (IR) magnetization preparation (Simon-
FOV
etti et al., 2000). Figure A11.3.5 shows a pulse sequence diagram that illustrates the
Slice location User define short axis locations
acquisition technique and acquisition parameters. between base and apex
The IR magnetization preparation should be repeated before each k-space acquisition. For Saturation pulses No
the parameters listed in Table A11.3.5, and a contrast agent dosage of 0.2 mmol/kg, one Slice series Interleaved
should use an inversion time (TI) of 150 msec as starting value. TI refers to the delay between ECG gating Yes
the inversion pulse and the acquisition of the first phase-encoded line of a segment, as Scan time 6–12 heart beats
illustrated in Fig. A11.3.5. The definition of TI can vary between different sequences and
is also manufacturer-dependent. It may, for example, also refer to the time between the
inversion pulse and the acquisition of the central k-space line of each segment. TI needs to
be adjusted by the user to null the signal from presumably normal myocardium, and to 13. If the patient can not hold his/her breath, the number of acquisitions (or number of
maximize the differential contrast enhancement of injured and non-viable myocardium. The excitations – NEX) needs to be set to a value of ≥3 to average out respiratory motion.
protocol produces best results if the patient holds his breath during image acquisition. The number of averages can be <3 if respiratory gating can be used.
12. A delay of at least 2 heartbeats should follow the acquisition of each k-space segment 14. Use the scout images acquired earlier to prescribe the slice positions for imaging of
to allow for relaxation of the magnetization before the next inversion-recovery delayed hyperenhancement.
preparation is applied. Without a delay the observable contrast enhancement is
Imaging the entire heart with contiguous slices in the short-axis orientation assures that
reduced.
the extent of infarcts or severe injury can be accurately assessed.
As with all breath-hold sequences, the adjustment of the total number of phase encoding
15. Inject the contrast agent and take note of time. Wait now ≥1 min before starting to
steps, the relaxation delay after acquisition of a k-space segment and the number of
averages should result in a total acquisition time that is within a comfortable breath-hold optimize the inversion time (TI). This requires acquiring images for a series of
duration. different TI values, and determining the TI that will result in nulling of signal from
Myocardial normal myocardium. TI can be set 5% to 10% higher than the optimal value to
Perfusion and Acquired Heart
Viability Disease

A11.3.14 A11.3.15
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Segmented Turbo FLASH
R R
T
ECG
FLASH relaxation delay FLASH
TI
seg 1 (1-2 beats) seg 2
RF
IR-pulse IR-pulse
Figure A11.3.5 Timing diagram for segmented, ECG-gated gradient echo sequence with inver-
sion recovery preparation for imaging of delayed enhancement. The inversion-recovery preparation
is repeated before acquisition of each segment of phase-encoding lines, and a sufficient delay on
the order of 2 to 3 heartbeats needs to be placed between the inversion pulse and the previous
segment acquisition. The inversion time (TI) is adjusted 1 to 2 min after injection of the contrast
agent to null the signal in normal myocardium.
Figure A11.3.6 Images of early (left) and delayed (right) contrast enhancement in a 65 year old
male patient 4 days after his first acute large anterior myocardial infarction (creatine kinase >7000).
In the images of the early enhancement a dark region in the antero-septal region indicates the
presence of a no-reflow zone. The image acquired with a 10 min delay shows hyperenhancement
in the same area and clearly delineates the extent of the infarction. The images were acquired with
an ECG-gated fast gradient echo sequence with an inversion recovery preparation for maximum
T1-weighting. Blood signal supression was used for better delineation of the endocardial border,
and images were acquired at end-diastole. Images courtesy of Drs. João Lima and Bernhard
Gerber, Johns Hopkins Medical Institution, Baltimore.
anticipate the slow, but steady decrease of contrast agent concentration (the half-time
for Gd-DTPA is ∼30 min).
16. Acquire images for slices in the short axis orientation from base to apex using a 6-mm
slice thickness and no gap, or a small gap of <0.2 of slice thickness. Ask the patient
to hold his/her breath for each image acquisition as the sequence for the slice positions
from base to apex is repeated.
For the detection of microvascular obstruction it is necessary to perform the acquisition of
images <5 to 10 min after the injection of contrast agent (Rochitte et al., 1998; Rogers et
al., 1999). The hypoenhancement characeteristic of microvascular obstruction will only
Myocardial persist for a limited period after which the contrast between non-infarcted tissue and the
Perfusion and
Viability zone with microvascular obstruction disappears. Consistent hyperenhancement through-
A11.3.16
Current Protocols in Magnetic Resonance Imaging
Figure A11.3.7 Images of delayed contrast enhancement obtained in a 49-year-old male with a
history of two myocardial infarctions—one myocardial infarction (MI) four years before MRI followed
by percutaneous transluminal coronary angioplasty (PTCA) and stent to his right coronary artery
(RCA). The images shown in the figure were acquired 2.5 months after a second, anterior wall MI
and subsequent coronary artery bypass graft (CABG) surgery. Gadolinium DTPA of 40 ml was used
and imaging was performed ∼30 min after contrast injection with a segmented turbo FLASH
sequence as shown in Fig. A11.3.5. Images provided by Dr. R. White, Cleveland Clinic Foundation.
out the entire infarct zone has been reported in studies where contrast enhancement was
measured 20 to 30 min after injection of the contrast agent (Kim et al., 1999). Figure
A11.3.6 shows images during early (<1 to 2 min post-injection) and delayed enhancement
(∼10 min post-injection) acquired in a patient with myocardial infarct in the anterior wall.
Figure A11.3.7 shows images in another patient acquired ∼30 min after contrast agent
injection with the sequence and parameter settings described in Table A11.3.5.
Gadopentetate dimeglumine (Gd-DTPA) has been the predominantly used contrast agent
for studies of myocardial viability. Hyperenhancement of injured and non-viable myocar-
dial tissue reflects an increased volume of distribution compared to normal myocardium.
With protein-binding contrast agents the hyperenhancement in injured and/or non-viable
myocardium may not only reflect an increase of the distribution volume but also a change
of the contrast agent relaxivity with binding to proteins released after myocardial injury.
Process data and view for sequence 3
17. View images of delayed contrast enhancement with an identical window-setting for
all slices. A simultaneous display of images for all slices is recommended for a visual
assessment of the extent of contrast hyperenhancement.
18. Software with simple planimetric measurement tools is useful to quantify the extent
of myocardial regions with contrast hyperenhancement. Determine the signal inten-
sity in the region with hyperenhancement and in a remote region. Measure the relative
area of hyperenhancement in each slice.
COMMENTARY
Background Information of coronary artery disease in the United States
Coronary artery diseases remain the most was $118.2 billion in 1999. While nuclear
common cause of death for Americans. It is medicine and echocardiography tests remain
estimated that this year 1,100,000 Americans the primary imaging modalities for assessing
will suffer a new or recurrent myocardial in- the presence of myocardial ischemia and viabil-
farction. Over 40% of people who experience ity, each test suffers from limitations of spatial
a myocardial infarction in a given year will die resolution and artifacts. First-pass MRI perfu- Acquired Heart
from it. It is estimated that the economic cost sion imaging is emerging as a very capable Disease
A11.3.17
Current Protocols in Magnetic Resonance Imaging Supplement 3
 alternative for non-invasive assessment. The
excellent spatial resolution of MRI allows for
the detection of subendocardial ischemia. This
is not possible with other tests and explains, in
part, why some studies reported better sensitiv-
ity and specificity with MRI than other non-in-
vasive imaging tests.
While the first pass of gadolinium contrast
agent is useful for the detection of ischemia,
late enhancement of the myocardium identifies
injured myocardium, infarct, and scar. It can be
used to effectively assess viable myocardium.
Late enhancement combined with the detection
of wall motion abnormalities by MR cine ap-
pears to be highly effective in establishing a
cardiac cause for chest pain in an emergency-
room setting.
Critical Parameters and
Troubleshooting
First pass image quality is limited by signal-
to-noise ratio. Thus, it is advantageous to use a
phase-array torso coil. As the sensitivity profile
of these coils is less at the lateral wall of the left
ventricle than the interventricular septum or
anterior wall, signal in the lateral wall needs to
be normalized not to mistake it for ischemia.
This can be corrected by acquiring calibration
images with the surface coil or the body coil as
receive elements, respectively. Calibration im-
ages should be acquired for the same slice
locations and fields of view as used for the
perfusion study. Comparison of the signal in-
tensity in the two calibration images for iden-
tical region of interest (ROI) locations allows
accurate determination of the scaling factors
that normalize the signal intensity curves to a
common standard. Susceptibility changes with
the passage of contrast through the heart can
produce apparent subendocardial perfusion de-
fects that are artifactual. These appear predomi-
nantly in the read direction and are more severe
with longer TE. These artifacts tend not to per-
sist as long as true ischemic changes and can
be accounted for with experience.
Better SNR is achievable at higher doses of
gadolinium-based contrast agent. However, the
high concentration of contrast media in the
blood will result in saturation of signal and
eliminates the potential for quantitative analy-
sis. The authors find that 0.03 to 0.05 mmol/kg
of Gd-DTPA provides adequate SNR with
minimal saturation effects. It is possible to
repeat the injection if necessary to either in-
crease the coverage of the left ventricle or to
Myocardial
Perfusion and obtain a stress set of images in addition to the
Viability
A11.3.18
Supplement 3
baseline. Antecubital injections at 5 to 8 ml/sec
provide a satisfactory bolus in most patients.
Adequate coverage of the left ventricle for
first pass imaging requires at least three 1-cm
slices in the short axis. The number of slices
that can be obtained in a sequence depends on
both the minimum TR and the heart rate. Like-
wise, the spatial resolution (number of phase
encoding steps) is inversely related to the tem-
poral resolution. Thus, for fast heart rates, it
may be necessary to obtain an image of every
other heart beat in order to obtain adequate
spatial resolution.
Optimal detection of myocardial ischemia
depends on achieving maximal hyperemia.
This may be accomplished with either adeno-
sine or dipyramidole. It is vital to monitor the
blood pressure to avoid hypotension. Addition-
ally, it is important to remain in close commu-
nication with the patient to assess chest pain or
dyspnea. If any of these conditions present, the
drug is immediately discontinued and appro-
priate therapy is given.
A higher dose of contrast agent is commonly
used for studies of enhancement. Good results
have been obtained with 0.1 mmol/kg of Gd-
DTPA. It is recommended to image 3 to 15 min
after injection.
Anticipated Results
Visual inspection of first pass perfusion im-
ages can identify ischemic zones in many pa-
tients. As indicated above, it is important to
assure that a dark zone is not an artifact by
verifying that it persists for more than a few
seconds. Even greater sensitivity may be
achieved by the use of quantitative methods that
model the tissue perfusion.
Delayed enhancement of infarcted myocar-
dium appears to be robust and a number of
groups have excellent results with this simple
test. There does remain some controversy as to
whether all hyperenhanced myocardium is in
fact non-viable (Rogers et al., 1999). Neverthe-
less, in the appropriate setting, this is a very
useful test.
Data processing and viewing of perfusion
images
Region of interest analysis is very useful to
generate time courses for the signal intensity
that facilitate the interpretation of regional con-
trast enhancement (Wilke et al., 1993).
Changes in signal intensity from image to im-
age are assumed to reflect changes in contrast
agent residue concentration but this interpreta-
Current Protocols in Magnetic Resonance Imaging
140
LV
120
slope
100
80 peak ampl.
60 anterior
posterior
Signal intensity [a.u.]
40
20
0
0 5 10 15 20 25
Time (seconds)
Figure A11.3.8 Signal intensity curves from regions of interest (ROI) in a series of perfusion
images, that is also shown in Fig. A11.3.4. The signal intensity curves correspond to the basal slice
position. The dashed line represents the signal changes in for an ROI in the center of the left
ventricular (LV) cavity. The signal curves with open circles and triangles were obtained for transmural
ROI’s in the anterior and posterior segments, respectively. Signal curves were scaled to correct for
inhomogeneity of sensitivity profile of the receiver coil. Relevant parameters for assessing perfusion
in a semiquantitative manner are the up-slope during contrast agent wash-in, and the peak
amplitude. The definition of these two parameters is illustrated in the inset of this figure. Perfusion
in the inferior/posterior wall segment was abnormally low in this patient.
tion applies rigorously only under special con- more closely with relative changes of tissue
ditions (Burstein et al., 1991). The resulting blood flow when the bolus is intravenously
curves of signal intensity versus time can be injected (Bassingthwaigthe et al., 1993).
compared for different myocardial ROI’s to Calculation of semi-quantitative parameters
compare the kinetics of contrast agent wash-in needs, in general, to be performed “off-line” on
(Jerosch-Herold et al., 1999; Kroll et al., 1996). a personal computer or workstation with appro-
The initial increase in signal intensity during priate analytical software to analyze the signal
wash-in of the contrast agent represents the curve characteristics. Semi-quantitative perfu-
portion of the signal curves that is most sensi- sion parameters such as the up-slope of the
tive to changes in myocardial blood flow signal intensity curve are best determined from
(Jerosch-Herold et al., 1998; Kroll et al., 1996). a smooth approximation to the measured data.
Several semi-quantitative parameters may be This requires the use of some sort of model that
used to characterize the contrast agent up-take. is fit to the measured data. Figure A11.3.8
Examples are the up-slope of the signal curve shows an example where a model of the impulse
during the contrast agent wash-in, which is response was used for convolution with the
indicative of the rate of contrast agent up-take, measured “input” function in the left ventricle
and the peak amplitude during the first pass that to calculate a tissue curve. The model parame-
provides a relative measure of the peak concen- ters were adjusted to the measured data by using
tration of contrast agent in a tissue region. With a least squares fitting algorithm.
a compact bolus, as can be achieved by central A simple approach to obtaining a smooth
injection of the contrast agent, the maximum approximation to the measured data consists of
amplitude of the tissue curves is indicative of performing a nonlinear least squares fit of a
the level of tissue perfusion, i.e., blood flow. gamma-variate function to the first pass of the
This correlation of the peak amplitude with measured data. From the fit of the gamma-vari-
blood flow becomes weaker as the bolus dura- ate function to the measured data, only the Acquired Heart
tion is increased, and the up-slope correlates portion of the signal curve before appearance Disease
A11.3.19
Current Protocols in Magnetic Resonance Imaging
 of the recirculation can be used. The recircula-
tion of tracer often can not be clearly identified
in tissue curves and the reappearance of con-
trast agent in the left ventricle should then be
used as a “cut-off” criterion. Many graphing
and data analysis programs for personal com-
puters offer the option of fitting the data to a
user-defined function, such as a gamma-variate
function, with a non-linear least squares algo-
rithm. A rigorous analysis of the curves with
tracer kinetic models is possible if the MRI
protocol is set up such that the signal intensity
can provide an accurate measure of the contrast
agent concentration in blood and tissue regions.
Tracer kinetic modeling generally requires the
use of an arterial input function. The signal time
course for an ROI in the left ventricle is often
chosen as an approximation for this purpose.
For further details on modeling of myocardial
perfusion the reader is referred to the literature
(Jerosch-Herold and Wilke, 1997; Jerosch-
Herold et al., 1999; Kroll et al., 1996; Wilke
and Jerosch-Herold, 1998).
Acknowledgements
Betsy Wilson, RN and Robert Wilson, MD
provided valuable input on patient preparation,
patient monitoring and pharmacological
“stress” protocols.
Literature Cited
Bassingthwaigthe, J.B., Raymond, G.R., and Chan,
J.I. 1993. Principles of tracer kinetics. In Nuclear
Cardiology: State of the Art and Future Direc-
tions (B. L. Zaret and G. A. Beller, eds.) pp. 3-23.
Mosby-Year Book, St. Louis.
Burstein, D., Taratuta, E., and Manning, W.J. 1991.
Factors in myocardial “perfusion” imaging with
ultrafast MRI and Gd-DTPA administration.
Magn. Reson. Med. 20:299-305.
Ding, S., Wolff, S.D., and Epstein, F.H. 1998. Im-
proved coverage in dynamic contrast-enhanced
cardiac MRI using interleaved gradient-echo
EPI. Magn. Reson. Med. 39:514-519.
Epstein, F.H. and Arai, A.E. 2000. Optimization of
fast cardiac imaging using an echo-train readout.
J. Magn. Reson. Imaging 11:75-80.
Fischer, S.E. and Lorenz, C.H. 1997. Determining
heart muscle perfusion by magnetic resonance
tomography progressing to clinical application.
Radiologe 37:366-371.
Jerosch-Herold, M. and Wilke, N. 1997. MR first
pass imaging: Quantitative assessment of trans-
mural perfusion and collateral flow. Int. J. Card.
Imaging 13:205-218.
Jerosch-Herold, M., Wilke, N., and Stillman, A.E.
1998. Magnetic resonance quantification of the
Myocardial myocardial perfusion reserve with a Fermi func-
Perfusion and tion model for constrained deconvolution. Med.
Viability Phys. 25:73-84.
A11.3.20
Jerosch-Herold, M., Wilke, N., Wang, Y., Gong,
G.R., Mansoor, A.M., Huang, H., Gur-
chumelidze, S., and Stillman, A.E. 1999. Direct
comparison of an intravascular and an extracel-
lular contrast agent for quantification of myocar-
dial perfusion. Cardiac MRI Group. Int. J. Card.
Imaging 15:453-464.
Judd, R.M., Lugo-Olivieri, C.H., Arai, M., Kondo,
T., Croisille, P., Lima, J.A., Mohan, V., Becker,
L.C., and Zerhouni, E.A. 1995. Physiological
basis of myocardial contrast enhancement in fast
magnetic resonance images of 2-day-old reper-
fused canine infarcts. Circulation 92:1902-1910.
Kim, R.J., Chen, E.L., Lima, J.A., and Judd, R.M.
1996. Myocardial Gd-DTPA kinetics determine
MRI contrast enhancement and reflect the extent
and severity of myocardial injury after acute
reperfused infarction. Circulation 94:3318-
3326.
Kim, R.J., Fieno, D.S., Parrish, T.B., Harris, K.,
Chen, E.L., Simonetti, O., Bundy, J., Finn, J.P.,
Klocke, F.J., and Judd, R.M. 1999. Relationship
of MRI delayed contrast enhancement to irre-
versible injury, infarct age, and contractile func-
tion. Circulation 100:1992-2002.
Kroll, K., Wilke, N., Jerosch-Herold, M., Wang, Y.,
Zhang, Y., Bache, R.J., and Bassingthwaigthe,
J.B. 1996. Accuracy of modeling of regional
myocardial flows from residue functions of an
intravascular indicator. Am. J. Physiol.
40:H1643-H1655.
Lima, J.A., Judd, R.M., Bazille, A., Schulman, S.P.,
Atalar, E., and Zerhouni, E.A. 1995. Regional
heterogeneity of human myocardial infarcts
demonstrated by contrast-enhanced MRI. Poten-
tial mechanisms. Circulation 92:1117-1125.
Manning, W.J., Atkinson, D.J., Grossman, W.,
Paulin, S., and Edelman, R.R. 1991. First-pass
nuclear magnetic resonance imaging studies us-
ing gadolinium-DTPA in patients with coronary
artery disease. J. Am Coll. Cardiol. 18:959-965.
Pereira, R.S., Prato, F.S., Sykes, J., and Wisenberg,
G. 1999. Assessment of myocardial viability us-
ing MRI during a constant infusion of Gd-DTPA:
further studies at early and late periods of reper-
fusion. Magn. Reson. Med. 42:60-68.
Pereira, R.S., Prato, F.S., Wisenberg, G., and Sykes,
J. 1996. The determination of myocardial viabil-
ity using Gd-DTPA in a canine model of acute
myocardial ischemia and reperfusion. Magn.
Reson. Med. 36:684-693.
Reeder, S.B., Atalar, E., Faranesh, A.Z., and
McVeigh, E.R. 1999. Multi-echo segmented k-
space imaging: An optimized hybrid sequence
for ultrafast cardiac imaging. Magn. Reson. Med.
41:375-385.
Rochitte, C.E., Lima, J.A., Bluemke, D.A., Reeder,
S.B., McVeigh, E.R., Furuta, T., Becker, L.C.,
and Melin, J.A. 1998. Magnitude and time
course of microvascular obstruction and tissue
injury after acute myocardial infarction. Circu-
lation 98:1006-1014.
Rogers, W.J., Jr., Kramer, C.M., Geskin, G., Hu,
Y.L., Theobald, T.M., Vido, D.A., Petruolo, S.,
Current Protocols in Magnetic Resonance Imaging
and Reichek, N. 1999. Early contrast-enhanced
MRI predicts late functional recovery after
reperfused myocardial infarction. Circulation
99:744-750.
Rossum, A.C.v., Visser, F.C., Van Eenige, M.J.,
Sprenger, M., Valk, J., Verheugt, F.W., and Roos,
J.P. 1990. Value of gadolinium-dethylene-
triamine pentaacetic acid dynamics in magnetic
resonance imaging of acute myocardial in-
farction with occluded and reperfused coronary
arteries after thrombolysis. Am. J. Cardiol.
65:845-851.
Simonetti, O., Kim, R.J., Fieno, D.S., Hillenbrand,
H., Wu, E., Bundy, J.M., Finn, J.P., and Rudd,
R.M. 2000. An improved MRI technique for the
visualization of myocardial infarction. Radiol-
ogy 218:215-223.
Tong, C.Y., Prato, F.S., Wisenberg, G., Lee, T.Y.,
Carroll, E., Sandler, D., and Wills, J. 1993a.
Techniques for the measurement of the local
myocardial extraction efficiency for inert diffus-
ible contrast agents such as gadopentate di-
meglumine. Magn. Reson. Med. 30:332-336.
Tong, C.Y., Prato, F.S., Wisenberg, G., Lee, T.Y.,
Carroll, E., Sandler, D., Wills, J., and Drost, D.
1993b. Measurement of the extraction efficiency
and distribution volume for Gd-DTPA in normal
and diseased canine myocardium. Magn. Reson.
Med. 30:337-346.
Tsekos, N.V., Zhang, Y., Merkle, H., Wilke, N.,
Jerosch-Herold, M., Stillman, A., and Ugurbil,
K. 1995. Fast anatomical imaging of the heart
and assessment of myocardial perfusion with
arrhythmia insensitive magnetization prepara-
tion. Magn. Reson. Med. 34:530-536.
Wilke, N. and Jerosch-Herold, M. 1998. Assessing
myocardial perfusion in coronary artery disease
with magnetic resonance first-pass imaging.
Cardiol. Clin. 16:227-246.
Wilke, N., Jerosch-Herold, M., Stillman, A.E.,
Kroll, K., Tsekos, N., Merkle, H., Parrish, T., Hu,
X., Wang, Y., Bassingthwaigthe, J., et al. 1994.
Concepts of myocardial perfusion imaging in
magnetic resonance imaging. Magn. Reson. Q.
10:249-286.
Wilke, N., Jerosch-Herold, M., Wang, Y., Huang, Y.,
Christensen, B.V., Stillman, A.E., Ugurbil, K.,
McDonald, K., and Wilson, R.F. 1997. Myocar-
dial perfusion reserve: Assessment with mul-
Current Protocols in Magnetic Resonance Imaging
tisection, quantitative, first-pass MR imaging.
Radiology 204:373-384.
Wilke, N., Simm, C., Zhang, J., Ellermann, J., Ya,
X., Merkle, H., Path, G., Ludemann, H., Bache,
R.J., and Ugurbil, K. 1993. Contrast-enhanced
first pass myocardial perfusion imaging: Corre-
lation between myocardial blood flow in dogs at
rest and during hyperemia. Magn. Reson. Med.
29:485-497.
Wu, K.C., Kim, R.J., Bluemke, D.A., Rochitte, C.E.,
Zerhouni, E.A., Becker, L.C., and Lima, J.A.
1998a. Quantification and time course of mi-
crovascular obstruction by contrast-enhanced
echocardiography and magnetic resonance im-
aging following acute myocardial infarction and
reperfusion. J. Am. Coll. Cardiol. 32:1756-1764.
Wu, K.C., Zerhouni, E.A., Judd, R.M., Lugo-
Olivieri, C.H., Barouch, L.A., Schulman, S.P.,
Blumenthal, R.S., and Lima, J.A. 1998b. Prog-
nostic significance of microvascular obstruction
by magnetic resonance imaging in patients with
acute myocardial infarction. Circulation 97:765-
772.
Internet Resources
http://www.heartmri.com
This web site is addressed at users of GE scanners
and provides specifics on sequences and techniques
for imaging of myocardial perfusion and viability.
http://www.drad.umn.edu/cvmr/home/html
The authors’ web site has a document in Adobe
Acrobat format with specific instructions on how to
perform myocardial perfusion studies on a Siemens
Vision scanner.
http://nsr.bioeng.washington.edu
This is a useful site for readers interested in state-
of-the-art tracer kinetic analysis that has been ap-
plied for analysis of MRI perfusion data. The
National Simulation Resource is an NIH-funded
resource.
Contributed by Michael Jerosch-Herold and
Arthur E. Stillman
University of Minnesota
Minneapolis, Minnesota
Acquired Heart
Disease
A11.3.21
Cardiac Function Evaluation with Cine MRI UNIT A11.4 Table A11.4.1 Equipment Parameters for Cine MRI

of the Heart Coil type Quadrature phase array torso coil

Gradient coil strength 24 mT/m (or higher if the system permits)
MRI is considered to be the gold standard for the calculation of hemodynamic parameters BASIC Cardiac gating Yes
of cardiac function such as ejection fraction (EF), end diastolic volume (EDV), end PROTOCOL Peripheral gating For monitoring only
systolic volume (ESV), stroke volume (SV), and cardiac mass. With MRI, cine loops of Respiratory gating No
the heart are acquired over several heartbeats in order to perform enough image encoding Respirator If required by patient
steps for each cardiac phase. The acquisition of image data for each cardiac phase is Oxygen If required by patient
synchronized with the heart cycle by gating of the encoding steps with the patients Breath-holding Optional
electrocardiogram (ECG). Recently, it has also become feasible to acquire cine loops in Motion cushions Can be used for patient comfort
real time, although the temporal resolution is not optimal, even with current state-of-the- Use of contrast agents No
art MR scanners. Nevertheless, in patients without regular rhythm, these new real-time Infusion pump Required for pharmacological stress
techniques have started to offer an acceptable alternative. Monitoring equipment Heart rate, oxygen saturation, and blood
pressure should be monitored with
Despite the recognized strengths of cardiac MRI for the assessment of cardiac function, MRI-compatible equipment
its use still lags far behind the use of echocardiography. Besides issues of cost and
availability of dedicated cardiovascular MRI scanners, the lack of training in this area
may also have contributed to a delay in adopting cardiac MRI for cardiac function phases of the cardiac cycle have been sampled adequately for image reconstruction.
evaluation. This unit is meant as a practical guide to some useful MRI protocols for cardiac Retrospective gating allows for coverage of the entire cardiac cycle. Some manufacturers
function evaluation. The parameters of these protocols are based on experience on a allow for rejection of data acquired from premature ventricular contractions with their
Siemens 1.5 T Sonata scanner and should be altered accordingly for different field retrospective gating software. It is not possible, however, to obtain the data in one
strengths and manufacturers. breath-hold duration, as these pulse sequences tend to be lengthier than prospectively-
gated acquisitions. Generally, prospective triggering is adequate. However, it is important
Imaging of cardiac function is currently best performed with cardiac-gated pulse se- to obtain as many phases as possible without spilling into the next cardiac cycle.
quences that split the image acquisition for each frame into a number of segments that Allowances must be made for change in heart rate during the data acquisition. Usually, it
are acquired over, typically, 10 to 20 heartbeat cycles. Still, there are many options to is sufficient to cover less than the R-to-R interval by for example, 100 msec of the cardiac
consider when doing these functional studies: breath hold versus free breathing; prospec- cycle. The entire cine protocol with acquisition of localizers for planning of cine
tive triggering versus retrospective gating; and volumetric data sets versus biplanar acquisitions takes ∼15 min.
approaches. Other issues that factor in are the patient’s heart rate or rhythm, degree of
functional impairment, the presence of valvular disease, and the need to assess for jets Table A11.4.1 lists the hardware necessary to perform cardiac function evaluation with
from shunts or valve dysfunction. There is no one unique way to do these studies, as cine MRI.
trade-offs must be made to accommodate a particular clinical situation.
NOTE: Be sure the technologists and nurses have immediate access to any emergency
The sharpest images are obtainable with breath holding and use of a segmented k-space equipment that may be relevant to a given study, or that may be needed for a particular
approach (Schulen et al., 1996). The more image encoding steps (i.e., lines of k-space) patient such as crash carts or oxygen.
acquired per heartbeat, the shorter the necessary breath-hold duration. However, temporal
resolution per cardiac phase will be reduced and there will be loss of myocardial-blood- Set up patient and equipment
pool contrast because of saturation effects, noticeable in particular for later phases of the 1. Take standard precautions as with any other MRI exam when preparing the patient
cardiac cycle. The effective temporal resolution can be improved using view-sharing for an MRI examination of cardiac function. Check for possible counter-indications
interpolation, whereby some image-encoding steps are shared for the reconstruction of and for ferromagnetic implants. For patients with heart disease, monitor heart rate,
images for adjacent cardiac phases. It is best to have an effective temporal resolution of blood pressure, and other vital signs while the patient undergoes the MRI examina-
≤40 msec, particularly for rapid heart rates. tion. Emergency equipment such as a defibrillator, crash cart, and emergency medi-
cations for treating cardiac arrests (e.g., epinephrine) should be nearby.
Gating of the image acquisition can be performed prospectively. The repetition of the
Generally, standard screening forms (see APPENDIX 1) are used for all patients scanned in
encoding steps in the pulse sequence is then controlled by a physiologic marker, e.g., the a magnetic resonance system.
peak of the R-wave in the electrocardiogram. (The R-wave peak in the electrocardiogram
coincides with the closing of the mitral valve, i.e., with the end of diastole.) By contrast, In patients with ferromagnetic implants, if in doubt as to the exact composition of the items,
with retrospective gating, the pulse sequence continuously acquires data, and the physi- it is best to exclude patients with any metal implants; see Shellock (1996) for a discussion
of what implants may be safely scanned using magnetic resonance.
ological gating signal is recorded independently and is not used to time the image
acquisition. During image reconstruction, each acquired phase encoding is assigned to a Patients may be accompanied into the magnet room by a friend or family member, who can
phase of the cardiac cycle according to the physiological marker. Retrospectively gated Cardiac Function sit in the room during the scan and comfort the patient as needed. This companion must
Evaluation with be screened as well to ensure the absence of loose metal objects on the body or clothing.
cardiac cine sequences require some redundancy in the acquired data to assure that all Acquired Heart Cine MRI of the
Disease Heart
Contributed by Arthur E. Stillman and Michael Jerosch-Herold A11.4.1 A11.4.2
Current Protocols in Magnetic Resonance Imaging (2003) A11.4.1-A11.4.19
Copyright © 2003 by John Wiley & Sons, Inc. Supplement 8 Supplement 8 Current Protocols in Magnetic Resonance Imaging
 2. If the procedure is a research protocol, have the patient sign any necessary consent Table A11.4.2 Primary Clinical Imaging Parameters for Sequence 1 (Pilot Scan)
forms.
Patient position Supine
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
Scan type Short TR gradient echo with
that might be found in clothing. steady-state free precession.
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating. Imaging plane (orientation) Transverse and coronal, pseudo-long
5. Explain to the patient the image protocol and what he or she should expect during axis, oblique short axis, or long axis
(four chamber)
the exam:
Central slice or volume center Laser light centered on coil center
a. Ascertain whether the patient can hold his/her breath for ∼10 to 15 sec. Echo time (TE) 2.3 msec
b. Provide the patient with ear plugs or headphones and let the patient know, in Repeat time (TR) 5 msec
advance, that the pulse sequences will cause a loud noise in the magnet. Instruct Flip angle (FA) 50°
the patient on how he/she can communicate with the person controlling the MRI Fields of view (FOVx, FOVy) 400 mm, 340–400 mm
scanner during the entire duration of the scan. Resolution (Δx, Δy) 1.56 mm, ∼2.43–2.86 mm
c. Instruct the patient not to move and to maintain a regular breathing pattern while Number of data points collected (Nx, Ny) 256, ∼140
he or she hears the scanner acquiring images. Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 8 mm
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
Number of slices 3 for transverse and 3 for coronal
6. Position the patient supine on the patient bed. Provide the patient with a cushion under views, 3 for pseudo-long axis views,
5–6 for oblique short-axis views, or 1
the legs for comfort.
for four-chamber view
7. Place the ECG electrodes around the heart with ≤2 to 3 in. of separation between Slice gap ≤2 mm
them. This results in an improved quality of the ECG signal and more reliable Number of acquisitions (Nacq) 1
triggering of the pulse sequence. Align the ECG wires parallel to the magnetic flux Swap read and phase encoding Swap if necessary (phase encoding
lines of the static magnetic field. Conductive cables should not cross inside the bore, direction is anterior–posterior)
and they should also be kept away from the sides of the bore. These precautions are Slice locations Not applicable
necessary to minimize the risk of burns due to radio frequency power deposition. Saturation pulses Not applicable
Optimal ECG triggering is vital to obtaining a quality MRI examination. Nevertheless, ECG gating Yes
optimal ECG triggering is often problematic in cardiac MRI examinations due to degra- Scan time ∼1 sec/slice, total time is ∼6 sec
dation of the ECG signal by superimposed voltages that are generated by blood flow (transverse and coronal), ∼3 sec
through the great vessels while the patient is positioned in a strong magnetic field. Special (pseudo-long axis), or 5-6 sec (oblique
MR-compatible physiologic recorders, which can be interfaced with the MRI scanner, often short axis), or ∼1 sec (long axis,
use customized ECG electrodes and proprietary filtering techniques to reduce artifacts in four-chamber)
the ECG.
The ECG electrodes may be placed on the chest or back. Placing them on the chest is best 9. Use a dedicated cardiac coil for these studies to maximize the achievable signal-to-
for monitoring, but in this position the ECG signal is more susceptible to artifacts from
noise ratio. Center the coil at the level of the heart, or over the patient’s heart, if a
breathing motion, as well as image artifacts from the leads affecting the right ventricle.
The dry rub method improves electrode contact to the skin and reduces contact resistance. relatively small surface coil is used.
If the baseline of the ECG moves up and down with respiration, then the ECG leads need 10. Instruct the patient on the procedure for breath holding during the MRI examination.
to be moved up higher on the chest. Asking the patient to avoid a deep breathing motion Determine whether the patient is able to hold his/her breath for ∼10 to 15 sec at a
can help. Reducing the distance between the ECG leads can reduce interference from time by practicing breath holding with the patient. When patient preparations have
switching of the gradient coils.
been finished, position the patient bed such that the centering light is aligned over
The dry-rub method consists in the use of a slightly abrasive material, such as a clean, dry the patient’s heart, then move the table to have the patient’s heart in the magnet center.
gauze pad, and rubbing it on the skin to remove skin oils and perspiration. This should
allow proper adhesion of the ECG electrodes to the skin, which is essential for consistent 11. Patients with claustrophobia may require sedation to tolerate the procedure. If this is
ECG readings during the MRI study. The use of standard alcohol wipes may be necessary the case, follow institutional guidelines for conscious sedation.
to remove an excess of oils, lotions, or perspiration.
Sequence 1: Scout imaging of the heart
8. Wrap a blood pressure cuff around the patient’s arm to monitor the blood pressure. 12. Acquire a multi-plane scout (producing 3 transverse and 3 coronal slices) with a short
MRI-compatible monitoring systems use a pump that inflates and deflates the blood TR gradient echo sequence, preferably with steady-state free precession to obtain
pressure cuff at predefined time intervals or when the technologist manually initiates optimal signal-to-noise ratio. Typical sequence parameters for such a pulse sequence
a blood pressure measurement. Place a fiber-optic sensor on the patient’s finger to are listed in Table A11.4.2 and the resulting images are shown in Figure A11.4.1. Ask
measure the arterial oxygen saturation. Take note of the baseline heart rate, systemic Cardiac Function the patient to take a breath in and then hold his/her breath at end-expiration. Obtain
blood pressure, and oxygen saturation. Repeat these measurements regularly during Evaluation with transverse and coronal scout images using sequence parameters listed in Table
Acquired Heart Cine MRI of the
examination. Disease Heart A11.4.2.

A11.4.3 A11.4.4
Current Protocols in Magnetic Resonance Imaging Supplement 8 Supplement 8 Current Protocols in Magnetic Resonance Imaging
 A
B
Figure A11.4.1 Coronal (A) and transverse (B) scout images acquired with single-shot steady-state
free precession imaging technique while the patient holds his or her breath. The transverse scout image
is used to position the slice for the pseudo-long-axis localizer. The slice position and orientation are
shown in the image on B as a dark gray bar.
A B C Sequence 2: Segmented cine imaging
Figure A11.4.2 The image in A shows a pseudo-long vertical long-axis view of the heart. Use this
scout image to position the slices for a series of short-axis views near the base of the heart. The
resulting short-axis views are shown in images B and C. The slice orientation for cine acquisition in a
four-chamber, long-axis view should pass through the center of the left ventricle, and intersect with the
antero-lateral tip of the right ventricle, as shown by the dark bar in image C.
Oblique pseudo-long-axis scout
13. Prescribe on a transverse view through the middle of the left ventricle and acquire an
image for this view as a localizer as shown in Figure A11.4.1. Ask the patient to take
a breath in and then hold his/her breath at end-expiration. Obtain images using the
same sequence parameters as in Table A11.4.2 with the orientation set to be the
pseudo-long axis, as shown by the dark bar in Figure A11.4.1B. An example of the
image obtained with this slice orientation appears in Figure A11.4.2A.
Oblique short-axis scout
14. Prescribe on the resulting pseudo-long-axis view in order to obtain a set of double-
oblique short-axis views near the base of the heart. Select the short-axis slice just
below the valve plane (Fig. A11.4.2A) and prescribe a slice that passes through the
middle of the left ventricle (LV) and intersects with the antero-laberal tip of the right
ventricle. Ask the patient to take a breath in and then hold his/her breath at end-ex-
Current Protocols in Magnetic Resonance Imaging
End diastole End systole
Figure A11.4.3 Two frames from a cine acquisition with steady state free precession sequence
for a four-chamber long-axis view of the heart with the long-axis slice orientation prescribed as
shown in Figure A11.4.2C.
piration. Obtain images (Fig. A11.4.2B and A11.4.2C) using parameters listed in
Table A11.4.2 with orientation set to oblique short axis.
Four-chamber view
15. Prescribe a slice as indicated in Figure A11.4.2C. Obtain images using the parameters
listed in Table A11.4.2 with orientation set to be the long axis four-chamber view
(same orientation as shown in Fig. A11.4.3). The resulting image gives a long-axis
four-chamber view of the hear.
For patients with cardiac arrhythmias, one might have to resort to a real-time imaging
sequence. This means that the acquisition of each cine frame is no longer broken up
into segments, but rather acquired in one single shot. Often, the real-time imaging
techniques require extensive interpolation of imaging data from adjacent cardiac
phases to produce a smooth cine loop. For details about setting up the real-time
acquisition, see sequence 3.
16. The authors recommend that a cine of a long-axis, four-chamber view (horizontal
long axis) be acquired first (Fig. A11.4.3). For this purpose, copy the slice position
and orientation of the scout four-chamber view obtained with sequence 1 in step 15.
In the long-axis view, the slice orientation results in the less flow-driven enhancement of
the signal intensity in the ventricular cavity. This can be countered by choosing a thinner
slice thickness.
17. Adjust the sequence parameters with the orientation set to be the double-oblique long
(four-chamber) axis (either Table A11.4.3 or Table A11.4.4; depending on the
capabilities of the scanner and the patient’s heart rate). Acquire a minimum of 15
cardiac phases for either cine loop. Adjust the number of phase-encoding lines per
cardiac phase correspondingly.
This setting is also referred to as the number of lines per segment, in reference to the fact
that the acquisition of the images is broken up into segments and for each segment, a number
of k-space lines is acquired. For example, for a TR = 7 msec per phase-encoding step and
a heart rate of 70 beats/min, one should allot ∼40 to 50 msec to each cardiac phase, and
with a TR = 7 msec, ∼5 to 7 lines (i.e., 40/7 to 50/7) can be acquired per segment.
The optimal flip angle that produces the best contrast on cine images between the
myocardium and the LV chamber depends, among other parameters, on the slice thickness
and the repetition time (TR). It is worthwhile to gather some initial experience with cine
Cardiac Function
Evaluation with MR imaging of the heart in a volunteer to optimize the flip angle and other sequence
Acquired Heart Cine MRI of the parameters. It is impossible to recommend fixed settings for all pulse sequence parameters
Disease Heart for users of MRI platforms from different vendors. The overall goal of this exam should be
A11.4.5 A11.4.6
Supplement 8 Supplement 8 Current Protocols in Magnetic Resonance Imaging
 Table A11.4.3 Primary Clinical Imaging Parameters for Sequence 2 Table A11.4.4 Primary Clinical Imaging Parameters for Sequence 2 (Bright-Blood
(Steady-State Free Precession Cine Imaging) Cine Gradient Echo)

Patient position Supine Patient position Supine

Scan type Segmented k-space cine gradient echo Scan type Segmented k-space cine gradient echo
with view sharing Imaging plane (orientation) Double oblique, long- (four-chamber),
Imaging plane (orientation) Double oblique, short- or or long- or short-axis view
(four-chamber) axis view Central slice or volume center Mid-level of left ventricle
Central slice or volume center Mid-level of left ventricle Echo time (TE) 4.8 msec with flow compensation
Echo time (TE) 1.5–1.6 msec 2.7 msec without flow compensation
Receiver bandwidth (RBW) 238 kHz (∼930 Hz/pixel) Receiver bandwidth (RBW) 50 kHz (~197 Hz/pixel)
Number of lines per segment 9 Number of lines per segment 5–7
Repeat time (TR) 27 msec (TR per k-space segment, Repeat time (TR) 40–56 msec (temporal resolution)
temporal resolution) Delay time (TD) 0 msec
Delay time (TD) 0 msec Flip angle (FA) 20°
Flip angle (FA) 60° Fields of view (FOVx, FOVy) 300–360 mm, 300r–360r mm, r = 3/4
Fields of view (FOVx, FOVy) 300–360 mm, 300r–360r mm, with r (rectangular field of view) depending
≤0.75 (rectangular field of view), on body habitus
depending on body habitus Resolution (Δx, Δy) 1.17–1.41 mm, 2.21–2.65 mm
Resolution (Δx, Δy) 1.17–1.41 mm, 2.17–2.61 mm Number of data points collected (Nx, Ny) 256, 136r, with r = 3/4 (rectangular
Number of data points collected (Nx, Ny) 256, 138r, with r ≤3/4 (rectangular field of view) depending on body
field of view) depending on body habitus
habitus Display matrix (Dx, Dy) 256, 256
Display matrix (Dx, Dy) 256, 256 Slice thickness (Δz) 6–8 mm
Slice thickness (Δz) 6–8 mm Number of slices 1
Number of slices 1 Slice gap N/A
Slice gap N/A Number of acquisitions (Nacq) 1 with breath-holding, 3 without
Number of acquisitions (Nacq) 1 breath-holding
Swap read and phase encoding Swap if wrap-around is reduced Swap read and phase encoding Swap if wrap around is reduced
Slice location Long axis of heart from mitral valve Slice location Long axis of heart from mitral valve
center to apical tip (the heart is center to apical tip (the heart is
covered from base to apex for the covered from base to apex for the
short-axis view) short-axis view)
Saturation pulses No Saturation pulses No
Number of cardiac phases (R-to-R interval) × 90%/TR (at least 15) Number of cardiac phases (R-to-R interval) × 90%/TR (at least 15)
ECG gating Yes ECG gating Yes
Scan time ∼9 sec Scan time <17 sec

to (a) assure complete coverage of the cardiac cycle by adjustment of the number of cardiac
phases, (b) good image quality that provides excellent contrast at the endocardial border,
and (c) complete coverage of the heart from base to apex. While image quality depends on
some aspects not entirely under the control of the MRI operator such as patient cooperation,
complete temporal and spatial resolution is the operator’s responsibility. Failure to meet
these requirements can render the acquired data useless for LV function analysis.
18. Ask the patient to take a breath in and then hold his/her breath at end-expiration. Start
the scan after the patient has begun to hold his/her breath (Fig. A11.4.4).
Irregular gradient noise occurs when the patient has an irregular cardiac rhythm or if the
acquisition window is very close to the R-to-R duration. In the latter case, the very slight
variations in heart rate, e.g., due to breath-holding, cause the sequence to be triggered
randomly every heartbeat or every second heartbeat. Irregular triggering prevents estab-
lishment of a steady state and results in ghosting artifacts.
Fat saturation requires a very good shim to prevent inadvertent suppression of water signal.
Generally, there is little benefit in the use of fat saturation when the objective is obtaining
functional data.

Breath-holding at end-expiration seems to result in a more reproducible diaphragm 20. After acquisition of a set of images for a slice position, play all cardiac phases as a
position, than if the patient stops breathing at end-inspiration. cine loop and make sure that image quality is sufficient and that the cine loop covers
the cardiac cycle as required. Repeat the acquisition if necessary, or otherwise move
19. During the acquisition of the images, one should listen to the noise from the switching to the next slice position.
gradients and make sure that the noise follows the cardiac rhythm. Do not forget to
tell the patient when he or she can start to breathe again, as any negligence or lack of Cardiac Function Ghosting artifacts propagating along the phase-encoding direction are a tell-tale sign of
Evaluation with breathing-induced motion. The acquisition of these cine-loops should be repeated if
courtesy may result in less cooperation from the patient. Acquired Heart Cine MRI of the
Disease
ghosting is found.
Heart

A11.4.7 A11.4.8
Current Protocols in Magnetic Resonance Imaging Supplement 8 Supplement 8 Current Protocols in Magnetic Resonance Imaging
 Table A11.4.5 Primary Clinical Imaging Parameters for Sequence 3
(Phase-Contrast Cine of Aorta)

Patient position Supine

Scan type Fast gradient echo with phase-contrast
flow velocity encoding
Imaging plane (orientation) Close to transverse orientation,
perpendicular intersection with
ascending aorta
Central slice or volume center At level of bifurcation of pulmonary
trunk
Echo time (TE) 3.6 msec
Repeat time (TR) 20 msec (temporal resolution)
Flip angle (FA) 15°
Figure A11.4.4 Regurgitant jet (highlighted by the arrow) in a patient with myocardial infarction Fields of view (FOVx, FOVy) 300 mm, 244 mm
and acute mitral valve regurgitation that developed with severe ventricular failure. The cine images Resolution (Δx, Δy) 1.17 mm, 1.89 mm
were acquired with a steady-state free precession (SSFP) sequence. With SSFP sequences, Number of data points collected (Nx, Ny) 256, ~129
turbulent flow causes pronounced signal loss and is well visualized.
Display matrix (Dx, Dy ) 256, 256
Slice thickness (Δz) 5 mm
Number of slices 1
21. If it seems that the patient is uncooperative or simply unable to hold his/her breath,
Slice gap Not applicable
attempt to reduce the required breath-hold duration or switch to a nonbreath-hold
Number of acquisitions (Nacq) 2
acquisition. To reduce the breath-hold duration, increase the number of phase
Swap read and phase encoding Swap if necessary for
encoding steps per cardiac phase (i.e., number of lines per segment). This will reduce
anterior–posterior as the phase
the total number of cardiac phases in the cine loop, i.e., it reduces temporal resolution. encoding direction
Try to further reduce the field of view, as long as the aliasing (wrap-around artifact) Slice locations Ascending aorta
does not obstruct view of the heart. For non-breath-hold acquisitions, all image Saturation pulses Not applicable
parameters, except the number of acquisitions, can be left unchanged. To average out Number of cardiac phases Between 25 and 35
breathing motion, average at least 3 acquisitions. Ask patient to breathe regularly. Vascular options Velocity encoding sensitivity of ±150
22. Obtain short-axis images by positioning the acquisition off the diastolic frame of the cm/sec
horizontal long axis (four-chambered long axis) cine image. The heart should be Scan time ∼129 heartbeats per acquisition (with
129 phase-encoding steps and 25–35
covered from base to apex, starting at a position above the mitral valve plane. Adjust cardiac phases). The total scan time is
the position such that the slice position is a number that is easy to increment. For ∼258 heartbeats.
example, instead of a slice position of 83.4 mm, use a slice position of 80 mm with
8-mm increments to cover the heart with consecutive breath holds. In cases of low
signal-to-noise ratio, choose a thicker slice for short axis images than for long axis Sequence 3: Phase-contrast flow velocity imaging of aorta
images (e.g., 8 mm for short axis images versus 6 mm for the long axis images). In the presence of aortic regurgitation, the stroke volumes calculated by volumetric
23. Ask the patient to hold his/her breath. To obtain the first slice, run sequence 2 according analysis will be different from the effective (forward) flow volume. See Commentary for
to either Table A11.4.3 or Table A11.4.4, with the orientation set to be the double- volumetric analysis. For determination of the effective stroke volume, it then becomes
oblique short axis. Ask the patient to breathe and hold his/her breath again. Run necessary to supplement these measurements with phase-contrast flow velocity measure-
sequence 2 according to either Table A11.4.3 or Table A11.4.4 to obtain the second ments, using sequence 3.
slice. Repeat this procedure until the entire heart is covered (from base to apex). 24. Load sequence 3 (Table A11.4.5) on the console. Position the slice at the level of the
bifurcation of the pulmonary trunk (Fig. A11.4.5). Intersect the slice with the
To assure complete coverage of the heart, the short axis cine acquisitions should be
repeated until one reaches the apical tip of the heart. No bright signal from the ventricular ascending aorta at a right angle for accurate quantification blood flow. Enable the
blood pool should be visible in any of the cine frames for the most apical slice location. reconstruction of phase and magnitude images. The phase images provide the velocity
information.
The sum of slice thickness and the slice gap between adjacent slices should be equal to the
slice increment, e.g., if the slice increment is 8 mm and the slice thickness is 6 mm, then the Sequence 3 should be loaded with through-plane flow quantification, which means that
resulting slice gap between adjacent slices is 2 mm. flow is encoded for a direction perpendicular to the slice plane.
For through-plane flow encoding, a tilt of the image plane from an orientation orthogonal
to the vessel axis by an angle α will result in a reduction of the flow velocity estimate below
Cardiac Function
Evaluation with the true flow velocity by a factor cos(α).
Acquired Heart Cine MRI of the
Disease Heart

A11.4.9 A11.4.10
Current Protocols in Magnetic Resonance Imaging Supplement 8 Supplement 8 Current Protocols in Magnetic Resonance Imaging
 Table A11.4.6 Primary Clinical Imaging Parameters for Sequence 4 (Real-Time
Cine Sequence)
A B mi Flow vs time
537 Patient position Supine
Legend
478 data Scan type Single-shot, steady-state free
spline (+/- 1) precession sequence
419
Imaging plane (orientation) Double oblique, short-axis view
360
Central slice or volume center Mid-level of left ventricle
301 Echo time (TE) 1.15 msec
242 Receiver bandwidth (RBW) 179 kHz (1398 Hz/pixel)
183 Repeat time (TR) 1.4 msec
124 Delay time (TD) 0 msec
Flip angle (FA) 50°
65
Time
Fields of view (FOVx, FOVy) 300 mm, 300r mm, r = 3/4
6 (rectangular field of view) depending
20 90 160 230 300 370 440 510 580 650 720 (msec)
–53 on body habitus
Slice position: SP H40.4 Venc adjustment-150 : 150 Resolution (Δx, Δy) 2.34 mm, 4.41 mm
Number of data points collected (Nx, Ny) 128, 68r, with r = 3/4 (rectangular
field of view) depending on body
Figure A11.4.5 Panel A shows an image of the signal phase (after background correction) in a patient at the level of habitus
bifurcation of the pulmonary trunk. Flow was encoded in a direction perpendicular to the slice plane. The slice plane was
Display matrix (Dx, Dy) 128, 128
oriented to intersect the ascending aorta at a right angle. The phase of the signal is encoded on a gray scale, and the phase
is directly proportional to the flow velocity. Panel B shows a graph of aortic flow versus the delay time after the R-wave. The Slice thickness (Δz) 10 mm
flow is calculated from the phase images by integrating the flow velocity over the aortic cross-section to obtain flow in units Number of slices 1
of ml/sec, in this case. Flow analysis software is available as an option on most MRI scanners. Slice gap N/A
Number of acquisitions (Nacq) 1
Swap read and phase encoding Swap if wrap-around is reduced
Saturation pulses No
25. Adjust the number of cardiac phases to anywhere between 25 and 35, depending on Number of cardiac phases (R-to-R interval) × 150%/(TR × Ny)
the patient’s heart rate. Perform a magnetic field shim before running the sequence, ECG gating No
as magnetic field inhomogeneities primarily alter the phase of the signal, and can Scan time 2 heartbeats
cause erroneous flow velocity estimates.
26. It is generally not possible to complete the flow velocity measurement within one
breath-hold. Instruct the patient to breathe normally, and use two averages (2 Sequence 4: Real-time cine sequence
acquisitions or excitations) to average out the breathing motion. Inform the patient In the absence of a regular heart rhythm, ECG-triggering of a cine acquisition will yield
that the total scan time will take ∼240 to 260 heartbeats, i.e., ∼4 to 5 min. Run unsatisfactory results. In this situation, a very rapid imaging sequence allowing real-time
sequence 3 according to Table A11.4.5. acquisition may be used (Setser et al., 2000). Table A11.4.6 provides the details for very
fast steady-state free precession (SSFP) imaging that meets the necessary requirements.
27. Load the images into a flow-quantification software program, draw a region of
interest along the inside border of the aorta and let the program calculate the flow 28. Load sequence 3 (based on parameters in Table A11.4.6) and disable ECG triggering.
through the region-of-interest. Make sure that the velocity encoding sensitivity was Trim the field of view to the smallest possible size. Due to the reduced number of
correct to avoid aliasing in the graph of flow versus cardiac phase. Repeat the phase-encoding steps the spatial resolution in the phase-encoding direction will be
measurement, if necessary, with a larger velocity encoding range. too low unless the FOVy is on the order of 300 mm. Use a rectangular field of view
As a result of the image acquisition, two sets of images should be obtained: images showing if possible.
the signal magnitude, and images showing the signal phase. Anatomical structures can be The resolution in the phase-encoding direction should be below 4 mm.
easily identified on the magnitude images, while the phase images are used for the flow
velocity estimate. 29. Instruct the patient to breathe normally. Set the number of images (or number of
cardiac phases) to be acquired such that they cover a couple of heart intervals. Run
A velocity encoding (Venc) sensitivity of 150 cm/sec is recommended for the initial
measurement. In the presence of a stenosis, peak flow velocities will significantly exceed sequence 4 according to the parameters given in Table A11.4.6.
this range and require an increase of the Venc range from 150 cm/sec to somewhere ∼300 The temporal resolution is on the order of 70 msec with ∼50 phase-encoding steps (Ny).
cm/sec, depending on stenosis severity. Velocities outside the Venc range will appear on
the graph as discontinuities, i.e., velocities outside the Venc range are folded back or aliased 30. Repeat sequence 4 with parameters according to Table A11.4.6, at least for a number
into the Venc range. Cardiac Function
Evaluation with of short-axis views from base to apex of the heart, if a quantitative estimate of
Acquired Heart Cine MRI of the ventricular parameters is required.
Disease Heart

A11.4.11 A11.4.12
Current Protocols in Magnetic Resonance Imaging Supplement 8 Supplement 8 Current Protocols in Magnetic Resonance Imaging
 Post-processing
160
For determination of the ventricular volumes, ejection fraction, and regional wall motion,
LV ejection fraction = 80.32 % the cine images need to be loaded into a program for image segmentation, i.e., the
endocardial and epicardial borders need to be identified at least on the end-diastolic (ED)
140 and end-systolic (ES) frames of the cine loops.

31. As a first step, select cine loops that provide coverage from base to apex and load
120 them into the analysis software.

ES
The program will “blindly” calculate chamber volumes based on the slice positions for
100 which data have been loaded, which leads to incorrect results if the slice coverage is
insufficient. For cine loops in the short axis view, one should include at the base of the heart
only those cardiac phases of the cardiac cycle where the endocardial volume falls below
80 the valve plane. This means that one excludes those frames in the cine loop from the
calculation of volumes where the slice position is in or above the valve plane. Generally,

LV volume (ml)
such exclusion is performed by default for those frames where the user did not draw
60 endocardial and epicardial borders.
32. If all images in a cine loop are segmented, then obtain a curve showing the variation
40 of the ventricular chamber volume over the cardiac cycle. If EDV and ESV denote,
end-diastolic and end-systolic volumes, respectively, use the formula (EDV-
Normal ESV)/EDV for calculation of the ejection fraction, with the result expressed in
20
percent.
0 100 200 300 400 500 600 700 800 900
Figure A11.4.6 shows two examples of left ventricular volume versus time curves for a
normal volunteer and a patient with congestive heart failure (CHF), respectively. In the
LV ejection fraction = 17.6 %
600 CHF patient, the ejection fraction is very low, and the chamber volumes are significantly
enlarged, both signs of a failing heart.
All global function parameters can be derived from such a volume versus time curve. The
560 analysis of the images can be reduced to segmentation of ED and ES frames if only
parameters such as ejection fractions are needed, as they only require ED and ES volumes.
Identification of the ES frame in a cine loop is generally straightforward but there are rare
instances where dyskinesis can lead to enlargement of a portion of the ventricle during
520 systole.

LV volume (ml)
COMMENTARY
480
Background Information
MRI is considered to be the gold standard
for calculating hemodynamic parameters of
440 cardiac function such as ejection fraction (EF),
end-diastolic volume (EDV), end-systolic vol-
ume (ESV), stroke volume (SV), and cardiac
CHF mass. While this accuracy (Matheijssen et al.,
400
1996) is not often needed for clinical patient
0 100 200 300 400 500 600 700
management, it is not uncommon that a patient
Delay after R-wave (msec) has poor acoustic windows for echocardiogra-
phy. Gated blood pool nuclear medicine scans
can also provide hemodynamic data, but per-
Figure A11.4.6 Example of variation of left ventricular volume over the cardiac cycle for (A) a forms poorly in the right ventricle and cannot
healthy volunteer, and (B) a patient with congestive heart failure (CHF). The CHF patient had an provide the richness of anatomical and func-
enlarged ventricle (i.e., large volume), and a very low ejection fraction (EF). Because of the low EF, tional information of MRI. Thus, MRI is a
the curve in (B) is relatively flat. Ventricular volumes were calculated by Simpson’s rule from a set
stand-alone examination or a cardiac MRI pro-
of short-axis images. The endocardial border had been traced on each cine frame to obtain a
complete LV volume versus time curve.
vides additional information for some other
purpose.
Cardiac Function
Evaluation with The contrast between the myocardium and
Acquired Heart Cine MRI of the the blood pool is of crucial importance for
Disease Heart
delineation of the endocardial border and de-
termination of ventricular volumes. Myocar-
dial-blood pool contrast can be improved with
the use of contrast agents. These may be ex-
tracellular agents such as Gd-DTPA (Pennell et
al., 1993), however, blood pool agents, such as
iron oxide ultra-small particles (USPIO), or
albumin-binding, gadolinium-based contrast
agents, appear to be better suited for this pur-
pose (Stillman et al., 1997). On scanners with
high-performance gradients, true FISP (fast im-
aging with steady state free precession) pro-
vides excellent myocardial-blood pool contrast
(Barkhausen et al., 2001; Plein et al., 2001;
Thiele et al., 2001). It is, however, sensitive to
off-resonance effects and turbulence. It is criti-
cal to have consistent breath holding. Gener-
ally, the most reproducible diaphragm position
is obtainable at end expiration. Oxygen by a
nasal cannula can be used to improve the

A11.4.13 A11.4.14
Current Protocols in Magnetic Resonance Imaging Supplement 10 Supplement 10 Current Protocols in Magnetic Resonance Imaging
breath-hold duration. A few minutes of instruc-
tion can also improve reproducibility. For non-
cooperative patients, it is best to permit free
quiet breathing and obtain an average over
several acquisitions to minimize respiratory
artifact. A respiratory navigator can be helpful
if available.
When time is of the essence, a simple bi-
plane method may be used to calculate left
ventricular volumes (Cranney et al., 1990).
This assumes an ellipsoidal model for the left
ventricle that is reasonably true for normal
hearts, but becomes less valid as the shape of
the heart changes as seen with myocardial in-
farction. Both two-chamber and four-chamber
views are obtained and the area bounded by the
endocardial borders is obtained at end diastole
(assumed to be represented by the first cardiac
phase) and end systole. The volume is calcu-
lated as:
volume = (8A2A4)/(3πLmin)
where A2 and A4 are the two-chamber and four-
chamber areas, respectively, and Lmin is the
length of the shorter long axis. Geometric mod-
els for the right ventricle have also been devel-
oped, although less fully evaluated (Czegledy
and Katz, 1993).
A more robust and accurate method uses a
series of short axis images from the base of the
heart to the apex. This method assumes no
geometric models that may be invalid in the
presence of disease. Volumes are determined
by calculating the area bounded by the endo-
cardium in each section. These are added and
multiplied by the slice thickness to calculate
the volume (Simpson’s rule method). Because
of through-plane motion, it is not uncommon
for all, or part, of a systolic phase image to
contain the atrium while the end-diastolic phase
image represents the ventricle. Since the base
of the heart contributes so greatly to the overall
volume, it is critical to measure the entire ven-
tricle without including the atrium. Normally,
the ventricle thickens during systole whereas
the atrium remains thin. This can be used as a
guide to include only that portion that thickens
in the volume measurement.
The effective ejection fraction is decreased
in the presence of either mitral or aortic regur-
gitation. The effective ejection fraction may be
calculated by the use of a velocity map of the
ascending aorta. This is obtained in a cross
section at the level of the right pulmonary artery
taking care to avoid areas with sternal wire or
clip artifact. For the determination of aortic
valve regurgitation, it was found that the slice
Current Protocols in Magnetic Resonance Imaging
position is critical with best accuracy when it
falls approximately between the aortic valve
and the coronary ostia (Chatzimavroudis et al.,
1997). Pixel-phase intensity is proportional to
the velocity of blood flow. Generally, a value
of velocity encoding of 150 cm/sec is adequate
unless there is significant aortic stenosis. A
nonsegmented pulse sequence is preferred so
that there is optimal temporal resolution. The
mean flow is calculated as the integral over the
cross-sectional area (S) for the velocity (v) as:
H H
q = o v ⋅ ds
∫ of this method depends on a regular heart
The effective (forward) stroke volume then
may be calculated as:
R − to − R interval
SVeffective = ∫o q dt
where the time integral is over the cardiac cycle
(one R-wave to the next). This is more similar
to the value that would be obtained from ther-
modilution methods than from the cine data in
that valvular incompetence is implicitly ac-
counted for. For simplicity, an assumption is
made that there is only mitral regurgitation
(isolated aortic insufficiency can be similarly
defined). The mitral regurgitant volume (VMR)
is
VMR = SVtotal − SVeffective = (EDV −
ESV) − SVeffective
Where the total stroke volume, SVtotal can
be calculated from the short axis cine set as
described in the Post-Processing section. Alter-
natively, if only two-chamber and four-cham-
ber views have been acquired, then the above
“biplane” formula can be used to calculate ESV
and EDV. The mitral regurgitant fraction (MR)
is calculated as
MR = VMR/SVtotal
and the effective ejection fraction (EFeffective) as
EFeffective = SVeffective/EDV.
While this may seem complicated, the net
forward- and reverse-flow volumes over a car-
diac cycle are provided by standard evaluation
software so the remaining calculations are
straightforward.
The combination of both aortic and mitral
regurgitation limits the accuracy of the calcu-
lated respective regurgitant volumes, as does
the presence of an intracardiac shunt. Never-
theless, the effective EF should be a reliable
number. In this case, one can assume that sys-
tole is defined as the period of forward flow and
that the mitral valve should be closed. There is
no aortic regurgitation during systole. The dif-
ference between SVtotal and the forward-flow
volume is then taken to be VMR.
Occasionally, VMR can be negative. This
occurs when there are dyskinetic segments or
dyssychrony of left ventricular function so that
an ESV is difficult to define. In such a case, it
is not possible to calculate VMR by this method.
The calculation of EFeffective is still valid and
preferred over the use of the ejection fraction
calculated solely from the cine data. Reliability
rhythm and consistent breath holding.
Steady-state free precession imaging
Acquisition of image data during steady-
state free precession method represents one of
the most significant recent developments for
the assessment of cardiac hemodynamics with
MRI (Bundy et al., 1999). A variety of acro-
nyms have been coined by different equipment
manufacturers for this pulse sequence type,
e.g., trueFISP (Siemens), Fiesta (GE), and Bal-
anced Fast Field Echo (Philips). The essential
characteristic of this technique is the goal to
avoid any spoiling of transverse magnetization
(see UNIT B5.2 and Duerk et al., 1998). In the
steady state, a radio frequency excitation pulse
in an SSFP sequence rotates a fraction of the
longitudinal magnetization into the transverse
plane, and at the same time returns the trans-
verse magnetization that was refocused at the
end of the previous repeat time (TR) to the
longitudinal direction. Careful refocusing of
the transverse magnetization after slice-selec-
tive excitation, phase-encoding, and echo-read-
out is essential for this technique. In SSFP
imaging, the trajectory of the magnetization
components is controlled by the radio fre-
quency (RF) pulses and gradient waveforms
without relying on longitudinal relaxation be-
tween RF pulses and replenishing the longitu-
dinal magnetization. Therefore, saturation by
repeated RF excitation is avoided, and the im-
age contrast is dominated by the T2/T1 relaxa-
tion times. With SSFP imaging, one can theo-
retically apply 90° radio frequency pulses in-
stead of small flip angle pulses, which produces
a significantly higher signal-to-noise ratio in
the images than with conventional gradient
echo images. For reasons of RF deposition and
contrast, smaller flip angles close to 45° are
more appropriate.
Cardiac Function In conventional gradient echo imaging with
Evaluation with
Acquired Heart Cine MRI of the prospective triggering, the pulse sequence is
Disease Heart
A11.4.15 A11.4.16
Supplement 8 Supplement 8
suspended once the last k-space line of the
phase-encoding steps has been collected. The
sequence resumes once an R-wave peak has
been detected on the ECG. By contrast, with
steady-state free precession sequences, the se-
quence runs continuously until the end of the
entire image acquisition to avoid any distur-
bance to the steady-state equilibrium. In SSFP
imaging, the preservation of coherence of the
transverse magnetization is still limited by the
T2* decay. Any phase errors introduced, e.g.,
by off-resonance shifts, are accumulated over
the image acquisition. SSFP imaging is, there-
fore, more sensitive than conventional gradi-
ent-echo imaging to off-resonance shift and
magnetic-field homogeneities. Homogeneous
magnetic fields by shimming are necessary to
obtain good results, much more so than with
conventional gradient echo imaging.
SSFP imaging appears to produce the most
striking improvements in image quality by
comparison to conventional gradient echo im-
aging when studies are performed on patients
with compromised cardiac function, e.g., in
patients with heart failure. In conventional gra-
dient echo cine MRI, the contrast between myo-
cardium and blood pool is, to a large extent, the
result of differences in magnetization satura-
tion in myocardial tissue and blood flowing in
and out of the imaging slice. With compromised
pumping function, the inflow of unsaturated
magnetization is reduced and this results in a
corresponding loss of contrast between myo-
cardium and blood pool. Figure A11.4.7 shows
an example in the same patient of conventional
gradient echo images and SSFP images. The
differences in contrast between conventional
gradient echo images and SSFP images are
most pronounced for long-axis views, as the
bright-blood effect is reduced when the pre-
dominant direction of flow is parallel to the
slice plane (long-axis view), instead of perpen-
dicular to the image plane (short-axis view).
SSFP sequences use relatively short echo times
which may in some cases cause an observer to
miss a jet, or overemphasize the jet. Standard
gradient echo sequences, which use a longer
TE than SSFP sequences, generally allow better
visibility of dephasing. Thus, jets caused by
valvular disease or intracardiac shunts are bet-
ter seen using standard gradient echo sequences
(Pereles et al., 2001).
Post-processing software
Currently available software packages im-
pose the requirement that the heart be covered
Current Protocols in Magnetic Resonance Imaging

A B
C D
Figure A11.4.7 Comparison of cine acquisition with steady-steady free precession (SSFP)
technique (A, B); with “conventional” gradient-echo sequence (C, D). The SSFP technique affords
better contrast between blood pool and myocardium. Poorer contrast in the cases of (C) and (D)
can lead to an underestimate of ventricular volume and an overestimate of myocardial mass.
with parallel slices from base to apex. Volumes suspend breathing motion during image acqui-
and myocardial mass are calculated with Simp- sition. Irregular gating of the pulse sequence
son’s rule. This yields good accuracy but re- results in an irregular noise pattern from the
quires a relatively large number of acquisitions. switching gradients, and on the images, the
Several long-axis views of the heart, rotated heart appears blurred. Unreliable gating with
around a central axis could, in principle, yield ECG can often be remedied by switching the
similar accuracy, but would require a more order of the ECG leads (electronically) or, as a
complex reconstruction of the ventricular last resort, by repositioning the leads on the
shape. Dulce et al. (1993) already showed that chest. With an image acquisition triggered by
a combination of two long-axis views and one the R-wave of the ECG, the first frame of the
to two short-axis views yield results in close cine loop coincides with the beginning of the
agreement with a significantly larger number isovolumetric contraction.
of co-planar short-axis views (Dulce et al.,
1993). Despite these encouraging results, the Specific absorption rate with steady-state
current software packages generally cannot ac- free precession imaging
commodate more flexible acquisition strategies A potential disadvantage of SSFP imaging
with multiple rotated and shifted image planes is the radio frequency power deposition due to
(Bloomgarden et al., 1997). high flip-angle pulses. The software on the
scanner limiting the tissue absorption of radiof-
Critical Parameters and requency power may prevent a pulse sequence
Troubleshooting from running. The user then has to reduce the
flip angle, and in practice, this means that the
ECG tracings SSFP sequences are used with flip angles con-
Cardiac motion artifacts in the images can siderably <90°, and more on the order of 40° Cardiac Function
Evaluation with
generally be traced to either unreliable gating to 60°. Acquired Heart Cine MRI of the
of the sequence or the inability of the patient to Disease Heart
A11.4.17 A11.4.18
Current Protocols in Magnetic Resonance Imaging Supplement 8 Supplement 8
Motion and flow artifacts
Breathing motion during image acquisitions
that require breath-holding of the patient results
in ghost images replicated along the phase-en-
coding direction. Ghosting will be most notice-
able from regions that produce very high signal
intensities such as subcutaneous fat. If the pa-
tient is not able to hold his or her breath, then
more than one acquisition should be acquired
to average out the breathing motion, although
breathing motion will still cause blurring of the
endocardial and epicardial borders.
Bright-blood gradient echo imaging has tra-
ditionally employed flow-compensated gradi-
ent waveforms to reduce signal loss due to flow.
By switching off flow compensation, one can
substantially reduce the echo time, and this can
be used as an advantage to speed up the image
acquisition without producing a significant re-
duction in image quality for typical rest heart
rates (<80 beats per minute) if the echo time is
on the order of 2 to 3 msec. Alternatively, one
can put the TE reduction to use by reducing the
TR , and thereby improve the temporal resolu-
tion, or acquire multiple slices during each
breath-hold.
SSFP sequences are more sensitive to flow
and motion than conventional FLASH (fast low
angle shot) cine MRI. The initial experiences
with this technique suggest that the advantages
of SSFP over conventional gradient echo-im-
aging may be partially lost at heart rates above
∼130 to 140 beats per minute. At high heart
rates, as encountered during inotropic stimula-
tion with dobutamine, the user may consider
switching back to conventional gradient echo
imaging if the image quality is unsatisfactory.
Overestimation of stroke volume
Although a “stack” of cine loops in the
short-axis view for slice positions from base to
apex are sufficient for the evaluation of ven-
tricular function, one should not skip over the
acquisition of long-axis cine loops. A qualita-
tive evaluation of aortic insufficiency can be
performed in the long-axis view, by assessing
the width and depth of the jet. Cardiac cine
should be augmented by phase-contrast meas-
urements for a more precise assessment of the
forward stroke volume and the degree of aortic
insufficiency.
Anticipated Results
Lorenz et al. (1999) reported normal ranges
of left and right ventricular mass and ventricu-
lar volumes measured with cine magnetic reso-
nance imaging from a study involving 75
healthy subjects. This report, and others that
may appear in the literature with the more
widespread adoption of the cine MRI, will
constitute a valuable database against which to
evaluate cardiac pathologies.
Systematic comparisons of ventricular vol-
umes calculated from conventional gradient
echo images and SSFP images show that SSFP
imaging gives higher values for the cavity vol-
umes and smaller values for myocardial mass.
The most accepted explanation for this discrep-
ancy is that stasis near the ventricular wall will
result on conventional gradient echo images in
reduced signal intensity near the endocardial
border that is easily mistaken for a myocardial
signal. In areas with pronounced trabecula-
tions, the reduced brightness of the signal will
create a bias towards extending the endocardial
border into the blood pool. On high quality
SSFP images, one can easily identify areas of
the myocardial wall with extensive endocardial
trabeculae, something that is not easily recog-
nizable with bright-blood gradient echo imag-
ing. For segmentation of the images, the exact
results for ventricular volumes will depend on
whether the contours are drawn through the
“middle” of the trabeculae, or whether they are
excluded. For follow-up studies, as in therapy
trials, it becomes important to stick to one and
the same prescription for image segmentation.
The use of inotropic agents such as dobu-
tamine in combination with echocardiography
is quite common to detect wall-motion abnor-
malities due to ischemia. A comparison of wall
motion studies performed with MRI versus
ultrasound in an important study by Nagel et
al. (1999) has demonstrated a significantly
higher diagnostic accuracy of stress MRI in
comparison to stress echocardiography. Moni-
toring patients during dobutamine stress re-
mains an area of concern that has impeded the
wider use of MRI for stress wall motion studies.
Near real-time feedback to the MRI operator of
the cine loops can provide some degree of
safety in preventing excessive ischemic insults,
a dangerous degradation of global ventricular
function, and/or cardiac arrest.
Literature Cited
Barkhausen, J., Ruehm, S.G., Goyen, M., Buck, T.,
Laub, G., and Debatin, J.F. 2001. MR evaluation
of ventricular function: True fast imaging with
steady-state precession versus fast low-angle
shot cine MR imaging: Feasibility study. Radiol-
ogy 219:264-269.
Bloomgarden, D.C., Fayad, Z.A., Ferrari, V.A.,
Chin, B., Sutton, M.G., and Axel, L. 1997.
Global cardiac function using fast breath-hold
Current Protocols in Magnetic Resonance Imaging
 MRI: Validation of new acquisition and analysis
techniques. Magn. Reson. Med. 37:683-692.
Bundy, J., Simonetti, O., Laub, G., and Finn, J.P.
1999. Segmented trueFISP cine imaging of the
heart. Seventh Meeting of the International So-
ciety for Magnetic Resonance in Medicine.
Philadelphia.
Chatzimavroudis, G.P., Walker, P.G., Oshinski, J.N.,
Franch, R.H., Pettigrew, R.I., and Yoganathan,
A.P. 1997. The importance of slice location on
the accuracy of aortic regurgitation measure-
ments with magnetic resonance phase velocity
mapping. Ann. Biomed. Eng. 25:644-652.
Cranney, G.B., Lotan, C.S., Dean, L., Baxley, W.,
Bouchard, A., and Pohost, G.M. 1990. Left ven-
tricular volume measurement using cardiac axis
nuclear magnetic resonance imaging. Validation
by calibrated ventricular angiography. Circula-
tion 82:154-163.
Czegledy, F.P. and Katz, J. 1993. A new geometric
description of the right ventricle. J. Biomed. Eng.
15:387-391.
Duerk, J.L., Lewin, J.S., Wendt, M., and Petersilge,
C. 1998. Remember true FISP? A high SNR,
near 1-second imaging method for T2- like con-
trast in interventional MRI at .2 T. J. Magn.
Reson. Imaging 8:203-208.
Dulce, M.C., Mostbeck, G.H., Friese, K.K., Caputo,
G.R., and Higgins, C.B. 1993. Quantification of
the left ventricular volumes and function with
cine MR imaging: Comparison of geometric
models with three-dimensional data. Radiology
188:371-376.
Lorenz, C.H., Walker, E.S., Morgan, V.L., Klein,
S.S., and Graham, T.P. Jr. 1999. Normal human
right and left ventricular mass, systolic function,
and gender differences by cine magnetic reso-
nance imaging. J. Cardiovasc. Magn. Reson.
1:7-21.
Matheijssen, N.A., Baur, L.H., Reiber, J.H., van der
Velde, E.A., van Dijkman, P.R., van der Geest,
R.J., de Roos, A., and van der Wall, E.E. 1996.
Assessment of left ventricular volume and mass
by cine magnetic resonance imaging in patients
with anterior myocardial infarction intra-ob-
server and inter-observer variability on contour
detection. Int. J. Card. Imaging 12:11-19.
Nagel, E., Lehmkuhl, H.B., Bocksch, W., Klein, C.,
Vogel, U., Frantz, E., Ellmer, A., Dreysse, S., and
Fleck, E. 1999. Noninvasive diagnosis of is-
chemia-induced wall motion abnormalities with
the use of high-dose dobutamine stress MRI:
Comparison with dobutamine stress echo-
cardiography. Circulation 99:763-770.
Pennell, D.J., Underwood, S.R., and Longmore,
D.B. 1993. Improved cine MR imaging of left
ventricular wall motion with gadopentetate di-
meglumine. J. Magn. Reson. Imaging 3:13-19.
Current Protocols in Magnetic Resonance Imaging
Pereles, F.S., Kapoor, V., Carr, J.C., Simonetti, O.P.,
Krupinski, E.A., Baskaran, V., and Finn JP. 2001.
Usefulness of segmented trueFISP cardiac pulse
sequence in evaluation of congenital and ac-
quired adult cardiac abnormalities. Am. J.
Roentgenol. 177:1155-1160.
Plein, S., Bloomer, T.N., Ridgway, J.P., Jones, T.R.,
Bainbridge, G.J., and Sivananthan, M.U. 2001.
Steady-state free precession magnetic resonance
imaging of the heart: Comparison with seg-
mented k-space gradient-echo imaging. J. Magn.
Reson Imaging 14:230-236.
Schulen, V., Schick, F., Loichat, J., Helber, U., Hup-
pert, P.E., Laub, G., and Claussen, C.D. 1996.
Evaluation of k-space segmented cine sequences
for fast functional cardiac imaging. Invest. Ra-
diol. 31:512-522.
Setser, R.M., Fischer, S.E., and Lorenz, C.H. 2000.
Quantification of left ventricular function with
magnetic resonance images acquired in real
time. J. Magn. Reson. Imaging 12:430-438.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Stillman, A.E., Wilke, N., and Jerosch-Herold, M.
1997. Use of an intravascular T1 contrast agent
to improve MR cine myocardial-blood pool defi-
nition in man. J. Mag. Reson. Imaging 7:765-
767.
Thiele, H., Nagel, E., Paetsch, I., Schnackenburg,
B., Bornstedt, A., Kouwenhoven, M., Wahl, A.,
Schuler, G., and Fleck, E. 2001. Functional car-
diac MR imaging with steady-state free preces-
sion (SSFP) significantly improves endocardial
border delineation without contrast agents. J.
Magn. Reson. Imaging 14:362-367.
Internet Resources
http://www.medis.nl
The software package MASS (Magnetic Resonance
Analytical Software System) has been developed for
the quantitative analysis of multi-slice/multi-car-
diac phase left and right ventricular function.
http://www.siemensmedical.com
Leonardo is an analysis software package from Sie-
mens Medical Systems for evaluating cardiac mag-
netic resonance cine images and assessment of
ventricular functions.
Contributed by Arthur E. Stillman and
Michael Jerosch-Herold
University of Minnesota
Minneapolis, Minnesota
Acquired Heart
Disease
A11.4.19
Supplement 8
Aortic Dissection UNIT A12.1
Aortic dissection can be diagnosed using magnetic resonance imaging (MRI), computed
tomography (CT), or transesophageal echo (TEE) with similar sensitivity and specificity
(Sommer et al., 1996). CT requires iodinated contrast agent but examination is quick;
while TEE is the most invasive imaging modality. With MRI, it is difficult to monitor the
patient closely and motion artifacts can be problematic in uncooperative patients. The
patient should be screened as to which modality is best suited for the patient by the given
clinical situation. MRI is most suitable as a problem solving tool in relatively stable
patients or patients with contraindication to iodinated contrast agent.
The core components of the imaging protocol are ECG (electrocardiograph)-gated black
blood HASTE (half Fourier single shot turbo spin echo), and contrast-enhanced MR
angiography (MRA). When no abnormal findings are seen in black blood HASTE or cine
sequences, it is somewhat controversial whether contrast-enhanced MRA is necessary. In
this instance, a time of flight (TOF) sequence may be used instead of contrast-enhanced
MRA (see Alternate Protocol). At the authors’ institution, a contrast-enhanced MRA
sequence is routinely performed in all patients with suspected aortic dissection. The
authors use contrast-enhanced MRA to exclude other diagnosis such as penetrating
atherosclerotic ulcer or vasculitis, which could explain symptoms such as chest pain or
differences in blood pressure of upper extremities.
The ECG-gated cine gradient echo sequence provides a view of a single level in multiple
phases of the cardiac cycle, which can be used to evaluate the aortic valve.
A typical MRI examination takes 30 to 45 min. In unstable patients or patients who require
emergent surgery, the examination should be tailored to save time. In some instances,
black blood HASTE sequences may be all that is needed when the extent of aortic
dissection is well delineated. Alternatively, one could choose contrast-enhanced MR
angiography to answer the question at hand.
The parameters are based on the authors’ experience using a Siemens 1.5 T Vision or
Symphony and should be altered accordingly for different field strengths and manufac-
turers.
IMAGING OF THORACIC AORTA TO RULE OUT AORTIC DISSECTION BASIC
PROTOCOL
Table A12.1.1 lists the hardware necessary to perform the procedure, along with appro-
priate parameters. The available gradient strength will depend on the scanner, and the
echo times given below may need to be varied accordingly (the smaller the gradient
strength, the longer the echo time for a particular scan). Higher gradient strengths with
imaging sequences that allow gated turbo spin-echo and cine gradient echo imaging are
preferable.
Table A12.1.1 Equipment Parameters for Aortic Dissection
Coil type Torso phased array coil and body coil (see 3-D
contrast-enhanced MRA)
Gradient coil strength 25 mT/m (or whatever the system permits)
Cardiac gating Yes
Peripheral gating No
Oxygen Yes, 2 liters via nasal cannula for most patients
(to ensure breath-holding >15 sec)
Power injector Yes
Use of contrast agents Yes Acquired Aortic
Disease
Contributed by Naoki Takahashi and Vamsidhar Narra A12.1.1
Current Protocols in Magnetic Resonance Imaging (2001) A12.1.1-A12.1.16
Copyright © 2001 by John Wiley & Sons, Inc.
 NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as a crash cart or oxygen.
Materials
Normal saline (0.9% NaCl), sterile
Gadolinium-based MR contrast agent (e.g., Magnevist, Omniscan, or Prohance)
Set up patient and equipment
1. Interview (screen) the patient to ensure that he or she has no contraindications such
as cardiac pacemakers or defibrillators or other implants containing ferromagnetic
materials. Also be sure to find out if the patient has any health conditions that may
require the presence of special emergency equipment during the scanning procedure,
or necessitate any other precautions.
Generally, standard screening forms are used for all patients scanned in a magnetic
resonance system.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(1996) for a discussion of what implants may be safely scanned using magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
2. If the procedure is a research protocol, have the patient sign any necessary consent
form.
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating
and image artifacts.
6. Have the patient mount onto the table. Place ECG leads according to manufacturer’s
guidelines on the chest wall. Placing leads on the back may reduce motion artifacts
related to breathing. Make sure that the ECG tracing shows positive R-waves.
7. Place the phased array coil on the chest.
8. If needed, place a pillow or other support under the knees to make the patient more
comfortable.
9. Place a 22-G i.v. catheter in the antecubital fossa. Use of a power injector is necessary
for administration of contrast agent.
10. Use the centering light to position the patient with the center at the patient’s mid-chest
and put him or her into the center of the magnet.
11. Once the patient has been centered in the magnet, check again to be sure that the ECG
tracing demonstrates positive R-waves for suitable triggering. If not, then with the
patient still positioned within the magnet, toggle the lead polarity until a suitable
tracing is obtained.
If the resulting trace is still not satisfactory, then bring the patient out of the magnet, check
the lead connections, and if necessary, reposition the leads until a satisfactory tracing is
obtained.
12. Load the injector with a double dose (0.2 mmol/kg) of gadolinium contrast agent.
This is usually 40 ml or less. The concentration of gadolinium contrast agent is usually 0.5
M.
Do not inject the contrast agent at this time.
Sequence 1: Rapid three-plane positioning scout
13. For localization of subsequent acquisitions, run the system’s three-plane scout scan
according to Table A12.1.2.
Table A12.1.2 Primary Clinical Imaging Parameters for Sequence 1 (Pilot
Scan)

5. Inform the patient about what will occur during the procedure, what he or she will Patient position Supine
experience while in the magnet, and how to behave, including the following: Scan type Gradient echo
Imaging plane (orientation) Three-plane: transverse, coronal,
a. If earphones or headphones are used to protect the ears from the loud sounds and sagittal
produced by the gradients, the patient will be asked to wear these, but will be able Central slice or volume center Upper chest
to communicate with you at any time during the imaging. Echo time (TE) 3.1 msec
b. The patient will be given a safety squeeze-bulb or similar equipment to request Repeat time (TR) 6.5 msec
assistance at any time (demonstrate how this works). Flip angle (FA) 80°
c. For good results the patient should not talk, and should avoid or minimize other Fields of view (FOVx, FOVy) 450 mm, 450 mm
movement, during each scan—i.e., as long as the banging sounds continue. Resolution (Δx, Δy) 1.76 mm, 2.34 mm
Between scans, talking is allowed in most cases, but should be avoided when Number of data points collected (Nx, Ny) 256, 192
comparative positional studies are being performed; the patient will be informed Display matrix (Dx, Dy) 256, 256
when this is the case. Slice thickness (Δz) 5 mm
Number of slices 7
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
Slice gap Variable
e. The patient will be asked to hold his or her breath for ∼15 to 25 sec for most of Number of acquisitions (Nacq) 1
the sequences. Assess the need for supplementary oxygen to improve breath-hold- Swap read and phase encoding No
ing capacity, and if necessary, administer 2 liters oxygen via nasal cannula. Advise Slice locations Variable
the patient of the importance of not moving during the acquisition periods and of Saturation pulses No
not taking deep breaths during the nonbreath-hold acquisitions. Scan time 8 sec
Acquired Aortic
Aortic Dissection Disease

A12.1.2 A12.1.3
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Sequence 2: 2-D transverse black blood half Fourier single shot turbo spin echo Table A12.1.4 Primary Clinical Imaging Parameters for Sequence 3 (Coronal
(black blood HASTE) and Sagittal Black Blood HASTE)
The black blood HASTE sequence utilizes a nonselective 180° presaturation pulse and a
Patient position Supine
selective 180° pulse to reduce signal from blood flow. Vessels with flow appear as signal
Scan type Single shot fast spin echo
void. Obtain black blood HASTE images in transverse, coronal, and oblique sagittal
Imaging plane (orientation) Coronal or oblique sagittal (see
planes. An oblique sagittal plane is prescribed along the plane of the aortic arch to cover
Basic Protocol, step 17)
the entire thoracic aorta.
Central slice or volume center Center of heart
14. Position a series of transverse slices off the scout image from the lung apices to below Echo time (TE) 43 msec
the diaphragm. Two or three sets are necessary to cover the entire thorax. Repeat time (TR) Infinity
Delay time (TD) 180 mseca
15. Run sequence 2 according to Table A12.1.3. Flip angle (FA) 180°b
This is a nonbreath-hold sequence. Fields of view (FOVx, FOVy) 450 mm, 450 mm
Resolution (Δx, Δy) 1.76 mm, 2.81 mm
Sequence 3: 2-D coronal and sagittal black blood half Fourier single shot turbo spin Number of data points collected (Nx, Ny) 256, 160
echo (black blood HASTE) Display matrix (Dx, Dy) 256, 256
16. Position a series of coronal slices off the transverse images to ensure coverage of the Slice thickness (Δz) 6 mm
entire thoracic aorta and run sequence 3 according to Table A12.1.4. Number of slices 20
Slice gap 1 mm
17. Position a series of oblique sagittal slices off the transverse images along the Number of acquisitions (Nacq) 1
orientation of the aortic arch and run sequence 3 according to Table A12.1.4. Swap read and phase encoding No
This is a nonbreath-hold sequence. Saturation pulses No
ECG gating Yes
Sequence 4: 2-D breath-hold cine gradient echo Scan time >47 sec
Gradient-echo cine images are obtained to assess the aortic valve and hemodynamics of aThe true trigger delay time is one R-to-R interval.
bThe system displays the flip angle of the refocusing pulse.
aortic dissection. Multiple phases of the heart cycle are imaged at a single slice position.

Table A12.1.3 Primary Clinical Imaging Parameters for Sequence 2 18. Choose the optimal line per segment acquisition and adjust the number of cardiac
(Transverse Black Blood HASTE) phases.

Patient position Supine The parameters may be adjusted depending on the patient’s breath-holding capability and
Scan type Single shot fast spin echo heart rate. Two parameters should be modified. One is the number of phase encoding lines
(Ny) collected for each excitation, which determines the minimal number of heart beats
Imaging plane (orientation) Transverse
required to obtain the single-level cine image. With 128 phase encoding steps, 15 heart
Central slice or volume center Center of heart beats are required for a 9-line per segment acquisition, 19 heart beats for a 7-line per
Echo time (TE) 43 msec segment acquisition, and 26 heart beats for a 5-line per segment acquisition. Choose the
Repeat time (TR) Infinity protocol with the optimal line per segment acquisition (in the Siemens’ protocol, this is
Delay time (TD) 180 mseca expressed by the number of heart beats required for imaging). Then adjust the number of
Flip angle (FA) 180°b cardiac phases. Number of cardiac phases times TR should be ≥100 msec shorter than
Fields of view (FOVx, FOVy) 350 mm, 350 mm minimal R-to-R interval. Finally, the number of phase encoding steps can be modified, but
Resolution (Δx, Δy) 1.37 mm, 2.19 mm this may change the number of heart beats required to obtain the image.
Number of data points collected (Nx, Ny) 256, 160 19. Position a single transverse slice at the level just above the aortic valve.
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 5 mm 20. Instruct the patient to hold their breath and run the sequence according to Table
Number of slices 20c A12.1.5.
Slice gap 0 If the patient cannot hold their breath for this sequence, use sequence 5 as an alternative.
Number of acquisitions (Nacq) 1
Swap read and phase encoding No 21. Position a single oblique coronal slice along the aortic root, so that the aortic valve
Saturation pulses No can be evaluated in profile and repeat step 20.
ECG gating Yes
Scan time >47 sec
Sequence 5: 2-D nonbreath-hold cine gradient echo (alternative for sequence 4)
aThe true trigger delay time is one R-to-R interval.
This is a nonbreath-hold gradient-echo cine sequence. When the patient is unable to hold
bThe system displays the flip angle of the refocusing pulse. their breath for the breath-hold cine sequence, use this sequence to assess the aortic valve.
cTwo to three sets of 20 slices are generally necessary to cover the entire chest. Acquired Aortic
Aortic Dissection Disease

A12.1.4 A12.1.5
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A12.1.5 Primary Clinical Imaging Parameters for Sequence 4 (Cine Table A12.1.6 Primary Clinical Imaging Parameters for Sequence 5 (Cine
Gradient Echo) Gradient Echo)

Patient position Supine Patient position Supine

Scan type Segmented k-space cine gradient Scan type Segmented k-space cine gradient
echo echo
Imaging plane (orientation) Transverse or oblique coronal (see Imaging plane (orientation) Oblique coronal (see Basic
Basic Protocol, step 21) Protocol, step 22)
Central slice or volume center Center of heart Central slice or volume center Center of heart
Echo time (TE) 6.1 msec Echo time (TE) 7 msec
Number of lines per segment 5, 7, or 9a Number of lines per segment 7
Repeat time (TR) 60-100 msecb (temporal Repeat time (TR) 80 msec (temporal resolution)
resolution) Delay time (TD) after R-wave 0 mseca
Delay time (TD) after R wave 0 msec Flip angle (FA) 25°
Flip angle (FA) 25° Fields of view (FOVx, FOVy) 300 mm, 300 mm
Fields of view (FOVx, FOVy) 300 mm, 300 mm Resolution (Δx, Δy) 1.17 mm, 2.34 mm
Resolution (Δx, Δy) 1.17 mm, 2.34 mm Number of data points collected (Nx, Ny) 256, 128
Number of data points collected (Nx, Ny) 256, 128 Display matrix (Dx, Dy) 256, 256
Display matrix (Dx, Dy) 256, 256 Slice thickness (Δz) 8 mm
Slice thickness (Δz) 8 mm Number of slices 1
Number of slices 1 Slice gap Not applicable
Slice gap Not applicable Number of acquisitions (Nacq) 3
Number of acquisitions (Nacq) 1 Swap read and phase encoding No
Swap read and phase encoding No Saturation pulses No
Saturation pulses No Number of cardiac phases 8–12b
Number of cardiac phases 8-15c ECG gating Yes
ECG gating Yes Scan time >2 min, 5 sec
Scan time ∼10–20 sec aThe true trigger delay time is one R-to-R interval.
aChoose the number of lines per segment according to the patient’s breath-hold capability. bNumber of cardiac phases multiplied by T should be ≥100 msec shorter than the minimal R-to-R
R
bT is dependent on the number of lines per segment collected.
R interval.
cNumber of cardiac phases multiplied by T should be ≥ 100 msec shorter than the minimal R-to-R
R
interval.
Sequence 7: 2-D transverse nonbreath-hold T1-weighted spin echo (alternative for
sequence 6)
The limitation of this sequence is that the acquisition time is ≥2 min to obtain single level
This is a nonbreath-hold T1-weighted sequence, which can be used in patients with limited
cine images.
breath-hold capacity.
22. Position a single oblique coronal slice along the aortic root, so that aortic valve can
24. Position a series of transverse slices through the selected area of the aorta that has
be evaluated in profile and run the sequence according to Table A12.1.6.
wall thickening or some irregularity, and run the sequence according to Table A12.1.8.
Sequence 6: 2-D transverse breath-hold single slice T1-weighted turbo spin echo
3-D contrast-enhanced MRA
(optional)
When aortic dissection is evident in the images from any of the previous sequences, the
T1-weighted turbo spin echo (TSE) is a sensitive sequence to detect intramural hematoma.
body coil should be used for a contrast-enhanced MRA to cover the entire aorta from
If there is thickening or irregularity of the aortic wall, obtain T1-weighted turbo spin echo
aortic arch to aortic bifurcation to delineate the extent of the dissection, site of entry, and
images through the area. Only one image can be obtained during one breath-hold, and
origin of the great vessels in relation to each lumen. When the dissection is not evident
multiple breath-holds are necessary to cover the area of interest. If the patient cannot hold
from previous sequences, phased array coil is still used and only thoracic aorta is imaged.
their breath, use a nonbreath-hold T1-weighted spin echo sequence (sequence 7) as an
alternative. A breath-hold technique is the preferred sequence because of reduced motion 25. If the entire aorta needs to be imaged, select the body coil from the scanner console.
artifact. Check the scout image to see if there is enough caudad coverage. If not, recenter the
patient over the xyphoid.
23. Position transverse slices through the selected area of aorta with wall thickening or
irregularity, instruct the patient to hold their breath and run the sequence according 26. If the patient is repositioned, repeat the system’s three-plane scout scan according to
to Table A12.1.7. Table A12.1.2.
Acquired Aortic
Aortic Dissection Disease

A12.1.6 A12.1.7
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A12.1.7 Primary Clinical Imaging Parameters for Sequence 6 (Single Table A12.1.9 Primary Clinical Imaging Parameters for Sequence 8 (Test
Slice Breath-Hold T1-Weighted TSE) Bolus)

Patient position Supine Patient position Supine

Scan type Fast spin echo Scan type Gradient echo
Imaging plane (orientation) Transverse Imaging plane (orientation) Transverse
Central slice or volume center Center of heart Central slice or volume center Mid chest
Echo time (TE) 30 msec Echo time (TE) 2.4 msec
Echo train length (ETL) 13 Repeat time (TR) 5.8 msec
Repeat time (TR) 700 mseca Inversion time (TI) 300 msec
Delay time (TD) after R-wave 150 msec Delay time (TD) 0 msec
Flip angle (FA) 160°b Flip angle (FA) 10°
Fields of view (FOVx, FOVy) 350 mm, 263 mm Fields of view (FOVx, FOVy) 400 mm, 300 mm
Resolution (Δx, Δy) 1.37 mm, 2.05 mm Resolution (Δx, Δy) 1.56 mm, 2.34 mm
Number of data points collected (Nx, Ny) 256, 128 Number of data points collected (Nx, Ny) 256, 128
Display matrix (Dx, Dy) 256, 256 Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 5 mm Slice thickness (Δz) 10 mm
Number of slices 1 Number of slices 1
Slice gap Not applicable Slice gap Not applicable
Number of acquisitions (Nacq) 1 Number of acquisitions (Nacq) 1
Swap read and phase encoding No Number of repetitions 50
Saturation pulses No Swap read and phase encoding No
ECG gating Yes Saturation pulses No
Scan time >7 sec ECG gating No
aT should be 50 to 100 msec shorter than minimal R-to-R interval.
R
Scan time 52 sec
bThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this

sequence is 90°.
Sequence 8: Test bolus
Prior to a full injection of the contrast agent, a test bolus is performed to determine the
Table A12.1.8 Primary Clinical Imaging Parameters for Sequence 7 contrast travel time from the peripheral vein to the thoracic aorta.
(Nonbreath-Hold T1-Weighted Spin Echo)
27. Position a single transverse slice at the level of the mid-thoracic aorta.
Patient position Supine
Scan type Spin echo 28. Set up the injector for a 2-ml gadolinium injection followed by 15 ml normal saline
Imaging plane (orientation) Transverse at the rate of 2 ml/sec.
Central slice or volume center Center of heart
29. As one is starting the injector, run sequence 8 according to Table A12.1.9. The system
Echo time (TE) 30 msec
will then acquire a single slice every second for 50 sec at the same location.
Repeat time (TR) 540 mseca
Delay time (TD) after R-wave 150 msec 30. By scrolling the images, determine the time to peak signal of the contrast bolus arrival
Flip angle (FA) 160°b from the start of contrast agent injection.
Fields of view (FOVx, FOVy) 300 mm, 263 mm
Resolution (Δx, Δy) 1.17 mm, 2.05 mm Sequence 9: 3-D precontrast MRA (mask)
Number of data points collected (Nx, Ny) 256, 128 The precontrast sequence is obtained as a mask for subtraction and to ensure the coverage
Display matrix (Dx, Dy) 256, 256 of the aorta and to check the image quality.
Slice thickness (Δz) 5 mm
Number of slices 10 31. Position the slab in a sagittal or oblique sagittal plane along the aortic arch. By using
Slice gap 1 mm an oblique sagittal slab, the slab thickness (thus, number of slices/partitions) can be
Number of acquisitions (Nacq) 2 reduced to shorten the image time.
Swap read and phase encoding No 32. If the thoracic and abdominal aorta need to be imaged, change the coil configuration
Saturation pulses Superior and inferior to body coil.
ECG gating Yes
Scan time >2 min, 21 sec 33. Instruct the patient to hold their breath and run sequence 9 according to Table
aT should be 50 to 100 msec less than minimal R-to-R interval. A12.1.10.
R
bThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this Acquired Aortic
Aortic Dissection Disease
sequence is 90°.
A12.1.8 A12.1.9
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A12.1.10 Primary Clinical Imaging Parameters for Sequence 9 (3-D Table A12.1.11 Primary Clinical Imaging Parameters for Sequence 10 (3-D
MRA, Mask) Contrast-Enhanced MRA)

Patient position Supine Patient position Supine

Scan type 3-D gradient echo Scan type 3-D gradient echo
Imaging plane (orientation) Sagittal or oblique sagittal Imaging plane (orientation) Sagittal or oblique sagittal
Central slice or volume center Mid-chest or xyphoida Central slice or volume center Mid-chest or xyphoida
Echo time (TE) 1.8 msec Echo time (TE) 1.8 msec
Repeat time (TR) 4.6 msec Repeat time (TR) 4.6 msec
Flip angle (FA) 30° Flip angle (FA) 30°
Fields of view (FOVx, FOVy) 390 mm, 390 mm or 480 mm, 480 Fields of view (FOVx, FOVy) 390 mm, 390 mm or 480 mm, 480
mmb mmb
Resolution (Δx, Δy) 1.52 mm, 2.29 mm or 0.94 mm, Resolution (Δx, Δy) 1.52 mm, 2.29 mm or 0.94 mm,
2.82 mm 2.82 mm
Number of data points collected (Nx, Ny) 256, 170 or 512, 170c Number of data points collected (Nx, Ny) 256, 170 or 512, 170c
Display matrix (Dx, Dy) 256, 256 or 512, 512 Display matrix (Dx, Dy) 256, 256 or 512, 512
Slice thickness (Δz) 3 mm (1.5 mm after interpolation) Slice thickness (Δz) 3 mm (1.5 mm after interpolation)
Number of slices 40 (80 after interpolation) Number of slices 40 (80 after interpolation)
Slice gap 0 mm Slice gap 0 mm
Number of acquisitions (Nacq) 1 Number of acquisitions (Nacq) 1
Swap read and phase encoding No Number of repetitions 2, with 8-sec interval between
ZIPd 2 Yes these two measures
Saturation pulses No Swap read and phase encoding No
ECG gating No ZIP 2 Yes
Scan time 25 sec Saturation pulses No
aCenter at mid-chest if phased array coil is used and at xyphoid if body coil is used. ECG gating No
b390 mm FOV for phased array coil and 480 mm FOV for body coil. Scan time 25 sec (the total sum time
c256, 170 matrix for phased away coil and 512, 170 matrix for body coil. including two measurements is 58
dZerofill interpolation process. This process produces twice the number of slices by interpolating the sec)
data. Also see UNIT A7.4. aCenter at mid chest if phased array coil is used and at xyphoid if body coil is used.
b390 mm FOV for phased array coil and 480 mm FOV for body coil.
c256, 170 matrix for phased array coil and 512, 170 matrix for body array coil.
34. Check the images to ensure the coverage of the entire aorta. Minimal wrapping
artifact does not interfere with the diagnostic ability of the study. The degree of
respiration artifact should be checked. 38. Change the coil configuration as described in step 32.
If the patient cannot hold their breath long enough, consider shortening the scan time at 39. Start the injector and instruct the patient to hold their breath at the 5-sec mark before
the cost of spatial resolution. the calculated scan delay from step 36. When the scan delay is reached, start sequence
10 according to Table A12.1.11.
Sequence 10: 3-D contrast-enhanced MRA
35. Set up the injector for a 38-ml gadolinium injection followed by 15 ml normal saline Data processing and viewing for sequences 9 and 10
at the rate of 2 ml/sec. 40. Create a subtraction data set from the pre- and post-contrast MRA. Display these data
The amount of contrast agent is dependent on patient’s body weight.
sets using a maximal intensity projection (MIP). For multi-plannar reconstruction
(MPR), use the nonsubtracted original data sets.
36. Scan delay from the start of injection can be calculated based on time to peak of
contrast arrival determined by the test bolus injection, injection duration, and scan
TIME OF FLIGHT (TOF) ALTERNATE
duration using the following formula: PROTOCOL
When no abnormal findings are seen in black blood HASTE or cine sequences, it is
scan delay = (time to peak) + (injection duration)/2 − (scan duration)/2. somewhat controversial whether contrast-enhanced MRA is necessary. In this instance,
The scan time is typically 25 sec and injection duration 19 sec (38 ml at 2 ml/sec). If the a time of flight (TOF) sequence may be used instead of a contrast-enhanced MRA.
time to peak at test injection is 15 sec, the scan delay would be 12 sec (= 15 + 19/2 − 25/2).
Set up patient and equipment
37. Position the slab exactly the same as in sequence 9. 1. Use the same equipment and the same patient setup as for the previous method except
for venous access or contrast agent (see Basic Protocol, steps 1 to 11, excluding step
The position should be copied from history so that subtraction can be performed later. Two
measures (two repetitions) are performed in this sequence, and an ∼8-sec interval should 9). No venous access or contrast agent is necessary.
Acquired Aortic
Aortic Dissection be allowed for the patient to breathe between the first and second measures. 2. Run sequences 1 to 7 as in Basic Protocol, steps 13 to 24. Disease

A12.1.10 A12.1.11
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Sequence 11: 2-D transverse breath-hold gradient echo time of flight (TOF) Table A12.1.13 Primary Clinical Imaging Parameters for Sequence 12
This is a breath-hold bright blood TOF sequence. Vessels with flowing blood appear bright (Transverse Gradient Echo TOF)
because of the time-of-flight effect.
Patient position Supine
3. Position transverse slices from lung apex to diaphragm. Instruct the patient to hold Scan type Segmented k-space gradient echo
their breath and run sequence 11 according to Table A12.1.12. Imaging plane (orientation) Transverse
Only 1 to 3 slices can be obtained in a single breath-hold. Central slice or volume center Center of heart
Echo time (TE) 10 msec
Multiple breath-holds are required to cover the entire aorta. Number of lines per segment 13
Sequence 12: 2-D transverse nonbreath-hold gradient echo time of flight (TOF) Repeat time (TR) 333 msec (temporal resolution)
(alternative for sequence 11) Delay time (TD) after R-wave 200 msec
This is a nonbreath-hold TOF sequence as an alternative for sequence 11 for patients who Flip angle (FA) 50°
cannot hold their breath. Fields of view (FOVx, FOVy) 300 mm, 263 mm
Resolution (Δx, Δy) 1.17 mm, 2.02 mm
4. Position a series of transverse slices from the lung apex to diaphragm and run Number of data points collected (Nx, Ny) 256, 130
sequence 12 according to Table A12.1.13. Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 5 mm
Sequence 13: 2-D transverse breath-hold cine gradient echo (same as Basic
Number of slices 30
Protocol, sequence 4) Slice gap 0
Cardiac pulsation artifact may be severe with the TOF technique (sequence 12) in some Number of acquisitions (Nacq) 2
slices near the heart. In those instances, use the cine gradient echo sequence to clear the Swap read and phase encoding No
cardiac pulsation artifact. Saturation pulses No
5. Repeat Basic Protocol, step 18 and position a single transverse slice over the area of ECG gating Yes
interest where pulsation artifact was severe by the TOF technique (sequence 12). Scan time >5 min, 43 sec

6. Instruct the patient to hold their breath and run sequence 13 according to Table
A12.1.5.
COMMENTARY
Background Information Critical Parameters and
Table A12.1.12 Primary Clinical Imaging Parameters for Sequence 11 Aortic dissection is the most common Troubleshooting
(Transverse Gradient Echo TOF)
acute emergency affecting the aorta. Un-
treated aortic dissection is often fatal, and Customization of gated MR imaging
Patient position Supine parameters
immediate diagnosis is necessary.
Scan type Segmented k-space gradient echo Aortic dissection can be diagnosed using MR imaging of the aorta requires optimiza-
Imaging plane (orientation) Transverse MRI, CT, or transesophageal echo (TEE) tion of imaging parameters according to a pa-
Central slice or volume center Center of heart with similar accuracy. MRI is most suitable tient’s heart rate and breath-hold capability. For
Echo time (TE) 10 msec in relatively stable patients or patients with a a turbo spin echo sequence, the TR should be
Number of lines per segment 13 contraindication to iodinated contrast agent. shorter than the patient’s minimal R-to-R inter-
Repeat time (TR) 333 msec (temporal resolution) The classification of aortic dissection is val (also see the discussion under the same
Delay time (TD) after R-wave 200 msec important in treatment decision making. Type subsection in UNIT A11.1).
Flip angle (FA) 50° A aortic dissections (dissections involving For cine gradient echo images, multiple
Fields of view (FOVx, FOVy) 300 mm, 263 mm ascending aorta) usually require emergency phases of the cardiac cycle are imaged at a
Resolution (Δx, Δy) 1.17 mm, 2.02 mm surgery because of the risk of complication, single slice position. The parameters may be
Number of data points collected (Nx, Ny) 256, 130 including rupture into pericardial space and adjusted depending on the patient’s breath-
resultant tamponade, coronary artery occlu- holding capability and heart rate. Two parame-
Display matrix (Dx, Dy) 256, 256
sion, and aortic valvular regurgitation. Type ters should be modified. One is the number of
Slice thickness (Δz) 5 mm
B aortic dissections (dissection not involving phase encoding lines collected for each excita-
Number of slices 1–3 ascending aorta) are usually treated medi- tion, which determines the minimal number of
Slice gap 0 mm cally. heart beats required to obtain the single-level
Number of acquisitions (Nacq) 1 Intramural hematoma and penetrating cine image. With 128 phase encoding steps, 15
Swap read and phase encoding No atherosclerotic ulcer are other conditions that heart beats are required for a 9-line per segment
Saturation pulses No affect the aorta and require immediate medi- acquisition, 19 heart beats for a 7-line per seg-
ECG gating Yes cal attention. ment acquisition, and 26 heart beats for a 5-line
Scan time ∼6–18 sec per segment acquisition. Choose the protocol
Acquired Aortic
Aortic Dissection Disease

A12.1.12 A12.1.13
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 with optimal line per segment acquisition (us-
ing the Siemens’ system, these protocols are
A B expressed by the number of heart beats required
for imaging). Then adjust the number of cardiac
phases. The number of cardiac phases multi-
plied by TR should be at least 100 to 150 msec
shorter than the minimal R-to-R interval. Fi-
nally, the number of phase encoding steps can
be modified, but this may change the number
of heart beats required to obtain the image.
Breath-hold versus nonbreath-hold
technique
Occasionally, despite a patient’s effort to
hold their breath, substantial motion artifacts
are seen on MR images. These artifacts can also
be seen when the patient is not able to hold their
breath long enough to complete the sequence.
In these instances, use nonbreath-hold se-
quences as an alternative. Breath-hold se-
quences usually give better images, and should
initially be tried, if the patient can hold their
breath long enough.
Customizing 3-D contrast-enhanced MRA
C D Subclavian artery
for patients who are not able to hold their
breath
Typically, an MRA sequence is ∼25 sec
long, but the patient may not be able to sustain
their breath for the entire sequence. In this
situation, shorten the scan time to 15 to 20 sec
by reducing the FOV along the phase encoding
directions or slab thickness and/or by reducing
the phase encoding steps or number of parti-
tions. If the patients can hold their breath for
only ∼10 sec, instruct the patient to hold their
Figure A12.1.1 Aortic dissection. Transverse (A) and coronal (B) ECG-gated black blood images
demonstrate an intimal flap (arrows) in the descending aorta. (C) Transverse ECG-gated TOF image
demonstrates an intimal flap (arrow) in the descending aorta. (D) Multi-planar reconstruction image
from the Gadolinium-enhanced MRA shows an intimal flap (arrow) in the descending aorta. The
relation between the origin of left subclavian artery (arrowhead) and the flap is well depicted.
Figure A12.1.2 Intramural hematoma. Transverse breath-hold ECG-
gated T1-weighted turbo spin echo image shows a crescent area of
Aortic Dissection hyperintensity (arrows) in the wall of the descending aorta.
A12.1.14
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
breath during the middle third of the scan,
granted noncentric k-space filling is used. If the
patient cannot hold their breath at all, have the
patient breathe shallowly through the scan, and
obtain the scan at a routine or high resolution
setting.
Anticipated Results
Identification of an intimal flap in the aortic
lumen is key to the diagnosis of aortic dissec-
tion. With the spin-echo technique, the intimal
flap appears as an intermediate signal curvilin-
ear structure in the background of low signal
from blood (Fig. A12.1.1A and A12.1.1B; Si-
monetti et al., 1996; Krinsky et al., 1997). With
a contrast-enhanced MRA or TOF technique,
the intimal flap appears as an intermediate to
low signal curvilinear structure in the back-
ground of high signal blood (Fig. A12.1.1C and
A12.1.1D; Prince et al., 1996). When the false
lumen is thrombosed, the intimal flap is not
visible. However, the thrombosed lumen ap-
pears as a crescent area of high signal intensity
on the spin echo technique, which is indistin-
guishable from intramural hematoma. Very
slow flow in the lumen can mimic a thrombosed
lumen on noncontrast techniques. In this in-
stance, contrast-enhanced MRA is useful in
differentiating thrombosed lumen versus slow
flow.
One should be careful not to mistake pulsa-
tion or motion artifact for an intimal flap. These
artifacts tend to extend outside the lumen of the
aorta. This artifact occurs only in the phase
encoding direction (UNIT B7.3), and usually, the
source of artifact can be identified. On the 3-D
Acquired Aortic
Disease
A12.1.15
Supplement 12
 contrast-enhanced MRA sequence, these arti-
facts could arise from out-of-plane structures
such as the main pulmonary artery, since there
is a second phase encoding gradient along the
z-axis (slab direction). In instances where a
linear structure is seen in the aortic lumen and
one is not certain if this is a flap, use the cine
sequence at the same level to differentiate it
from an artifact.
The extent of aortic dissection should be
carefully evaluated. Contrast-enhanced MRA
is the best sequence for this purpose since it is
a 3-D technique and has the highest spatial and
contrast resolution in the shortest possible time.
Use of multi-planar reconstruction is important
in the evaluation of the extent of aortic dissec-
tion and also to access the origin of the vessel
in relation to true and false lumens (Fig.
A12.1.1D).
Intramural hematoma appears as a crescent
of high signal intensity on the spin echo tech-
nique. T1-weighted spin echo (sequence 6 or 7)
is a sensitive sequence to identify the intramural
hematoma (Fig. A12.1.2; Murray et al., 1997).
With contrast-enhanced MRA or TOF, intra-
mural hematomas can be overlooked (Krinsky
et al., 1997). Penetrating atherosclerotic ulcer
appears as a focal outpouching of aortic lumen
into the aortic wall.
Literature Cited
Krinsky, G.A., Rofsky, N.M., Decorato, D.R., We-
inreb, J.C., Earls, J.P., Flyer, M.A., Galloway,
Aortic Dissection
A12.1.16
Supplement 12
A.C., and Colvin, S.B. 1997. Thoracic aorta:
Comparison of gadolinium-enhanced three-di-
mensional MR angiography with conventional
MR imaging. Radiology 202:183-193.
Murray, J.G., Manisali, M., Flamm, S.D., Van Dyke,
C.W., Lieber, M.L., Lytle, B.W., and White, R.D.
1997. Intramural hematoma of the thoracic aorta:
MR image findings and their prognostic impli-
cations Radiology 204:349-355.
Prince, M.R., Narasimham, D.L., Jacoby, W.T., Wil-
liams, D.M., Cho, K.J., Marx, M.V., and Deeb,
G.M. 1996. Three-dimensional gadolinium-en-
hanced MR angiography of the thoracic aorta.
AJR Am. J. Roentgenol. 166:1387-1389.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Simonetti, O.P., Finn, J.P., White, R.D., Laub, G.,
and Henry, D.A. 1996. “Black blood” T2-
weighted inversion-recovery MR imaging of the
heart. Radiology 1996:49-57.
Sommer, T., Fehske, W., Holzknecht, N., Smekal,
A.V., Keller, E., Lutterbey, G., Kreft, B., Kuhl,
C., Gieseke, J., Abu-Ramadan, D., and Schild, H.
1996. Aortic dissection: a comparative study of
diagnosis with spiral CT, multiplanar transeso-
phageal echocardiography, and MR imaging.
Radiology 199:347-352.
Contributed by Naoki Takahashi and
Vamsidhar Narra
Mallinckrodt Institute of Radiology
Washington University Medical Center
St. Louis, Missouri
Current Protocols in Magnetic Resonance Imaging
Aortic Aneurysm and Pseudoaneurysm UNIT A12.2
Assessment
Although contrast angiography is still considered the “gold standard” for the evaluation
of the aorta and its major branches, magnetic resonance angiography (MRA) has quickly
gained popularity as an imaging tool for the assessment of the entire aorta. MRA serves
as an alternative imaging modality that can be utilized in patients with impaired renal
function and with allergies to iodinated contrast medium (iodinated contrast medium is
required in contrast angiography and computed tomography, CT). The purpose of this
unit is to present fundamental MRA techniques useful in the evaluation of the thoracic
and abdominal aorta based on experience on a 1.5 T GE LX scanner.
IMAGING OF THORACIC AORTA BASIC
PROTOCOL 1
The sequences used in the evaluation of the thoracic aorta include black-blood T1-
weighted or T2-weighted images as well as white-blood gadolinium-enhanced short TR
gradient echo images. For the black-blood sequence, the authors use cardiac-gated double
inversion recovery fast spin echo imaging in the transverse plane. However, cardiac-gated
T1-weighted transverse spin echo images are an acceptable alternative. The authors
perform gadolinium-enhanced imaging as a standard dynamic 3-D acquisition in the
sagittal oblique plane. The entire Basic Protocol 1 will take ∼30 min to perform.
Table A12.2.1 outlines the equipment parameters necessary for imaging of the thoracic
aorta.
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.
Materials
Normal saline (0.9% NaCl), sterile
Gadolinium-based MR contrast agent (e.g., Magnevist, Omniscan, or Prohance)
Set up patient and equipment
1. Screen patients for ferromagnetic materials, such as cardiac pacemakers, cochlear
implants, and intracranial aneurysm clips, which preclude them from MR imaging.
Know the patient’s medical and surgical history, such as respiratory or cardiac
conditions or presence of spinal hardware, which is critical to optimize the patient’s
MRI exam.
Generally, standard screening forms (APPENDIX 1) are used for all patients scanned in a
magnetic resonance system.
Table A12.2.1 Equipment Parameters for Thoracic Aorta Imaging
Coil type Body coil
Gradient coil strength 40 mT/m, but minimum 20 mT/m is
required
Cardiac gating Yes
Peripheral gating No
Respiratory gating Yes
Oxygen 2-liter nasal cannula if required by patient
Use of contrast agents Yes, 40 ml gadolinium chelate i.v. Acquired Aortic
Disease
Contributed by Maricela Contreras, Rajesh S. Amin, and E. Kent Yucel A12.2.1
Current Protocols in Magnetic Resonance Imaging (2001) A12.2.1-A12.2.14
Copyright © 2001 by John Wiley & Sons, Inc. Supplement 3
 The presence of any ferromagnetic metals may be a health hazard to the patient when he Table A12.2.2 Primary Clinical Imaging Parameters for Sequence 1 (Pilot Scan)
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock Patient position Supine, feet first
(1996) for discussion of what implants may be safely scanned using magnetic resonance. Scan type Short TR gradient echo
Patients may be accompanied into the magnet room by a friend or family member, who can Imaging plane (orientation) Three planes (transverse, sagittal, and
sit in the room during the scan and comfort the patient as needed. This companion must coronal)
be screened as well to ensure the absence of loose metal objects on the body or clothing. Central slice or volume center Center of chest
Echo time (TE) 1.6 msec (or minimum)
2. If the procedure is a research protocol, have the patient sign any necessary consent Repeat time (TR) 8.1 msec (or minimum)
form. Flip angle (FA) 30°
3. Have the patient remove all jewelry and change into a gown to eliminate any metal Fields of view (FOVx, FOVy) 400 mm, 400 mm
that might be found in clothing. Resolution (Δx, Δy) 1.56 mm, 3.13 mm
Number of data points collected (Nx, Ny) 256, 128
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating Display matrix (Dx, Dy) 256, 256
and image artifacts. Slice thickness (Δz) 10 mm
5. Inform the patient about what will occur during the procedure, what he or she will Number of slices 9/plane
experience while in the magnet, and how to behave, including the following: Slice gap 2 mm
Number of acquisitions (Nacq) 2
a. If earphones or headphones are used to protect the ears from the loud sounds Swap read and phase encoding No
produced by the gradients, the patient will be asked to wear these, but will be able Slice locations Variable
to communicate with you at any time during the imaging. Saturation pulses None
b. The patient will be given a safety squeeze-bulb or similar equipment to request Scan time 57 sec
assistance at any time.
c. For good results the patient should not talk, and should avoid or minimize
swallowing or other movement, during each scan, i.e., as long as the banging Table A12.2.3 Primary Clinical Imaging Parameters for Sequence 2 (Double
sounds continue. Between scans, talking and swallowing are allowed in most Inversion Recovery Fast Spin Echo)
cases, but should be avoided when comparative positional studies are being
performed; the patient will be informed when this is the case. Patient position Supine, feet first
Scan type Double inversion recovery fast spin
d. Nevertheless, the patient may call out at any time if he or she feels it necessary. echo
6. Instruct the patient how to hold his or her breath at end expiration for 20 to 30 sec. Imaging plane (orientation) Transverse
Supplemental oxygen may facilitate breath-holding. Anxiolytics, such as Valium or Central slice or volume center Center of chest
Xanax, may be useful in patients who are anxious or claustrophobic. Echo time (TE) 40 msec
Receiver bandwidth (RBW) 62.5 kHz
7. Have the patient mount onto the table in the supine position with feet first in the Echo train length (ETL) 32
magnet. Either before or right after the patient lies down, set up any triggering devices Repeat time (TR) 2 R-to-R intervals
or other monitoring equipment that are to be used. Delay time (TD) Minimum (∼12 msec)
Flip angle (FA) 90°
8. Place electrocardiogram (ECG) leads on the patient in accordance with need.
Fields of view (FOVx, FOVy) 300 mm, 300 mm
9. Place a 20- to 22-G i.v. catheter, preferably in an antecubital vein, for the injection Resolution (Δx, Δy) 1.17 mm, 1.56 mm
of gadolinium-based MR contrast agent during imaging. The patient’s arms are Number of data points collected (Nx, Ny) 256, 192
positioned above his or her head to decrease wraparound artifact. Display matrix (Dx, Dy) 512, 512
Slice thickness (Δz) 10 mm
10. If needed, place a pillow or other support under the knees to make the patient more
Number of slices ∼20 (or as needed)
comfortable. Slice gap None
11. Use the centering light to position the patient and put him or her into the center of Number of acquisitions (Nacq) 2
the magnet. Swap read and phase encoding No
Slice locations Aortic arch to diaphragm
The patient’s chest should be centered in the body coil. If only the thoracic aorta is to be
ZIP 512 Yes
imaged, the nipple line can be used as a landmark. Otherwise, the xyphoid can be used as
Slice series Sequential
an anatomical landmark if the entire aorta is to be imaged.
Aortic ECG gating Yes
Aneurysm and Scan time 8 min, 8 sec
Pseudoaneurysm Acquired Aortic
Assessment Disease

A12.2.2 A12.2.3
Supplement 3 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 3
 Table A12.2.4 Primary Clinical Imaging Parameters for Sequence 3 (Optional Table A12.2.5 Primary Clinical Imaging Parameters for Sequence 4 (3-D
Cardiac-Gated Transverse T1-Weighted Spin Echo for Black-Blood Imaging) Gadolinium-Enhanced Short TR Gradient Echo)

Patient position Supine, feet first Patient position Supine, feet first
Scan type T1-weighted spin echo Scan type 3-D short TR gradient echo
Imaging plane (orientation) Transverse Imaging plane (orientation) Sagittal oblique
Central slice or volume center Center of chest Central slice or volume center Center of chest
Echo time (TE) 20 msec Echo time (TE) About ≤3 msec
Repeat time (TR) Depends on heart rate (1 R-to-R Receiver bandwidth (RBW) 31 kHz
interval) Repeat time (TR) About ≤10 msec
Delay time (TD) Minimum Flip angle (FA) 35°
Flip angle (FA) 90° Fields of view (FOVx, FOVy) 400 mm, 400r mm, with r = 3/4
Fields of view (FOVx, FOVy) 300 mm, 300 mm (rectangular field of view)
Resolution (Δx, Δy) 1.17 mm, 1.88 mm Resolution (Δx, Δy) 1.56 mm, 1.56 mm
Number of data points collected (Nx, Ny) 256, 160 Number of data points collected (Nx, Ny) 256, 256r, with r = 3/4 (rectangular
Display matrix (Dx, Dy) 256, 256 field of view)
Slice thickness (Δz) 8 mm Display matrix (Dx, Dy) 512, 512
Number of slices 20 Slice thickness (Δz) 2.6 mm
Slice gap 2 mm Number of slices 32
Number of acquisitions (Nacq) 2 Slice gap None
Swap read and phase encoding Yes Number of acquisitions (Nacq) 0.5 (half Fourier)
Read direction Anterior to posterior Read direction Superior to inferior
Slice locations Aortic arch to diaphragm Slice locations Aortic arch to diaphragm
Saturation pulses None ZIP 512 Yes
Slice series Sequential ZIP 2 Yes
ECG gating Yes Saturation pulses Interleaved fat saturation pulse
Scan time ∼4 min, 36 sec Slice series Sequential
ECG gating No
Scan time 38 sec
Once this step has been performed, so long as the patient does not move on the table, the
table itself can be moved and then replaced in the same position as before without
jeopardizing the positioning of one scan relative to another. The number of slices acquired will depend on the heart rate. For example, for a heart rate
of 75 beats per min, the number of slices acquired will be ∼24 slices.
Sequence 1: Three-plane localizer pilot scan
12. Run sequence 1 according to Table A12.2.2. Sequence 4: Sagittal oblique 3-D gadolinium-enhanced short TR gradient echo
Three-plane localizer pilot scan is obtained using a short TR gradient echo sequence. (infusion)
Patient may hold his or her breath in end expiration, if possible. 15. Leave patient in the magnet and inject the gadolinium-based contrast agent, then flush
the line with 10 ml saline.
Sequence 2: Transverse double inversion recovery fast spin echo (black-blood 16. Determine the 3-D gadolinium-enhanced MRA volume position from the pilot scan.
sequence) Begin the volume coverage 1 to 2 cm above the arch and extend caudally to the
13. Instruct the patient breath normally. Run sequence 2 according to Table A12.2.3. diaphragm. Use the aortic arch as the axis of the plane.
Double inversion recovery fast spin echo sequence is used for black-blood imaging and is
An automatic dose timing module such as Smartprep (GE) or fluoroscopic triggering
implemented with cardiac gating and nonbreath-holding. Black-blood imaging enables
(Phillips), if available, can be used to track the bolus of gadolinium in the descending aorta.
accurate measurement of true aortic diameter, including patent and thrombosed portions
of the lumen. Images are obtained as sequential slices with interpolation to 512 by 512 17. Set injection rate of contrast agent at 2.5 ml/sec for a total of 40 ml.
display matrix. The triggering window is 20% of the R-to-R interval with minimum delay
after the QRS waves. Slice are obtained during free breathing. A dose of 0.1 mmol/kg of contrast agent is usually given.
The authors use automatic triggering to detect the arrival of contrast agent in a region of
Sequence 3: Transverse T1-weighted spin echo (optional black-blood sequence) interest in the aorta and to determine the time between injection and data acquisition for
14. Run sequence 3 according to Table A12.2.4. the central lines of the k-space (centric reordering). This allows for synchronization
between the acquisition of central lines of k-space and the arterial phase of the exam.
If patient is unable to hold his or her breath, the T1-weighted transverse spin echo images
Aortic can be obtained with cardiac gating and respiratory compensation. Images are sequentially If the reader wishes to run a test bolus sequence, see UNIT A10.1 or UNIT A12.1.
Aneurysm and
Pseudoaneurysm acquired from the level of the diaphragm to the aortic arch. Also, the read direction is Acquired Aortic
Assessment changed to anterior to posterior. Disease

A12.2.4 A12.2.5
Supplement 3 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 3
 18. Start the injection of the contrast agent. When the bolus of gadolinium is detected in Table A12.2.7 Primary Clinical Imaging Parameters for Sequence 5 (Pilot Scan
the aorta, instruct the patient to hold his or her breath. Wait 6 sec and run sequence Abdominal Aorta)
4 according to Table A12.2.5.
Patient position Supine, feet first
The arterial phase images are obtained during the first breath-hold period. Scan type Short TR gradient echo
An interleaved fat saturation pulse may be used to suppress background signal, if available. Imaging plane (orientation) Three planes (transverse, sagittal, and
Spectrally fat-selective saturation pulses, such as SPECIAL (Spectral Inversion at Lipids, coronal)
GE Medical Systems) can be used with a time of 22 msec before applying the rf pulse. Central slice or volume center Center of abdomen (at level of
Alternatively, a precontrast image may be obtained and mask subtraction performed. umbilicus)
Echo time (TE) 1.6 msec
In addition to acquiring sequential slices with interpolation to 512 by 512 display matrix Repeat time (TR) 8.1 msec
in the plane of the slice, slices are reconstructed at half the interval in the slice select
Flip angle (FA) 30°
direction acquired.
Fields of view (FOVx, FOVy) 400 mm, 400 mm
19. Instruct the patient to breathe and take another breath-hold. Six seconds after the Resolution (Δx, Δy) 1.56 mm, 3.13 mm
previous scan, run sequence 4 again to obtain the venous phase images. Number of data points collected (Nx, Ny) 256, 128
Display matrix (Dx, Dy) 256, 128
BASIC IMAGING OF ABDOMINAL AORTA Slice thickness (Δz) 10 mm
PROTOCOL 2 Number of slices 9/plane
For patients with suspected abdominal aortic aneurysm, transverse and sagittal black- Slice gap 2 mm
blood images are necessary followed by coronal 3-D gadolinium-enhanced MRA on the Number of acquisitions (Nacq) 2
abdominal aorta. The patient set-up for this exam is the same as in Basic Protocol 1. The Swap read and phase encoding No
equipment parameters (see Table A12.2.6) for abdominal aorta imaging are also similar, Slice locations Variable
except no cardiac-gating is required. The entire Basic Protocol 2 will take ∼30 min to Saturation pulses None
perform. Scan time 57 sec

Set up patient and equipment

1. Repeat Basic Protocol 1, steps 1 to 11. Table A12.2.8 Primary Clinical Imaging Parameters for Sequence 6 (T1-Weighted
Spin Echo)
Sequence 5: Three-plane localizer pilot scan
2. Run three-plane localizer pilot scan using the short TR gradient echo sequence Patient position Supine, feet first
according to Table A12.2.7. Patient may hold his or her breath in end expiration. Scan type T1-weighted spin echo
Imaging plane (orientation) Transverse
The body coil is used to ensure that both the aorta and iliac vessels are imaged. The pilot Central slice or volume center Center of abdomen (at level of
scan should include both kidneys and major vessels of the abdominal aorta (proximal celiac umbilicus)
and superior mesenteric arteries, and renal arteries). In tall patients, the field of view may Echo time (TE) 20 msec
need to be increased to 48 cm.
Repeat time (TR) 600-800 msec
Flip angle (FA) 90°
Sequence 6: T1-weighted spin echo (black-blood sequence)
Fields of view (FOVx, FOVy) 400 mm, 400 mm
3. Run sequence 6 according to Table A12.2.8.
Resolution (Δx, Δy) 1.56 mm, 2.50 mm
As in the thoracic aorta, transverse T1-weighted spin echo images are acquired through Number of data points collected (Nx, Ny) 256, 160
the abdominal aorta for black-blood imaging. This allows accurate measurement of the Display matrix (Dx, Dy) 256, 256
aortic diameter and facilitates evaluation of the lumen for thrombus and patency. Respi- Slice thickness (Δz) 10 mm
ratory compensation can also be utilized.
Number of slices 15-21
Slice gap 2 mm
Number of acquisitions (Nacq) 2
Table A12.2.6 Equipment Parameters for Abdominal Aorta Imaging
Swap read and phase encoding No
Coil type Body coil Scan time ∼3-4 min
Gradient coil strength At least 20 mT/m
Cardiac gating No
Peripheral gating No
Respiratory gating Yes
Aortic Oxygen 2 liter nasal cannula if required by patient
Aneurysm and Use of contrast agents Yes, 40 ml gadolinium chelate
Pseudoaneurysm Acquired Aortic
Assessment intravenously Disease

A12.2.6 A12.2.7
Supplement 3 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 3
 Table A12.2.9 Primary Clinical Imaging Parameters for Sequence 7 (3-D 6. Start the injector (contrast timing module). After 10 sec (see above), instruct the
Gadolinium-Enhanced Short TR Gradient Echo) patient to hold his or her breath and run sequence 7 according to Table A12.2.9.
Patient position Supine, feet first An interleaved fat saturation pulse may be used to suppress background signal.
Scan type 3-D short TR Gradient Echo Images are also obtained as sequential slices with interpolation to a 512 by 512 display
Imaging plane (orientation) Coronal matrix. Slices are also reconstructed at half the interval (or l.3-mm increments).
Central slice or volume center Center of abdomen (at level of
umbilicus)
ANEURYSMS FOLLOW-UP ALTERNATE
Echo time (TE) About ≤3 msec
PROTOCOL
Receiver bandwidth (RBW) 62 kHz Generally, the initial assessment of true and false aneurysms in patients with no contra-
Repeat time (TR) About ≤10 msec indications to iodinated i.v. contrast can be performed with CT. However, in the presence
Flip angle (FA) 35° of a contraindication or in the follow-up assessment of aortic aneurysms, MRI can be a
Fields of view 400 mm, 400r mm, with r = 3/4 useful alternative. In the situation of follow-up assessment, once the aneurysm has been
(rectangular field of view) thoroughly evaluated, MRI utilizing a black-blood technique, such as gradient echo
Resolution (Δx, Δy) 1.56 mm, 2.34 mm sequences (FLASH, fast low angle shot-type imaging) or DIR (double inversion recovery)
Number of data points collected (Nx, Ny) 256, 128 sequences in the transverse plane, can provide adequate imaging to monitor for size
Display matrix (Dx, Dy) 512, 512 variations in a given interval time. Although contrast enhanced MRA is advantageous in
Slice thickness (Δz) 2.6 mm the initial assessment of aneurysms, imaging without contrast would be more cost-effec-
Number of slices 48 tive in the follow-up evaluation.
Slice gap None
Number of acquisitions (Nacq) 0.5 (half Fourier) 1. Repeat Basic Protocol 1, steps 1 to 11 (skip step 9).
Swap read and phase encoding No
2. Repeat sequences 1 to 2.
Slice locations Diaphragm to common femoral arteries
ZIP 512 Yes
ZIP 2 Yes COMMENTARY
Saturation pulses Interleaved fat saturation pulse Background Information with Marfan’s, Ehrler’s-Danlos, and others (see
Slice series Sequential UNIT A10.1,
Alternate Protocol 2).
Scan time 38 sec Thoracic aortic aneurysm
Thoracic aortic aneurysms (TAA) and Aortic aneurysms and pseudoaneurysms
pseudoaneurysms can have several etiologies A false aneurysm, or pseudoaneurysm, in-
4. Choose the imaging plane (orientation) to be sagittal. Run sequence 6 again according including atherosclerosis, cystic medial degen- volves disruption of one or more of the arterial
to the rest parameters in Table A12.2.8. eration, trauma, infection, connective tissue wall layers. Pseudoaneurysms are often de-
disorders, and vasculitis (Figure A12.2.1). The scribed as saccular with blood generally con-
Sequence 7: Coronal 3-D gadolinium-enhanced short TR gradient echo (infusion) majority of TAA can be attributed to tained by the surrounding tissue or the outer
The coronal 3-D gadolinium-enhanced short TR gradient echo sequence is the primary atherosclerosis, which typically involves the adventitia. The neck of the pseudoaneurysm
sequence for demonstrating the entire abdominal aorta, its major vessels and the iliac descending aorta, although, any of the possible can be variable in size. Although trauma is the
arteries. The 3-D MRA volume position is determined from the pilot scan. The volume etiologies for TAA can also involve the ascend- most common etiology of thoracic aortic
position must be anterior enough to include the origins of the celiac and superior ing aorta (Figure A12.2.2). The normal ascend- pseudoaneurysms, they can also arise as com-
ing aorta should be <3.0 cm in diameter and the plications from cardiac surgery concerning
mesenteric arteries and the common femoral arteries, if possible. The posterior border of
descending aorta should be <2.5 cm. Generally, valve replacements. Within the thorax or abdo-
the 3-D volume should include both iliac arteries, renal arteries, and kidneys.
the risk of rupture is low if the aortic aneurysm men pseudoaneurysms can form subsequent to
For optimal timing of the gadolinium bolus in the abdominal aorta, a Smartprep or similar is <5 cm in diameter (Hirose et al., 1992). certain vascular surgical procedures such as
Aneurysms are usually monitored until they bypass grafting, particularly at cannulation or
contrast timing module can be used. This is especially crucial during the arterial phase of
reach 6 cm in diameter, although the threshold anastomotic sites. Pseudoaneurysms can be-
the bolus. In the presence of an abdominal aortic aneurysm, a delay of 10 sec can be used
for surgery is highly dependent on the patient’s come secondarily infected or be the result of
with the contrast timing module. Otherwise, a delay of 5 or 6 sec will suffice if the aorta underlying physical condition and life expec- infection in the post-operative period. The lu-
is of normal caliber. The gadolinium injection rate for the aorta should be 2.5 ml/sec for tancy. MR imaging including contrast en- men of pseudoaneurysms may contain patent
a total of 40 ml. hanced imaging can provide adequate preop- to-and-from flow or be completely thrombosed
erative information such as branch vessel in- and calcified. Surgery is advocated for most
Images are acquired in the arterial and venous phases using the method as described in volvement, aneurysm extent, characteristics, pseudoaneurysms regardless of their size.
Basic Protocol 1, steps 18 and 19. and location (Krinsky et al., 1997; Figure An abdominal aortic aneurysm (AAA) is
A12.2.3). Cine imaging can be used to identify described as dilation of the abdominal aorta
5. Repeat Basic Protocol 1, step 15. Determine the slice positions as described above.
Aortic the neck and source of the contained rupture in with a wall-to-wall diameter >3 cm or localized
Aneurysm and
Repeat Basic Protocol 1, step 17. pseudoaneurysms. Cine imaging might also be dilation of the vessel by >50% of its normal
Pseudoaneurysm included to assess the aortic valve in patients diameter (Johnston et al., 1991). The incidence Acquired Aortic
Assessment Disease

A12.2.8 A12.2.9
Supplement 3 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 3
 Figure A12.2.1 Pseudoaneurysm. Sagittal oblique maximum intensity projection (MIP) of the
proximal aorta reveals a saccular pseudoaneurysm in the region of the isthmus.

Figure A12.2.3 Penetrating ulcer. Coronal MIP image of a contrast-enhanced MRA (CEMRA)
demonstrates a markedly tortuous thoracic aorta with a focal aneurysm in the aortic arch and a
large penetrating ulcer in the proximal descending thoracic aorta. An infrarenal abdominal aortic
aneurysm is also present.

of AAA continues to increase largely due to
earlier detection and aging of the population.
Most AAA’s are detected incidentally. In those
patients who are symptomatic, the common
manifestations include abdominal pain, ab-
dominal mass, and distal embolization.
AAA are often described as either fusiform
(concentric dilation) or saccular (asymmetric
dilation). Suprarenal, juxtarenal, and infrarenal
are terms used to describe the location and
Aortic extent of AAA. The most common cause of
Aneurysm and Figure A12.2.2 Ascending aortic aneurysm. Transverse double inversion recovery image dem-
Pseudoaneurysm onstrates a 6 cm aneurysm in the ascending aorta at the level of the pulmonary artery. The etiology AAA is atherosclerosis with associated degen-
Assessment of this aneurysm was Takayasu’s Arteritis. eration. Other less common causes include
chronic dissections, infection, trauma, inflam-
mation, and vasculitis. Although CT angiogra-
phy and ultrasound have played significant
roles in the detection and surveillance of AAA,
contrast enhanced MRA (CEMRA) continues
to emerge as a powerful tool in the preoperative
and postoperative assessment of AAA (Schoen-
berg et al., 1999; Grist, 2000). CEMRA can
provide high-resolution three-dimensional im-
aging without ionizing radiation or the risks
associated with iodine-based radiographic con-
trast media. MRA provides key preoperative Acquired Aortic
information regarding the proximal and distal Disease

A12.2.10 A12.2.11
Supplement 3 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 3
 Figure A12.2.4 Abdominal aortic aneurysm with accessory right renal artery. Coronal MIP image Figure A12.2.5 Coronal CEMRA of the distal aorta reveals a fusiform distal abdominal aortic
of the abdominal aorta demonstrates an infrarenal aneurysm without extension into the common aneurysm with extension into the common iliac arteries.
iliac arteries. A subtle small accessory right renal artery is also seen arising from the distal
abdominal aorta.
Critical Parameters and mated detection, patients with slow flow may
extent of the aneurysm (Figure A12.2.4). It also Most vascular surgeons agree that aneu- Troubleshooting require up to a 10-sec delay between detection
provides information concerning the number rysms <4 cm in diameter should be monitored Volume coverage is critical, especially when of bolus and scan initiation. Routinely obtain-
and patency of the renal arteries, patency of the and that aneurysms >5 cm should be repaired the iliac arteries are tortuous. This can be reme- ing a second phase helps. Venous phase imag-
mesenteric arteries, presence of an inflamma- (Crawford and Hess, 1989). Intervention on died by increasing the slice thickness as needed. ing is generally important if there is a problem
tory aneurysm, and the presence of obstructive AAA in the 4 to 5 cm range depends on the The MIPs can overestimate stenosis. There- with arterial phase timing (not uncommon in
iliofemoral arterial disease (Figure A12.2.5). patient’s underlying medical condition and ex- fore, it is important to review the source images aneurysms) and for seeing false lumen of dis-
The presence of such disease or variants, as well pected life-expectancy, as well as whether the to obtain precise measurements and determine section.
as location and extent of disease, affects surgi- AAA is increasing in size on serial imaging the degree of stenosis, especially in the renal
cal approach and technique. Precise assessment studies. MRI can also be useful in the postop- arteries and iliac arteries. Anticipated Results
of the infrarenal neck and the status of the iliac erative evaluation for complications such as The timing of the contrast bolus is a crucial The goals of MR imaging of the thoracic and
vessels is even more important now with the anastomotic pseudoaneurysm formation and part of MRA. Ringing artifacts can occur if the abdominal aorta in patients with suspected
Aortic advent of percutanous aortic stent-grafts for aortoenteric fistula. CEMRA should be useful central k-space data are acquired during the aneurysms are detection, surveillance, preop-
Aneurysm and treatment of AAA. for follow-up of stent-grafts for potential leaks, rising portion of the contrast-enhancement erative and postoperative assessment. With new
Pseudoaneurysm although this will require additional study. curve during arterial enhancement. With auto- advances in technique and the emerging role of Acquired Aortic
Assessment Disease

A12.2.12 A12.2.13
Supplement 3 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 3
 contrast enhanced MRA, the sensitivity of this
valuable imaging tool approaches that of con-
ventional angiography. Certainly the lack of
ionizing radiation and nephrotoxicity provides
an advantage over CT angiography. The ability
to identify the extent of the aneurysm, branch
vessel involvement, and congenital variants
clearly provides more information than sonog-
raphy. The major role of CEMRA in the evalu-
ation of TAA and AAA is likely in the preop-
erative assessment. For example, extension of
the aneurysm into the iliac vessels would re-
quire placement of a bifurcated graft. Identify-
ing the superior most extent of an TAA or AAA
and its relationship to the left subclavian artery
origin or the renal artery origin is important
because their involvement alters the surgical
approach.
Literature Cited
Crawford, E.S. and Hess, K.R. 1989. Abdominal
aortic aneurysm. N. Engl. J. Med. 321:1040-
1042.
Grist, T. 2000. MRA of the abdominal aorta and
lower extremities. J. Mag. Reson. Imaging
11:32-43.
Hirose, Y., Hamada, S., Takamiya, M., Imakita, S.,
Naito, H., and Nishimura, T. 1992. Aortic aneu-
rysm: Growth rates measured by CT. Radiology
185:249-252.
Johnston, K.W., Rutherford, R.B., Tilson, M.D.,
Shah, D.M., Hollier, L., and Stanley, J.C. 1991.
Suggested standards for reporting on aortic
aneurysms. J. Vasc. Surg. 13:452-458.
Krinsky, G.A., Rofsky, N.M., DeCorato, D.R., We-
inreb, J.C., Earls, J.P., Flyer, M.A., Galloway,
A.C., and Colvin, S.B. 1997. Thoracic aorta:
Comparison of gadolinium-enhanced three-di-
mensional MR angiography with conventional
MR imaging. Radiology 202:183-193.
Schoenberg, S.O., Wunsch, C., Knopp, M.V., Essig,
M., Hawighorst, H., Laub, G., Prince, M.R.,
Aortic
Aneurysm and
Pseudoaneurysm
Assessment
A12.2.14
Supplement 3
Allenberg, J.R., and Van Kaick, G. 1999. Ab-
dominal aortic aneurysm. Detection of multi-
level vascular pathology by time-resolved multi-
phase 3D gadolinium MR angiography: Initial
report. Invest. Radiol. 34:648-659.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Key References
Fenlon, H.M. and Yucel, E.K. 1999. Advances in
abdominal, aortic, and peripheral contrast-en-
hanced MR angiography. In Magnetic Reso-
nance Imaging Clinics of North America: New
Techniques in Body MR Imaging. W. B. Sauders
Company, Philadelphia.
Reviews basic principles and new MRA techniques
in the evaluation of the aorta and peripheral arter-
ies.
Prince, M.R., Grist, T.M., and Debatin, J.K., 1999.
3D Contrast MR Angiography. 2nd Ed. Sprin-
ger-Verlag, New York.
Provides practical information on the technical as-
pects and clinical applications of MR angiography.
Yucel, E.K., Anderson, C.M., Edelman, R.R., Grist,
T.M., Baum, R.A., Manning, W.J., Culebras, A.,
and Pearce, W. 1999. Magnetic resonance
angiography: Update on applications for extrac-
ranial arteries (AHA Scientific Statement). Cir-
culation 100:2284-2301.
The American Heart Association Science Advisory
and Coordinating Committee reviews the current
clinical applications and provides recommenda-
tions for the use of MRA in the extracranial vessels.
Potential future clinical applications for MRA are
also presented.
Contributed by Maricela Contreras, Rajesh S.
Amin, and E. Kent Yucel
Brigham & Women’s Hospital
Boston, Massachusetts
Current Protocols in Magnetic Resonance Imaging
Pulmonary Embolism UNIT A13.1
There have been a number of approaches taken to image the pulmonary vasculature. Black
blood spin echo and/or gradient echo techniques have been applied for detection of
pulmonary embolisms and deep vein thrombosis. On the other hand, the entire vascular
tree can be visualized with the bright blood magnetic resonance angiography (MRA)
technique, 2-D time-of-flight (TOF), and 3-D contrast-enhanced MRA. Although non-
contrast-enhanced 3-D TOF MRA techniques have been described for imaging the
pulmonary vessels, adequate vascular signal-to-noise ratio without respiratory- or cardio-
vascular-related motion artifacts requires respiratory gating (or navigator echo respiratory
compensation) as well as electrocardiogram (ECG) triggering. This leads to a time
consuming scan, the results of which are often not reproducible on most MRI systems.
This method will not be discussed. The parameters provided in this unit are from the
authors’ experience with the Siemens 1.5T Vision Scanner. These parameters may need
to be altered depending on the field strength and equipment manufacturer.
IMAGING THE PULMONARY ARTERIES WITH BLACK BLOOD SPIN BASIC
ECHO AND GRADIENT ECHO TECHNIQUES PROTOCOL 1
The sequences described in this section are based on the authors’ experience with a
Siemens 1.5T Vision Scanner, but are expected to be equally applicable to machines from
other manufacturers. It should be noted that with current experience employing these
sequences, small thrombi beyond the interlobar vessels are very difficult to diagnose with
confidence. The turbo spin echo (fast spin echo) techniques, described herein, are optional
at the authors’ imaging facility, but provide higher resolution imaging of the pulmonary
arterial wall than the gradient-echo techniques. They are frequently omitted due to time
constraints (3 to 4 min) and the presence of artifacts in patients who are not stable and
who have dyspnea. Table A13.1.1 lists the hardware required for the sequences. The entire
examination time is 45 min, which includes patient set-up with ECG gating hardware.
NOTE: Be sure that technicians and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.
Set up patient and equipment
1. Interview (screen) the patient to ensure that he or she has no contraindications such
as cardiac pacemakers or other implants containing ferromagnetic materials. Also,
be sure to find out if the patient has any health conditions that may require the
Table A13.1.1 Hardware Requirements for Black Blood Imaging of Pulmonary Arteries
Coil type Body phased array coil preferred over
quadrature body coil
Maximum gradient coil strength 25 mT/m (or whatever the system permits)
Cardiac triggering Yes
Peripheral vital sign monitoring Optional (blood pressure, heart rate,
peripheral O2 saturation, and CO2
capnometry)
Respiratory gating No
Respirator If required by patient
Pulmonary
Oxygen Usually administered in these patients (2-3 Artery,
liters/min) by nasal cannula Mediastinum,
Pleura, and Lung
Contributed by Kostaki G. Bis and Anil N. Shetty A13.1.1
Current Protocols in Magnetic Resonance Imaging (2001) A13.1.1-A13.1.12
Copyright © 2001 by John Wiley & Sons, Inc. Supplement 2
 Table A13.1.2 Imaging Parameters for Sequence 1 (Localizers) Table A13.1.3 Imaging Parameters for Sequence 2 (Black Blood Multi-Cardiac
Phase Multi-Slice Turbo Spin Echo)a
Patient position Supine
Scan type 2-D gradient echo Patient position Supine
Imaging plane (orientation) Transverse, sagittal, and coronal Scan type Turbo spin echo
Central slice or volume center Laser light centered at Imaging plane (orientation) Transverse
mid-sternum Central slice or volume center Top slice of first stack at top of
Echo time (TE) 6.0 msec aortic arch. Second stack below
Repeat time (TR) 15.0 msec first stack with one slice overlap.
Flip angle (FA) 30° Echo time (TE) 12 msec
Fields of view (FOVx, FOVy) 450 mm, 450 mm Echo train length (ETL) 3 (the echo spacing is 12.4 msec)
Resolution (Δx, Δy) 1.76 mm, 3.52 mm Repeat time (TR) Set 100 msec less than R-to-R
interval but not >750 msec.
Number of data points collected (Nx, Ny) 256a, 128
Alternatively, use 80% of R-to-R
Display matrix (Dx, Dy) 256, 256 interval.
Slice thickness (Δz) 10 mm Delay time (TD) after R-wave 50 msec
Number of slices 3 Flip angle (FA) 180°
Slice gap Not applicable Fields of view (FOVx, FOVy) 400 mmb, 300 mm
Number of acquisitions (Nacq) 1 Resolution (Δx, Δy) 1.56 mm, 1.92 mm
Swap read and phase encoding No Number of data points collected (Nx, Ny) 256, 156
Slice location Not applicable Display matrix (Dx, Dy) 256, 256
Saturation pulses Not applicable Slice thickness (Δz) 6 mm
Slice series Not applicable Number of slices ∼13 (80% of R-to-R
Scan time 16 secb interval/duration of ETL)
aOversampling
Slice gap 1.2 mm (20%)
bSequential 2-D-TOF sequence takes about 6 sec per slice.
Number of acquisitions (Nacq) 5
Swap read and phase encoding No
presence of special emergency equipment during the scanning procedure, or neces- Saturation pulses Yes, parallel, i.e., superior and
sitate any other precautions. inferior, TR may increase
Scan time >2 min, 43 sec
Generally, standard screening forms (APPENDIX 1) are used for all patients scanned in a aUse black blood preparation pulse.
magnetic resonance system. bOversampling.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants, see Shellock b. The patient will be given a safety squeeze-bulb or similar equipment to request
(1997) for discussion of what implants may be safely scanned using a magnetic resonance assistance at any time (demonstrate how this works). It is provided in any event
imager. requiring urgent intervention.
Patients may be accompanied into the magnet room by a friend or a family member, who c. For good results the patient should not talk, and should avoid or minimize other
can sit in the room during the scan and comfort the patient as needed. This companion movement, during each scan—i.e., as long as the banging sounds continue.
must be screened as well to ensure the absence of loose metal objects on the body or Between scans, talking is allowed in most cases, but should be avoided when
clothing. comparative positional studies are being performed; the patient will be informed
2. If the procedure is a research protocol, have the patient sign any necessary consent when this is the case.
forms. d. Nevertheless, the patient may call out at any time during the scanning procedure
3. Have the patient remove all jewelry and change into a gown to eliminate any metal if he or she feels it necessary.
that might be found in clothing. e. The patient is provided with instructions on being very still during the acquisition
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating of black blood spin echo sequences. He or she is also informed on holding his or
and image artifacts. her breath on inspiration for the breath-holding turbo spin echo and breath-hold
gradient echo sequences.
5. Inform the patient about what will occur during the procedure, what he or she will
experience while in the magnet, and how to behave, including the following: 6. Have the patient lie down supine on the table over the posterior phased-array coil.
a. If earphones or headphones are used to protect the ears from the loud sounds Place the ECG triggering leads appropriately over the chest. Either before or right
produced by the gradients, the patient will be asked to wear these, but will be able after the patient lies down, set up any ECG triggering devices, or other monitoring Pulmonary
to communicate with you at any time during the imaging procedure. equipment, that are to be used. Artery,
Pulmonary Mediastinum,
Embolism Pleura, and Lung

A13.1.2 A13.1.3
Supplement 2 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 2
 Table A13.1.4 Imaging Parameters for Breath-Holding Turbo Spin Echo Table A13.1.5 Imaging Parameters for Breath-Holding Cine MRI
Sequencea
Patient position Supine
Patient position Supine Scan type Segmented k-space gradient echo
Scan type Turbo spin echo with echo sharing
Imaging plane (orientation) Transverse Imaging plane (orientation) Transverse
Central slice or volume center Centered at the central pulmonary Central slice or volume center Begin first slice at
arteries where an accurate aortoco-pulmonic window and
assessment of pulmonary arterial then proceed inferiorly with ∼10
wall thickness is needed or in breath holding acquisitions
suspicious areas Echo time (TE) 4.8 msec
Echo time (TE) 76 msec Number of lines per segment 5 or 7
Echo train length (ETL) 23 Repeat time (TR) 80.0 msec (temporal resolution)
Repeat time (TR) 800 msec (based on R-to-R Delay time (TD) 0 msec
interval, keep nearly equal to Flip angle (FA) 20°
R-to-R interval)
Fields of view (FOVx, FOVy) 350 mma, 263 mm
Delay time (TD) 500 msec
Resolution (Δx, Δy) 1.37 mm, 2.09 mm
Flip angle (FA) 160°b
Number of data points collected (Nx, Ny) 256, 126
Fields of view (FOVx, FOVy) 400 mmc, 300 mm
Display matrix (Dx, Dy) 256, 256
Resolution (Δx, Δy) 1.56 mm, 2.17 mm
Slice thickness (Δz) 6-8 mm
Number of data points collected (Nx, Ny) 256, 138
Number of slices 1/breath-hold
Display matrix (Dx, Dy) 256, 256
Slice gap 1.2-1.6 mm (20%)
Slice thickness (Δz) 5-7 mm
Number of acquisitions (Nacq) 1
Number of slices 1/breath-hold
Swap read and phase encoding No
Slice gap Not applicable
Saturation pulses No
Number of acquisitions (Nacq) 1
Number of cardiac phases 2n − 1 (two phases per TR)b
Swap read and phase encoding No
ECG gating Yes
Saturation pulses No
Scan time 15 sec/breath-hold
Scan time >9 sec aOversampling.
aBlack blood preparation pulse.
bDue to interleaved segmentation with shared echo, 2 phases per T result in twice as many cardiac
bThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this R
phases, n = (80% of R-to-R interval/TR), approximately equal to 10.
sequence is 180°.
cOversampling.
The TR is set to be ∼80% of the patient’s R-to-R interval, and a trigger delay of 50 msec is
7. If needed, place a pillow or other support under the knees to make the patient more also used. Generally, only two stacks of 2-D slices are obtained, beginning from the top of
comfortable. the aortic arch. Parallel (superior and inferior) saturation bands are also employed.

8. Use the centering light to position the patient (at mid-sternum) and put him or her
into the center of the magnet.
Once this step has been performed, so long as the patient does not move on the table, the
table itself can be moved and then replaced in the same position as before without
jeopardizing the positioning of one scan relative to another.
9. Do not sedate the patient.
Sedation is not appropriate for patients with pulmonary embolisms.
Sequence 1: Rapid three-plane positioning scouts
10. Run the system’s pilot scan according to the imaging parameters given in Table
A13.1.2 to insure the correct location of the patient’s chest.
Three orthogonal planes are obtained.
Sequence 2: Multi-slice multi-cardiac phase turbo spin echo (nonbreath-hold)
11. Bring the multi-slice multi-cardiac phase turbo spin echo sequence onto the console.
Pulmonary
Embolism
12. Run sequence 2 according to Table A13.1.3.
Sequence 3: Breath-holding turbo spin echo
This optional sequence can be employed on suspicious areas after reviewing the transverse
multi-slice, multi-cardiac phase turbo spin echo images. In patients with pulmonary
thrombo-embolism, this is usually not performed since the patient will tire and abort the
examination. This can, however, be applied for better depiction of the pulmonary arterial
wall for assessment of pulmonary wall thickness, especially if there are many motion
artifacts related to respiration on sequence 2.
13. Instruct the patient to hold his or her breath and run sequence 3 according to Table
A13.1.4.
Sequences 4 and 5: Imaging the pulmonary arteries with dynamic 2-D time-of-flight
cine MRI
Currently, segmented gradient echo sequences are employed for this form of white blood
imaging and can be applied as a breath-holding protocol whereby one slice is obtained
per breath-holding period or whereby multiple slices can be obtained over a nonbreath- Pulmonary
Artery,
holding period. The multi-slice nonbreath-holding protocol is applied to patients who Mediastinum,
Pleura, and Lung

A13.1.4 A13.1.5
Supplement 2 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 2
 Table A13.1.6 Imaging Parameters for Multi-Slice Nonbreath-Hold Cine MRI Although this is the preferred MRA technique for imaging the pulmonary vessels for
pulmonary emboli, many centers, including the authors’, primarily use computed to-
Patient position Supine mography (CT) angiography since it is more readily available and has higher spatial
Scan type Segmented k-space gradient echo resolution. MRA, therefore, is reserved at the authors’ institution for patients that have a
with sequential mode contraindication to iodinated contrast agents such as, severe allergic reaction or renal
Imaging plane (orientation) Transverse insufficiency. In addition, MRA is frequently preferred in the patient who has severe
Central slice or volume center 5 slices are positioned with the pulmonary hypertension.
top slice through the
aortico-pulmonic window. Two The scan time is on the order of 16 to 19 sec, making it easier to perform while the patient
stacks are obtained by shifting the is holding his or her breath. The entire examination time is 15 min.
stack exactly 1-slice thickness to
fill in the gap. Materials
Echo time (TE) 6.8 msec
Normal saline (0.9 % NaCl), sterile
Number of lines per segment 7
Extravascular contrast agent (i.e., Magnevist, Omniscan or ProHance)
Repeat time (TR) 40.0 msec (temporal resolution)
20-G angio-catheter needle
Delay time (TD) 0 msec
Power injector
Flip angle (FA) 30°
Fields of view (FOVx, FOVy) 350 mma, 263 mm
Set up patient and equipment
Resolution (Δx, Δy) 1.37 mm, 1.99 mm
1. Repeat Basic Protocol 1, steps 1 to 7 (skip step 5e).
Number of data points collected (Nx, Ny) 256, 132
Display matrix (Dx, Dy) 256, 256 2. Establish an i.v. line in the antecubital fossa with a 20-G angio-catheter needle.
Slice thickness (Δz) 7-8 mm Connect the angio-catheter to the saline at a rate to keep the vein open.
Number of slices 5 per stack
It is preferred, however, to connect a power injector for more reproducible delivery of
Slice gap 7-8 mm (one slice thickness) contrast. At the authors’ institution, a Medrad power injector is used.
Number of acquisitions (Nacq) 3
Swap read and phase encoding No 3. Repeat Basic Protocol 1, steps 8 and 9.
Saturation pulses No 4. Run sequence 1 (see Basic Protocol 1).
Number of cardiac phases ∼20× (80% of R-to-R interval/TR)
Slice series Interleaved From this, the coronal 3-D gradient echo sequence can be positioned frequently. If there is
ECG gating Yes not enough anatomical detail for accurate placement, the same positioning scouts can be
Scan time 8 min, 50 sec applied in a transverse fashion through the main pulmonary artery to unfold the main
pulmonary artery and its bifurcation.
aOversampling.

Sequence 6: 3-D MRA

cannot hold their breath. Although both techniques provide white blood imaging for
detection of low signal intensity thrombo-emboli, both are time consuming and predomi-
nantly employed for the detection of major central thrombo-embolic pulmonary disease
down to the proximal lobar branches. The application of these techniques is primarily in
situations where 3-D contrast enhanced MRA sequences are not available. The multi-slice
nonbreath-holding cine protocol is performed with two interleaved stacks, each stack
consisting of five slices. The breath-holding protocol employs the acquisition of ten
different slices with ten different breath-holding acquisitions.
14. Instruct the patient to hold his or her breath and run sequence 4 according to Table
A13.1.5.
15. Optional: If the patient cannot hold his or her breath, run sequence 5 according to
Table A13.1.6.
BASIC CONTRAST ENHANCED PULMONARY MRA
PROTOCOL 2
Contrast enhanced pulmonary MRA protocols are currently preferred over the black blood
and white blood techniques already described. This protocol is conducted over a breath-
hold period and, as such, it is the most rapid protocol for imaging more of the pulmonary
Pulmonary vascular bed. The imaging acquisition plane can be applied parallel to the vascular course
Embolism without the associated in-plane saturation effects that are seen with TOF techniques.
5. Instruct the patient to hold his or her breath and run the rapid 3-D gradient echo scan
according to Table A13.1.7 (with one repetition) before i.v. contrast administration
in the coronal plane.
If the patient is unable to hold his or her breath for the longer coronal acquisition,
individual sagittal acquisitions with lesser partitions (shorter scan time) are performed in
the right and left pulmonary artery distributions. These are run as a mask for subsequent
subtraction of contrast enhanced data in the same orientations.
6. Inject the extravascular contrast agent, flush the line with saline, and instruct the
patient to hold his or her breath and run sequence 6 according to Table A13.1.7.
A dose of 0.1 mmol/kg of contrast agent is usually given at a rate of 2 to 3 ml/sec followed
by a saline flush consisting of 15 ml at 2-3 ml/sec.
7. At the end of the first repetition, allow the patient to breathe for 8 sec, then ask the
patient to hold his or her breath on inspiration, and run the second repetition of the
MRA sequence according to Table A13.1.7. This technique ensures that the bolus is
not missed.
Utilizing this timing, the pulmonary arteries are usually visualized but there is often some Pulmonary
venous contamination. After injecting the contrast agent, the images display both the Artery,
pulmonary arteries and the pulmonary veins, the former with greater signal. For patients Mediastinum,
Pleura, and Lung

A13.1.6 A13.1.7
Supplement 2 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 2
 with suspected or proven poor cardiac function and low cardiac output, the initial 3-D Post process the 3-D gradient echo sequence
sequence measurement can be delayed by an additional 5 to 10 sec. This protocol is 8. Use the mask image and data prior to i.v. contrast administration, and subtract them
employed at the authors’ institution avoiding a timing bolus. Alternatively, however, a 2-ml from the post-contrast enhanced data. Use the subtracted images to perform maxi-
test bolus may be performed to time the acquisition of the first arterial pass accurately (see
mum intensity projection.
Table A10.1.10).
All source data prior to subtraction are reviewed for presence or absence of filling defects.
This can also be subjected to multi-planar reformation for further evaluating the pulmo-
Table A13.1.7 Imaging Parameters for 3-D Contrast Enhanced MRA for nary vascular bed. At the authors’ institution, typically, a transverse multi-planar reforma-
Breath-Holding tion with a slice thickness of 3 mm in a contiguous fashion through the central pulmonary
arteries is employed. The original source data and multi-planar reformations are reviewed
Patient position Supine followed by the maximum intensity projections.
Scan type 3-D gradient echo with sinc
interpolation. Interpolation
requires zero filling of high COMMENTARY
k-space data and then subjecting
Background Information problem of in-plane saturation from slow flow
to fast Fourier transform.
Magnetic resonance angiography has been is eliminated. These factors improve the signal-
Imaging plane (orientation) Coronal, if the patient is able to
utilized to evaluate the pulmonary vasculature, to-noise ratio and can allow imaging of the
hold their breath using the thicker
noninvasively. Both 2-D and 3-D time-of-flight peripheral pulmonary arterial system to the
coronal 3-D slab. Alternatively, a
techniques have been performed without the subsegmental level. Another important factor
thinner sagittal 3-D slab (with
use of intravenous contrast agents. More re- in the improvement of the signal-to-noise ratio
lesser partitions) may be
employed to separately image the cently, however, MRA of the pulmonary vas- is the development and use of phased array
right and left central pulmonary culature has been performed with the use of surface coils (Hatabu et al., 1992). Cardiac
arteries for reducing the imaging intravascular paramagnetic contrast agents (Si- gating has also been utilized to improve the
time in patients who have monetti et al., 1996). signal-to-noise ratio in techniques depending
difficulty in holding their breath. Magnetic resonance imaging of the chest on in-flow effects and this holds true for tech-
Central slice or volume center The coronal slab is centered at the has been challenging because of problems of niques utilizing an intravascular paramagnetic
pulmonary artery bifurcation. The artifacts from respiratory and cardiac motion. contrast agent. Cardiac gating allows for acqui-
sagittal 3-D slab, when Techniques to deal with respiratory motion sition of information during diastole, which is
performed, is placed to include include respiratory gating/compensation when dephasing from moving spins of blood is
the main pulmonary artery and (Westbrook and Kaut, 1993), pseudogating minimized.
respective ipsilateral vascular bed. (Laub and Kaiser, 1988), and the use of single An important consideration of any pulmo-
Echo time (TE) As short as possible. Currently, breath-holding acquisitions (Shetty et al., nary MRA technique is the ability to image the
1.8 msec 1995). Cardiac gating or triggering (Simonetti pulmonary arteries of both lungs in a reason-
Receiver bandwidth (RBW) 390 Hz et al., 1996) has been used to decrease or elimi- able time period. The 3-D gradient echo tech-
Repeat time (TR) 4.6 msec nate cardiac motion artifact. Motion at the pro- nique described in this unit utilizes gradients,
Flip angle (FA) 25° ton spin level of moving blood has also been a which allow for use of short TR and short TE.
problem. The dephasing effects of moving The combination of short TR and TE, intravas-
Fields of view (FOVx, FOVy) 450 mm, 338 mm
spins within a voxel can result in signal loss cular contrast, and phased array surface coils
Resolution (Δx, Δy) 0.88 mm, 1.72 mm
with techniques relying on in-flow effect of allows imaging of the pulmonary arterial tree
Number of data points collected (Nx, Ny) 512, 196
TOF techniques. Gradient moment rephasing with high signal-to-noise ratio and spatial reso-
Display matrix (Dx, Dy) 256, 256 (velocity compensation; Wielopolski, 1993) lution in a reasonable single breath-hold period.
Slice thickness (Δz) 4.55-5.45 mm (coronal) or and use of cardiac gating have been used to
3.18-4.55 mm (sagittal) decrease or eliminate this problem. Critical Parameters and
Number of slices 22 interpolated to 44 (coronal); Another challenge to pulmonary MRA is Troubleshooting
16-18 interpolated to 32-36 obtaining satisfactory vascular signal-to-noise The preferred imaging techniques to evalu-
(sagittal) ratio (SNR). The air-containing lungs produce ate pulmonary embolism (PE) are of the breath-
Slab thickness 100-120 mm (coronal) or 51-82 magnetic susceptibility artifacts that can de- holding type. Most patients presenting with PE
mm (sagittal) crease the SNR of pulmonary vessels. Time-of- are severely dyspneic and are not able to hold
Slice gap 0 flight techniques relying on in-flow effect suf- their breath >14 to 16 sec. Therefore, when
Number of acquisitions (Nacq) 1 fer from signal loss secondary to in plane satu- considering different imaging options that in-
Number of repetitions 2 ration effects. The use of intravenous volve breath holding, care must be given in
Swap read and phase encoding No gadolinium, however, can markedly shorten the optimizing sequence parameters to reduce im-
ECG gating No T1 of blood (Wielopolski et al., 1993) and, aging time under ≤16 sec.
Scan timea 18 sec per repetition, 44 sec total thereby, markedly increase the signal intensity However, for high spatial resolution, it is
(coronal); 14 sec per repitition, 36 of moving blood independent of in-flow effects imperative to use a higher acquisition matrix Pulmonary
sec total (sagittal) needed in TOF techniques. In addition, by giv- which increases scan time. A better balance is Artery,
Pulmonary Mediastinum,
Embolism aThe scan delay between two repetitions is 8 sec. ing all blood an intrinsically bright signal, the achieved by changing phase encoding lines and Pleura, and Lung

A13.1.8 A13.1.9
Supplement 2 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 2
 A A

Figure A13.1.2 Chronic pulmonary thrombo-embolism. Coronal (A) and sagittal (B) contrast
enhanced 3-D gradient echo partitions display low signal organized thrombus (arrows) within the
roof of the left pulmonary artery.

Figure A13.1.1 Acute pulmonary thrombo-embolism. Coronal contrast enhanced 3-D gradient
echo partitions display central low signal filling defects (arrows) in the left (A) and right (B) interlobar
pulmonary arteries.
Pulmonary
Artery,
Pulmonary Mediastinum,
Embolism Pleura, and Lung

A13.1.10 A13.1.11
Supplement 2 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 2
 slice partitions to reduce scan time. In addition,
providing oxygen and hyperventilation prior to
the examination does help to increase the pa-
tients’ breath-holding capacity even if they are
sick.
Actual application of techniques for imag-
ing pulmonary arteries and their branches may
require some modification of the pulse se-
quence parameters based on patient circulation
time and anatomical region of interest. For
example, the RF flip angle may be increased if
the imaging interest lies in observing the main
pulmonary artery with a sufficiently high blood
flow. Use of the lowest TR and TE always helps
and is dictated by the system manufacturer.
Even with breath-holding techniques, it is
not uncommon to see respiratory related arti-
facts. These are the primary sources of ghosting
artifacts and tend to obscure the details of vas-
cular anatomy. It is important to educate pa-
tients in holding their breath in a consistent
manner.
The bolus arrival time can also be assessed
using “smart prep” developed by Foo et al.
(1997). The smart prep technique uses a small
sampling volume placed downstream in a ves-
sel to sense the arrival of contrast agent. It then
triggers the acquisition of a measurement scan
to correspond to the arrival of gadolinium
within the region. Other techniques such as
fluoroscopic triggering, developed by Riederer
et al. (1988) or 3-D-TRICKS (time-resolved
imaging of contrast kinetics), developed by
Korosec et al. (1996) may also be used. We used
the test bolus approach because of its simplic-
ity.
Anticipated Results
Pulmonary thrombus or embolus presents
itself as intermediate to high signal on the black
blood sequences. On cine gradient echo se-
quences, thrombus and/or embolus are usually
depicted with low signal intensity compared to
the higher signal of moving spins of the blood
pool. Acute thrombus usually presents as an
intraluminal filling defect, typically more
round in morphology. Chronic embolus or
thrombus, due to its incorporation within the
aortic wall with subsequent endothelialization,
presents as crescentic thickening of the pulmo-
nary arterial wall. Similarly, contrast enhanced
3-D MRA (Figs. A13.1.1 and A13.1.2) will
depict the emboli as low signal intensity filling
defects. In patients with pulmonary hyperten-
Pulmonary
Embolism
A13.1.12
Supplement 2
sion due to chronic thrombo-pulmonary dis-
ease, the central pulmonary arteries may be
enlarged and this may be associated with pe-
ripheral arterial pruning.
Literature Cited
Foo, T.K.F., Manojkumar, S., Prince, M.R., and
Chenvert, T.L. 1997. Automated detection of
bolus arrival and initiation of data acquisition in
fast, three-dimensional gadolinium-enhanced
MR angiography. Radiology 203:275-280.
Hatabu, H., Gefter, W.B., Listerud, J., Hoffman,
E.A., Axel, L., McGowan, III J.C., Palevsku,
H.I., Hayes, C.E., Konishi, J., and Kressel, H.Y.
1992. Pulmonary MR angiography utilizing
phased-array surface coils. J. Comput. Assist.
Tomogr. 16:410-417.
Kovosec, F.R., Frayne, R., Grist, T.M., and Mis-
tretta, C.A. 1996. Time-resolved contrast-en-
hanced 3-D-MR angiography. Magn. Reson.
Med. 36:345-351.
Laub, G.A. and Kaiser, W.A. 1988. MR angiography
with gradient motion refocusing. J. Comput. As-
sist. Tomogr. 12:377-382.
Riederer, S.J., Tasciyan, T., Farzaneh, F., Lee, N.J.,
Wright, R.C., and Herfkin, R.J. 1988. MR fluo-
roscopy: Technical feasibility. Magn. Reson.
Med. 8:1-15.
Shellock, F.G. 1997. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Shetty, A.N., Shirkhoda, A., Bis, K.G., and Alcan-
tara, A. 1995. Contrast-enhanced three dimen-
sional MR angiography in a single breath-hold:
A novel technique. A.J.R. 165:1290-1292.
Simonetti, O.P., Finn, J.P., White, R.D., Bis, K.G.,
Shetty, A.N., Tkach, J., Flamm, S., and Laub, G.
1996. ECG-triggered breath-held gadolinium-
enhanced 3D MRA of the thoracic vasculature.
In Proceedings of the International Society for
Magnetic Resonance in Medicine, Vol 2. pp. 703.
New York.
Westbrook, C. and Kaut, C. 1993. MRI in Practice.
pp. 154-155. Blackwell Scientific Publications,
Oxford.
Wielopolski, P.A. 1993. Pulmonary Arteriography.
In MRI Clinics of North America, Vol 1, No 2.
W.A. Saunders, Philadelphia.
Wielopolski, P.A., Haacke, E.M., and Adler, L.P.
1993. Evaluation of the pulmonary vasculature
with three-dimensional magnetic resonance im-
aging techniques. M.A.G.M.A. 1:21-34.
Contributed by Kostaki G. Bis and
Anil N. Shetty
William Beaumont Hospital
Royal Oak, Michigan
Current Protocols in Magnetic Resonance Imaging
Deep Vein Thrombosis Detection UNIT A13.2
Although magnetic resonance angiography (MRA) is a robust and noninvasive technique
for evaluating the vascular system with respect to deep venous thrombosis (DVT) of the
lower extremities, there are several competing modalities, which primarily involve duplex
sonography and currently, delayed contrast enhanced spiral computed tomography (CT)
following a pulmonary CT angiography study. At the authors’ institution, patients with
suspected pulmonary embolisms and lower extremity DVTs are primarily worked up with
lower extremity duplex sonography and spiral (helical) computed tomography of the
pulmonary vasculature. Lower extremity MR venography (MRV) is performed primarily
in patients where a pelvic source of emboli is suspected and if a duplex sonogram was
not performed. In these patients, the groin is evaluated down to the proximal calf and this
is supplemented with imaging of the pelvis.
In patients where the pulmonary MRA study, when performed, is equivocal or unremark-
able, a lower extremity MRV study is also performed. Multi-station imaging to include
the thorax, pelvis, and lower extremities in a patient with suspected pulmonary embolus
is certainly time consuming since the patient, including body phased array coils, needs
to be repositioned at each respective region of interest. These protocols may be as long
as 2 hr, which is prohibitively long in this patient population. Fortunately, with the
development of stepping MRI tables, the entire multi-station examination to include the
chest, pelvis, and lower extremities can be performed much faster. At the authors’
institution, a manually retrofitted stepping MRI table known as the stepping kinematic
imaging platform (SKIP, Magnetic Moments) is employed. The SKIP device allows the
user to employ the existing body phased array coil for improved spatial resolution and
signal-to-noise as compared to a quadrature body coil.
Venous imaging within the pelvis and lower extremities is primarily performed with
transverse 2-D TOF (time of flight) techniques employing an arterial traveling saturation
band. Although 3-D contrast enhanced MRA techniques utilizing extracellular MRI
contrast agents such as Omniscan, ProHance, or Magnevist can be employed, the venous
signal is, frequently, not as robust as that on 2-D TOF sequences. Lower venous signal
on 3-D contrast enhanced MRA is due to the filtering effect at the capillary bed whereby
50% of the contrast concentration is reduced due to the first pass. In the future, however,
implementation of blood-pool agents, such as MS-325 (EPIX Medical) will likely be the
preferred agent for evaluating deep venous thrombosis. Finally, MRV within the pelvis
can be supplemented with T1- and T2-weighted sequences, when the MRV is abnormal,
for differentiating intrinsic from extrinsic disease and in characterizing extrinsic pathol-
ogy.
IMAGING THE PELVIC AND LOWER EXTREMITY VEINS BASIC
PROTOCOL 1
The sequence described herein is based on the authors’ experience on a Siemens 1.5T
Vision scanner, however, the protocol is likely equally applicable to machines from other
manufacturers. Currently, this technique is the protocol of choice since contrast enhanced
3-D MRA techniques employing a blood-pool agent have not yet been tested and have
not been FDA (Food and Drug Administration)-approved.
In patients where pulmonary MRA is performed, this protocol is performed after the
pulmonary examination has been completed. If the stepping kinematic imaging platform
(SKIP) is utilized, the patient is simply manually repositioned to the pelvic station. If a Pulmonary
stepping table is not available, the patient is taken out of the magnet bore isocenter and Artery,
Mediastinum,
Pleura, and Lung
Contributed by Kostaki G. Bis and Anil N. Shetty A13.2.1
Current Protocols in Magnetic Resonance Imaging (2002) A13.2.1-A13.2.11
Copyright © 2002 by John Wiley & Sons, Inc. Supplement 4
 Table A13.2.1 Hardware Requirements for Deep Vein Thrombosis 5. Inform the patient about what will occur during the procedure, what he or she will
Detection experience while in the magnet, and how to behave, including the following:
Coil type Body phased array coil preferred over a. If earphones or headphones are used to protect the ears from the loud sounds
quadrature body coil produced by the gradients, the patient will be asked to wear these, but will be able
Maximum gradient coil 25 mT/m (or whatever the system permits) to communicate with you at any time during the imaging.
strength
b. The patient will be given a safety squeeze-bulb or similar equipment to request
Cardiac triggering No
assistance at any time (demonstrate how this works).
Peripheral vital sign Optional (blood pressure, heart rate, and
monitoring peripheral O2 saturation) c. For good results, the patient should not talk and should avoid other movement,
Respiratory gating No during each scan—i.e., as long as the banging sounds continue.
Respirator If required by patient d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
Oxygen Usually administered in these patients
(2-3 liters/min by nasal cannula) 6. Establish an i.v. line in the antecubital fossa employing a 20-G angiocatheter needle.
Connect the angiocatheter to the saline at a rate to keep the vein open.
the phased array coil and patient is repositioned by centering the coil at the mid-point Preferably, however, it is connected to a power injector for more reproducible delivery of
between the umbilicus and symphysis pubis. The equipment parameters are listed in Table contrast agent. At the authors’ institution, a MedRad power injector (MedRad) is used.
A13.2.1. 7. Have the patient lie down on the table, with feet first, over the posterior phased array
The entire examination should take ∼40 min. coil.
8. If needed, place a pillow or other support under the knees to make the patient more
NOTE: Be sure that technologists and nurses have immediate access to any emergency
comfortable.
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen. 9. Use the centering light to position the patient (halfway point between umbilicus and
symphysis pubis) and put him or her into the center of the magnet.
Materials
Once this step has been performed, so long as the patient does not move on the table, the
Normal saline (0.9% NaCl), sterile
table itself can be moved and then replaced in the same position as before without
Extracellular contrast agents (e.g., Magnevist, Omniscan, or ProHance) jeopardizing the positioning of one scan relative to another.
Set up patient and equipment 10. If the patient is unable to hold still, provide an appropriate sedative.
1. Interview (screen) the patient to ensure that he or she has no contraindications such
If the patient requires sedation, peripheral monitoring (blood pressure, heart rate, periph-
as cardiac pacemakers or other implants containing ferromagnetic materials. Also, eral O2 saturation, and CO2 capnometry) is required.
be sure to find out if the patient has any health conditions that may require the presence
of special emergency equipment during the scanning procedure, or necessitate any Sequence 1: Rapid three-plane positioning scouts
other precautions. 11. Run the system’s pilot scan to insure the correct location of the patient’s pelvis
Generally, standard screening forms (APPENDIX 1) are used for all patients scanned in a according to the imaging parameters given in Table A13.2.2.
magnetic resonance system. Three orthogonal planes are obtained.
The presence of any ferromagnetic metals may be a health hazard to the patient when he These positioning scouts are performed to validate the patient’s position and to provide
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact anatomical landmarks for subsequent sequences.
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(1996) for a discussion of what implants may be safely scanned using magnetic resonance.
Sequence 2: 2-D time-of-flight
Patients may be accompanied into the magnet room by a friend or family member, who can 12. Prescribe five imaging stacks in an overlapping fashion per region of interest to cover
sit in the room during the scan and comfort the patient as needed. This companion must a distance of 35.0 cm. Run sequence 2 according to Table A13.2.3.
be screened as well to ensure the absence of loose metal objects on the body or clothing.
NOTE: Table A13.2.3 only covers 70 mm.
2. If the procedure is a research protocol, have the patient sign any necessary consent
forms. 13. After obtaining five stacks at the pelvic station, reposition the patient to the mid-thigh
level for imaging the superficial femoral veins down to the popliteal vein. Repeat step
3. Have the patient remove all jewelry and change into a gown to eliminate any metal 12.
that might be found in clothing.
Each imaging stack requires 2 min of acquisition time with ∼1 to 2 min of reconstruction
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating time. Typically, the vasculature is covered down to the popliteal trifurcation, since at most
Pulmonary
Deep Vein and image artifacts. institutions, deep venous thrombosis within the calf is typically not sought for since it is Artery,
Thrombosis not treated. Mediastinum,
Detection Pleura, and Lung

A13.2.2 A13.2.3
Supplement 4 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 4
 Table A13.2.2 Imaging Parameters for Sequence 1 (Localizers) Data processing and viewing for sequence 2
14. Use these five stacks acquired at each region of interest to obtain maximum intensity
Patient position Supine projection images. Use the source data to review for filling defects (see UNIT A13.1,
Scan type 2-D gradient echo Anticipated Results).
Imaging plane (orientation) Transverse, sagittal, and coronal
Central slice or volume center Laser light centered at a point halfway Sequence 3: Bolus test assessment
between umbilicus and symphisis pubis 15. Review the rapid three-plane positioning scouts and position the transverse bolus test
Echo time (TE) 6.0 msec slice at the midpoint (iliac arteries) within the region of interest. Inject 2 ml contrast
Repeat time (TR) 15.0 msec agent at a rate of 1.5 ml/sec followed by a saline flush of 15 to 20 ml at a similar rate
Flip angle (FA) 30° of 1.5 ml/sec.
Fields of view (FOVx, FOVy) 450 mm, 450 mm
Resolution (Δx, Δy) 1.76 mm, 3.52 mm 16. Perform a transverse bolus test assessment study by running sequence 3 according
Number of data points collected (Nx, Ny) 256a, 128 to imaging parameters in Table A13.2.4. Review the image data and note the time
Display matrix (Dx, Dy) 256, 256 that it takes for the contrast agent to arrive at the imaging slice position. This is the
Slice thickness (Δz) 10 mm bolus arrival time that can be used as the scan delay time.
Number of slices 3
Slice gap Not applicable Sequence 4: 3-D contrast-enhanced MRV
Number of acquisitions (Nacq) 1 The pilot scouts (sequence 1) are reviewed and the coronal 3-D contrast enhanced MRV
Swap read and phase encoding No sequence is appropriately placed to evaluate the course of the common and external iliac,
Scan time 10 sec common femoral and superficial femoral and finally, popliteal veins. Without a stepping
aOversampling. MRI table, the three regions of interest (pelvis, thighs, and legs) are evaluated with three
separate injections, however, with an MRI stepping table (see Basic Protocol 2), two
regions of interest (pelvis and thighs), can be evaluated with just one contrast agent
Table A13.2.3 Imaging Parameters for Lower Extremity 2-D TOF MRV
injection utilizing the bolus chase protocol. At this time, however, implementation of 3-D
Patient positioning Supine contrast MRV with extracellular MRV contrast agents is not promoted due to the lower
Scan type 2-D gradient echo vascular-to-background contrast of the venous signal. This protocol, however, can be
Imaging plane (orientation) Transverse employed as a rapid screening procedure for evaluating the presence or absence of gross
Central slice or volume center The cranial initial 2-D slab is abnormalities. In the future, it will likely be replaced employing the same sequence but
positioned at a point just above the with a blood pool agent such as MS-325, whereby arterial pass imaging is followed by
iliac crest to cover the lower inferior
vena cava. The lower stack typically Table A13.2.4 Imaging Parameters for Test Bolus Assessment
covers a point just below the inguinal
region. Patient position Supine
Echo time (TE) 7.0 msec Scan type 2-D gradient echo
Repeat time (TR) 608 msec Imaging plane (orientation) Transverse
Flip angle (FA) 70° Central slice or volume center Slice centered at midpoint within the
Fields of view (FOVx, FOVy) 350 mma, 263 mm first region of interest
Resolution (Δx, Δy) 1.37 mm, 2.31 mm Echo time (TE) 2.4 msec
Number of data points collected (Nx, Ny) 256, 114 Repeat time (TR) 5.8 msec
Display matrix (Dx, Dy) 256, 256 Inversion time (TI) 300 msec
Slice thickness (Δz) 3 mm Flip angle (FA) 8°
Number of slices 32 Fields of view (FOVx, FOVy) 350 mm, 263 mm
Slice gap −0.81 mm (−27%) Resolution (Δx, Δy) 1.37 mm, 2.05 mm
Number of acquisitions (Nacq) 1 Number of data points collected (Nx, Ny) 256, 128
Swap read and phase encoding No Display matrix (Dx, Dy) 256, 256
Saturation pulses Traveling superior saturation pulse for Slice thickness (Δz) 8–10 mm
arterial saturation Number of slices 1
Scan time 1 min, 57 sec Slice gap Not applicable
aOversampling. Number of acquisitions (Nacq) 1
Number of repetitions 60
Swap read and phase encoding No
Pulmonary
Deep Vein Saturation pulses No Artery,
Thrombosis Scan time 60 seca Mediastinum,
Detection Pleura, and Lung
aEach acquisition time is 1 sec, and the total scan time is 60 sec.

A13.2.4 A13.2.5
Supplement 4 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 4
 Table A13.2.5 Imaging Parameters for 3-D Contrast Enhanced MRV of Lower IMAGING THE PELVIC AND LOWER EXTREMITY VEINS WITH A BASIC
Extremities and Pelvis STEPPING TABLE PROTOCOL 2

Patient position Supine If a stepping table is available, sequence 4 can be run differently. During a time difference
Scan type 3-D gradient echo with sync of 5 sec, the table is manually repositioned from the first region of interest (pelvis) to the
interpolation second region of interest (thigh) and the bolus is chased to image the arterial anatomy at
Imaging plane (orientation) Coronal the second region of interest. After this, a 30- to 45-sec delay is employed followed by
Central slice or volume center Coronal slab is centered at the iliac, two additional measurements, one is at the second region of interest and the other is at
femoral and popliteal vessels the first region of interest. The time difference between the last two measurements is also
Echo time (TE) As short as possible (currently, 1.8 5 sec to allow for the table stepping to be repositioned.
msec)
Receiver bandwidth (RBW) 390 Hz Set up patient and equipment
Repeat time (TR) 4.6 msec 1. Repeat Basic Protocol 1, steps 1 to 11 and 15 to 17.
Flip angle (FA) 25°
2. Repeat Basic Protocol 1, step 17, with the number of repetitions set to 4 in Table
Fields of view (FOVx, FOVy) 450 mm, 338 mm
A13.2.5.
Resolution (Δx, Δy) 0.88 mm, 1.72 mm
Number of data points collected (Nx, Ny) 512, 196 3. Reposition the stepping table from the pelvis to the thigh area by moving the table
Display matrix (Dx, Dy) 256, 256 by one field of view.
Slice thickness (Δz) 4.55–5.45 mm
This is done in 5 sec so it is still at the arterial phase.
Number of slices 22 (interpolated to 44)
Slab thickness 100–120 mm 4. Obtain images according to Table A13.2.5.
Slice gap 0 mm
The scan time will be 21 sec in Table A13.2.5.
Number of acquisitions (Nacq) 1
Number of repetitions 2 5. Wait for 30 to 45 sec and repeat step 4.
Swap read and phase encoding No
This is the venous phase at the thigh.
Scan timea 21 sec
aThe time delay between two measurements is 60 sec and the total scan time is 102 sec. 6. Reverse step 3, i.e., reposition the stepping table from the thigh area to the pelvis area
by one field of view.

steady state image acquisition to obtain the venous anatomy. The arterial first pass This is done in 5 sec.
anatomy is then used as a mask for subtracting from the steady state image acquisition 7. Repeat step 4.
data to yield venous-only data. The imaging parameters are similar to those described for
pulmonary MRA (UNIT A13.1) but with slight modifications as listed below in Table A13.2.5.
COMMENTARY
Without a stepping table, each region of interest is imaged with two measurements:the Background Information
first measurement will display the arterial anatomy; the second, the venous anatomy Primary indications for performing venous
during steady state. A time delay between two measurements is ∼60 sec. Subsequently, MR angiography include: (1) detection and
the patient and imaging coils are repositioned to the next region of interest (e.g., thighs) exclusion of deep venous thrombosis, (2)
and the protocol is repeated as in the pelvis. saphenous venous mapping to determine its
suitability in coronary bypass, and (3) monitor-
17. Leaving the patient in the magnet, inject the contrast agent, flush the line with 20 ml ing post-thrombotic changes. Current develop-
saline, wait for the bolus arrival time, and then run sequence 4 according to Table ments in magnetic resonance imaging, espe-
A13.2.5. cially time-of-flight (TOF) techniques and con-
trast enhanced MR venography have been
A dose of 0.1 mmol/kg of contrast agent is usually given. successful in the study of venous pathology
18. Pull the patient out of the magnet, reposition the coil at the thigh, center the patient noninvasively without the need for catheter-
into the magnet, and repeat sequences 1, 3, and 4 (steps 11, 15 to 17). based angiography.
The imaging techniques are broadly classi-
19. Pull the patient out of the magnet, reposition the coil at the legs, center the patient fied into two main categories; (1) bright blood
into the magnet, and repeat sequences 1, 3, and 4 (steps 11, 15 to 17). technique and (2) black blood technique. Bright
blood signal can be either due to flow-related
enhancement such as in TOF techniques or
flow-induced phase shifts such as in phase
Deep Vein contrast techniques. The black blood tech-
Thrombosis nique, on the other hand, is based on nulling
Detection
the flow signal by using gradients to induce
phase shifts. Among the available TOF tech-
niques, a non-ECG triggered sequential 2-D
TOF with a traveling saturation band is well
suited for venous MRA in the pelvis and lower
extremity. A traveling saturation band is placed
superior to the slice so that arterial signal is
saturated. With a sequential mode of data ac-
quisition, the slice acquisition follows the
course of the flow providing uniform signal
across each slice.
In the pelvic region, 2-D TOF techniques
have been very accurate when compared to
conventional venography (Evans et al., 1993;
Carpenter et al., 1993). Furthermore, it is robust
and easy to implement. However, deep small
veins are often missed or not clearly visualized
due to insufficient inflow effects. Under these
circumstances, subtraction techniques have Pulmonary
been useful by eliminating background signal Artery,
to improve small vessel signal (Lebowitz et al., Mediastinum,
Pleura, and Lung

A13.2.6 A13.2.7
Supplement 4 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 4
 A B A B

Figure A13.2.1 Acute nonocclusive DVT. The MIP image (A) from a 2-D TOF MRV study shows a low intensity
defect involving the left common iliac vein. This is a pitfall of extrinsic compression of the vein from the overlying
common iliac artery as demonstrated on the transverse 2-D TOF source data (B) at that location. A nonocclusive
low signal filling defect consistent with acute thrombus is seen in the left proximal superficial femoral vein. This is
best visualized by reviewing the transverse 2-D TOF source data (C) and can be overlooked if only the MIP images
are viewed.

1997). Another disadvantage with the 2-D TOF
is the loss of signal due to in-plane saturation
when vessels course parallel to the imaging
plane. Contrast enhanced angiography, on the
other hand, relies on reducing T1-relaxation,
consequently, increasing the blood signal. The
reduction in T1 that results in signal enhance-
Deep Vein
Thrombosis ment is independent of flow direction and,
Detection hence, is independent of acquisition plane ori-
entation. Recently, blood pool agents have been
used to image the pelvis and lower extremities
(Saeed et al., 1998).
Critical Parameters and
Troubleshooting
Lower extremity MRV imaging is generally
tolerated by patients as it does not require
patients holding their breath during the scan.
Figure A13.2.2 Chronic lower extremity DVT. MIP images from the proximal (A), mid (B), and distal (C) stations of a
2-D TOF MRV study are shown. The stepping kinematic imaging platform (SKIP) was employed for rapid repositioning
of the patient relative to the phased array coils when proceeding from one station to the next. Occlusive disease with
collateral veins are shown at all stations.
Pulmonary
Artery,
Mediastinum,
Pleura, and Lung

A13.2.8 A13.2.9
Supplement 4 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 4
 Short TR gradient echo sequences have been
widely applied to clinical MRV.
A B
Saturation Artifacts
Magnetic resonance angiography without
the use of a contrast agent works on the princi-
ple of signal improvement due to inflow of fresh
unsaturated spins into the imaging volume. The
conventional 2-D TOF approach, without con-
trast and with a traveling saturation band
(placed superiorly), has a problem associated
with motion and in-plane saturation. In areas
where vessels become tortuous and parallel to
the plane of acquisition (such as the common
iliac area), blood spins will be saturated due to
repeated application of RF pulses. The in-plane
saturation is a well known phenomenon and
causes severe loss of signal during imaging
when vessels course parallel to the excitation
plane and is routinely seen in noncontrast en-
hanced methods. Contrast enhanced 3-D MRV
techniques, on the other hand, do improve ves-
sel visualization due to improved relaxivity of
blood spins. Actual application of the sequence
presented in this unit may require some modi-
fications depending on the region of interest
C D covered.
Anticipated Results
Nonocclusive thrombus on 2-D TOF MRV
presents as a low signal intensity-filling defect.
This frequently is detected on source data and
can be frequently missed if only maximum
intensity projections (MIPs) are reviewed. Oc-
clusive thrombus with collaterals, however, is
better appreciated when reviewing the MIPs.
Examples of nonocclusive and occlusive
thromboses are depicted with 2-D TOF MRV
(Figs. A13.2.1 and A13.2.2). Arterial first-pass
3-D contrast enhanced pulmonary and lower
Figure A13.2.3 Pulmonary 3-D contrast enhanced MRA followed by venous steady state imaging. The coronal
contrast enhanced (Omniscan, Amersham Health) 3-D MRA MIP image of the proximal station to include the pulmonary
vessels is shown (A) followed by the pelvic and thigh arterial anatomy (B). A single peripheral i.v. injection of contrast
was employed along with stepping kinematic imaging platform (SKIP) for MRI table stepping. After stepping the table
back to the proximal station, 30 sec later, a steady state image acquisition was made at both imaging stations again.
The arterial phase data from (A) and (B) were used as a mask and subtracted from the steady state data. The MIP
images of the subtracted data yield only venous anatomy at both stations (C, D).
A13.2.10
Supplement 4 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
extremity imaging followed by steady state
imaging to demonstrate the lower extremity
venous anatomy is shown in Fig. A13.2.3.
Blood pool contrast enhanced MRV
The availability of blood pool agents in the
future is likely to expand beyond the current
experimental use of MS-325. Similar protocols
that are employed above with extracellular con-
trast agents can be applied for more robust
imaging of the veins in this patient population.
Literature Cited
Evans, A.J., Sostman, H.D., Knelson, M.H.,
Spritzer, C.E., Newman, G.E., Paine, S.S., and
Beam, C.A. 1993. Detection of deep venous
thrombosis: Prospective comparison of MR im-
aging with contrast venography. A.J.R. Am. J.
Roentgenol. 161:131-135.
Carpenter, J.P., Holland, G.A., Baum, R.A., Owen,
R.S., Carpenter, J.T., and Cope, C. 1993. Mag-
netic resonance venography for the detection of
deep venous thrombosis: Comparison with con-
trast venography and duplex Doppler ultra-
sonography. J. Vasc. Surg. 18:734-739.
Lebowitz, J.A., Rofsky, N.M., Krinsky, G.A., and
Weinreb, J.C. 1997. Gadolinium-enhanced body
MR venography with subtraction technique.
A.J.R. 169:755-758
Saeed, M., Wendland, W.F., Engelbrecht, M.,
Sakuma, H., and Higgins, C.B. 1998. Value of
blood pool contrast agents in magnetic reso-
nance angiography of the pelvis and lower ex-
tremities. Eur. Radiol. 8:1047-1053.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. (F.G. Shellock, ed.).
Lippincott-Raven, Philadelphia.
Contributed by Kostaki G. Bis and Anil N.
Shetty
William Beaumont Hospital
Royal Oak, Michigan
Pulmonary
Artery,
Mediastinum,
Pleura, and Lung
A13.2.11
Supplement 4
Brachial Plexus
Patients are usually referred to MR because of typical brachial plexus symptoms, which
include pain, numbness or weakness of the shoulder or arm. MR is the primary imaging
modality used to diagnose the cause of brachial plexopathy.
The Basic Protocol is the core component of this procedure. When a specific cause of
brachial plexopathy is suspected, modification of the Basic Protocol may be necessary.
For example, if the patient is suspected of having avulsion injury from trauma, use
Alternate Protocol 1. If neoplastic involvement of the brachial plexus is suspected or any
abnormalities are seen, use of intravenous contrast material is helpful (Alternate Protocol
2). When the patient has symptoms of thoracic outlet syndrome and vascular compression
is suspected, contrast enhanced MRA (magnetic resonance angiography) can be added
(Alternate Protocol 3).
The Basic Protocol takes ∼30 min, and Alternate Protocols take ∼45 min.
The parameters suggested here are based on experience with the Siemens 1.5 T Vision or
Symphony and 1.0 T Impact scanner and should be altered accordingly for different field
strengths and manufacturers. The flip angle of all turbo spin echo (fast spin echo)
sequences in the tables represents the refocusing pulse. The first pulse of turbo spin echo
sequences is 90°.
IMAGING OF BRACHIAL PLEXUS
The brachial plexus has a relatively long and oblique course. The brachial plexus is formed
by C5 to T1 nerve roots, and runs along the subclavian and axillary arteries. Images are
obtained in transverse, coronal, and sagittal planes. A T1-weighted sequence is best for
delineating the anatomy with little motion artifact, and T2-weighted images with fat
suppression are helpful in characterizing the lesion. Flow-sensitive gradient echo se-
quences may be used when it is difficult to differentiate vascular structures from the lesion.
Table A14.1.1 lists the hardware necessary to perform the procedure, along with appro-
priate parameters. The available gradient strength will depend on the scanner, and the
echo times given in other tables may need to be varied accordingly (the smaller the
gradient strength, the longer the echo time for a particular scan).
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.
Set up patient and equipment
1. Interview (screen) the patient to ensure that he or she has no contraindications such
as cardiac pacemakers or defibrillator or other implants containing ferromagnetic
materials. Also be sure to find out if the patient has any health conditions that may
Table A14.1.1 Equipment Parameters for Brachial Plexus
Coil type Torso phased-array coil or large flex coil
Gradient coil strength 25 mT/m (or whatever the system permits)
Cardiac gating No
Peripheral gating No
Use of contrast agent Yes, only used in Alternate Protocols 2 and 3
Contributed by Naoki Takahashi and Vamsidhar Narra
Current Protocols in Magnetic Resonance Imaging (2001) A14.1.1-A14.1.17
Copyright © 2001 by John Wiley & Sons, Inc.
UNIT A14.1 require the presence of special emergency equipment during the scanning procedure,
or necessitate any other precautions.
Generally standard screening forms are used for all patients scanned in a magnetic
resonance system.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(1996) for discussion of what implants may be safely scanned using magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
2. If the procedure is a research protocol, have the patient sign any necessary consent
form.
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating
and image artifacts.
5. Inform the patient about what will occur during the procedure, what he or she will
experience while in the magnet, and how to behave, including the following:
a. If earphones or headphones are used to protect the ears from the loud sounds
produced by the gradients, the patient will be asked to wear these, but will be able
BASIC to communicate with you at any time during the imaging.
PROTOCOL b. The patient will be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this works).
c. For good results the patient should not talk, and should avoid or minimize
swallowing or other movement, during each scan—i.e., as long as the banging
sounds continue. Between scans, talking and swallowing are allowed in most
cases, but should be avoided when comparative positional studies are being
performed; the patient will be informed when this is the case.
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
All the sequences are nonbreath-hold except for MRA sequence in Alternate Protocol 3 for
suspected thoracic outlet syndrome.
6. Help the patient mount onto the table. Either before or right after the patient lies down,
set up other monitoring equipment that is to be used.
7. Place the phased array coil high on the chest and neck. Place a towel as a block around
the neck to support the coil.
A large flex coil can also be used, but it is more difficult to position properly and has a
smaller FOV (field of view).
8. If needed, place a pillow or other support under the knees to make the patient more
comfortable.
9. Use the centering light to position the patient with the center at the patient’s sternum
and put him or her into the center of the magnet.
Sequence 1: Rapid three-plane positioning scout
10. For localization of subsequent acquisitions, run the system’s three-plane scout scan
according to Table A14.1.2.
Chest Wall Brachial Plexus
A14.1.1 A14.1.2
Current Protocols in Magnetic Resonance Imaging
 Table A14.1.2 Primary Clinical Imaging Parameters for Pilot Scan (Sequence 1) Table A14.1.3 Primary Clinical Imaging Parameters for 2-D Coronal
T1-Weighted Turbo Spin Echo (Sequence 2)
Patient position Supine
Scan type Gradient echo Patient position Supine
Imaging plane (orientation) Three-plane: transverse, coronal, Scan type Fast spin echo
and sagittal Imaging plane (orientation) Coronal
Central slice or volume center Upper chest Central slice or volume center Midline
Echo time (TE) 3.1 msec Echo time (TE) 12 msec
Repeat time (TR) 6.5 msec Echo train length (ETL) 3
Flip angle (FA) 80° Repeat time (TR) 875 msec
Fields of view (FOVx, FOVy) 450 mm, 450 mm Flip angle (FA) 160°a
Resolution (Δx, Δy) 1.76 mm, 2.34 mm Fields of view (FOVx, FOVy) 375 mm, 375 mm
Number of data points collected (Nx, Ny) 256, 192 Resolution (Δx, Δy) 0.73 mm, 1.46 mm
Display matrix (Dx, Dy) 256, 256 Number of data points collected (Nx, Ny) 512, 256b
Slice thickness (Δz) 5 mm Display matrix (Dx, Dy) 512, 512
Number of slices 7 Slice thickness (Δz) 5 mm
Slice gap Variable Number of slices 15
Number of acquisitions (Nacq) 1 Slice gap 1 mm
Swap read and phase encoding No Number of acquisitions (Nacq) 3
Slice locations Variable Swap read and phase encoding No
Saturation pulses No Saturation pulses Inferior and superior
Scan time 8 sec Scan time 4 min, 54 sec
aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this

sequence is 90°.
Sequence 2: 2-D Coronal T1-weighted turbo spin echo scan of bilateral brachial b30% phase oversampling.
plexus
Use the images from this coronal T1-weighted turbo spin echo sequence for an overview Table A14.1.4 Primary Clinical Imaging Parameters for 2-D Transverse
of the bilateral brachial plexus. This is the only sequence obtained for the contralateral T1-Weighted Spin Echo (Sequence 3)
side. The image of the contralateral brachial plexus would serve as a normal control to
check for symmetry. Patient position Supine
Scan type Spin echo
11. Position a series of coronal slices off the scout images to ensure coverage of the
Imaging plane (orientation) Transverse
bilateral brachial plexus by including both acromioclavicular joints from side to side,
Central slice or volume center Symptomatic brachial plexus
and mid C4 to below axilla from top to bottom. From anterior to posterior, start from
Echo time (TE) 14 msec
slightly anterior to the subclavian and axillary artery to slightly posterior to the neural
Repeat time (TR) 500 msec
foramina of C4 to T1.
Flip angle (FA) 90°
12. Run sequence 2 according to Table A14.1.3. Fields of view (FOVx, FOVy) 200 mm, 200 mm
Resolution (Δx, Δy) 0.78 mm, 1.04 mm
Sequence 3: 2-D Transverse T1-weighted spin echo sequence of symptomatic
Number of data points collected (Nx, Ny) 256, 192a
brachial plexus
Display matrix (Dx, Dy) 256, 256
13. Position a series of transverse slices off the coronal T1-weighted images to ensure
Slice thickness (Δz) 5 mm
coverage of the symptomatic brachial plexus by including mid vertebral body to the
Number of slices 20
acromioclavicular joint from side to side, and mid C4 to axilla from top to bottom.
Slice gap 1 mm
14. Run sequence 3 according to Table A14.1.4. Number of acquisitions (Nacq) 2
Swap read and phase encoding No
Sequence 4: 2-D Sagittal T1-weighted spin echo sequence of symptomatic brachial
Saturation pulses No
plexus
Scan time 3 min, 53 sec
15. Position a series of sagittal slices off the coronal T1-weighted images from the anterior
a20% phase oversampling.
scalene muscle to slightly lateral to the acromioclavicular joint.
16. Run sequence 4 according to Table A14.1.5.

Chest Wall Brachial Plexus

A14.1.3 A14.1.4
Current Protocols in Magnetic Resonance Imaging Supplement 11 Supplement 11 Current Protocols in Magnetic Resonance Imaging
 Table A14.1.5 Primary Clinical Imaging Parameters for 2-D Sagittal Table A14.1.7 Primary Clinical Imaging Parameters for 2-D Sagittal Inversion
T1-Weighted Spin Echo (Sequence 4) Recovery Turbo Spin Echo (Sequence 6)

Patient position Supine Patient position Supine

Scan type Spin echo Scan type Inversion recovery turbo spin echo
Imaging plane (orientation) Sagittal Imaging plane (orientation) Sagittal
Central slice or volume center Symptomatic brachial plexus Central slice or volume center Symptomatic brachial plexus
Echo time (TE) 14 msec Echo time (TE) 30 msec
Repeat time (TR) 600 msec Echo train length (ETL) 7
Flip angle (FA) 90° Repeat time (TR) 5400 msec
Fields of view (FOVx, FOVy) 200 mm, 200 mm Inversion time (TI) 155 msec
Resolution (Δx, Δy) 0.78 mm, 1.04 mm Flip angle (FA) 160°
Number of data points collected (Nx, Ny) 256, 192a Fields of view (FOVx, FOVy) 200 mm, 200 mm
Display matrix (Dx, Dy) 256, 256 Resolution (Δx, Δy) 0.78 mm, 1.04 mm
Slice thickness (Δz) 5 mm Number of data points collected (Nx, Ny) 256, 192a
Number of slices 20 Display matrix (Dx, Dy) 256, 256
Slice gap 1 mm Slice thickness (Δz) 5 mm
Number of acquisitions (Nacq) 2 Number of slices 20
Swap read and phase encoding No Slice gap 1 mm
Saturation pulses Inferior over the heart Number of acquisitions (Nacq) 1
Scan time 4 min, 39 sec Swap read and phase encoding No
a20% phase oversampling. Saturation pulses Inferior over the heart
Scan time 2 min, 59 sec
a20% phase oversampling.
Table A14.1.6 Primary Clinical Imaging Parameters for 2-D Sagittal
T2-Weighted Turbo Spin Echo with Fat Suppression (Sequence 5)
Table A14.1.8 Primary Clinical Imaging Parameters for 2-D Sagittal Flow
Patient position Supine Sensitive Gradient Echo (Sequence 7)
Scan type Fast spin echo
Imaging plane (orientation) Sagittal Patient position Supine
Central slice or volume center Symptomatic brachial plexus Scan type 2-D gradient echo
Echo time (TE) 99 msec Imaging plane (orientation) Sagittal
Echo train length (ETL) 11 Central slice or volume center Symptomatic brachial plexus
Repeat time (TR) 3980 msec Echo time (TE) 9 msec
Flip angle (FA) 160°a Repeat time (TR) 30 msec
Fields of view (FOVx, FOVy) 200 mm, 200 mm Flip angle (FA) 30°
Resolution (Δx, Δy) 0.78 mm, 1.04 mm Fields of view (FOVx, FOVy) 200 mm, 200 mm
Number of data points collected (Nx, Ny) 256, 192b Resolution (Δx, Δy) 0.78 mm, 1.04 mm
Display matrix (Dx, Dy) 256, 256 Number of data points collected (Nx, Ny) 256, 192a
Slice thickness (Δz) 5 mm Display matrix (Dx, Dy) 256, 256
Number of slices 20 Slice thickness (Δz) 5 mm
Slice gap 1 mm Number of slices 20
Number of acquisitions (Nacq) 3 Slice gap 1 mm
Swap read and phase encoding No Number of excitations (NEX) 20b
Saturation pulses Fat suppression Number of acquisitions (Nacq) 1
Scan time 4 min, 16 sec Swap read and phase encoding No
aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this Flow compensation Yes
sequence is 90°. Saturation pulses No
b20% phase oversampling.
Scan time 2 min, 20 sec
a20% phase oversampling.
bThe number of concatenation is set to be 20. This means that only one slice will be excited during a

given repeat time.

Chest Wall Brachial Plexus

A14.1.5 A14.1.6
Current Protocols in Magnetic Resonance Imaging Supplement 11 Supplement 11 Current Protocols in Magnetic Resonance Imaging
Sequence 5: 2-D Sagittal T2-weighted turbo spin echo sequence with fat suppression Table A14.1.10 Primary Clinical Imaging Parameters for 2-D Transverse
of symptomatic brachial plexus T2-Weighted Turbo Spin Echo with Fat Suppression (Sequence 9)
17. Positioning is the same as sequence 4. Run sequence 5 according to Table A14.1.6.
Patient position Supine
The T2-weighted sequence is useful in characterizing the lesions in the brachial plexus. Scan type Turbo spin echo
Sequence 6: 2-D Sagittal inversion recovery turbo spin echo sequence of Imaging plane (orientation) Transverse
symptomatic brachial plexus (optional) Central slice or volume center Symptomatic brachial plexus
If the fat suppression is not adequately obtained with sequence 5, inversion recovery turbo Echo time (TE) 99 msec
spin echo sequence is an alternative sequence. Echo train length (ETL) 11
Repeat time (TR) 4200 msec
18. Positioning is same as sequence 4. Run sequence 6 according to Table A14.1.7. Flip angle (FA) 160°
Sequence 7: 2-D Sagittal flow sensitive gradient echo sequence of symptomatic Fields of view (FOVx, FOVy) 200 mm, 200 mm
brachial plexus (optional) Resolution (Δx, Δy) 0.78 mm, 1.04 mm
This sequence is an optional sequence, and can be performed to help differentiate vascular Number of data points collected (Nx, Ny) 256, 192a
structures from lesions. Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 5 mm
19. Positioning is the same as sequence 4. Run sequence 7 according to Table A14.1.8. Number of slices 20
Slice gap 1 mm
IMAGING OF BRACHIAL PLEXUS FOR AVULSION INJURY ALTERNATE Number of acquisitions (Nacq) 3
For patients with suspected avulsion injury of the nerve root, the Basic Protocol should PROTOCOL 1 Swap read and phase encoding No
be modified and a coronal T2-weighted sequence of bilateral brachial plexus as well as a Saturation pulses Fat suppression
transverse T2-weighted sequence of the symptomatic side are added. The sagittal T2- Scan time 4 min, 41 sec
weighted turbo spin-echo sequences in the Basic Protocol can be omitted. a20% phase oversampling.

Set up patient and equipment

1. Use the same equipment and the same patient setup as for the previous method (see Sequence 8: 2-D Coronal T2-weighted turbo spin echo sequence of bilateral brachial
Basic Protocol, steps 1 to 9). plexus
3. Position a series of coronal slices off the scout images to ensure coverage of the
2. Run the same sequences 1 to 4 in Basic Protocol (see Basic Protocol, steps 10 to 16). bilateral brachial plexus by including both acromioclavicular joints from side to side,
and mid C4 to below axilla from top to bottom. From anterior to posterior, start from
Table A14.1.9 Primary Clinical Imaging Parameters for 2-D Coronal slightly anterior to the subclavian and axillary artery to slightly posterior to the neural
T2-Weighted Turbo Spin Echo (Sequence 8)
foramina of C4 to T1.
Patient position Supine This is a high-resolution sequence and fat suppression is not used.
Scan type Fast spin echo
Imaging plane (orientation) Coronal 4. Run sequence 8 according to Table A14.1.9.
Central slice or volume center Midline
Sequence 9: 2-D Transverse T2-weighted turbo spin echo sequence of symptomatic
Echo time (TE) 99 msec
Echo train length (ETL) 13
brachial plexus (with fat suppression)
Repeat time (TR) 4200 msec
5. Position a series of transverse slices off the coronal images to ensure coverage of the
Flip angle (FA) 160°a
symptomatic brachial plexus by including mid vertebral body to the acromioclavicu-
Fields of view (FOVx, FOVy) 375 mm, 375 mm
lar joint from side to side, and mid C4 to axilla from top to bottom.
Resolution (Δx, Δy) 0.73 mm, 1.39 mm 6. Run sequence 9 according to Table A14.1.10.
Number of data points collected (Nx, Ny) 512, 270b
Display matrix (Dx, Dy) 512, 512
ALTERNATE IMAGING OF BRACHIAL PLEXUS FOR NEOPLASTIC INVOLVEMENT
Slice thickness (Δz) 5 mm
PROTOCOL 2
Number of slices 15 If neoplastic involvement of the brachial plexus is suspected, the Basic Protocol should
Slice gap 1 mm be modified. Pre- and post-contrast T1-weighted spin-echo with fat suppression should
Number of acquisitions (Nacq) 2 be performed. The post-contrast sequence increases the sensitivity to detect abnormalities
Swap read and phase encoding No and help characterize the lesions. The image plane should be chosen which best depicts
Saturation pulses Inferior and superior the abnormality.
Scan time 3 min, 59 sec
aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this

sequence is 90°.
b30% phase oversampling. Chest Wall Brachial Plexus

A14.1.7 A14.1.8
Current Protocols in Magnetic Resonance Imaging Supplement 11 Supplement 11 Current Protocols in Magnetic Resonance Imaging
Materials Sequence 11: 2-D T1-weighted spin echo sequence with fat suppression of
Normal saline (0.9% NaCl), sterile symptomatic brachial plexus after contrast agent injection
Gadolinium-based MR contrast agent (e.g., Magnevist, Omniscan, or Prohance) 6. Leaving the patient in the magnet, administer 0.1 mmol/kg gadolinium contrast
22-G intravenous catheter intravenously. Flush the catheter with 5 ml normal saline. Run sequence 11 according
Power injector (optional) to Table A14.1.11.
The image plane and position should be same as sequence 10.
Set up patient and equipment
1. Use the same equipment and the same patient setup as for the previous method (see
Basic Protocol, steps 1 to 8). ALTERNATE CONTRAST-ENHANCED MRA FOR THORACIC OUTLET SYNDROME
PROTOCOL 3
2. Place a 22-G intravenous catheter in the antecubital fossa. Establish an intravenous When thoracic outlet syndrome with vascular compromise is suspected, contrast-en-
line from which the contrast agent can be injected, and attach this line securely to the hanced MRA can be added to the Basic Protocol. Two series of MRA images are obtained
patient so that movement into or out of the magnet will not pull at the patient’s arm. with the patient’s arm in neutral position and above the shoulder. Double-dose gadolinium
contrast (e.g., 40 ml) is divided into half for each series. After the test bolus, one measure
It is preferable to insert the line prior to imaging and to leave the patient in the magnet, for a pre-contrast mask and two measures for post-contrast MRA are obtained for each
with no intervening motion between the scans run before contrast agent injection and those series.
run after injection.
Materials
3. Use the centering light to position the patient with the center at the patient’s sternum
and put him or her into the center of the magnet. Normal saline (0.9% NaCl), sterile
Gadolinium-based MR contrast agent (e.g., Magnevist, Omniscan, or Prohance)
4. Run the same sequences 1 to 7 in Basic Protocol (see Basic Protocol, steps 10 to 19). Power injector (for the MRA technique)
Sequence 10: 2-D T1-weighted spin echo sequence with fat suppression of Set up patient and equipment
symptomatic brachial plexus before contrast agent injection 1. Use the same equipment and the same patient setup as for the previous method (see
5. Choose the image plane which best depicts the lesion and run sequence 10 according Basic Protocol, steps 1 to 8).
to Table A14.1.11. If no lesions are seen, choose transverse plane.
2. Place a 22-G intravenous catheter in the antecubital fossa of the asymptomatic arm.
Establish an intravenous line from which the contrast agent can be injected, and attach
this line securely to the patient so that movement into or out of the magnet will not
Table A14.1.11 Primary Clinical Imaging Parameters for 2-D T1-Weighted pull at the patient’s arm.
Spin Echo with Fat Suppression (Sequences 10 and 11)
Load the injection with a double dose (0.2 mmol/kg) of gadolinium contrast agent (this is
Patient position Supine usually 40 ml or less).
Scan type Spin echo It is preferable to insert the line prior to imaging and to leave the patient in the magnet,
Imaging plane (orientation) The one that can best depict the with no intervening motion between the scans run before contrast agent injection and those
lesion run after injection.
Central slice or volume center Symptomatic brachial plexus When the contrast material is injected from the symptomatic side, the vessels may appear
Echo time (TE) 12 msec artificially narrowed and simulate disease on the MRA sequence because of the high-con-
Repeat time (TR) 410 mseca centration of contrast material and the accompanying T2 shortening effect. To avoid this
Flip angle (FA) 90° artifact, the catheter should be placed in the asymptomatic arm.
Fields of view (FOVx, FOVy) 200 mm, 200 mm 3. Use the centering light to position the patient with the center at the patient’s sternum
Resolution (Δx, Δy) 0.78 mm, 1.04 mm and put him or her into the center of the magnet.
Number of data points collected (Nx, Ny) 256, 192b
Display matrix (Dx, Dy) 256, 256
4. Use the same sequences 1 to 7 as in the Basic Protocol (see Basic Protocol, steps 10
Slice thickness (Δz) 5 mm
to 19).
Number of slices 20
Sequence 12: Test bolus
Slice gap 2 mm
5. Position a single transverse slice at the level of the aortic arch on screen.
Number of acquisitions (Nacq) 3
Swap read and phase encoding No 6. Set up the injector for injection of 2 ml of gadolinium followed by 15 ml normal
Saturation pulses Fat suppression saline at the rate of 2 ml/sec.
Scan time 6 min, 35 sec 7. As you start the injector, start sequence 12 according to Table A14.1.12.
aT of 600 msec or less is preferred, if possible. This can be achieved in systems which can concatenate
R The sequence will acquire a single slice every second for 50 sec at the same location.
the acquisition. If the system does not allow concatenation of the sequence, two sets of 10 slices can
be obtained separately. 8. By scrolling the images, determine the time to the peak of contrast bolus arrival from
b20% phase oversampling and a phase partial Fourier factor of 11/16.
the start of contrast agent injection.
Chest Wall Brachial Plexus

A14.1.9 A14.1.10
Current Protocols in Magnetic Resonance Imaging Supplement 11 Supplement 11 Current Protocols in Magnetic Resonance Imaging
Sequence 13: 3-D Pre-contrast MRA with the arm in neutral position (mask)
The pre-contrast sequence is obtained as a mask for subtraction and to ensure the coverage
and image quality.
9. Position the slab in coronal plane to cover the aortic arch and bilateral subclavian
arteries. Run sequence 13 according to Table A14.1.13 under breath-hold in inspira-
tion.
10. Check the images to ensure the coverage of the bilateral subclavian arteries and aortic
arch.
Minimal wrapping artifact does not interfere with the diagnostic ability of the study. The
degree of respiration artifact should be checked. If the patient cannot hold their breath long
enough, consideration should be given to shorten the scan time at the cost of spatial
resolution.

Sequence 14: 3-D Contrast-enhanced MRA with the arm in neutral position
Table A14.1.12 Primary Clinical Imaging Parameters for Test Bolus (Sequence 12) 11. Set up the injector for injection of half of the gadolinium (∼19 ml) followed by 15
ml normal saline at the rate of 2 ml/sec.
Patient position Supine
Scan type Gradient echo At this point, 21 ml of gadolinium is used (2 ml of gadolinium was used in the test bolus,
sequence 12). The other half of the gadolinium will be used in sequence 16, step 17.
Imaging plane (orientation) Transverse
Central slice or volume center Aortic arch 12. Scan delay from the start of injection can be calculated based on time to peak of
Echo time (TE) 2.4 msec contrast agent arrival determined by test bolus injection, injection duration, and scan
Repeat time (TR) 5.8 msec duration using following formula:
Inversion time (TI) 300 msec
Flip angle (FA) 15° scan delay = time to peak + (injection duration)/2 − (scan duration)/2.
Fields of view (FOVx, FOVy) 400 mm, 300 mm The scan time is typically 25 sec and injection duration is 9.5 sec (19 ml at 2 ml/sec). If the
Resolution (Δx, Δy) 1.56 mm, 2.34 mm time to peak at test injection is 15 sec, the scan delay would be 7 sec (= 15 + 9.5/2 − 25/2).
Number of data points collected (Nx, Ny) 256, 128
13. Position the slab exactly the same as in sequence 13.
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 10 mm The position should be copied from “history” so that subtraction can be performed later.
Number of slices 1 Two measures are performed in this sequence, and ∼8-sec interval should be allowed for
Slice gap Not applicable the patient to breathe between the first and second measures.
Number of acquisitions (Nacq) 50 14. Start the injector, and after the scan delay as calculated in step 12, run sequence 14
Swap read and phase encoding No according to Table A14.1.14 using a breath-hold in inspiration.
Saturation pulses No
Scan time 50 sec Sequence 15: 3-D Pre-contrast MRA with the arm above the shoulder (mask)
Even though the contrast agent has been injected during the previous scan, the word
Table A14.1.13 Primary Clinical Imaging Parameters for 3-D MRA, Mask “pre-contrast” is still used here to distinguish this sequence from sequence 16.
(Sequence 13)
15. Change the patient’s arm position to above the shoulder.
Patient position Supine
16. Repeat steps 9 and 10 in this alternate protocol.
Scan type 3-D gradient echo
Imaging plane (orientation) Coronal
Sequence 16: 3-D Contrast-enhanced MRA with the arm above the shoulder
Central slice or volume center Midline
17. Set up the injector for the injection of the remaining gadolinium (∼19 ml), followed
Echo time (TE) 1.8 msec
by 15 ml normal saline at the rate of 2 ml/sec.
Repeat time (TR) 4.6 msec
Flip angle (FA) 25° 18. Repeat steps 11 to 14 in this alternate protocol.
Fields of view (FOVx, FOVy) 375 mm, 375 mm
Sequences 15 and 16 are exactly the same as sequences 13 and 14 except for the arm
Resolution (Δx, Δy) 1.46 mm, 2.21 mm
position.
Number of data points collected (Nx, Ny) 256, 170
Display matrix (Dx, Dy) 256, 256 Data processing and viewing for sequences 13 to 16
Slice thickness (Δz) 3 mm (1.5 mm after interpolation) 19. Create subtraction data set from the pre- and post-contrast MRA sequences for each
Number of slices 34 (68 after interpolation) arm position. Subtraction is essential for the second run (sequences 14 and 16) to
Slice gap 0 mm eliminate venous contamination from the first run. Display these data sets in maximal
Number of acquisitions (Nacq) 1 intensity projection (MIP) and multi-planar reconstruction (MPR).
Swap read and phase encoding No
ZIP 2 Yes
Saturation pulses No
Scan time 21 sec
Chest Wall Brachial Plexus

A14.1.11 A14.1.12
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A14.1.14 Primary Clinical Imaging Parameters for 3-D Contrast
Enhanced MRA (Sequence 14)
A
Patient position Supine
Scan type 3-D gradient echo
Imaging plane (orientation) Coronal
Central slice or volume center Midline
Echo time (TE) 1.8 msec
Repeat time (TR) 4.6 msec
Flip angle (FA) 25°
Fields of view (FOVx, FOVy) 375 mm, 375 mm
Resolution (Δx, Δy) 1.46 mm, 2.21 mm
Number of data points collected (Nx, Ny) 256, 170
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 3 mm (1.5 mm after interpolation)
Number of slices 34 (68 after interpolation)
Slice gap 0 mm
Number of acquisitions (Nacq) 1
Number of repetitions 2 (with 8-sec time interval in
between)
Swap read and phase encoding No
ZIP 2 Yes
Saturation pulses No
Scan time 21 sec (50 sec with two B
repetitions)

COMMENTARY
Background Information
The brachial plexus provides motor and sen-
sory innervation to the upper extremity. A va-
riety of imaging techniques including com-
puted tomography (CT) and myelography have
been used in the past for assessment of the
brachial plexus. Currently, MR is considered
the modality of choice because of its multi-pla-
nar capability and good soft tissue contrast
(Posniak et al., 1993).
The brachial plexus is formed by the C5 to
T1 nerve roots, with inconstant contributions
from C4 and T2 nerve roots. The brachial
plexus courses anterolaterally to pass between
the anterior and middle scalene muscles, where
the three trunks are formed. Each trunk is di-
vided into two to four divisions, which run
along the superior aspect of the subclavian
artery in the supraclavicular fossa. Three cords
are formed from divisions, which run around
the axillary artery in the axilla.
Symptoms of brachial plexopathy include
numbness, pain or weakness involving multiple
root levels and multiple peripheral nerves. Bra-
chial plexopathy may be caused by various
pathologic entities including trauma, primary
neurogenic tumors or secondary neoplasms,
radiation injury, and neuritis.
Critical Parameters and
Troubleshooting
Fat suppression is used with the T2-weighted
turbo spin echo sequence. Fat suppression may
not be effective in some patients partly due to
field inhomogeneity near the surface of the
body. In those instances, use an inversion re-
covery sequence (sequence 6) to achieve fat
suppression.
Some authors advocate the use of ECG gat-
ing or respiratory compensation to eliminate
cardiac pulsatility or respiratory motion arti-
fact. The protocols described herein do not call
for the use of these techniques. If necessary,
Figure A14.1.1 Normal brachial plexus. (A) Coronal and (B) sagittal T1-weighted images dem-
ECG gating or respiratory compensation can
onstrate normal brachial plexus (arrows), which is a thin linear structure surrounded by fat and
be used. running along the subclavian artery (arrow heads).
The only breath-hold sequence is the con-
trast-enhanced MRA sequence for suspected
thoracic outlet syndrome with vascular com-
pression. If the patient cannot hold his or her
breath for the period of the sequence (typically
22 to 25 sec), decrease the number of phase
encoding steps (Ny, or number of acquisitions)
to reduce the scan time. Chest Wall Brachial Plexus

A14.1.13 A14.1.14
Current Protocols in Magnetic Resonance Imaging Supplement 2 Supplement 2 Current Protocols in Magnetic Resonance Imaging
 Anticipated Results
On T1-weighted images, the normal brachial
A plexus appears as low signal fine linear struc-
ture that runs along the subclavian and axillary
artery (Fig. A14.1.1A, A14.1.1B). T1-weighted
images are best for delineation of anatomy
because of good soft tissue contrast between fat
and brachial plexus. T2-weighted images and
post-contrast T1-weighted images are good at
characterizing lesions.
Brachial plexopathy may result from pene-
trating or nonpenetrating trauma. Nonpenetrat-
ing injuries include avulsion injury, compres-
sion by adjacent fractures, or hematomas. On
MR, avulsion injury is recognized by a thick-
ened, folded appearance of the brachial plexus
and pseudomeningocele formation.
Primary neurogenic neoplasms are usually
intermediate in signal intensity on the T1-
weighted image and bright in signal intensity

C B

Figure A14.1.2 Brachial plexus involved by metastatic breast carcinoma. (A) Sagittal T1-weighted
image shows an isointense soft tissue mass (arrow) encasing the brachial plexus and subclavian
artery (arrow head). (B) On inversion recovery image, the mass is hyper-intense (arrow). The
subclavian artery (arrow head) is encased by the mass. (C) After administration of gadolinium
contrast, the mass shows enhancement (T1-weighted image with fat saturation) (arrow). Note the Figure A14.1.3 Radiation fibrosis involving brachial plexus. (A) Coronal and (B) sagittal T1-
encasement of the subclavian artery (arrow head) by the mass. weighted images demonstrate soft tissue strands and loss of fat plane surrounding the right brachial
Chest Wall Brachial Plexus plexus (arrows). Compare with the normal brachial plexus on the left in the coronal image.

A14.1.15 A14.1.16
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
on the T2-weighted image, and fusiform to
round in shape. After gadolinium administra-
tion, such neoplasms are typically enhanced.
Involvement of the brachial plexus by sec-
ondary tumor is much more common than pri-
mary neurogenic tumors. The most common
secondary tumors to involve brachial plexus are
lung and breast carcinomas (Wittenberg and
Adkins, 2000). The brachial plexus may be
involved by tumors by direct extension as in
Pancoast tumor (lung carcinoma) or indirectly
from lymph node metastasis (Fig. A14.1.2).
Radiation brachial plexopathy can occur 18
months to 10 years after the radiation therapy.
It should be differentiated from recurrence of
the tumor. Evidence of edema (high signal on
the T2-weighted image) or fibrosis (low signal
on the T2-weighted image) depending on the
phase of radiation injury and lack of mass lesion
are the hallmark of MR findings (Fig. A14.1.3)
(Glazer et al., 1985; Bowen et al., 1996; Thya-
garajan et al., 1995).
Brachial plexus neuritis can be caused by
various different etiologies or can be idiopathic.
On MR, the brachial plexus is thickened and
high in signal intensity on T2-weighted images.
Thoracic outlet syndrome occurs because of
compression of brachial plexus or subclavian
artery in the scalene triangle or costoclavicular
space. Underlying skeletal abnormalities such
as a cervical rib or a fracture deformity of
clavicle or first rib can be responsible for the
compression. Other causes include fibromus-
cular band or spasm of scalene muscle. The
compression is often positional and subtle,
these abnormalities may be difficult to identify.
When vascular compression is present, a con-
trast enhanced MRA sequence may show
change in the caliber of the subclavian artery
in the position that produces the symptom.
Literature Cited
Bowen, B.C., Verma, A., Brandon, A.H., and
Fiedler, J.A. 1996. Radiation induced brachial
Current Protocols in Magnetic Resonance Imaging
plexopathy: MR imaging with clinical correla-
tion. Am. J. Neuroradiol. 17:1932-1936.
Glazer, H.S., Lee, J.K.T., Levitt, R.G., Heiken, J.P.,
Ling, D., Totty, W.G., Balfe, D.M., Emani, B.,
Wasserman, T.H., and Murphy, W.A. 1985. Ra-
diation fibrosis: Differentiation from recurrent
tumor by MR imaging—work in progress. Radi-
ology 156:721-726.
Posniak, H.V., Olson, M.C., Dudiak C.M., Wis-
niewski, R., and O’Malley, C. 1993. MR imaging
of the brachial plexus. Am. J. R Roentgenol.
161:373-379.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Thyagarajan, D. Cascino, T., and Harms, G. 1995.
Magnetic resonance imaging in brachial
plexopathy of cancer. Neurology 45:421-427.
Wittenberg, K.H. and Adkins, M.C. 2000. MR im-
aging of nontraumatic brachial plexopathies: fre-
quency and spectrum of findings. Radiographics
20:1023-1032.
Key References
Posniak et al., 1993. See above.
Covers overview of MR appearances of various
diseases involving brachial plexus.
Shellock, 1996. See above.
Covers a number of important patient management
issues related to MR imaging, including recom-
mended safety procedures, a list of metallic implants
that have been tested for MR compatibility, and a
list of other sources on MR safety.
Wittenberg and Adkins, 2000. See above.
Covers overview of MR appearances of various
diseases involving brachial plexus.
Contributed by Naoki Takahashi and
Vamsidhar Narra
Mallinckrodt Institute of Radiology
Washington University Medical Center
St. Louis, Missouri
Chest Wall
A14.1.17
MRI of the Liver UNIT A15.1
MRI provides comprehensive information on the full range of liver diseases, including
congenital abnormalities, benign and malignant focal liver lesions, and diffuse liver
disease. We employ a protocol incorporating various types of T1-and T2-weighted se-
quences, including transverse and coronal data acquisition, and the routine use of
intravenous gadolinium.
LIVER IMAGING BASIC
PROTOCOL
The high signal-to-noise ratio (SNR) obtained at high field strength makes it possible to
image the entire liver during a single breath-hold sequence. The sequences described
herein are based on the authors’ experience with a Siemens 1.5 T Vision scanner, but are
expected to be equally applicable to machines from other manufacturers.
Scanning a patient or volunteer is a joint effort among technologists, nurses, and
physicians, with the technologist normally responsible for following proper scanning
protocols and techniques. Unless otherwise specified, in what follows the person to whom
directions are given is assumed to be the technologist. Table A15.1.1 lists the hardware
necessary to perform the procedure, along with appropriate parameters.
The following ten sequences comprise the liver imaging protocol. This protocol employs
multiple data acquisitions and early and late post-gadolinium imaging. However, most
sequences are acquired in a single breath hold and the entire exam time is less than 15
min. Most sequences require the patient to be able to suspend respiration for ∼25 sec. It
is imperative that there be clear communication between the technologist and the patient
throughout the exam. This protocol results in consistent, reproducible image quality that
is effective for evaluating the full spectrum of liver diseases.
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.
Materials
Normal saline (0.9% NaCl), sterile, 40 ml minimum
Extravascular contrast agent (e.g., Magnevist, Omniscan, or Prohance), volume by
patient weight
Table A15.1.1 Equipment Specifications for Imaging of the Liver
Coil type Circular polarized body phased array coil
Field strength 1.5 T
Gradient strength 24 mT/m (or whatever the system
permits, but minimum of 24 mT/m for
sequences 2 and 4)
Knee cushion Yes
Use of contrast agents Yes
Pulse oximeter If patient requires sedation
Power injector Yes
Normal saline Yes
35-in. extension tubing Yes
Liver
Contributed by Laurie Fisher, Richard Semelka, and Kathy Wilber A15.1.1
Current Protocols in Magnetic Resonance Imaging (2001) A15.1.1-A15.1.14
Copyright © 2001 by John Wiley & Sons, Inc.
 Set up equipment and patient
1. Interview (screen) the patient to assess for contraindications such as cardiac pace-
maker, implanted mechanical devices, and/or ferromagnetic materials. Also, deter-
mine if the patient has any health conditions that may require the presence of special
emergency equipment during the scanning procedure, or if he or she will need
sedation medication necessitating the use of appropriate monitoring equipment.
Generally standard screening forms are used for all patients scanned in a magnetic
resonance system.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(1996) for discussion of what implants may be safely scanned using magnetic resonance.
The presence of ferromagnetic materials in the globe of the eye is contraindicated for MRI.
Patients with prior metal exposure to the eye should have plain X rays of the orbital area
to ensure that all metal has been removed, prior to placing them in the magnetic field.
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
2. Explain the procedure to the patient and record relevant clinical history. Ensure that
the patient understands what is expected and ask if there are any questions; answer
appropriately. If the procedure is a research protocol, have the patient sign any
necessary consent form.
3. Request that the patient change into a gown to eliminate any metal that might be found
in clothing. Ask the patient to remove all personal effects such as jewelry, hearing
aids, glasses, etc., prior to entering the MRI scan room.
All personal belongings should be secured during the examination.
4. Fill a 20-ml syringe with normal saline and attach to saline filled extension tubing
(35-in.). Obtain intravenous (i.v.) access utilizing a 22-G angiocatheter and attach
saline-prepared extension tubing and syringe. This will allow you to flush the
extension tubing with ∼10 ml of saline while the patient waits to be imaged, and, in
cases in which a power injector is not available, this will allow preparation for bolus
injection (see Note below). Secure the position of the angiocatheter with tegraderm
or tape.
Obtaining i.v. access prior to entering the scan room will promote patient throughput and
eliminate “dead” time of starting the i.v. while the patient is on the exam table. Follow
power injector manufacturer guidelines with regard to appropriate gauge needle/angio-
catheter to be used; this will depend on chosen injection/flow rates.
NOTE: If you do not have access to a power injector you will still be able to perform
dynamic liver imaging, as the extension tubing will allow the saline syringe to be placed
at the foot of the patient table during precontrast imaging. In this case, you will need to
draw up the contrast agent in another syringe. When you are ready to bolus-inject the
contrast agent, simply disconnect the saline syringe and connect the syringe filled with
contrast agent; once you have injected the bolus of contrast agent, reconnect the saline
syringe and bolus an appropriate volume of flush, usually ∼10 ml. Alternatively, to
eliminate the need of switching syringes, incorporate the use of a 3-way stopcock.
5. Set up the exam room by securing the circularly polarized (CP) body array coil onto
the table and providing a clean exam table.
6. Set up the power injector as specified by the manufacturer.
A minimum of 40 ml normal saline should be drawn up to ensure sufficient saline is
available to keep the vein open (KVO) throughout the exam. To determine the amount of
MRI of the Liver
A15.1.2
Current Protocols in Magnetic Resonance Imaging
contrast agent to be used, reference the contrast agent packet insert and draw up the amount
indicated per kg of patient weight. There is no need to double dose.
7. Escort the patient to the MR examination room and ask the patient to lie down
accordingly with respect to the exam to be performed. Help the patient mount onto
the table. Either before or right after the patient lies down, set up any triggering
devices or other monitoring equipment that is to be used.
8. Connect the extension tubing, secured to the syringe, to the power injector extension
tubing.
9. Inform the patient about what will occur during the procedure, what he or she will
experience while in the magnet, and how to behave, including the following:
a. If earphones or headphones are used to protect the ears from the loud sounds
produced by the gradients, the patient will be asked to wear these, but will be able
to communicate with you at any time during the imaging.
b. The patient will be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this works).
c. For good results the patient should not talk, and should avoid or minimize
swallowing or other movement, during each scan—i.e., as long as the banging
sounds continue. Between scans, talking and swallowing are allowed in most
cases, but should be avoided when comparative positional studies are being
performed; the patient will be informed when this is the case.
Additionally, review breath-holding instructions with the patient, and provide the patient
with an approximate time that the examination will take.
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
10. Secure the top portion of the CP-body array coil to prevent it from moving side-to-side
during breath-holding imaging sequences.
Usually the manufacturer provides straps that are directly attached to the coil.
11. If needed, place a pillow or other support under the knees to make the patient more
comfortable.
12. Use the laser light to position the patient, and to center the coil (see Table A15.1.2).
Then, advance the patient table to isocenter.
Once this step has been performed, so long as the patient does not move on the table, the
table itself can be moved and then replaced in the same position as before without
jeopardizing the positioning of one scan relative to another.
13. If the patient is unable to hold still, provide an appropriate sedative.
14. Program the power injector for a contrast agent and saline injection rate of 2 ml/sec.
Total volume of saline following contrast agent injection should be programmed for
10 ml. Program a scan delay of 18 sec (contrast agent is injected, scan is initiated 18
sec after contrast and saline are delivered). Arm the power injector and keep the vein
open.
Do not inject the contrast agent at this time!
Sequence 1: Three-plane positioning scout
15. To validate the patient’s position and to have a reference to prescribe successive
imaging sequences, acquire a three-plane orthogonal scout sequence. See Table
A15.1.2 for specific parameters.
Liver
A15.1.3
Current Protocols in Magnetic Resonance Imaging
 Most MR scanners can be programmed to acquire the scout automatically after coil tuning Table A15.1.2 Imaging Parameters for Three-Plane Positioning Scout (Sequence 1)
or after the patient has been placed in isocenter (for systems that do not require tuning).
Patient position Supine
Sequence 2: Half-acquisition (partial Fourier) turbo spin echo coronal Scan type Gradient echo
16. Display both the coronal and transverse scout images in two separate quadrants on Imaging plane (orientation) Sagittal, transverse, and coronal
the scan monitor. Change imaging parameters to those listed in Table A15.1.3. Central slice or volume center Laser light centered approximately one
Position slices to center of the transverse scout, ensuring that the entire liver is hand width above the inferior rib
covered. margin
Echo time (TE) 6 msec
17. Instruct the patient to remain motionless and to breathe normally, as the scan will Repeat time (TR) 15 msec
begin and last for ∼40 sec. Flip angle (FA) 30°
Fields of view (FOVx, FOVy) 450 mm, 450 mm
Sequence 3: Gradient echo coronal Resolution (Δx, Δy) 1.76 mm, 3.52 mm
18. Duplicate the setup in step 17. Change imaging parameters to those listed in Table Number of data points collected (Nx, Ny) 256, 128
A15.1.4. Position slices to center of the transverse scout, ensuring that the entire liver Display matrix (Dx, Dy) 256, 256
is covered. Slice thickness (Δz) 10 mm
19. Instruct the patient to take in a deep breath and exhale, take in another deep breath, Number of slices 3
and hold it. Initiate the scan. Slice gap Not applicable
Number of acquisitions (Nacq) 1
Sequence 4: Half-acquisition turbo spin echo transverse with fat saturation (fat Swap read and phase encoding No
suppression) Slice location Not applicable
20. Display both the coronal and transverse scout images in two separate quadrants on Saturation pulses Not applicable
the scan monitor. Change imaging parameters to those listed in Table A15.1.5. Scan time 16 sec
Position slices to middle of liver on coronal scout, ensuring that the entire liver is
covered.
Perform system shim as recommended by manufacturer as this is a fat saturation sequence. Table A15.1.3 Imaging Parameters for Half-Acquisition Turbo Spin Echo
(Sequence 2)
21. Instruct the patient to remain motionless and to breathe normally, as the scan will
begin and last for ∼40 sec. Patient position Supine
Scan type Half acquisition turbo spin echo
Sequence 5: Turbo inversion recovery transverse Imaging plane (orientation) Coronal
22. Display the midline slice of the gradient echo coronal image (sequence 3) and the Central slice or volume center Slices posted on transverse scout;
transverse scout image in two separate quadrants on the scan monitor. Change center to liver
imaging parameters to those listed in Table A15.1.6. Position slices to cover from the Echo time (TE) 90 msec
top of the liver through mid liver. This sequence will need to be repeated in order to Repeat time (TR) 4.4 msec (note: the true TR is infinite,
obtain coverage through the entire liver. 4.4 msec represents the echo spacing)
Delay time (TD) 1500 msec
It is imperative that the slices be prescribed off of the breath-hold gradient echo coronal Flip angle (FA) 150°
image, as this is also a breath-held imaging sequence. Otherwise, the slice location will Fields of view (FOVx, FOVy) 400 mm, 400 mm
not be accurate relative to the reference image if a non-breath-held image is used.
Resolution (Δx, Δy) 1.56 mm, 2.08 mm
23. Instruct the patient to take in a deep breath and exhale, take in another deep breath Number of data points collected (Nx, Ny) 256, 192 (using half Fourier)
and hold it, and initiate the scan. Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 8–10 mm
24. Scan the slices inferiorly to cover the remaining liver that was not imaged, making it
Number of slices 20
the second run. You should overlap the slices by at least one slice to ensure complete
Slice gap 1.6–2 mm
coverage. Repeat step 23. Number of acquisitions (Nacq) 1
Swap read and phase encoding No
Sequence 6: Gradient echo transverse (out-of-phase)
Slice location Centered to cover entire liver
25. Display the midline slice of the gradient echo coronal image (sequence 3) and the
Saturation pulses No
transverse scout image in two separate quadrants on the scan monitor. Change
Slice series Interleaved
imaging parameters to those listed in Table A15.1.7. Position slices to cover entire
Scan time 40 sec
liver.

MRI of the Liver Liver

A15.1.4 A15.1.5
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A15.1.4 Imaging Parameters for Gradient Echo Coronal (Sequence 3) Table A15.1.6 Imaging Parameters for Turbo Inversion Recovery (Sequence 5)

Patient position Supine Patient position Supine

Scan type Gradient echo Scan type Echo train inversion recovery
Imaging plane (orientation) Coronal Imaging plane (orientation) Transverse
Slice or volume center Slices posted on transverse scout; Central slice or volume center Slices centered upper liver
center to liver Echo time (TE) 76 msec
Echo time (TE) 4.1 msec Repeat time (TR) 5110 msec
Repeat time (TR) 140 msec Inversion time (TI) 170 msec
Flip angle (FA) 80° Flip angle (FA) 160°
Fields of view (FOVx, FOVy) 400 mm, 400 mm Fields of view (FOVx, FOVy) 350 mm, 263 mm
Resolution (Δx, Δy) 1.56 mm, 3.12 mm Resolution (Δx, Δy) 1.37 mm, 2.65 mm
Number of data points collected (Nx, Ny) 256,128 Number of data points collected (Nx, Ny) 256, 99
Display matrix (Dx, Dy) 256, 256 Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 8–10 mm Slice thickness (Δz) 8–10 mm
Number of slices 20 Number of slices 13
Slice gap 1.6–2 mm Slice gap 1.6–2 mm
Number of acquisitions (Nacq) 1 Number of acquisitions (Nacq) 1
Swap read and phase encoding No Swap read and phase encoding No
Slice location Centered to cover entire liver Slice locations First slice above top of liver (first run);
Saturation pulses No mid liver down (second run)
Slice series Interleaved Saturation pulses No
Scan time 18 sec Slice series Interleaved
Scan time 20 sec

Table A15.1.5 Imaging Parameters for Half-Acquisition Turbo Spin Echo with Fat
Saturation (Sequence 4) Table A15.1.7 Imaging Parameters for Gradient Echo Out-of-Phase (Sequence 6)

Patient position Supine Patient position Supine

Scan type Half acquisition turbo spin echo Scan type Gradient echo
Imaging plane (orientation) Transverse Imaging plane (orientation) Transverse
Central slice or volume center Slices posted on coronal; center to liver Central slice or volume center Slices posted on coronal; center to liver
Echo time (TE) 90 msec Echo time (TE) 2.2 msec
Repeat time (TR) 4.4 msec (note: the true TR is infinite, Repeat Time (TR) 140 msec
4.4 msec represents the echo spacing) Flip angle (FA) 80°
Delay time (TD) 1500 msec Fields of view (FOVx, FOVy) 350 mm, 263 mm
Flip angle (FA) 150° Resolution (Δx, Δy) 1.37 mm, 2.05 mm
Fields of view (FOVx, FOVy) 350 mm, 263 mm Number of data points collected (Nx, Ny) 256, 128
Resolution (Δx, Δy) 1.37 mm, 1.37 mm Display matrix (Dx, Dy) 256, 256
Number of data points collected (Nx, Ny) 256, 192 using half Fourier Slice thickness (Δz) 8–10 mm
Display matrix (Dx, Dy) 256, 256 Number of slices 20
Slice thickness (Δz) 8–10 mm Slice gap 1.6–2 mm
Number of slices 20 Number of acquisitions (Nacq) 1
Slice gap 1.6–2 mm Swap read and phase encoding No
Number of acquisitions (Nacq) 1 Slice location Centered to cover entire liver
Swap read and phase encoding No Saturation pulses No
Slice location Centered to cover entire liver Slice series Interleaved
Saturation pulses Yes, superior and inferior to slices Scan time 19 sec
Fat suppression Yes
Slice series Interleaved
Scan time 40 sec

MRI of the Liver Liver

A15.1.6 A15.1.7
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 It is imperative that the slices be prescribed off of the breath-hold gradient echo coronal It is imperative that the slices be prescribed off of the breath-hold gradient echo coronal
image, as this is also a breath-held imaging sequence. Otherwise, the slice location will image, as this is also a breath-held imaging sequence. Otherwise, the slice location will
not be accurate relative to the reference image if a non-breath-held image is used. not be accurate relative to the reference image if a non-breath-held image is used.
26. Instruct the patient to take in a deep breath and exhale, to take in another deep breath 28. Instruct the patient to take in a deep breath and exhale, to take in another deep breath
and hold it, and initiate the scan. and hold it, and initiate the scan.
Sequence 7: Gradient echo transverse
Sequence 8: Gradient echo transverse—immediate post contrast
27. Display the midline slice of the gradient echo coronal image (sequence 3) and the
NOTE: See patient setup section for specific instructions on preparation for contrast
transverse scout image in two separate quadrants on the scan monitor. Change
injection. This preparation must be done prior to placing the patient in the scanner.
imaging parameters to those listed in Table A15.1.8. Position slices to cover entire
liver (Fig. A15.1.1). 29. Repeat step 27.
It is imperative that the slices be prescribed off of the breath-hold gradient echo coronal
Table A15.1.8 Imaging Parameters for Gradient Echo Transverse (Sequence 7) image as this is also a breath-held imaging sequence. Otherwise, the slice location will not
be accurate relative to the reference image if a non-breath-held image is used.
Patient position Supine
Scan type Gradient echo 30. Explain to the patient that you will now be injecting the contrast agent and he or she
Imaging plane (orientation) Transverse may feel a cool sensation in his or her arm. Initiate the injection. Do not begin
Central slice or volume center Slices posted on coronal; center to liver scanning until the 18-sec scan delay has expired. However, breathing instructions
Echo time (TE) 4.5 msec should be delivered when 10 sec of delay are remaining (see step 31).
Repeat time (TR) 140 msec If you do not have access to a power injector and are “hand” injecting, follow step 30.
Flip angle (FA) 80° However, after you have completed the bolus contrast injection, reattach the saline-filled
Fields of view (FOVx, FOVy) 350 mm, 263 mm syringe and flush with 10 ml of saline. Begin breathing instructions after 5 ml of the saline
Resolution (Δx, Δy) 1.37 mm, 2.05 mm has been injected, then proceed to initiate the scan. The process of switching syringes must
Number of data points collected (Nx, Ny) 256, 128 be completed as quickly as possible and thus, the suggestion of incorporating the use of a
Slice thickness (Δz) 8–10 mm 3-way stopcock.
Number of slices 18 31. When there are 10 sec of delay remaining, instruct the patient to take in a deep breath
Slice gap 1.6–2 mm and exhale, then take in another deep breath and hold it Initiate the scan
Number of acquisitions (Nacq) 1
Swap read and phase encoding No
Slice locations Centered to cover entire liver
Saturation pulses No Table A15.1.9 Imaging Parameters for Gradient Echo with Fat Saturation—90
Slice series Interleaved Sec Delay (Sequence 10)
Scan time 18 sec
Patient position Supine
Scan type Gradient echo
Imaging plane (orientation) Transverse
Central slice or volume center Slices posted on coronal; center to liver
Echo time (TE) 4.1 msec
Repeat time (TR) 147.2 msec
Flip angle (FA) 80°
Fields of view (FOVx, FOVy) 350 mm, 263 mm
Resolution (Δx, Δy) 1.37 mm, 2.05 mm
Number of data points collected (Nx, Ny) 256, 128
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 8–10 mm
Number of slices 20
Slice gap 1.6–2 mm
Number of acquisitions (Nacq) 1
Swap read and phase encoding No
Slice location Centered to cover entire liver
Fat saturation Yes
Slice series Interleaved
Figure A15.1.1 Coronal breath-hold image used to plan transverse breath-hold
MRI of the Liver Scan time 19 sec Liver
sequences.

A15.1.8 A15.1.9
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Sequence 9: Gradient echo transverse (45-sec delay after sequence 8) 35. Once 90 sec has expired, instruct the patient to take in a deep breath and exhale, then
32. Repeat step 27 (see sequence 7). take in another deep breath and hold it. Initiate the scan.
33. Once 45 sec has expired, instruct the patient to take in a deep breath and exhale, then
take in another deep breath and hold it. Initiate the scan COMMENTARY
Background Information the same time, we consider it imperative that
Sequence 10: Gradient echo transverse with fat saturation (90 sec delay after By using multiple data acquisitions and each individual sequence be short, such that it
injection) early and late post-gadolinium images (Figs. can be acquired in a breath hold. This ensures
34. Display the midline slice of the gradient echo coronal image (sequence 3) and the A15.1.2 through A15.1.5), accurate evaluation that the entire study can be performed in 15 min
transverse scout image in two separate quadrants on the scan monitor. Change of liver disease can be made (Semelka et al., or less (Semelka et al., 1999). Short study times
imaging parameters to those listed in Table A15.1.9. Position slices to cover entire 1992, 1993, 1994, 1997, 1999; Low et al., 1993, permit good patient throughput and ensure that
liver (Fig. A15.1.1). 1999; Whitney et al., 1993; Larson et al., 1994; imaging quality will be good throughout the
Oi et al., 1996; Semelka and Kelekis, 1997; entire study. We have found that patient coop-
It is imperative that the slices are prescribed off of the breath-hold gradient echo coronal Yamashita et al., 1996; Rofsky et al., 1999;). At eration diminishes over the time course of
image as this is also a breath-held imaging sequence. Otherwise, the slice location will not
be accurate relative to the reference image if a non-breathheld image is used.

Figure A15.1.4 1-min post-gadolinium spoiled gradient echo transverse image

Figure A15.1.2 Unenhanced transverse spoiled gradient echo image. (portal-phase) demonstrates maximal hepatic enhancement.

Figure A15.1.3 Immediate post-gadolinium spoiled gradient echo image (arte- Figure A15.1.5 2-min post-gadolinium fat-suppressed spoiled gradient echo
rial phase) demonstrates the absence of gadolinium in hepatic veins and the transverse image (equilibrium-phase) demonstrates delayed contrast enhance-
MRI of the Liver presence of gadolinium in the hepatic arteries and portal veins. ment. Liver

A15.1.10 A15.1.11
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 4
 Table A15.1.10 Imaging Parameters for Non-Slice Selective Turbo Flasha longer examinations, and therefore image qual- immediate post-contrast spoiled gradient echo
ity deteriorates. A set protocol also makes it is superior to dynamic sequential or spiral CT
Patient position Supine possible to perform studies at any time during techniques (Semelka et al., 1992, 1993, 1997;
Scan type Inversion recovery prepared snap shot the operation of the MR center without the Low et al., 1993; Larson et al., 1994; Oi et al.,
gradient echo requirement of close physician supervision. 1996, Yamashita et al., 1996).
Imaging plane (orientation) Transverse This above-described protocol results in fast
Central slice or volume center Slices posted on coronal; center to liver examination times and consistent, reproducible Literature Cited
Echo time (TE) 4.2 msec image quality that is effective for evaluating the Larson, R.E., Semelka, R.C., Bagley, A.S., Molina,
full spectrum of liver diseases. P.L., Brown, E.D., and Lee, J.K. 1994. Hyper-
Repeat time (TR) 11 msec vascular malignant liver lesions; comparison of
Inversion time (TI) 360 msec various MR imaging pulse sequences and dy-
Delay time (TD) 4900 msec
Critical Parameters and namic CT. Radiology 192:393-399.
Troubleshooting Low, R.N., Francis, I.R., Sigeti, J.S., and Foo, T.K.
Flip angle (FA) 15° One of the most critical aspects of this 1993. Abdominal MR imaging: Comparison of
Fields of view (FOVx, FOVy) 350 mm, 263 mm protocol is to ensure that the initial scan per- T2-weighted fast conventional spin-echo, and
Resolution (Δx, Δy) 1.37 mm, 2.05 mm formed after gadolinium administration is contrast-enhanced fast multiplanar spoiled gra-
Number of data points collected (Nx, Ny) 256, 128 timed appropriately. We use the presence of dient-recalled imaging. Radiology 186:803-811.
Display matrix (Dx, Dy) 256, 256 contrast in the hepatic arteries and portal veins Low, R.N., Semelka, R.C., Worawattankul, S.,
Slice thickness (Δz) 8–10 mm with the absense of contrast in the hepatic veins Alzate, G.D., and Sigeti, J.S. 1999. Extra hepatic
abdominal imaging in patients with malignancy:
Number of slices 21 as a check for optimal enhancement (Semelka
Comparison of MR imaging and helical CT with
Slice gap 1.6–2 mm et al., 1992, 1993, 1997; Low et al., 1993; subsequent surgical correlation. Radiology
Number of acquisitions (Nacq) 1 Larson et al., 1994; Semelka and Kelekis, 210:625-632.
Swap read and phase encoding No 1997). It is also important that the spoiled Oi, H., Murakami, T., Kim, T., Matsushita, M.,
Slice location Centered to cover entire liver gradient echo sequence have a sufficient num- Kishimoto, H., and Nakamura, H. 1996. Dy-
ber of images (at least 14) to encompass the namic MR imaging and early-phase helical CT
Saturation pulses No for detection small intrahepatic metastases of
entire liver in one breath hold for the data
Slice series Interleaved hepatocellular carcinoma. AJR Am. J.
acquisition immediately following contrast
Scan time 2 min, 20 sec Roentgenol. 366:36-374.
administration.
aUse pre-contrast for patients unable to breath-hold. Rofsky, N.M., Lee, V.S., Laub, G., Pollack, M.A.,
The two major potential problems with this
Krinsky, G.A., Thomasson, D., Ambrosino,
protocol are that patients may not be able to M.M., and Weinreb, J.C. 1999. Abdominal MR
Table A15.1.11 Imaging Parameters for Slice Selective Turbo Flasha suspend respiration, and that timing of gadolin- imaging with a volumetric interpolated breath-
ium administration may be incorrect. The first hold examination. Radiology 212:876-884.
Patient position Supine problem can be addressed by substituting snap- Semelka, R.C. and Kelekis, N.L. 1997. Liver. In
Scan type Inversion recovery prepared snap shot shot T1-weighted sequences (e.g., inversion re- MRI of the Abdomen and Pelvis. A Text-Atlas.
gradient echo covery prepared snapshot gradient echo; see (R.C. Semelka, S.M. Ascher, and C. Reinhold,
eds.) pp. 19-135. Wiley-Liss, New York.
Imaging plane (orientation) Transverse Table A15.1.10 and Table A15.1.11) for the
breath hold T1-weighted spoiled gradient echo. Semelka, R.C., Shoenut, J.P., Kroeker, M.A., Green-
Central slice or volume center Slices posted on coronal; center to liver berg, H.M., Simm, F.C., Minuk, G.Y., Kroeker,
Echo time (TE) 4.2 msec Fast acquisition of the first and second post-
R.M., and Micflikier, A.B. 1992. Focal liver
Repeat time (TR) 11 msec gadolinium spoiled gradient echo sequences disease: Comparison of dynamic contrast-en-
compensate for the second problem. Alterna- hanced CT and T2-weighted fat suppressed,
Inversion time (TI) 300 msec
tively, other investigators have described using FLASH, and dynamic gadolinium-enhanced
Delay time (TD) 2 msec MR imaging at 1.5T. Radiology 184:687-694.
a test dose of gadolinium to determine timing
Flip angle (FA) 15° (Rofsky et al., 1999). Semelka, R.C., Cumming, M.J., Shoenut, J.P., Ma-
Fields of view (FOVx, FOVy) 350 mm, 263 mm gro, C.M., Yaffe, C.S., Kroeker, M.A., and
Resolution (Δx, Δy) 1.37 mm, 2.05 mm Greenberg, H.M. 1993. Islet cell tumors: Com-
Anticipated Results parison of dynamic contrast-enhanced CT and
Number of data points collected (Nx, Ny) 256, 128 Accurate detection and characterization of MR imaging with dynamic gadolinium enhance-
Display matrix (Dx, Dy) 256, 256 focal liver lesions and a positive effect on pa- ment and fat suppression. Radiology 186:799-
Slice thickness (Δz) 8–10 mm tient management can be expected with this 802.
Number of slices 21 protocol (Semelka et al., 1997). The greatest Semelka, R.C., Willms, A.B., Brown, M.A., Brown,
Slice gap 1.6–2 mm advantages over computed tomography (CT) E.D., and Finn, J.P. 1994. Comparison of breath-
are in the investigation of hypervascular lesions hold T1-weighted MR sequences for imaging of
Number of acquisitions (Nacq) 1 the liver. J. Magn. Reson. Imaging 4:759-765.
Swap read and phase encoding No such as hepatocellular carcinoma (Oi et al.,
1996; Yamashita et al., 1996; Semelka and Semelka, R.C., Worawattanakul, S., Kelekis, N.L.,
Slice location Centered to cover entire liver John, G., Woosley, J.T., Graham, M., and Cance,
Kelekis, 1997) and hypervascular metastases.
Saturation pulses No W.G. 1997. Liver lesion detection, charac-
Employing this MR protocol, superior re- terization, and effect on patient management;
Slice series Interleaved
sults for lesion detection and characterization comparison of single-phase spiral CT and cur-
Scan time 36 sec comparable to spiral CT have been demon- rent MR techniques. J. Magn. Reson. Imaging
MRI of the Liver
aUse post-contrast for patients unable to breath hold.
strated (Semelka et al., 1997). In particular, 7:1040-1047. Liver

A15.1.12 A15.1.13
Supplement 4 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Semelka, R.C., Balci, N.C., Op de Beeck, B., and
Reinhold, C. 1999. Evaluation of a 10-minute
comprehensive MR imaging examination of the
upper abdomen. Radiology 211:189-195.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Whitney, W.S., Herfkens, R.J., Jeffrey, R.B.,
McDonnell, C.H., Li, K.C., Van Dalsem, W.J.,
Low, R.N., Francis, I.R., Dabatin, J.F., Glazer,
G.M. 1993. Dynamic breath-hold multiplanar
spoiled gradient-recalled MR imaging with gad-
olinium enhancement for differentiating hepatic
hemangiomas from malignancies at 1.5T. Radi-
ology 189:863-870.
Yamashita, Y., Mitsuzaki, K., Yi, T., Ogata, I., Nishi-
haru, T., Urata, J., and Takahashi, M. 1996. Small
hepatocellular carcinoma in patients with
chronic liver damage: Prospective comparison of
MRI of the Liver
A15.1.14
detection with dynamic MR imaging and helical
CT of the whole liver. Radiology 200:79-84.
Key References
Shellock 1996. See above.
Covers a number of important patient management
issues related to MR imaging, including recom-
mended safety procedures, a list of metallic implants
that have been tested for MR compatibility, and a
list of other sources on MR safety.
Contributed by Laurie Fisher
Siemens Uptime Service Center
Cary, North Carolina
Richard Semelka and Kathy Wilber
University of North Carolina at Chapel Hill
Chapel Hill, North Carolina
Current Protocols in Magnetic Resonance Imaging
Hepatic MRI for GE Scanners UNIT A15.2
MRI techniques continue to evolve rapidly, but the basic components of a liver exami-
nation, and the clinical role of each component, have changed little. The three most
important considerations for choosing techniques for imaging the liver continue to be
contrast, motion artifact, and coverage. In particular, the approach to motion has evolved
from averaging to motion compensation to suspended respiration and/or subsecond
imaging.
This unit is written as an alternate to UNIT A15.1, which bases recommendations on
experience using Siemens MRI equipment. The authors’ experience has been acquired
primarily using General Electric equipment. However, these basic approaches to hepatic
MR imaging are quite similar, with most details relating to minor vendor-specific
differences.
LIVER IMAGING BASIC
PROTOCOL
Whenever possible, it is best to image the entire liver during a single breath-hold
sequence. The advantages of single breath-hold coverage include simplicity and avoid-
ance of misregistration. Single breath-hold imaging being essential for dynamic mul-
tiphasic imaging must be emphasized; the crucial arterial phase can only be acquired
during a brief interval of ∼20 sec. Whenever possible, patients should be triaged to MRI
machines that have this capability. The abdomen ranks along with the heart and vascular
system as body parts demanding the most up-to-date MRI hardware and software for
successful use.
Several years ago, successful use of MRI for imaging the liver usually involved close
involvement of the senior MR radiologist with each scanning procedure. During the ≥60
min that these exams often required, the radiologist would spend much of his/her time
behind the scanner discussing the details of next pulse sequence with the technologist
during the 5 to 10 min that would elapse during each sequence acquisition. Cases would
be read between exams.
Times have changed. Markedly increased caseloads and decreased reimbursements no
longer permit this luxury. Hepatic MRI can only become a viable procedure for routine
use if the execution of protocols can be trusted to a trained technologist, freeing the
radiologist to read cases and interact with referring clinicians. Fortunately, hardware and
software improvements now permit this, and the comprehensive hepatic MRI protocol
described here should not require >30 min of MRI scanner time. Continued improvements
in the scanner interface will allow these exams to be completed in ∼10 min; actual image
acquisition is <5 min, with all images acquired during breath holding. These sequences
require the patient to be able to suspend respiration for ∼25 sec. It is imperative that there
is clear communication between the technologist and the patient throughout the exam. If
patients are unable to cooperate with this, there are alternative pulse sequences that can
be used to yield images of acceptable quality.
Table A15.2.1 lists the hardware necessary to perform the procedure, along with appro-
priate parameters.
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.
Liver
Contributed by Donald G. Mitchell, Peter Natale, and George Holland A15.2.1
Current Protocols in Magnetic Resonance Imaging (2002) A15.2.1-A15.2.14
Copyright © 2002 by John Wiley & Sons, Inc. Supplement 4
 Table A15.2.1 Equipment Specifications Needed to Perform Imaging Obtaining i.v. access prior to entering the scan room will promote patient throughput and
Sequences eliminate “dead” time of starting the i.v. while the patient is on the exam table. Follow
power injector manufacturer guidelines with regard to appropriate needle gauge/angio-
Coil type Torso phased array coil catheter to be used; this will depend on the chosen injection/flow rates.
Manufacturer and system type GE Signa
If one does not have access to a power injector, one will still be able to perform dynamic
Field strength 1.5 T liver imaging. In this case, draw-up the contrast agent in one 20-ml syringe and saline in
Gradient strength 25 mT/m (or whatever the system permits) two others to flush the timing and examination boluses. Alternatively, larger syringes can
Knee cushion Yes be used, drawing 20-ml of contrast agent into one and 40-ml of saline into the other. To
Pulse oximeter If patient requires sedation eliminate the need of attaching and detaching syringes when switching from contrast agent
Power injector Yes, if available to flushing the saline, incorporate the use of a 3-way stopcock or Y-tubing.
Normal saline Yes
5. Set up the power injector as specified by the manufacturer.
35-in. extension tubing Yes
Use of contrast agents Yes A minimum of 40 ml normal saline should be drawn-up to ensure sufficient saline is
available to keep the vein open (KVO) throughout the exam. There is no need to use a
double dose for hepatic MRI. While the standard on-label dose is 0.1 mmol/kg, this can
Materials safely be rounded up to the nearest bottle size, or a dose of 20 ml can be used routinely.
Normal saline (0.9% NaCl), sterile, 100 ml minimum
6. Escort the patient to the MR examination room and ask them to lie down accordingly
Extravascular contrast agent (e.g., Magnevist, Omniscan, or Prohance), 20 ml for
with respect to the exam to be performed. Connect the extension tubing secured to
most patients
the syringe, to the power injector extension tubing. Review the following items with
22-G angiocatheter
the patient:
110-in. extension tubing
a. Provide earplugs or headphones to the patient to minimize the loud knocking noise
Set up patient and equipment that will be produced by the gradients but ensure them that they will still be able
1. Interview (screen) the patient to assess for contraindications such as cardiac pace- to hear you.
maker, implanted mechanical devices, and/or ferromagnetic materials. Also, deter-
b. Provide the patient with a safety squeeze-bulb and demonstrate how it works;
mine if the patient will need sedation medication necessitating the need for
explain to the patient when to use the squeeze-bulb (i.e., if they need assistance
appropriate monitoring equipment.
during the exam).
Generally, standard screening forms (APPENDIX 1) are used for all patients screened in a c. Explain to the patient that you will be talking to them between imaging sequences
magnetic resonance system. which will be when the loud knocking noise stops. Additionally, review breath-
The presence of ferromagnetic materials may be a health hazard to the patient while in the holding instructions with the patient.
magnetic field and/or adversely affect image quality. To determine the safety of scanning d. Explain to the patient that it is imperative that they remain motionless during the
such ferromagnetic materials see Shellock and Kanal (1996). loud knocking noise to ensure good results; also explain that they should not
The presence of ferromagnetic materials in the globe of the eye is contraindicated for MRI. reposition their body between imaging sequences.
Patients with prior metal exposure to the eye should have plain X-rays of the orbital area e. Nevertheless, the patient may call out at any time if he or she feels it necessary.
to ensure that all metal has been removed prior to placing them in the magnetic field.
f. Provide the patient with an approximate time that the examination will take.
Patients may be accompanied into the magnet room by a friend or family member, who can
g. Attach the power injector to the i.v., and set it to keep the line open.
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
7. Place the torso phased array coil on the scanning table. Secure the torso phased array
2. Request the patient to change into a gown to eliminate any metal that might be found coil to the patient, centering at the costal margin in most patients. Usually, straps are
in clothing. Ask the patient to remove all personal effects such as jewelry, hearing provided by the manufacturer that are directly attached to the coil. Place a foam pad
aids, glasses, etc., prior to entering the MRI scan room. between the patient’s abdomen and the coil to reduce near-field artifacts. Place a
All personal belongings should be secured during the examination. support under the patient’s knees to enhance patient comfort.

3. Explain the procedure to the patient and record relevant clinical history. Ensure that
the patient understands what is expected and ask if they have any questions; answer
appropriately. If this is a research protocol, have the patient sign any necessary
consent forms.
4. Obtain intravenous (i.v.) access utilizing a 22-G angiocatheter and attach to saline
filled extension tubing (110-in.), with saline set to slow infusion by gravity drip.
Secure position of angiocatheter with sterile, clear occlusive dressing.
Hepatic MRI for
GE Scanners
8. Place nasal canula for oxygen 2 liters, if available, to aid in suspended respiration.
Hyperventilation before breathholding is helpful. Most helpful is administration of
oxygen during the examination (Marks et al., 1997).
9. Using the laser light, center the patient’s upper abdomen to the coil.
10. Advance the patient table to isocenter.
Once this step has been performed, so long as the patient does not move on the table, the
table itself can be moved and then replaced in the same position as before without
jeopardizing the positioning of one scan relative to another.
Liver

A15.2.2 A15.2.3
Supplement 4 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 4
 11. Program the power injector for a contrast agent and saline injection rate of 2 ml/sec. Table A15.2.2 Imaging Parameters for Sequence 1 (Single Shot Fast Spin Echo)
Total volume of saline following contrast agent injection should be programmed for
20 ml. The scan delay in healthy young patients averages 18 sec (contrast agent is Patient position Supine
injected, scan is initiated 18 sec after the contrast agent and saline are delivered), but Scan type Single shot fast spin echo
this is highly variable in the elderly or those with cardiovascular disease. Imaging plane (orientation) Coronal
Variable bandwidth Yes
Although Fisher, Semelka, and Wilber (UNIT A15.1) recommend the routine use on an 18-sec Pulse sequence database (PSD) SSFSE
delay, since this is a good estimate for most healthy individuals, the authors strongly
Central slice or volume center Center to xyphoid
recommend determination of the optimum delay in each patient. This can either be done
with the automated bolus detection (SMARTPREP, General Electric) feature, or by use of Echo time (TE) 100 msec (effective)
a 2 ml timing bolus sequence, as per Earls et al. (1997) and described later in this unit. Receiver bandwidth (RBW) ±31 kHz
Both of these methods can be adapted to either hand or power injection. Repeat time (TR) Infinite
Flip angle (FA) 90°
12. Arm the power injector and keep the vein open. IMPORTANT: Do not inject the Fields of view (FOVx, FOVy) 360 mm, 360 mm
contrast agent. Resolution (Δx, Δy) 1.41 mm, 1.92 mm
Number of data points collected (Nx, Ny) 256, ∼100
Sequence 1: Coronal single shot fast spin echo (SSFSE) Display matrix (Dx, Dy) 256, 256
Although some technologists precede this with a three-plane orthogonal scout sequence, Slice thickness (Δz) 5 mm
the authors find that this can be safely omitted by most experienced technologists. The Number of slices 28
coronal single shot fast spin echo (SSFSE) sequence is nearly always of high quality and Slice gap 0
high information content (Fig. A15.2.1). These can be acquired using a 320- to 360-mm2 Number of excitations (NEX) 0.5 (half Fourier)
field of view (FOV) and 5- to 8-mm image slices. Using thicker image slices will reduce Swap read and phase encoding No
the number of slices and save imaging time. Using a large FOV allows additional imaging Slice location A70–P70
of the abdomen and much of the lower chest and upper pelvis in a single breathhold, and Saturation pulses None
provides added assurance against wrap-around artifact. Alternatively, a smaller FOV and Scan time 40 sec
thinner image slices (5 mm) can be acquired in two or more breathholds to provide greater
detail of biliary and pancreatic ducts. Each breath-hold should cover one slab (with a
certain number of slices) such that the scan time is ∼20 sec. Then move the region of
Obtaining images during suspended respiration will minimize misregistration and image
interest to a new slab. Slabs should be overlapped, not interleaved. blurring. However, it is important to be aware that the absence of visible motion artifact
does not indicate that the patient successfully suspended respiration; the multislice
13. Instruct the patient to remain motionless, and to take a deep breath in and blow it out
gradient echo images must be checked for this.
two times, then hold it. Run the sequence according to the parameters, except with
14 number of slices, listed in Table A15.2.2. Instruct the patient to breathe and hold Although some patients may hold their breath longer at full inspiration, more consistent
their breath, run the sequence to cover the remaining 14 slices. location and absence of motion are achieved if respiration is suspended at relaxed
expiration.
Observe the respiratory waveform to determine if the patient is complying with breath
holding instructions. Remember that poor compliance will not degrade these images
because of their relative lack of motion sensitivity.
14. Confirm positioning of the torso coil by making certain that all parts of the liver are
included within the sensitive volume of the coil. Reposition the patient and repeat
the sequence, if necessary.

Sequence 2: Heavily T2-weighted (TE = 180 msec) transverse single shot FSE
15. Using the coronal images, position slices to ensure that the entire liver is covered on
transverse images (Fig. A15.2.2). Ask the patient to hold his or her breath and run the
sequence according to the parameters listed in Table A15.2.3.

Sequence 3: Transverse gradient echo in-phase and out-of-phase

If possible, these should be obtained as a dual-echo sequence, with TEs of ∼2.3 msec
(out-of-phase) and 4.6 msec (in-phase; Fig. A15.2.3). If this is not available, separate
in-phase and out-of-phase acquisitions will be necessary. Although the TEs mentioned
above are ideal, software may be configured to yield slightly different TEs, with acceptable
results.
Hepatic MRI for Figure A15.2.1 T2-weighted coronal single shot FSE sequence used as a localizer, with an
GE Scanners effective TE of 100 msec. Arrow indicates a small cavernous hemangioma. Liver

A15.2.4 A15.2.5
Supplement 4 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 4
 Sequence 4: Timing bolus sequence (if automated bolus detection is not used)
If automated bolus detection (e.g., SMARTPREP) is not used, the delay between injection
and scanning should be determined by injecting a 2-ml test bolus, followed by a 20-ml
flush of saline (Earls et al., 1997). The size of the syringe used should be the same as that
used for injecting the bolus of gadolinium chelate, and the speed of injection should be
as identical as possible. The pulse sequence used for the test bolus on GE equipment
should be a single 20-mm sagittal slice short TR gradient echo sequence (60 multi-phases),

A B

Figure A15.2.2 Heavily T2-weighted transverse single shot FSE image, with effective TE = 180
msec. Arrow indicates the small cavernous hemangioma, with signal intensity intermediate between
that of spleen and cerebrospinal fluid.

Table A15.2.3 Imaging Parameters for Sequence 2 (Single Shot Fast Spin Echo)

Patient position Supine

Scan type Single shot fast spin echo
Imaging plane (orientation) Transverse
Variable bandwidth Yes
Pulse sequence database (PSD) SSFSE
Central slice or volume center Center to liver
Echo time (TE) 180 msec (effective)
Receiver bandwidth (RBW) ±31 kHz Figure A15.2.3 Spoiled gradient echo images: out-of-phase, with TE = 2.3 msec (A) and in-phase, with TE = 4.6 msec (B).
Repeat time (TR) Infinite
Flip angle (FA) 90° Table A15.2.4 Imaging Parameters for Sequence 3 Dual (Gradient Echo)
Fields of view (FOVx, FOVy) 320 mm, 240 mm
Resolution (Δx, Δy) 1.25 mm, 2.00 mm Patient position Supine
Number of data points collected (Nx, Ny) 256, ∼120 Scan type 2-D double gradient echo
Display matrix (Dx, Dy) 256, 256 Imaging plane (orientation) Transverse
Slice thickness (Δz) 5 mm Variable bandwidth Yes
Number of slices 24 Pulse sequence database (PSD) FGRE-dual
Slice gap 0 Central slice or volume center Center to liver
Number of excitations (NEX) 0.5 (half Fourier) Echo time (TE) 2.3 msec and 4.6 msec
Swap read and phase encoding No Receiver bandwidth (RBW) ±62 kHz
Slice location Entire liver Repeat time (TR) 200 msec
Saturation pulses Superior and inferior Flip angle (FA) 90°
Scan time 24 sec Fields of view (FOVx, FOVy) 320 mm, 240 mm
Resolution (Δx, Δy) 1.25 mm, 3.00 mm
Number of data points collected (Nx, Ny) 256, ∼80
16. Position slices to cover entire liver, as for sequence 2, adjusting levels if necessary. Display matrix (Dx, Dy) 256, 256
Ask the patient to hold his or her breath and run the sequence according to the Slice thickness (Δz) 5 mm
parameters listed in Table A15.2.4. Number of slices 24
Slice gap 0
17. Evaluate for image quality. Number of excitations (NEX) 0.5 (half Fourier)
This is the first motion-sensitive scan, and will therefore be the first to be degraded if breath Swap read and phase encoding No
holding is not successful. If there is substantial motion artifact, remind the patient about Slice location Entire liver
the importance of breath holding, and repeat the acquisition. Consider reducing coverage Saturation pulses Superior, inferior
or acquisition matrix, or increasing image thickness, to reduce acquisition time if neces- Scan time 24 sec
Hepatic MRI for sary.
GE Scanners Liver

A15.2.6 A15.2.7
Supplement 4 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 4
 Table A15.2.5 Imaging Parameters for Sequence 4 (Timing Bolus)

Patient position Supine A B

Scan type Gradient echo
Imaging plane (orientation) Sagittal
Variable bandwidth No
Pulse sequence database (PSD) FGRE
Central slice or volume center Upper abdominal aorta
Echo time (TE) Minimum (∼2.1 msec)
Receiver bandwidth (RBW) ±31 kHz
Repeat time (TR) 7 msec
Flip angle (FA) 70°
Fields of view (FOVx, FOVy) 320 mm, 240 mm
Resolution (Δx, Δy) 1.25 mm, 2.50 mm
Number of data points collected (Nx, Ny) 256, 96
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 20 mm
Number of slices 1
Slice gap N/A
Number of excitations (NEX) 1
Swap read and phase encoding No Figure A15.2.4 Selected images from sagittal timing bolus sequence, without (A) and during (B) aortic enhance-
ment (arrows) from the 2-ml bolus of gadolinium chelate.
Saturation pulses None
Multi-phase 60
Scan time 50 sec
A B

with fields of view 320 mm × 240 mm (3⁄4 fields of view), centered over the upper
abdominal aorta. Each phase should be obtained in ∼1 sec.
If the acquisition of the sequence begins when the flush begins, when the peak signal
intensity in the mid-aorta is shown, the delay time is 5 sec longer than the time it takes
an injection of the contrast agent to travel from the i.v. tubing port to the abdominal aorta.
This should correspond to the delay time between beginning of the major gadolinium
bolus and the beginning of the acquisition. For example, if it takes 25 sec for the
gadolinium to travel from the i.v. tubing port to the mid abdominal aorta, breath-holding
instructions should be given so that the actual pulse sequence begins 30 sec following the
beginning of the gadolinium injection.
C
18. Inject 2 ml of gadolinium chelate, followed immediately by a 20 ml flush of saline
with an injection of 2 ml/sec.
19. Run sequence 4 according to Table A15.2.5 at the beginning of the flush.
Breath holding is not necessary.
The aorta will have low signal intensity on all of these images, until the 2-ml bolus of
gadolinium arrives, at which point there will be a noticeable increase in aortic signal
intensity (Fig. A15.2.4). The time after the beginning of the acquisition is annotated in the
upper left hand corner of the image as the delay time (DT) and can also be determined by
noting the number of phases, and multiplying by 1 sec and then adding 5 sec.

Sequence 5: Dynamic multi-phasic pre/post contrast 3-D short TR gradient echo

Prescribe the volume to cover the liver. TR and TE should be as short as possible, and the
Figure A15.2.5 3-D fat-suppressed short TR gradient echo images with TR/TE = 4 msec/1.3 msec, and flip angle = 12°,
flip angle should be 12° to 15°. There will be visible contrast in the kidneys and collecting before (A), immediately after (B), and 45 sec after (C) gadolinium chelate injection. Through Fourier interpolation, images
Hepatic MRI for
systems if a timing bolus was used, but there will be no significant effect on the liver. are reconstructed every 2.5 mm. Arrows indicate the cavernous hemangioma, which does not show enhancement in A.
GE Scanners Obtain images (see Fig. A15.2.5) using the technique described above.
A15.2.8 A15.2.9
Supplement 4 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 4
 Table A15.2.6 Imaging Parameters for Sequence 5 (Dynamic 3-D Gradient Echo) important contrast-enhanced images earlier and allowing a longer interval before obtain-
ing the delayed post-contrast images.
Patient position Supine 24. Instruct the patient to hold his/her breath and run sequence 6 according to Table
Scan type 3-D gradient echo A15.2.7 (Fig. A15.2.6) with 13 slices. Allow the patient to breathe for 10 seconds,
Imaging plane (orientation) Transverse then instruct the patient to hold his/her breath and run the sequence for the additional
Variable bandwidth Yes 13 slices.
Pulse sequence database (PSD) LUFGRE3D
Central slice or volume center Center to liver Sequence 7: Transverse gradient echo with fat suppression
Echo time (TE) Minimum (∼1.5 msec) TR should be ∼20 msec in this sequence. These images show delayed extracellular space
Receiver bandwidth (RBW) ±62 kHz contrast, and all vessels are white because of the combined effects of gadolinium and time
Repeat time (TR) 4 msec
Flip angle (FA) 12°-15° Table A15.2.7 Imaging Parameters for Sequence 6 (Fast Spin Echo)
Fields of view (FOVx, FOVy) 320 mm, 240 mm
Resolution (Δx, Δy) 1.25 mm, 3.00 mm Patient position Supine
Number of data points collected (Nx, Ny) 256, ~80 Scan type Fast spin echo
Display matrix (Dx, Dy) 256, 256 Imaging plane (orientation) Transverse
Slice thickness (Δz) 5 mm Variable bandwidth Yes
Number of slices 64 Pulse sequence database (PSD) FRFSE-OPT
Slice gap −2.5 mm (50% overlap) Central slice or volume center Center to liver
Number of excitations (NEX) 0.5 (half Fourier) Echo time (TE) 82 msec (effective)
Swap read and phase encoding No Receiver bandwidth (RBW) ±62 kHz
Slice location Cover the entire liver Echo train length (ETL) 19
ZIP2 Yes Repeat time (TR) 2850 msec
Saturation pulses None Flip angle (FA) 90°
Fat suppression Yes, spectral inversion at lipids Fields of view (FOVx, FOVy) 320 mm, 240 mm
(SPECIAL) Resolution (Δx, Δy) 1.25 mm, 1.67 mm
Scan time 24 sec Number of data points collected (Nx, Ny) 256, ∼144
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 7 mm
20. Instruct the patient to hold his/her breath. Run the 3-D gradient echo sequence Number of slices 26
according to Table A15.2.6. Slice gap 0.5 mm
Number of excitations (NEX) 0.5 (half Fourier)
21. Explain to the patient that the contrast agent will now be injected and that they may
Swap read and phase encoding No
feel a cool sensation in their arm. Initiate the injection. IMPORTANT: Do not begin
Slice location Cover the liver
scanning until the scan delay (determined by the timing bolus procedure) has expired.
Saturation pulses Superior, inferior
However, breathing instructions should be initiated when ∼5 sec of delay are remain-
Fat suppression Yes
ing.
Scan time 46 sec
If a power injector is not accessible and one is “hand”-injecting, reattach the saline-filled
syringe and flush with 20 ml of saline. The process of switching syringes must be completed
as quickly as possible and therefore, it is recommended to use of a three-way stopcock.
22. Obtain a first set of post-contrast images according to Table A15.2.6 as soon as
possible. Repeat the same breath-holding procedure as in step 20. Allow the patient
∼5 sec to breathe and obtain a second identical set of post-contrast acquisitions.
Sequence 6: Moderately T2-weighted FSE with fat suppression
23. Prescribe the volume to cover the liver. Usually, this will require two separate
overlapping breath-holdings.
IMPORTANT: Do not obtain these as interleaved images. The “pause before scan” must
be set as “none.”
Other than causing lower signal intensity of kidneys and renal collecting structures, the
previously administered gadolinium will have little effect on the image, although there may
be slightly improved conspicuity of solid liver lesions (Jeong et al., 2001). Other advantages
Hepatic MRI for of performing this sequence after contrast agent administration include obtaining the Figure A15.2.6 Moderately T2-weighted FSE breath-hold image, with TR = 2850 msec, effective
GE Scanners Liver
TE = 82 msec. Arrow indicates hemangioma.
A15.2.10 A15.2.11
Supplement 4 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 4
 of flight (Fig. A15.2.7). Motion artifact is minimal because the short TR allows completion COMMENTARY
of each image in <2 sec.
Background Information Other Contrast Agents
25. Instruct the patient to hold his/her breath and run sequence 7 according to Table A comprehensive protocol for liver imaging With good single breath-hold whole-liver
A15.2.8. that includes in-phase and out-of-phase T1- dynamic technique, gadolinium chelates are
weighted images, moderately and heavily T2- generally considered preferable to other con-
The breath-hold instruction and the way to run this sequence is similar to step 20.
weighted, and early and late post-gadolinium trast agents for routine liver imaging. Manga-
images is described. Each sequence is acquired fodipir trisodium (Teslascan) or ferumoxides
in one or more breath holds. Total image acqui- (Feridex) may be considered in some cases,
sition time is, therefore, only a few minutes. especially those who have already had an
Most time is spent with patient set-up and with equivocal gadolinium-enhanced exam, or to
inter-sequence delays. By working efficiently, exclude additional tumors prior to curative sur-
the entire study can be performed in ≤30 min, gery or ablation.
and will decrease further with user interface
enhancements. Anticipated Results
At the authors’ institution, body MRI studies Most reports comparing MRI and computed
are obtained from several MRI instruments, tomography (CT) indicate that MRI has higher
including 8 GE Signa scanners at six separate sensitivity for detecting liver lesions. More im-
locations. Although referring clinicians are en- portantly, it is more specific for distinguishing
couraged to include ample clinical information, benign from malignant lesions. This is partly
their compliance with this is highly variable. A because the smaller volume of contrast agent
set protocol, as described here, ensures that the and paramagnetic effect of gadolinium contrast
full spectrum of liver diseases will be detected agents results in a stronger imaging effect than
and characterized without requiring physician that of iodinated contrast agents in CT. Addi-
Figure A15.2.7 Two-min post-gadolinium fat-suppressed short TR gradient echo transverse intervention. When necessary, attending radi- tionally, the use of moderately and heavily
image (extracellular-phase) demonstrates delayed contrast enhancement, including the cavernous ologists can be contacted by phone to give T2-weighted images provides benefits for de-
hemangioma (arrow). pertinent directions regarding troubleshooting tecting and characterizing, respectively, focal
or addressing unusual requests. liver lesions beyond what can be obtained using
contrast agents. Heavily T2-weighted breath-
Table A15.2.8 Imaging Parameters for Sequence 7 (Fat Suppressed Gradient
Echo) Critical Parameters and hold images, in particular, allow distinction
Troubleshooting between benign and malignant lesions in most
Patient position Supine The two major potential problems with this cases even before the contrast agent is injected
Scan type 2-D gradient echo protocol are that patients may not be able to (Ito et al., 1997).
suspend respiration or timing of gadolinium
Imaging plane (orientation) Transverse
administration may be incorrect. The first prob- Literature Cited
Variable bandwidth No
lem can be addressed by replacing sequence 5 Earls, J.P., Rofsky, N.M., Lee, V.S., DeCorato, D.R.,
Pulse sequence database (PSD) FSPGR with short TR (≤10 msec) single-slice T1- Krinsky, G.A., and Weinreb, J.C. 1997. Hepatic
Central slice or volume center Center to liver arterial-phase dynamic gadolinium-enhanced
weighted 2-D gradient echo sequences with a
Echo time (TE) Minimum (∼2.2 msec) MR imaging: Optimization with a test examina-
flip angle of ∼30°. The second problem is pre- tion and a power injector. Radiology 202:268-
Receiver bandwidth (RBW) ±16 kHz vented in most cases by using a test dose of 273.
Repeat time (TR) 20 msec gadolinium to determine timing (Earls et al., Ito, K., Mitchell, D.G., Outwater, E.K., Szklaruk, J.,
Flip angle (FA) 30° 1997), or by using an automated detection and Sadek, A.G. 1997. Hepatic lesions: Dis-
Fields of view (FOVx, FOVy) 320 mm, 240 mm method such as SMARTPREP. crimination of nonsolid, benign lesions from
Resolution (Δx, Δy) 1.25 mm, 2.00 mm Another problem occurs in patients with solid, malignant lesions with heavily T2-
weighted fast spin-echo MR imaging. Radiology
Number of data points collected (Nx, Ny) 256, 120 implanted metal, such as from surgical clips.
204:729-737.
Display matrix (Dx, Dy) 256, 256 Artifacts from metal on gradient echo images
Jeong, Y.Y., Mitchell, D.G., and Holland, G. 2001.
Slice thickness (Δz) 5 mm can be minimized by using the shortest possible
Liver lesion conspicuity on T2-weighted breath-
TE and avoiding fat suppression. TE on the 3-D hold fast spin-echo MR imaging before and after
Number of slices ∼30
short TR gradient echo sequence is minimized gadolinium administration: Initial experience.
Slice gap 0 Radiology 221:117-121.
by increasing the number of excitations (NEX)
Number of excitations (NEX) 1 from 0.5 to 1.0, and increasing the sampling Marks, B., Mitchell, D.G., and Simelaro, J.P. 1997.
Swap read and phase encoding No bandwidth from ±31 to ±62 kHz. The acquisi- Breath-holding in healthy and pulmonary-com-
Slice location Cover the liver tion time will be slightly longer, so adjustments promised populations: Effects of hyperventila-
Fat suppression Yes tion and oxygen inspiration. J. Magn. Reson.
in the number of image slices and/or acquisition
Imaging 7:595-597.
Scan time ∼60 sec matrix may be necessary.

Hepatic MRI for

GE Scanners Liver

A15.2.12 A15.2.13
Supplement 4 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 4
 Shellock, F.G. and Kanal, E. 1996. Magnetic Reso-
nance Bioeffects, Safety, and Patient Manage-
ment. Lippincott Williams and Wilkins, Phila-
delphia.
Key References
Shellock and Kanal, 1996. See above.
Covers a number of important patient management
issues related to MR imaging, including recom-
mended safety procedures, a list of metallic implants
that have been tested for MR compatibility, and a
list of other sources on MR safety.
Internet Resources
http://www.mri.tju.edu
A noncommercial site that lists all body MRI proto-
cols, continually updated, updated GE Signa scan-
Hepatic MRI for
GE Scanners
A15.2.14
Supplement 4
ners by the Thomas Jefferson University Depart-
ment of Radiology. Additionally, there are descrip-
tions and explanations of the various pulse
sequences, tips for problems solving, and examples
of clinical applications.
http://www.mrisafety.com
Managed by Frank Shellock, contains updated items
regarding MRI safety and compatibility.
Contributed by Donald G. Mitchell,
Peter Natale, and George Holland
Thomas Jefferson University
Philadelphia, Pennsylvania
Current Protocols in Magnetic Resonance Imaging
MRI of the Kidney UNIT A16.1
MRI provides comprehensive information on the full range of kidney diseases including
benign and malignant processes and diffuse renal parenchymal disease. We employ a set
protocol incorporating various types of sequences including transverse, coronal and
sagittal data acquisition, and the routine use of intravenous gadolinium.
IMAGING THE KIDNEY BASIC
PROTOCOL
The high signal-to-noise ratio (SNR) obtained at high field strength makes it possible to
image the kidneys during a single breath-hold sequence. The sequences described herein
are based on the authors’ experience with a Siemens 1.5 T Vision scanner, but are expected
to be equally applicable to machines from other manufacturers.
Scanning a patient or volunteer is a joint effort among technologists, nurses, and
physicians, with the technologist normally responsible for following proper scanning
protocols and techniques. Unless otherwise specified, in what follows the person to whom
directions are given is assumed to be the technologist. Table A16.1.1 lists the hardware
necessary to perform the procedure, along with appropriate parameters.
The following nine sequences comprise the kidney imaging protocol. This protocol
employs pre- and post-gadolinium T1 weighted imaging acquired utilizing nonsup-
pressed and fat suppressed sequences. Most sequences require the patient to be able to
suspend respiration for ∼25 sec. It is imperative that there be clear communication
between the technologist and the patient throughout the exam. This protocol results in
consistent, reproducible image quality that is effective for evaluating the full range of
renal diseases.
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.
Materials
Normal saline (0.9% NaCl), sterile, 40 ml minimum
Extravascular contrast agent (e.g., Magnevist, Omniscan, or Prohance), volume
prescribed by patient weight
Table A16.1.1 Equipment Specifications Imaging of the Kidney
Coil type Circular polarized body-phased array coil
Field strength 1.5 T
Gradient strength 24 mT/ m (or whatever the system
permits, but minimum of 24 mT/m for
sequences 3 and 4)
Knee cushion Yes
Use of contrast agents Yes
Pulse oximeter If patient requires sedation
Power injector Yes
Normal saline Yes
35-in. extension tubing Yes
Kidney
Contributed by Richard Semelka, Laurie Fisher, and Kathy Wilber A16.1.1
Current Protocols in Magnetic Resonance Imaging (2001) A16.1.1-A16.1.13
Copyright © 2001 by John Wiley & Sons, Inc.
 Set up equipment and patient
1. Interview (screen) the patient to assess for contraindications such as cardiac pace-
maker, implanted mechanical devices, and/or ferromagnetic materials. Also, deter-
mine if the patient has any health conditions that may require the presence of special
emergency equipment during the scanning procedure, of if he or she will need
sedation medication necessitating the use of appropriate monitoring equipment.
Generally standard screening forms are used for all patients scanned in a magnetic
resonance system.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(1996) for discussion of what implants may be safely scanned using magnetic resonance.
The presence of ferromagnetic materials in the globe of the eye is contraindicated for MRI.
Patients with prior metal exposure to the eye should have plain X rays of the orbital area
to ensure that all metal has been removed prior to placing them in the magnetic field.
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
2. Request that the patient change into a gown to eliminate any metal that might be found
in clothing. Ask the patient to remove all personal effects such as, jewelry, hearing
aids, glasses, etc., prior to entering the MRI scan room.
All personal belongings should be secured during the examination.
3. Explain the procedure to the patient and record relevant clinical history. Ensure that
the patient understands what is expected and ask them if they have any questions;
answer appropriately. If the procedure is a research protocol, have the patient sign
any necessary consent form.
4. Fill a 20 ml syringe with normal saline and attach to saline filled extension tubing
(35-in.). Obtain intravenous (i.v.) access utilizing a 22-G angiocatheter and attach
saline-prepared extension tubing and syringe. This will allow you to flush the
extension tubing while the patient waits to be imaged, and, in cases in which a power
injector is not available, this will allow preparation for bolus injection (see Note
below). Secure the position of the angiocatheter with tegraderm or tape.
Obtaining i.v. access prior to entering the scan room will promote patient throughput and
eliminate “dead” time of starting the i.v. while the patient is on the exam table. Follow
power injector manufacturer guidelines with regard to appropriate gauge needle/angio-
catheter to be used; this will depend on chosen injection/flow rates.
NOTE: If you do not have access to a power injector you will still be able to perform
dynamic renal imaging as the extension tubing will allow the saline syringe to be placed
at the foot of the patient table during pre-contrast imaging. In this case, you will need to
draw up the contrast agent in another syringe. When you are ready to bolus inject the
contrast agent, simply disconnect the saline syringe and connect the syringe filled with
contrast agent; once you have injected the bolus of contrast reconnect the saline syringe
and bolus an appropriate volume of flush, usually ∼10 ml. Alternatively, to eliminate the
need of switching syringes, incorporate the use of a 3-way stopcock.
5. Set up the exam room by securing the circularly polarized (CP) body array coil onto
the table and providing a clean exam table.
6. Set up the power injector as specified by the manufacturer.
MRI of the
Kidney
A16.1.2
Current Protocols in Magnetic Resonance Imaging
A minimum of 40 ml normal saline should be drawn up to ensure sufficient saline is
available to keep the vein open (KVO) throughout the exam. To determine the amount of
contrast agent to be used, reference the contrast packet insert and draw up the amount
indicated per kg of patient weight. There is no need to double dose.
7. Escort the patient to the MR examination room and ask the patient to lie down
accordingly with respect to the exam to be performed. Help the patient mount onto
the table. Either before or right after the patient lies down, set up any triggering
devices or other monitoring equipment that is to be used.
8. Connect the extension tubing secured to the syringe to the power injector extension
tubing.
9. Inform the patient about what will occur during the procedure, what he or she will
experience while in the magnet, and how to behave, including the following:
a. If earphones or headphones are used to protect the ears from the loud sounds
produced by the gradients, the patient will be asked to wear these, but will be able
to communicate with you at any time during the imaging.
b. The patient will be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this works).
c. For good results the patient should not talk, and should avoid or minimize
swallowing or other movement, during each scan—i.e., as long as the banging
sounds continue. Between scans, talking and swallowing are allowed in most
cases, but should be avoided when comparative positional studies are being
performed; the patient will be informed when this is the case. Explain to the patient
that one should not reposition one’s body between imaging sequences.
Additionally, review breath-holding instructions with the patient, and provide the patient
with an approximate time that the examination will take.
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
e. Position a support under the patient’s knees to enhance patient comfort.
10. Secure the top portion of the CP-body array coil to prevent it from moving side-to-side
during breath-holding imaging sequences.
Usually straps are provided by the manufacturer that are directly attached to the coil.
11. Use the laser light to position the patient, and to center the coil (see Table A16.1.2).
Then, advance the patient table to isocenter.
12. If the patient is unable to hold still, provide an appropriate sedative.
13. Program the power injector for a contrast agent and saline injection rate of 2 ml/sec.
Total volume of saline following contrast agent injection should be programmed for
10 ml. Program a scan delay of 18 sec (contrast agent is injected, scan is initiated 18
sec after contrast agent and saline are delivered). Arm the power injector and begin
the keep the vein open.
Do not inject the contrast agent at this time!
Sequence 1: Three-plane positioning scout
14. To validate the patient’s position and to have a reference to prescribe successive
imaging sequences, acquire a three-plane orthogonal scout sequence. See Table
A16.1.2 for specific parameters.
Most MR scanners can be programmed to acquire the scout automatically after coil tuning
or after the patient has been placed in isocenter (for systems that do not require tuning).
Kidney
A16.1.3
Current Protocols in Magnetic Resonance Imaging
 Table A16.1.2 Imaging Parameters for Scout Sequences (Sequences 1 and 2) Sequence 5: Transverse gradient echo with fat saturation (Fig. A16.1.2)
22. Display both the coronal and transverse scout images (use breath-hold images) in
Patient position Supine two separate quadrants on the scan monitor.
Scan type Gradient echo
Change imaging parameters to those listed in Table A16.1.5. Position slices to center of
Imaging plane (orientation) Sagittal, transverse, and coronal
the coronal scout ensuring that both kidneys are entirely covered.
Central slice or volume center Laser light centered approximately one
hand width above the inferior rib It is imperative that the slices are prescribed off of the breath-hold scout images as this is
margin a breath-held imaging sequence. Otherwise, the slice location will not be accurate relative
Echo time (TE) 6 msec to the reference image if a non-breath-held scout image is used.
Repeat Time (TR) 15 msec
Flip angle (FA) 30° Table A16.1.3 Imaging Parameters for Half-Acquisition Turbo Spin Echo
(Sequence 3)
Fields of view (FOVx, FOVy) 450 mm, 450 mm
Resolution (Δx, Δy) 1.76 mm, 3.52 mm Patient position Supine
Number of data points collected (Nx, Ny) 256, 128 Scan type Half acquisition turbo spin echo
Display matrix (Dx, Dy) 256, 256 Imaging plane (orientation) Coronal
Slice thickness (Δz) 10 mm Central slice or volume center Slices posted on transverse scout;
Number of slices 3 center to kidneys
Slice gap Not applicable Echo time (TE) 90 msec
Number of acquisitions (Nacq) 1 Repeat time (TR) 4.4 msec (note: The true TR is infinite;
Swap read and phase encoding No 4.4 msec represents the echo spacing)
Slice location Not applicable Delay time (TD) 1500 msec
Saturation pulses Not applicable Flip angle (FA) 150°
Scan time 16 sec Field of view (FOVx, FOVy) 400 mm, 400 mm
Resolution (Δx, Δy) 1.56 mm, 2.08 mm
Number of data points collected (Nx, Ny) 256, 192 (using half Fourier)
Sequence 2: Breath-hold three-plane positioning scout
Display matrix (Dx, Dy) 256, 256
15. To have a reference to prescribe successive breath-hold imaging sequences, acquire
Slice thickness (Δz) 8–10 mm
a second three plane orthogonal scout sequence. See Table A16.1.2 for specific
Number of slices 20
parameters.
Slice gap 1.6–2 mm
16. Instruct the patient to take in a deep breath and exhale, take in another deep breath Number of acquisitions (Nacq) 1
and hold it. Swap read and phase encoding No
Slice location Centered to cover both kidneys
17. Initiate the scan. Saturation pulses No
Slice series Interleaved
Sequence 3: Half-acquisition (partial Fourier) turbo spin echo coronal
Scan time 40 sec
18. Display both the coronal and transverse scout images (use non-breath-hold images)
in two separate quadrants on the scan monitor.
Change imaging parameters to those listed in Table A16.1.3. Position slices to center of
the transverse scout, ensuring that both kidneys are entirely covered.
19. Instruct the patient to remain motionless and to breathe normally as the scan will
begin and last for ∼40 sec.

Sequence 4: Half-acquisition turbo spin echo transverse (Fig. A16.1.1)

20. Display both the coronal and transverse scout images (use non-breath-hold images)
in two separate quadrants on the scan monitor.
Change imaging parameters to those listed in Table A16.1.4. Position slices to center of
the coronal scout, ensuring that both kidneys are entirely covered.
21. Instruct the patient to remain motionless and to breathe normally as the scan will
begin and last for ∼40 sec.

MRI of the
Kidney Kidney
Figure A16.1.1 Unenhanced transverse half-acquisition turbo spin echo image.
A16.1.4 A16.1.5
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 23. Instruct the patient to take in a deep breath and exhale, take in another deep breath Change imaging parameters to those listed in Table A16.1.6. Position slices to cover
and hold it. the kidneys.
24. Initiate the scan. It is imperative that the slices be prescribed off of breath-hold coronal image as this is a
breath-held imaging sequence. Otherwise, the slice location will not be accurate relative
Sequence 6: Transverse gradient echo (Fig. A16.1.3) to the reference image if a non-breath-held image is used.
25. Display the midline slice of the transverse gradient echo image (sequence 5) and the
26. Instruct the patient to take in a deep breath and exhale, take in another deep breath
breath-hold coronal scout image in two separate quadrants on the scan monitor.
and hold it.
Table A16.1.4 Imaging Parameters for Half-Acquisition Turbo Spin Echo 27. Initiate the scan.
(Sequence 4)
Table A16.1.5 Imaging Parameters for Gradient Echo with Fat Saturation
Patient position Supine (Sequence 5)
Scan type Half-acquisition turbo spin echo
Imaging plane (orientation) Transverse Patient position Supine
Central slice or volume center Slices posted on coronal; center to Scan type Gradient echo
kidneys Imaging plane (orientation) Transverse
Echo time (TE) 90 msec Central slice or volume center Slices posted on coronal; center to
Repeat time (TR) 4.4 msec (note: The true TR is infinite; kidneys
4.4 msec represents the echo spacing) Echo time (TE) 4.1 msec
Delay time (TD) 1500 msec Repeat time (TR) 147.2 msec
Flip angle (FA) 150° Flip angle (FA) 80°
Fields of view (FOVx, FOVy) 350 mm, 263 mm Fields of view (FOVx FOVy) 350 mm, 263 mm
Resolution (Δx, Δy) 1.37 mm, 1.37 mm Resolution (Δx, Δy) 1.37 mm, 1.83 mm
Number of data points collected (Nx, Ny) 256, 192 (using half Fourier) Number of data points collected (Nx, Ny) 256, 144
Display matrix (Dx, Dy) 256, 256 Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 8–10 mm Slice thickness (Δz) 8–10 mm
Number of slices 20 Number of slices 20
Slice gap 1.6–2 mm Slice gap 1.6–2 mm
Number of acquisitions (Nacq) 1 Number of acquisitions (Nacq) 1
Swap read and phase encoding No Swap read and phase encoding No
Slice location Centered to cover both kidneys Slice location Centered to cover both kidneys
Saturation pulses Yes, superior and inferior to slices Saturation pulses No
Fat suppression No Fat suppression Yes
Slice series Interleaved Slice series Interleaved
Scan time 40 sec Scan time 21 sec

MRI of the Figure A16.1.2 Unenhanced fat-suppressed transverse spoiled gradient echo
Kidney image. Kidney
Figure A16.1.3 Unenhanced transverse spoiled gradient echo image.
A16.1.6 A16.1.7
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Sequence 7: Transverse gradient echo—immediate post-contrast (Fig. A16.1.4) 29. Explain to the patient that you will now be injecting the contrast agent and he/she
NOTE: See patient setup section for specific instructions on preparation for contrast agent may feel a cool sensation in his/her arm. Initiate the injection. Do not begin scanning
injection. This preparation must be done prior to placing the patient in the scanner. until the 18 sec scan delay has expired. However, deliver breathing when 10 sec of
28. Repeat step 25. delay are remaining (see step 30).
If you do not have access to a power injector and are “hand” injecting, you can still use
NOTE: If possible, cover the liver in addition to the kidneys on the immediate post-contrast
step 29. However, after you have completed the bolus contrast injection, reattach the
imaging.
saline-filled syringe and flush with 10 ml of saline. Begin breathing instructions after 5 ml
It is imperative that the slices be prescribed off of the breath-hold coronal image as this is of the saline has been injected, then proceed to initiate the scan. The process of switching
a breath-held imaging sequence. Otherwise, the slice location will not be accurate relative syringes must be completed as quickly as possible and thus it is suggested that a 3-way
to the reference image if a non-breath-held image is used. stopcock be incorporated.
30. When there are 10 sec of delay remaining, instruct the patient to take in a deep breath
Table A16.1.6 Imaging Parameters for Gradient Echo (Sequence 6) and exhale, take in another deep breath and hold it.
Patient position Supine 31. Initiate the scan.
Scan type Gradient echo
Imaging plane (orientation) Transverse Sequence 8: Transverse gradient echo with fat saturation (45 sec delay after
Central slice or volume center Slices posted on coronal; center to injection; Fig. A16.1.5)
kidneys 32. Display the midline slice of the transverse gradient echo image (sequence 5) and the
Echo time (TE) 4.1 msec breath-hold coronal scout image in two separate quadrants on the scan monitor.
Repeat time (TR) 140 msec Change imaging parameters to those listed in Table A16.1.7. Position slices to cover
Flip angle (FA) 80° kidneys.
Fields of view (FOVx, FOVy) 350 mm, 263 mm
It is imperative that the slices are prescribed off of the breath-hold coronal image as this
Resolution (Δx, Δy) 1.37 mm, 2.05 mm is a breath-held imaging sequence. Otherwise, the slice location will not be accurate
Number of data points collected (Nx, Ny) 256, 128 relative to the reference image if a non-breath-held image is used.
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 8–10 mm 33. Once 45 sec has expired, instruct the patient to take in a deep breath and exhale, then
Number of slices 19 take in another deep breath and hold it.
Slice gap 1.6–2 mm 34. Initiate the scan.
Number of acquisition (Nacq) 1
Swap read and phase encoding No Sequence 9: Gradient echo sagittal
Slice location Centered to cover both kidneys 35. Display the midline slice of the transverse gradient echo image (sequence 5) and the
Saturation pulses No breath-hold sagittal scout image in two separate quadrants on the scan monitor.
Slice series Interleaved
Scan time 18 sec

Figure A16.1.5 Contrast-enhanced fat-suppressed spoiled gradient echo trans-

MRI of the Figure A16.1.4 Immediate post-gadolinium transverse spoiled gradient echo verse image.
Kidney image. Kidney

A16.1.8 A16.1.9
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A16.1.7 Imaging Parameters for Gradient Echo with Fat Saturation—45 Change imaging parameters to those listed in Table A16.1.8. Position slices to cover
Sec Delay (Sequence 8) the kidneys.
Patient position Supine It is imperative that the slices are prescribed off of breath-hold transverse and sagittal
Scan type Gradient echo images, as this is a breath-held imaging sequence. Otherwise, the slice location will not
Imaging plane (orientation) Transverse be accurate relative to the reference image if non-breath-held images are used.
Central slice or volume center Slices posted on coronal; center to 36. Instruct the patient to take in a deep breath and exhale, then take in another deep
kidneys breath and hold it.
Echo time (TE) 4.1 msec
Repeat time (TR) 147.2 msec 37. Initiate the scan.
Flip angle (FA) 80°
Fields of view (FOVx, FOVy) 350 mm, 263 mm COMMENTARY
Resolution (Δx, Δy) 1.37 mm, 1.83 mm
Number of data points collected (Nx, Ny) 256, 144 Background Information played in the same image. Breath-hold imag-
Appropriate MR imaging of the kidneys ing, as performed in this protocol, is useful in
Display matrix (Dx, Dy) 256, 256
requires the standard use of intravenous gad- the evaluation of cyst versus solid mass in the
Slice thickness (Δz) 8–10 mm
olinium (Balci et al., 1999; John et al., 1997; kidney. It allows for less motion artifact blur-
Number of slices 20 Hricak et al., 1988; Huch Boni et al., 1996; ring and less contribution of noise to the calcu-
Slice gap 1.6–2 mm Rofsky et al., 1991; Semelka et al., 1991; Se- lation of lesion signal, if region-of-interest
Number of acquisitions (Nacq) 1 melka et al., 1992; Semelka et al., 1993). Our measurements are made. Sagittal plane images
Swap read and phase encoding No standard renal protocol acquires the majority are a useful component to the protocol, as they
Slice location Centered to cover both kidneys of sequences in the transverse plane. We believe allow good determination of extra- or intrarenal
Saturation pulses No this minimizes the potential problem with par- masses in the region of the superior pole of the
Fat suppression Yes tial volume effects, that coronal or sagittal im- kidneys. The immediate post-contrast images
Slice series Interleaved ages may have, because the volume of kidney may also be acquired as an MR angiographic
Scan time 21 sec tissue at its edges is smaller in the transverse sequence, which thereby also provides infor-
plane than in other orthogonal planes. Coronal mation on the renal arteries.
images do have the attractive feature of display- In patients who are able to breath-hold for
Table A16.1.8 Imaging Parameters for Sagittal Gradient Echo (Sequence 9) ing the kidneys in an intravenous urographic the 20 sec required for spoiled gradient echo,
format, and the upper and lower poles are dis- MR appears to be marginally superior to CT
Patient position Supine
Scan type Gradient echo Table A16.1.9 Imaging Parameters for Non-Slice Selective Turbo Flash (Use
Imaging plane (orientation) Sagittal Pre-Contrast for Patients Unable to Breath-Hold)
Central slice or volume center Slices posted on transverse; center to
kidneys Patient position Supine
Echo time (TE) 4.1 msec Scan type Inversion recovery prepared snap shot
Repeat time (TR) 140 msec gradient echo
Flip angle (FA) 80° Imaging plane (orientation) Transverse
Fields of view (FOVx FOVy) 350 mm, 263 mm Central slice or volume center Slices posted on coronal; center to
Resolution (Δx, Δy) 1.37 mm, 2.05 mm kidneys
Number of data points collected (Nx, Ny) 256, 128 Echo time (TE) 4.2 msec
Display matrix (Dx, Dy) 256, 256 Repeat time (TR) 11 msec
Slice thickness (Δz) 8–10 mm Flip angle (FA) 15°
Number of slices 19 Fields of view (FOVx, FOVy) 350 mm, 263 mm
Slice gap 1.6–2 mm Resolution (Δx, Δy) 1.37 mm, 2.05 mm
Number of acquisitions (Nacq) 1 Number of data points collected (Nx, Ny) 256, 128
Swap read and phase encoding No Display matrix (Dx, Dy) 256, 256
Slice location Centered to cover both kidneys Slice thickness (Δz) 8–10 mm
Saturation pulses No Number of slices 21
Slice series Interleaved Slice gap 1.6–2 mm
Scan time 18 sec Number of acquisitions (Nacq) 1
Swap read and phase encoding No
Slice location Centered to cover kidneys
Saturation pulses No
MRI of the Slice series Interleaved
Kidney Kidney
Scan time 2 min, 20 sec
A16.1.10 A16.1.11
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A16.1.10 Imaging Parameters for Slice Selective Turbo Flash (Use (computed tomography) for the detection of ings on MR images. Am. J. Roentgenol.
Post-Contrast for Patients Unable to Breath-Hold) small renal masses and the distinction between 172:1495-1500.
cystic and solid lesions (Semelka et al., 1992; Hricak, H., Thoeni, R.F., Carroll, P.R., Demas, B.E.,
Patient position Supine Semelka et al., 1993). In general, the advan- Marotti, M., and Tanagho, E.A. 1988. Detection
Scan type Inversion recovery prepared snap shot and staging of renal neoplasms: A reassessment
tages of MR over CT for renal imaging include:
of MR imaging. Radiology 166:643-649.
gradient echo decreased incidence of allergic reaction to the
Imaging plane (orientation) Transverse Huch Boni, R.A., Debatin, J.F., and Krestin, G.P.
contrast agent; better enhancement of renal
1996. Contrast-enhanced MR imaging of the
Central slice or volume center Slices posted on coronal; center to parenchyma and tumors in the setting of kidneys and adrenal glands. Magn. Reson. Imag-
kidneys chronic renal failure (John et al., 1997; Rofsky ing Clinics of North America 4(1):101-131.
Echo time (TE) 4.2 msec et al., 1991); better evaluation of the presence John, G., Semelka, R.C., Burdeny, D.A., Kelekis,
Repeat time (TR) 11 msec and extent of tumor thrombus; better determi- N.L., Woosley, J.T., Kettritz, U., and Freeman,
Flip angle (FA) 15° nation of cyst versus solid renal lesions; and J.A. 1997. Renal cell cancer: Incidence of hem-
better evaluation of the liver in the setting of orrhage on MR images in patients with chronic
Fields of view (FOVx, FOVy) 350 mm, 263 mm renal insufficiency. J. Magn. Reson. Imaging.
1.37 mm, 2.05 mm suspected metastases. In addition, gadolinium
Resolution (Δx, Δy) 7:157-60.
does not appear to worsen renal runction, unlike
Number of data points collected (Nx, Ny) 256, 128 Rofsky, N.M., Weinreb, J.C., Bosniak, M.A., Libes,
iodine (John et al., 1997; Rofsky et al., 1991).
Display matrix (Dx, Dy) 256, 256 R.B., and Birnbaum, B.A. 1991. Renal lesion
Slice thickness (Δz) 8–10 mm characterization with Gadolinium-enhanced MR
Critical Parameters and imaging: Efficacy and safety in patients with
Number of slices 21 Troubleshooting renal insufficiency. Radiology 180:85-89.
Slice gap 1.6–2 mm The most critical component for determin- Semelka, R.C., Hricak, H., Stevens, S.K., Fingold,
Number of acquisitions (Nacq) 1 ing cyst versus solid renal tumor is that the R., Tomei, E., and Carroll, P.R. 1991. Combined
Swap read and phase encoding No patient must breath-hold well for 20 sec. If they gadolinium-enhanced and fat saturation MR im-
aging of renal masses. Radiology 178:803-809.
Slice location Centered to cover kidneys are not able to do so, they may be better served
Saturation pulses No with a CT or ultrasound study. In order to Semelka, R.C., Shoenut, J.P., Kroeker, M.A.,
MacMahon, R.G., and Greenberg, H.M. 1992.
Slice series Interleaved evaluate the superior and inferior margins of
Renal lesions: controlled comparison between
Scan time 36 sec renal lesions, imaging in two planes is useful. CT and 1.5T MR imaging with nonenhanced and
Our standard approach is to image in the trans- gadolinium-enhanced fat-suppressed spin-echo
verse plane supplemented with images ac- and breath-hold FLASH techniques. Radiology
quired in the sagittal plane. Other investigators 182:425-430.
Table A16.1.11 Imaging Parameters for Breathing Averaged Fat Suppressed
Spin Echo (Use for Patients Unable to Breath-Hold) image primarily in the coronal plane. Semelka, R.C., Shoenut, J.P., Magro, C.M.,
The most problematic artifact in renal imag- Kroeker, M.A., MacMahon, R., and Greenberg,
Patient position Supine H.M. 1993. Renal cancer staging: Comparison
ing, particularly when the question is cyst versus
of contrast-enhanced CT and gadolinium-en-
Scan type Spin echo solid renal lesion, is breathing artifact. As a rule, hanced fat-suppressed spin-echo and gradient-
Imaging plane (orientation) Transverse our recommendation is to perform CT or ultra- echo MR imaging. J. Magn. Reson. Imaging
Central slice or volume center Slices posted on coronal; center to sound in patients who cannot breath-hold satis- 3:597-602.
kidneys factorily. Nevertheless, it is possible to substi- Shellock, F.G. 1996. Pocket Guide to MR Proce-
Echo time (TE) 15 msec tute in breathing independent inversion recov- dures and Metallic Objects. Lippincott-Raven ,
ery prepared gradient echo (see Table A16.1.9 Philadelphia.
Repeat time (TR) 500 msec
Flip angle (FA) 90° and Table A16.1.10), or breathing-averaged fat
suppressed spin echo (see Table A16.1.11) for Key References
Fields of view (FOVx, FOVy) 350 mm, 263 mm Shellock, F.G. 1996.
the breath-hold spoiled gradient echo.
Resolution (Δx, Δy) 1.37 mm, 2.05 mm Covers a number of important patient management
Number of data points collected (Nx, Ny) 256, 128 issues related to MR imaging, including recom-
Anticipated Results mended safety procedures, a list of metallic implants
Display matrix (Dx, Dy) 256, 256 Employing the protocol described above, that have been tested for MR compatibility, and a
Slice thickness (Δz) 8–10 mm the full range of renal diseases may be evaluated list of other sources on MR safety.
Number of slices 10 well by MRI. MRI does not, however, show
Slice gap 1.6–2 mm calcium well, and since the majority of renal
Number of acquisitions (Nacq) 4 calculi contain calcium, MRI is not recom- Contributed by Richard Semelka and
Swap read and phase encoding No mended generally for renal stone disease. These Kathy Wilber
Slice location Centered to cover kidneys; may need patients are best evaluated by non-contrast spi- University of North Carolina–Chapel Hill
two sets ral CT. Chapel Hill, North Carolina
Saturation pulses No
Laurie Fisher
Fat suppression Yes Literature Cited Siemens Uptime Service Center
Balci, N.C., Semelka, R.C., Patt, R.H., Dubois, D.,
Slice series Interleaved Cary, North Carolina
Freeman, J.A., Gomez-Caminero, A., and
Scan time 4 min, 19 sec Woosley, J.T. 1999. Complex renal cysts: Find-
MRI of the
Kidney Kidney

A16.1.12 A16.1.13
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
MRI of the Adrenal Glands
MRI is an effective method to evaluate adrenal pathology. The authors employ a set
protocol incorporating various types of T1- and T2-weighted sequences including the use
of in-phase and out-of-phase gradient echo techniques and the routine use of intravenous
gadolinium.
ADRENAL IMAGING
The high signal-to-noise ratio (SNR) obtained at high field strength makes it possible to
image the adrenal glands during a single breath-hold sequence. The sequences described
herein are based on the authors’ experience with a Siemens 1.5 T Vision scanner, but are
expected to be equally applicable to machines from other manufacturers.
Scanning a patient or volunteer is a joint effort among technologists, nurses, and
physicians, with the technologist normally responsible for following proper scanning
protocols and techniques. Unless otherwise specified, in what follows, the person to whom
directions are given is assumed to be the technologist. Table A17.1.1 lists the hardware
necessary to perform the procedure, along with appropriate parameters.
The following ten sequences comprise the adrenal imaging protocol. This protocol
employs multiple data acquisitions and serial post-gadolinium imaging. Most sequences
require the patient to be able to suspend respiration for ∼25 sec. It is imperative that there
be clear communication between the technologist and the patient throughout the exam.
This protocol results in consistent, reproducible image quality that is effective for
evaluating the full spectrum of adrenal diseases.
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.
UNIT A17.1 patient will need sedation medication necessitating the use of appropriate monitoring
equipment.
A screening form is signed by each patient or legal guardian prior to bringing the patient
into the exam area.
The presence of ferromagnetic materials may be a health hazard to the patient while in the
magnetic field and/or adversely affect image quality. To determine the safety of scanning
such ferromagnetic materials see Shellock and Kanal (1996).
BASIC The presence of ferromagnetic materials in the globe of the eye is contraindicated for MRI.
PROTOCOL Patients with prior metal exposure to the eye should have plain X rays of the orbital area
to ensure that all metal has been removed prior to placing them in the magnetic field.
2. Request that the patient change into a gown and remove all personal effects such as
jewelry, hearing aids, and glasses, prior to entering the MRI scan room. All personal
belongings should be secured during the examination.
3. Explain the procedure to the patient and record relevant clinical history. Ensure that
the patient understands what is expected, ask if the patient has any questions, and
answer appropriately.
4. Fill a 20-ml syringe with normal saline and attach to a 35-in. saline-filled extension
tubing. Obtain i.v. access utilizing a 22-G angiocatheter and attach the saline-pre-
pared extension tubing and syringe. This allows flushing of the extension tubing while
the patient waits to be imaged. In cases where a power injector is not available, this
allows for preparation for bolus injection (see note below). Secure the position of
angiocatheter with tegraderm or tape.
Obtaining i.v. access prior to entering the scan room will promote patient throughput and
eliminate “dead” time of starting the i.v. while the patient is on the exam table. Follow
power injector manufacturer guidelines with regard to appropriate-gauge needle/angio-
catheter to be used; this will depend on chosen injection/flow rates.

Materials NOTE: If access to a power injector is not available, performing dynamic imaging is still
possible as the extension tubing will allow the saline syringe to be placed at the foot of the
Normal saline (0.9% NaCl), sterile, 40 ml minimum patient table during pre-contrast imaging. In this case, the contrast agent will need to be
Extravascular contrast agent (e.g., Magnevist, Omniscan, or Prohance), volume is drawn up by another syringe. When the bolus injection of the contrast agent is prepared,
prescribed by patient weight simply disconnect the saline syringe and connect the syringe filled with the contrast agent;
once the bolus of contrast agent is injected, reconnect the saline syringe and bolus an
Set up patient and equipment appropriate volume of flush, usually ∼10 ml. Alternatively, to eliminate the need of
1. Interview the patient to assess for contraindications such as cardiac pacemakers, switching syringes, incorporate the use of a 3-way stopcock.
implanted mechanical devices, and/or ferromagnetic materials. Also, determine if the
5. Set up the exam room by securing the circularly polarized (CP)-body array coil onto
the table and providing a clean exam table.
Table A17.1.1 Equipment Specifications Needed to Perform the Following
Imaging Sequences 6. Set up the power injector as specified by the manufacturer. A minimum of 40 ml
normal saline should be drawn-up to ensure sufficient saline is available to keep the
Coil type Circular polarized body phased array coil vein open (KVO) throughout the exam. To determine the amount of contrast agent to
Manufacturer and system type Siemens, Vision be used, reference the contrast packet insert and draw-up the amount indicated per
Field strength 1.5 T kilogram of patient weight. There is no need to double dose.
Gradient strength 24 mT/m (or whatever the system permits, but
minimum of 24 mT/m for sequences 2 and 4) 7. Escort the patient to the MR examination room and ask them to lie down accordingly
Knee cushion Yes with respect to the exam to be performed. Connect the extension tubing secured to
Use of contrast agents Yes the syringe to the power injector extension tubing. Review the following items with
Pulse oximeter If patient requires sedation the patient:
Power injector Yes
a. Provide earplugs or headphones to the patient to minimize the loud knocking noise
Normal saline Yes
that will be produced by the gradients, but ensure them that the patient will still
35" extension tubing Yes MRI of the
Adrenal Glands
be able to hear you.
Adrenal Glands
Contributed by Kathy Wilber, Richard Semelka, and Laurie Fisher A17.1.1 A17.1.2
Current Protocols in Magnetic Resonance Imaging (2001) A17.1.1-A17.1.12
Copyright © 2001 by John Wiley & Sons, Inc. Current Protocols in Magnetic Resonance Imaging
 b. Provide the patient with a safety squeeze-bulb and demonstrate how it works; Table A17.1.2 Imaging Parameters for Sequence 1 (Scout Sequence)
explain to the patient when to use the squeeze-bulb (i.e., if assistance is needed
during the exam). Patient position Supine
Scan type Gradient echo
c. Explain to the patient that you will be talking to them between imaging sequences
Imaging plane (orientation) Sagittal, transverse, and coronal
which will be when the loud knocking noise stops. Additionally, review breath
Central slice or volume center Laser light centered
holding instructions with the patient.
approximately one hand width
d. Explain to the patient that it is imperative to remain motionless during the loud above the inferior rib margin
knocking noise to ensure good results; also explain that one should not reposition Echo time (TE) 6 msec
one’s body between imaging sequences. Repeat time (TR) 15 msec
e. Position a support under the patient’s knees to enhance patient comfort. Flip angle (FA) 30°
Field of view (FOVx, FOVy) 450 mm, 450 mm
f. Provide the patient with an approximate time that the examination will take.
Resolution (Δx, Δy) 1.76 mm, 3.52 mm
8. Secure the top portion of the CP–body array coil to prevent it from moving side-to- Number of data points collected (Nx, Ny) 256, 128
side during breath-holding imaging sequences. Display matrix (Dx, Dy) 256, 256
Usually there are straps provided by the manufacturer that are directly attached to the coil. Slice thickness (Δz) 10 mm
Number of slices 3
9. Use the laser light to position the patient and to center the coil (see Table A17.1.2). Slice gap Not applicable
10. Advance the patient table to isocenter. Number of acquisitions (Nacq) 1
Swap read and phase encoding No
11. Program the power injector for a contrast agent and saline injection rate of 2 ml/sec.
Slice locations Not applicable
Total volume of saline following contrast agent injection should be programmed for
Saturation pulses Not applicable
10 ml. Program a scan delay of 18 sec (contrast agent is injected, scan is initiated 18
Scan time 16 sec
sec after contrast agent and saline are delivered). Arm the power injector and keep
the vein open. Do not inject the contrast agent.
Sequence 1: Three-plane positioning scout
Table A17.1.3 Imaging Parameters for Sequence 2 (Half-Acquisition Turbo
12. To validate the patient’s position and to have a reference to prescribe successive Spin Echo)
imaging sequences, acquire a three-plane orthogonal scout sequence using parame-
ters in Table A17.1.2. Patient position Supine
Most MR scanners can be programmed to acquire the scout automatically after coil tuning Scan type Half-acquisition turbo spin echo
or after the patient has been placed in isocenter (for systems that do not require tuning). Imaging plane (orientation) Coronal
Central slice or volume center Slices posted on transverse scout;
Sequence 2: Half-acquisition (partial Fourier) turbo spin echo (HASTE) coronal center to adrenals
13. Display both the coronal and transverse scout images in two separate quadrants on Echo time (TE) 90 msec
the scan monitor. Change imaging parameters to those listed in Table A17.1.3. Repeat time (TR) 4.4 mseca
Position slices to center of the transverse scout ensuring that the adrenals are covered. Delay time 1500 msec
14. Instruct the patient to remain motionless and to breathe normally as the scan will Flip angle (FA) 150°
begin and last for ∼40 sec. Field of view (FOVx, FOVy) 400 mm, 400 mm
Resolution (Δx, Δy) 1.56 mm, 2.08 mm
Sequence 3: Gradient echo coronal Number of data points collected (Nx, Ny) 256, 192 (using half Fourier)
15. Change imaging parameters to those listed in Table A17.1.4. Position slices to center Display matrix (Dx, Dy) 256, 256
of the transverse scout ensuring that the adrenals are covered. Slice thickness (Δz) 8–10 mm
16. Instruct the patient to take in a deep breath and exhale, take in another deep breath Number of slices 20
and hold it. Run the scan. Slice gap 1.6–2 mm
Number of acquisitions (Nacq) 1
Sequence 4: Half-acquisition turbo spin echo transverse with fat saturation (fat Swap read and phase encoding No
suppression) Slice locations Centered to adrenals
17. Display both the coronal and transverse scout images in two separate quadrants on Saturation pulses No
the scan monitor (see step 13). Change imaging parameters to those listed in Table Slice series Interleaved
A17.1.5. Position slices to cover from the diaphragm to the aortic bifurcation. Scan time 40 sec
18. Perform system shim as recommended by manufacturer, as this is a fat saturation aThe true T is infinite; 4.4 msec represents the echo spacing.
R
sequence.
MRI of the
Adrenal Glands Adrenal Glands

A17.1.3 A17.1.4
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A17.1.4 Imaging Parameters for Sequence 3 (Gradient Echo) 19. Instruct the patient to remain motionless and to breathe normally as the scan will
begin and last for ∼40 sec.
Patient position Supine
Scan type Gradient echo Sequence 5: Transverse gradient echo with fat saturation (Figure A17.1.1)
Imaging plane (orientation) Coronal 20. Display the coronal images (use breath-hold images from sequence 3) on the scan
Central slice or volume center Slices posted on transverse scout; monitor. Change imaging parameters to those listed in Table A17.1.6. Position slices
center to adrenals to center of the coronal gradient echo image ensuring that the adrenals are covered.
Echo time (TE) 4.1 msec
Repetition time (TR) 140 msec
Flip angle (FA) 80°
Field of view (FOVx, FOVy) 400 mm, 400 mm
Resolution (Δx, Δy) 1.56 mm, 3.13 mm
Number of data points collected (Nx, Ny) 256, 128
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 8–10 mm
Number of slices 20
Slice gap 1.6–2 mm
Number of acquisitions (Nacq) 1
Swap read and phase encoding No
Slice locations Centered to cover adrenals
Saturation pulses No
Slice series Interleaved
Scan time 18 sec

Figure A17.1.1 Unenhanced fat suppressed transverse spoiled gradient echo

Table A17.1.5 Imaging Parameters for Sequence 4 (Half-Acquisition Turbo image.
Spin Echo with Fat Saturation)
Table A17.1.6 Imaging Parameters for Sequence 5 (Gradient Echo with Fat
Patient position Supine Saturation)
Scan type Half-acquisition turbo spin echo
Imaging plane (orientation) Transverse Patient position Supine
Central slice or volume center Slices posted on coronal; center to Scan type Gradient echo
adrenals Imaging plane (orientation) Transverse
Echo time (TE) 90 msec Central slice or volume center Slices posted on coronal; center to
Repeat time (TR) 4.4 mseca adrenals
Delay time 1500 msec Echo time (TE) 4.1 msec
Flip angle (FA) 150° Repeat time (TR) 147.2 msec
Field of view (FOVx, FOVy) 350 mm, 263 mm Flip angle (FA) 80°
Resolution (Δx, Δy) 1.37 mm, 1.37 mm Field of view (FOVx, FOVy) 350 mm, 263 mm
Number of data points collected (Nx, Ny) 256, 192 (using half Fourier) Resolution (Δx, Δy) 1.37 mm, 2.05 mm
Display matrix (Dx, Dy) 256, 256 Number of data points collected (Nx, Ny) 256, 128
Slice thickness (Δz) 8–10 mm Display matrix (Dx, Dy) 256, 256
Number of slices 20 Slice thickness (Δz) 6 mm
Slice gap 1.6–2 mm Number of slices 20
Number of acquisitions (Nacq) 1 Slice gap 1.2 mm
Swap read and phase encoding No Number of acquisitions (Nacq) 1
Slice locations Centered to cover adrenals Swap read and phase encoding No
Saturation pulses Yes, superior and inferior to slices Slice locations Centered to cover adrenals
Fat suppression Yes Saturation pulses No
Slice series Interleaved Fat suppression Yes
Scan time 40 sec Slice series Interleaved
aThe true T is infinite; 4.4 msec represents the echo spacing.
R
Scan time 19 sec
MRI of the
Adrenal Glands Adrenal Glands

A17.1.5 A17.1.6
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 It is imperative that the slices are prescribed off of the breath-hold gradient echo images
from sequence 3 as this is a breath-held imaging sequence. Otherwise, the slice location
will not be accurate relative to the reference image if a nonbreath-held image is used.
21. Instruct the patient to take in a deep breath and exhale, take in another deep breath
and hold it. Run the scan.
Sequence 6: Transverse gradient echo (out-of-phase) (Figure A17.1.2)
22. Display the midline slice of the gradient echo coronal image (sequence 3) and the
previous transverse gradient echo image in two separate quadrants on the scan
monitor. Change imaging parameters to those listed in Table A17.1.7. Position slices
to cover the adrenals.
It is imperative that the slices are prescribed off of the gradient echo coronal image as this
is a breath-held imaging sequence. Otherwise, the slice location will not be accurate
relative to the reference image if a non-breath-held image is used.
23. Instruct the patient to take in a deep breath and exhale, take in another deep breath
and hold it. Run the scan.
Sequence 7: Transverse gradient echo
24. Display the midline slice of the gradient echo coronal image (sequence 3) and the
previous transverse gradient echo image in two separate quadrants on the scan
monitor. Change imaging parameters to those listed in Table A17.1.8. Position slices
to cover the adrenals.
It is imperative that the slices are prescribed off of the gradient echo coronal image as this
is a breath-held imaging sequence. Otherwise, the slice location will not be accurate
relative to the reference image if a non-breath-held image is used.
25. Instruct the patient to take in a deep breath and exhale, take in another deep breath
and hold it. Run the scan.

Sequence 8: Transverse gradient echo—immediate post contrast (Figure A17.1.3)

See patient set up section for specific instructions on preparation for contrast agent
injection. This preparation must be done prior to placing the patient in the scanner.
26. Repeat step 24.
It is imperative that the slices be prescribed off of the gradient echo coronal image as this
is a breath-held imaging sequence. Otherwise, the slice location will not be accurate
relative to the reference image if a non-breath-held image is used.
Figure A17.1.2 Out-of-phase transverse spoiled gradient echo image.
27. Explain to the patient that you will now be injecting the contrast agent and he/she
Table A17.1.7 Imaging Parameters for Sequence 6 (Gradient may feel a cool sensation in his/her arm. Initiate the injection. Do not begin scanning
Echo–Out-of-Phase)
Table A17.1.8 Imaging Parameters for Sequence 7 (Gradient Echo)
Patient position Supine
Scan type Gradient echo Patient position Supine
Imaging plane (orientation) Transverse Scan type Gradient echo
Central slice or volume center Slices posted on coronal; center to Imaging plane (orientation) Transverse
adrenals Central slice or volume center Slices posted on coronal; center to adrenals
Echo time (TE) 2.2 msec Echo time (TE) 4.5 msec
Repeat time (TR) 140 msec Repeat time (TR) 140 msec
Flip angle (FA) 80° Flip angle (FA) 80°
Field of view (FOVx, FOVy) 350 mm, 263 mm Field of view (FOVx, FOVy) 350 mm, 263 mm
Resolution (Δx, Δy) 1.37 mm, 2.05 mm Resolution (Δx, Δy) 1.37 mm, 2.05mm
Number of data points collected (Nx, Ny) 256, 128 Number of data points collected (Nx, Ny) 256, 128
Display matrix (Dx, Dy) 256, 256 Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 8–10 mm Slice thickness (Δz) 8–10 mm
Number of slices 20 Number of slices 18
Slice gap 1.6–2 mm Slice gap 1.6–2 mm
Number of acquisitions (Nacq) 1 Number of acquisitions (Nacq) 1
Swap read and phase encoding No Swap read and phase encoding No
Slice locations Centered to adrenals Slice locations Centered to cover adrenals
Saturation pulses No Saturation pulses No
Slice series Interleaved Slice series Interleaved
MRI of the
Scan time 19 sec Adrenal Glands Adrenal Glands Scan time 18 sec

A17.1.7 A17.1.8
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 until the 18 sec scan delay has expired. However, breathing instructions should be It is imperative that the slices are prescribed off of the gradient echo coronal image as this
delivered when 10 sec of delay are remaining (see step 28). is a breath-held imaging sequence. Otherwise, the slice location will not be accurate
relative to the reference image if a non-breath-held image is used.
If you do not have access to a power injector and are “hand” injecting, you can still use
step 27. However, after you have completed the bolus contrast agent injection, reattach the 30. Once 45 sec has expired, instruct the patient to take in a deep breath and exhale, take
saline-filled syringe and flush with 10 ml of saline. Begin breathing instructions after 5 ml in another deep breath and hold it. Run the scan.
of the saline has been injected then proceed to initiate the scan. The process of switching
syringes must be completed as quickly as possible and thus, the suggestion of incorporating
the use of a 3-way stopcock.
28. When there are 10 sec of delay remaining, instruct the patient to take in a deep breath
and exhale, take in another deep breath and hold it. Run the scan.

Sequence 9: Transverse gradient echo—45 sec delay after injection (Figure A17.1.4)
29. Repeat step 24.

Figure A17.1.5 Post-contrast fat suppressed transverse spoiled gradient echo

image.

Table A17.1.9 Imaging Parameters for Sequence 10 (Gradient Echo with Fat
Saturation, 90 Sec Delay)

Figure A17.1.3 Immediate post contrast transverse spoiled gradient echo Patient position Supine
image. Scan type Gradient echo
Imaging plane (orientation) Transverse
Central slice or volume center Slices posted on coronal; center to adrenals
Echo time (TE) 4.1 msec
Repeat time (TR) 147.2 msec
Flip angle (FA) 80°
Field of view (FOVx, FOVy) 350 mm, 263 mm
Resolution (Δx, Δy) 1.37 mm, 2.05mm
Number of data points collected (Nx, Ny) 256, 128
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 8–10 mm
Number of slices 20
Slice gap 1.6–2 mm
Number of acquisitions (Nacq) 1
Swap read and phase encoding No
Slice locations Centered to cover adrenals
Saturation pulses No
Fat suppression Yes
Figure A17.1.4 One-minute post gadolinium spoiled gradient echo transverse MRI of the Slice series Interleaved
image. Adrenal Glands Adrenal Glands Scan time 19 sec

A17.1.9 A17.1.10
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
Sequence 10: Transverse gradient echo with fat saturation (90 sec delay after Mayo-Smith, W.W., Lee, M.J., McNicholas, aging: Comparison of techniques. Radiology
M.M.J., Hahn, P.F., Boland, G.W., and Saini, S. 192:41-46.
injection) (Figure A17.1.5) 1995. Characterization of adrenal masses (<5
31. Display the midline slice of the gradient echo coronal image (sequence 3) and the Shellock, F.G. and Kanal, E. 1996. Magnetic Reso-
cm) by use of chemical shift MR imaging: Ob- nance: Bioeffects, Safety, and Patient Manage-
previous transverse gradient echo image in two separate quadrants on the scan server performance versus quantitative meas- ment. Lippincott, Williams & Wilkins, Philadel-
monitor using parameters in Table A17.1.9. Position slices to cover the adrenals. ures. Am. J. Roentgenol. 165:91-95. phia.
McNicholas, M.M.J., Lee, M.J., Mayo-Smith, Tsushima, Y., Ishizaka, H., and Matsumoto, M.
It is imperative that the slices are prescribed off of the gradient echo coronal image as this W.W., Hahn, P.F., Boland, G.W., and Mueller, P. 1993. Adrenal masses: Differentiation with
is a breath-held imaging sequence. Otherwise, the slice location will not be accurate 1995. An imaging algorithm for the differential chemical shift, fast low-angle shot MR imaging.
relative to the reference image if a non-breath-held image is used. diagnosis of adrenal adenomas and metastases. Radiology 186:705-709.
Am. J. Roentgenol. 165:1453-1459.
32. Once 90 sec has expired, instruct the patient to take in a deep breath and exhale, take Mitchell, D.G., Grovello, M., Matteucci, T., Peter- Key References
in another deep breath and hold it. Run the scan. son, R.O., and Miettinen, M.M. 1992. Benign Shellock and Kanal, 1996. See above.
adenocortical masses: Diagnosis with chemical
shift MR imaging. Radiology 185:345-351. Covers a number of important patient management
COMMENTARY issues related to MR imaging, including recom-
Outwater, E.K., Siegelman, E.S., Radecki, P.D., Pic- mended safety procedures, a list of metallic implants
coli, C.W., and Mitchell, D.G. 1995. Distinction that have been tested for MR compatibility, and a
Background Information is employed for this detection. Pheochromocy- between benign and malignant adrenal masses:
The most common indication for MR of the tomas are moderately high in signal on the fat list of other sources on MR safety.
Value of T1-weighted chemical-shift MR imag-
adrenals is to distinguish adenoma from metas- suppressed T2-weighted HASTE and are gen- ing. Am. J. Roentgenol. 165:579-583.
tases (Mitchell et al., 1992; Tsushima et al., erally higher in signal than other adrenal Outwater, E.K., Siegelman, E.S., Huang, A.B., and
1993; Lee et al., 1994; Reinig et al., 1994; masses and higher in signal than the majority Birnbaum, B.A. 1996. Adrenal masses: Correla- Contributed by Kathy Wilber and
Bilbey et al., 1995; Korobkin et al., 1995, 1996; of retroperitoneal structures. As fat suppressed tion between CT attenuation value and chemical Richard Semelka
Mayo-Smith et al., 1995; McNicholas et al., T2-weighted HASTE is breathing independent, shift ratio at MR imaging with in-phase and University of North Carolina
opposed-phase sequences. Radiology 200:749- Chapel Hill, North Carolina
1995; Outwater et al., 1995, 1996). The above there is no requirement that the patient be able 752.
described protocol employs gadolinium con- to suspend respiration in order to search for Laurie Fisher
Reinig, J.W., Stutley, J.E., Leonhardt, C.M., Spicer,
trast in order that the full range of adrenal pheochromocytomas. K.M., Margolis, M., and Caldwell, C.B. 1994. Siemens Uptime Service Center
diseases may be evaluated. The use of gadolin- Differentiation of adrenal masses with MR im- Cary, North Carolina
ium is not, however, essential if the clinical Anticipated Results
question is just the characterization of a known In patients who are able to suspend respira-
adrenal mass such as adenoma or metastasis. tion for 20 sec, successful imaging of the full
The next most common indication is the detec- range of adrenal abnormalities is achieved. This
tion of pheochromocytoma (Lee et al., 1994). single protocol is designed to image the various
Rather than having separate protocols for ade- types of disease processes that affect the adre-
noma versus metastases and pheochromocy- nal glands. The combination of in-phase and
toma search, we combine both indications into out-of-phase images distinguish between be-
a single protocol to maintain study simplicity. nign adenomas and metastases.

Critical Parameters and Literature Cited

Troubleshooting
The most critical sequences are in-phase and
out-of-phase spoiled gradient echo (Mitchell et
al., 1992; Tsushima et al., 1993; Lee et al.,
1994; Reinig et al., 1994; Bilbey et al., 1995;
Korobkin et al., 1995; Mayo-Smith et al., 1995;
McNicholas et al., 1995; Outwater et al., 1995).
It is essential that patients be able to hold their
breath in order to achieve optimal results. In the
event that patients are unable to breath-hold,
contrast computed tomography (CT) should be
performed. Results with Hounsfield unit meas-
urements on non-contrast CT and observed loss
of signal from in-phase to out-of-phase MRI
are comparable for the characterization of ad-
renal masses as adenomas (Korobkin et al.,
1996; Outwater et al., 1996).
In looking for pheochromocytomas, the fat
suppressed T2-weighted HASTE sequence is
the sequence (sequence 4) in the protocol that
Bilbey, J.H., McLoughlin, R.F., Kurkjian, P.S.,
Wilkins, G.E.L., Chan, N.H.L., Schmidt, N., and
Singer, J. 1995. MR imaging of adrenal masses:
Value of chemical-shift imaging for distinguish-
ing adenoma from other tumors. Am. J.
Roentgenol. 164:637-642.
Korobkin, M., Lombardi, T.J., Aisen, A.M., Francis,
I.R., Quint, L.E., and Dunnick, N.R. 1995. Char-
acterization of adrenal masses with chemical
shift and gadolinium-enhanced MR imaging.
Radiology 197:411-418.
Korobkin, M., Giordano, T.J., Brodeur, F.J., Francis,
I.R., Siegelman, E.S., Quint, L.E., Dunnick,
N.R., Heiken, J.P., and Wang, H.H. 1996. Adre-
nal adenomas: Relationship between histologic
lipid and CT and MR findings. Radiology
200:743-747.
Lee, M.J., Mayo-Smith, W.M., Hahn, P.F., Gold-
berg, M.A., Boland, G.W., Saini, S., and Papani-
colaou, N. 1994. State-of-the-art MR imaging of
the adrenal gland. RadioGraphics 14:1015-
1029. MRI of the
Adrenal Glands Adrenal Glands

A17.1.11 A17.1.12
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
MRI of the Pancreas UNIT A18.1 emergency equipment during the scanning procedure, or if he or she will need
sedation medication necessitating the use of appropriate monitoring equipment.
MRI provides comprehensive information on the full range of pancreatic diseases. We
Generally standard screening forms are used for all patients scanned in a magnetic
employ a set protocol incorporating various types of sequences including transverse and resonance system.
coronal data acquisitions, and with the routine use of intravenous gadolinium.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
IMAGING THE PANCREAS BASIC composition of the items, it is best to exclude patients with any metal implants; see Shellock
PROTOCOL (1996) for discussion of what implants may be safely scanned using magnetic resonance.
The high signal-to-noise ratio (SNR) obtained at high field strength makes it possible to
image the pancreas during a single breath-hold sequence. The sequences described herein The presence of ferromagnetic materials in the globe of the eye is contraindicated for MRI.
Patients with prior metal exposure to the eye should have plain X rays of the orbital area
are based on the authors’ experience with a Siemens 1.5 T Vision scanner, but are expected to ensure that all metal has been removed prior to placing them in the magnetic field.
to be equally applicable to machines from other manufacturers.
Patients may be accompanied into the magnet room by a friend or family member, who can
Scanning a patient or volunteer is a joint effort among technologists, nurses, and sit in the room during the scan and comfort the patient as needed. This companion must
physicians, with the technologist normally responsible for following proper scanning be screened as well to ensure the absence of loose metal objects on the body or clothing.
protocols and techniques. Unless otherwise specified, in what follows, the person to whom
directions are given is assumed to be the technologist. Table A18.1.1 lists the hardware 2. Request that the patient change into a gown to eliminate any metal that might be found
necessary to perform the procedure, along with appropriate parameters. in clothing. Ask the patient to remove all personal effects such as, jewelry, hearing
aids, glasses, etc., prior to entering the MRI scan room.
The following ten sequences comprise the pancreatic imaging protocol. Most sequences
require the patient to be able to suspend respiration for ∼20 sec. It is imperative that there All personal belongings should be secured during the examination.
be clear communication between the technologist and the patient throughout the exam. 3. Explain the procedure to the patient and record relevant clinical history. Ensure that
This protocol results in consistent, reproducible image quality that is effective for the patient understands what is expected and ask them if they have any questions;
evaluating the full spectrum of pancreatic diseases. answer appropriately. If the procedure is a research protocol, have the patient sign
NOTE: Be sure that technologists and nurses have immediate access to any emergency any necessary consent form.
equipment that may be relevant to a given study, or that may be needed for a particular
4. Fill a 20-ml syringe with normal saline and attach to saline filled extension tubing
patient, such as crash carts or oxygen.
(35-in.). Obtain intravenous (i.v.) access utilizing a 22-G angiocatheter and attach
Materials saline-prepared extension tubing and syringe. This will allow you to flush the
extension tubing while the patient waits to be imaged, and, in cases in which a power
Normal saline (0.9% NaCl), sterile, 40 ml minimum
injector is not available, this will allow preparation for bolus injection (see Note
Extravascular contrast agent (e.g., Magnevist, Omniscan, or Prohance), volume
prescribed by patient weight below). Secure the position of the angiocatheter with tegraderm or tape.
Obtaining i.v. access prior to entering the scan room will promote patient throughput and
Set up equipment and patient eliminate “dead” time of starting the i.v. while the patient is on the exam table. Follow
1. Interview (screen) the patient to assess for contraindications such as cardiac pace- power injector manufacturer guidelines with regard to appropriate gauge needle/angio-
maker, implanted mechanical devices, and/or ferromagnetic materials. Also, deter- catheter to be used; this will depend on chosen injection/flow rates.
mine if the patient has any health conditions that may require the presence of special NOTE: If you do not have access to a power injector, you will still be able to perform
dynamic imaging as the extension tubing will allow the saline syringe to be placed at the
Table A18.1.1 Equipment Specifications: Imaging of the Pancreas
foot of the patient table during pre-contrast imaging. In this case, you will need to draw
up the contrast agent in another syringe. When you are ready to bolus inject the contrast
Coil type Circularly polarized body phased array coil agent, simply disconnect the saline syringe and connect the syringe filled with contrast
agent; once you have injected the bolus of contrast agent, reconnect the saline syringe and
Manufacturer and system type Siemens Vision
bolus an appropriate volume of flush, usually ∼10 ml. Alternatively, to eliminate the need
Field strength 1.5 T of switching syringes, incorporate the use of a 3-way stopcock.
Gradient strength 24 mT/ m (or whatever the system permits,
but minimum of 24 mg/m for sequences 3 5. Set up the exam room by securing the circularly polarized (CP)-body array coil onto
and 4) the table and providing a clean exam table.
Knee cushion Yes
Use of contrast agents Yes
6. Set up the power injector as specified by the manufacturer.
Pulse oximeter If patient requires sedation A minimum of 40 ml normal saline should be drawn up to ensure sufficient saline is
Power injector Yes available to keep the vein open (KVO) throughout the exam. To determine the amount of
Normal saline Yes contrast agent to be used, reference the contrast agent packet insert and draw up the amount
35-in. extension tubing Yes indicated per kg of patient weight. There is no need to double dose.

MRI of the
7. Escort the patient to the MR examination room and ask the patient to lie down
Pancreas Pancreas accordingly with respect to the exam to be performed. Help the patient mount onto
Contributed by Kathy Wilber, Richard Semelka, and Laurie Fisher A18.1.1 A18.1.2
Current Protocols in Magnetic Resonance Imaging (2001) A18.1.1-A18.1.13
Copyright © 2001 by John Wiley & Sons, Inc. Current Protocols in Magnetic Resonance Imaging
 the table. Either before or right after the patient lies down, set up any triggering Table A18.1.2 Imaging Parameters for Scout Sequence (Sequences 1 and 2)
devices or other monitoring equipment that is to be used.
Patient position Supine
8. Connect the extension tubing secured to the syringe to the power injector extension Scan type Gradient echo
tubing. Imaging plane (orientation) Sagittal, transverse, and coronal
9. Inform the patient about what will occur during the procedure, what he or she will Central slice or volume center Laser light centered
approximately one hand width
experience while in the magnet, and how to behave, including the following:
above the inferior rib margin
a. If earphones or headphones are used to protect the ears from the loud sounds Echo time (TE) 6 msec
produced by the gradients, the patient will be asked to wear these, but will be able Repeat time (TR) 15 msec
to communicate with you at any time during the imaging. Flip angle (FA) 30°
b. The patient will be given a safety squeeze-bulb or similar equipment to request Field of view (FOVx, FOVy) 450 mm, 450 mm
assistance at any time (demonstrate how this works). Resolution (Δx, Δy) 1.76 mm, 3.52 mm
Number of data points collected (Nx, Ny) 256, 128
c. For good results the patient should not talk, and should avoid or minimize
Display matrix (Dx, Dy) 256, 256
swallowing or other movement, during each scan—i.e., as long as the banging
Slice thickness (Δz) 10 mm
sounds continue. Between scans, talking and swallowing are allowed in most
Number of slices 3
cases, but should be avoided when comparative positional studies are being
Slice gap Not applicable
performed; the patient will be informed when this is the case. Explain to the patient Number of acquisitions (Nacq) 1
that one should not reposition one’s body between imaging sequences. Swap read and phase encoding No
Additionally, review breath-holding instructions with the patient, and provide the patient Slice locations Not applicable
with an approximate time that the examination will take. Saturation pulse Not applicable
d. Nevertheless, the patient may call out at any time if he or she feels it necessary. Scan time 16 sec
e. Position a support under the patient’s knees to enhance patient comfort.
16. Instruct the patient to take in a deep breath and exhale, take in another deep breath
10. Secure the top portion of the CP-body array coil to prevent it from moving side-to-side and hold it.
during breath-holding imaging sequences.
17. Initiate the scan.
Usually straps are provided by the manufacturer that are directly attached to the coil.
11. Use the laser light to position the patient, and to center the coil (see Table A18.1.2). Sequence 3: Half-acquisition (partial Fourier) turbo spin echo coronal
Then advance the patient table to isocenter. 18. Display both the coronal and transverse scout images (use non-breath-hold images)
in two separate quadrants on the scan monitor. Change imaging parameters to those
12. If the patient is unable to hold still, provide an appropriate sedative. listed in Table A18.1.3. Position slices to center of the transverse scout, ensuring that
13. Program the power injector for a contrast agent and saline injection rate of 2 ml/sec. the pancreas and liver are covered.
Total volume of saline following contrast agent injection should be programmed for 19. Instruct the patient to remain motionless and to breathe normally as the scan will
10 ml. Program a scan delay of 18 sec (contrast agent is injected, scan is initiated 18 begin and last for ∼40 sec.
sec after contrast agent and saline are delivered). Arm the power injector and keep
the vein open. Sequence 4: Half-acquisition turbo spin echo transverse (Fig. A18.1.1)
Do not inject the contrast agent at this time! 20. Display both the coronal and transverse scout images (use non-breath-hold images)
in two separate quadrants on the scan monitor. Change imaging parameters to those
Sequence 1: Three-plane positioning scout listed in Table A18.1.4. Position slices to center of the coronal scout, ensuring that
14. To validate the patient’s position and to have a reference to prescribe successive the pancreas and liver are entirely covered.
imaging sequences, acquire a three-plane orthogonal scout sequence. See Table
21. Instruct the patient to remain motionless and to breathe normally as the scan will
A18.1.2 for specific parameters.
begin and last for ∼40 sec.
Most MR scanners can be programmed to acquire the scout automatically after coil tuning
or after the patient has been placed in isocenter (for systems that do not require tuning). Sequence 5: Transverse gradient echo with fat saturation (fat suppression) (Fig.
A18.1.2)
Sequence 2: Breath-hold three—plane positioning scout 22. Display both the coronal and transverse scout images (use breath-hold images) in
15. To have a reference to prescribe successive breath-hold imaging sequences, acquire two separate quadrants on the scan monitor. Change imaging parameters to those
a second three-plane orthogonal scout sequence. See Table A18.1.2 for specific listed in Table A18.1.5. Position slices to center of the coronal scout, ensuring that
parameters. the pancreas is entirely covered.
MRI of the
Pancreas Pancreas

A18.1.3 A18.1.4
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 It is imperative that the slices are prescribed off of the breath-hold scout images as this is Table A18.1.4 Imaging Parameters for Half-Acquisition Turbo Spin Echo
a breath-held imaging sequence. Otherwise, the slice location will not be accurate relative (Sequence 4)
to the reference image if a non-breath-held image is used.
Patient position Supine
23. Instruct the patient to take in a deep breath and exhale, take in another deep breath
Scan type Half-acquisition turbo spin echo
and hold it.
Imaging plane (orientation) Transverse
24. Initiate the scan. Central slice or volume center Slices posted on coronal; center to
pancreas
Table A18.1.3 Imaging Parameters for Half-Acquisition Turbo Spin Echo Echo time (TE) 90 msec
(Sequence 3) Repeat time (TR) 4.4 msec (note: The true TR is
infinite; 4.4 msec represents the
Patient position Supine echo spacing)
Scan type Half-acquisition turbo spin echo Delay time (TD) 1500 msec
Imaging plane (orientation) Coronal Flip angle (FA) 150°
Central slice or volume center Slices posted on transverse scout; Fields of view (FOVx, FOVy) 350 mm, 263 mm
center to pancreas Resolution (Δx, Δy) 1.37mm, 1.37 mm
Echo time (TE) 90 msec Number of data points collected (Nx, Ny) 256, 192 (using half Fourier)
Repeat time (TR) 4.4 msec (note: The true TR is Display matrix (Dx, Dy) 256, 256
infinite; 4.4 msec represents the
Slice thickness (Δz) 8–10 mm
echo spacing)
Number of slices 20
Delay time (TD) 1500 msec
Slice gap 1.6–2 mm
Flip angle (FA) 150°
Number of acquisitions (Nacq) 1
Fields of view (FOVx FOVy) 400 mm, 400 mm
Swap read and phase encoding No
Resolution (Δx Δy) 1.56 mm, 2.08 mm
Slice location Centered to cover the pancreas
Number of data points collected (Nx, Ny) 256, 192 (using half Fourier)
and liver
Display matrix (Dx, Dy) 256, 256
Saturation pulses Yes, superior and inferior to slices
Slice thickness (Δz) 8–10 mm
Fat suppression No
Number of slices 20
Slice series Interleaved
Slice gap 1.6–2 mm
Scan time 40 sec
Number of acquisitions (Nacq) 1
Swap read and phase encoding No
Slice location Centered to cover the pancreas
and liver
Saturation pulses No
Slice series Interleaved
Scan time 40 sec

Figure A18.1.2 Unenhanced fat suppressed transverse spoiled gradient echo

image.

MRI of the
Pancreas Pancreas
Figure A18.1.1 Unenhanced transverse half-acquisition turbo spin echo image.
A18.1.5 A18.1.6
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
Sequence 6: Transverse gradient echo (out-of-phase) (Fig. A18.1.3)
25. Display the coronal and transverse breath-hold scout images (sequence 2) in two
separate quadrants on the scan monitor. Change imaging parameters to those listed
in Table A18.1.6. Position slices to cover the pancreas and liver.
It is imperative that the slices bee prescribed off of the breath-hold coronal scout as this is
a breath-held imaging sequence. Otherwise, the slice location will not be accurate relative
to the reference image if a non-breath-held image is used.
26. Instruct the patient to take in a deep breath and exhale, take in another deep breath
and hold it.
27. Initiate the scan.
Sequence 7: Transverse gradient echo (Fig. A18.1.4)
28. Display the breath-hold coronal and transverse scout images (sequence 2) in two
separate quadrants on the scan monitor. Change imaging parameters to those listed
in Table A18.1.7. Position slices to cover the pancreas and liver.

Table A18.1.5 Imaging Parameters for Gradient Echo with Fat Saturation
(Sequence 5) Table A18.1.6 Imaging Parameters for Gradient Echo—Out-of-Phase
(Sequence 6)
Patient position Supine
Scan type Gradient echo Patient position Supine
Imaging plane (orientation) Transverse Scan type Gradient echo
Central slice or volume center Slices posted on coronal; center to Imaging plane (orientation) Transverse
pancreas Central slice or volume center Slices posted on coronal; center to
Echo time (TE) 4.1 msec the pancreas
Repeat time (TR) 147.2 msec Echo time (TE) 2.2 msec
Flip angle (FA) 80° Repeat Time (TR) 140 msec
Fields of view (FOVx, FOVy) 350 mm, 263 mm Flip angle (FA) 80°
Resolution (Δx, Δy) 1.37 mm, 2.05 mm Fields of view (FOVx, FOVy) 350 mm, 263 mm
Number of data points collected Resolution (Δx, Δy) 1. 37 mm, 2.05 mm
(Nx, Ny) 256, 128 Number of data points collected (Nx, Ny) 256, 128
Display matrix (Dx, Dy) 256, 256 Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 6 mm Slice thickness (Δz) 8–10 mm
Number of slices 20 Number of slices 20
Slice gap 1.2 mm Slice gap 1.6–2 mm
Number of acquisitions (Nacq) 1 Number of acquisitions (Nacq) 1
Swap read and phase encoding No Swap read and phase encoding No
Slice location Centered to cover the pancreas Slice location Centered to cover the pancreas
Saturation pulses No and liver
Fat suppression Yes Saturation pulses No
Slice series Interleaved Slice series Interleaved
Scan time 19 sec Scan time 19 sec

MRI of the
Pancreas Pancreas
Figure A18.1.3 Out-of-phase transverse spoiled gradient echo image. Figure A18.1.4 Unenhanced transverse spoiled gradient echo image.
A18.1.7 A18.1.8
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A18.1.7 Imaging Parameters for Gradient Echo (Sequence 7)

Patient position Supine

Scan type Gradient echo
Imaging plane (orientation) Transverse
Central slice or volume center Slices posted on coronal; center to
the pancreas
Echo time (TE) 4.1 msec
Repeat time (TR) 140 msec
Flip angle (FA) 80°
Fields of view (FOVx, FOVy) 350 mm, 263 mm
Resolution (Δx, Δy) 1.37 mm, 2.05mm
Number of data points collected (Nx, Ny) 256, 128
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 8–10 mm
Number of slices 18
Slice gap 1.6–2 mm
Figure A18.1.5 Immediate post contrast transverse spoiled gradient echo im-
Number of acquisitions (Nacq) 1 age.
Swap read and phase encoding No
Slice location Centered to cover the pancreas
and liver
Saturation pulses No
Slice series Interleaved
Scan time 18 sec

It is imperative that the slices are prescribed off of the breath-hold coronal scout image as
this is a breath-held imaging sequence. Otherwise, the slice location will not be accurate
relative to the reference image if a non-breath-held image is used.
29. Instruct the patient to take in a deep breath and exhale, take in another deep breath
and hold it.
30. Initiate the scan.

Sequence 8: Transverse gradient echo—Immediate postcontrast (Fig. A18.1.5)

NOTE: See patient set up section for specific instructions on preparation for contrast
injection. This preparation must be done prior to placing the patient in the scanner. Figure A18.1.6 Post contrast fat suppressed transverse spoiled gradient echo
image.
31. Repeat step 28.
It is imperative that the slices be prescribed off of the breath-hold coronal scout image as 33. When there are 10 sec of delay remaining, instruct the patient to take in a deep breath
this is a breath-held imaging sequence. Otherwise, the slice location will not be accurate and exhale, take in another deep breath and hold it.
relative to the reference image if a non-breath-held image is used. 34. Initiate the scan.
32. Explain to the patient that you will now be injecting the contrast agent and he/she
may feel a cool sensation in his/her arm. Initiate the injection. Do not begin scanning Sequence 9: Transverse gradient echo—45 sec delay after sequence 8
until the 18 sec scan delay has expired. However, deliver breathing instructions when 35. Repeat step 28.
10 sec of delay are remaining (see step 33). 36. Once 45 sec has expired, instruct the patient to take in a deep breath and exhale, take
If you do not have access to a power injector and are “hand” injecting, you can still use
in another deep breath and hold it.
step 32. However, after you have completed the bolus contrast injection, reattach the 37. Initiate the scan.
saline-filled syringe and flush with 10 ml of saline. Begin breathing instructions after 5 ml
of the saline has been injected, then proceed to initiate the scan. The process of switching Sequence 10: Transverse gradient echo with fat saturation (90 sec delay after
syringes must be completed as quickly as possible and thus it is suggested that a 3-way injection; Fig. A18.1.6)
stopcock be incorporated.
38. Display the midline slice of the breath-hold coronal and transverse scout images
MRI of the
Pancreas Pancreas (sequence 2) in two separate quadrants on the scan monitor. Change imaging

A18.1.9 A18.1.10
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A18.1.8 Imaging Parameters for Gradient Echo with Fat Saturation—90 Table A18.1.9 Imaging Parameters for Breathing Averaged Fat Suppressed
Sec Delay (Sequence 10) Spin Echo (Use if Fat Suppressed Spoiled Gradient Echo Sequence Is
Unavailable)
Patient position Supine
Scan type Gradient echo Patient position Supine
Imaging plane (orientation) Transverse Scan type Spin echo
Central slice or volume center Slices posted on coronal; center to Imaging plane (orientation) Transverse
the pancreas Central slice or volume center Slices posted on coronal; center to
Echo time (TE) 4.1 msec pancreas
Repeat time (TR) 147.2 msec Echo time (TE) 15 msec
Flip angle (FA) 80° Repeat time (TR) 500 msec
Fields of view (FOVx, FOVy) 350 mm, 263 mm Flip angle (FA) 90°
Resolution (Δx, Δy) 1.37 mm, 2.05mm Fields of view (FOVx, FOVy) 350 mm, 263 mm
Number of data points collected (Nx, Ny) 256, 128 Resolution (Δx, Δy) 1.37 mm, 2.05 mm
Display matrix (Dx, Dy) 256, 256 Number of data points collected (Nx, Ny) 256, 128
Slice thickness (Δz) 6 mm Display matrix (Dx, Dy) 256, 256
Number of slices 20 Slice thickness (Δz) 8–10 mm
Slice gap 1.2 mm Number of slices 10
Number of acquisitions (Nacq) 1 Slice gap 1.6–2 mm
Swap read and phase encoding No Number of acquisitions (Nacq) 4
Slice location Centered to cover entire pancreas Swap read and phase encoding No
and liver Slice location Centered to cover pancreas
Saturation pulses No Saturation pulses No
Fat suppression Yes Fat suppression Yes
Slice series Interleaved Slice series Interleaved
Scan time 19 sec Scan time 4 min, 19 sec

parameters to those listed in Table A18.1.8. Position slices to cover the pancreas and
liver. echo sequence. The critical aspect of this se- Particular strengths of this approach are in the
quence is that the timing of gadolinium must detection of small pancreatic ductal adenocar-
It is imperative that the slices are prescribed off of the gradient echo coronal image as this be during the hepatic arterial dominant phase cinomas, ampullary tumors and islet cell tu-
is a breath-held imaging sequence. Otherwise, the slice location will not be accurate (or capillary phase) of enhancement (Semelka mors, and staging of pancreatic neoplasms in
relative to the reference image if a non-breath-held image is used. et al., 1993a). To ensure appropriate timing, general.
39. Once 90 sec has expired, instruct the patient to take in a deep breath and exhale, then contrast should be present in the hepatic arteries
take in another deep breath and hold it. and portal veins and not yet present in the Literature Cited
hepatic veins (Semelka et al., 1993a). The pa- Gabata, T., Matsui, O., Kadoya, M., Yoshikawa, J.,
40. Initiate the scan. tient must be able to breath-hold for 20 sec to Miyayama, S., Takashima, T., Nagakawa, T.,
Kayahara, M., and Nonomura, A. 1994. Small
achieve optimal results for studying the pan-
pancreatic adenocarcinomas: Efficacy of MR
COMMENTARY creas, as required for the T1-weighted gradient imaging with fat suppression and gadolinium
echo sequence. enhancement. Radiology 193:683-688.
Background Information intense on non-contrast T1-weighted fat sup- Fat-suppressed spin echo (see Table Kelekis, N.L., Semelka, R.C., Molina, P.L., and
MRI using a combination of non-contrast pressed images due to the presence of aqueous A18.1.9) can be substituted for the fat sup- Doer, M.E. 1995. ACTH-secreting islet cell tu-
T1-weighted fat suppressed imaging and imme- protein in the ascini of the glandular pancreas pressed spoiled gradient echo, if this latter se- mor: Appearances on dynamic gadolinium-en-
diate post-gadolinium spoiled gradient echo (Mitchell et al., 1995; Semelka et al., 1991; quence is not available on the scanner. Patients hanced MRI. J. Magn. Reson. Imag. 13:641.
has been shown to be effective at exhibiting the Semelka et al., 1993a; Winston et al., 1995). who are unable to breathhold or breathe in a Mitchell, D.G., Winston, C.B., Outwater, E.K., and
full range of pancreatic disease (Gabata et al., Particular strengths of MRI include the detec- regular fashion are better examined by CT Ehrlich S.M. 1995. Delineation of pancreas with
1994; Kelekis et al., 1995; Mitchell et al., 1995; tion of small ductal adenocarcinomas and islet MR imaging: Multi-observer comparison of five
(computed tomography).
Semelka et al., 1991; Semelka and Ascher, cell tumors (Gabata et al., 1994; Kelekis et al., pulse sequences. J. Magn. Reson. Imag. 5:193-
199.
1993; Semelka et al., 1993a,b; Semelka et al., 1995; Semelka et al., 1991; Semelka et al., Anticipated Results
1996; Winston et al., 1995). Normal pancreas 1993a,b; Semelka et al., 1996). Semelka, R.C. and Ascher, S.M. 1993. MRI of the
In a patient who can suspend respiration and pancreas: State of the art. Radiology 188:593-
enhances in a moderately intense, uniform using an MR protocol that includes non-con- 602.
fashion on immediate post-gadolinium images, Critical Parameters and trast T1-weighted fat suppressed spoiled gradi- Semelka, R.C., Kroeker, M.A., Shoenut, J.P.,
reflecting a uniform network of capillary ves- Troubleshooting ent echo and immediate post contrast spoiled Kroeker, R., Yaffe, C.S., and Micflikier, A.B.
sels (Semelka et al., 1991; Semelka and Ascher, The most important data acquisition is the gradient echo imaging, one can successfully 1991. Pancreatic disease: prospective compari-
MRI of the
1993). The normal pancreas is also moderately immediate post gadolinium spoiled gradient Pancreas Pancreas evaluate a full range of pancreatic diseases. son of CT, ERCP, and 1.5T imaging with dy-

A18.1.11 A18.1.12
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 namic gadolinium enhancement and fat suppres-
sion. Radiology181:785-791.
Semelka, R.C., Cummings, M., Shoenut, J.P., Yaffe,
C.S., Kroeker, M.A., and Greenberg, H.M.
1993a. Islet cell tumors: A comparison of detec-
tion by dynamic contrast-enhanced CT and MR
imaging with dynamic gadolinium enhancement
and fat suppression. Radiology186:799-802.
Semelka, R.C., Shoenut, J.P., Kroeker, M.A., and
Micflikier, A.B. 1993b. Chronic pancreatitis:
MR imaging features before and after admini-
stration of gadopentetate dimeglumine. J. Magn.
Reson. Imag. 3:79-82.
Semelka, R.C., Kelekis, N.L., Molina, P.L., Sharp,
T., and Calvo, B. 1996. Pancreatic masses with
indeterminate findings on spiral CT, Is there a
role for MRI? J. Magn. Reson. Imag. 6:585-588.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Winston, C.B., Mitchell, D.G., Outwater, E.K., and
Ehrlich, S.M. 1995. Pancreatic signal intensity
Current Protocols in Magnetic Resonance Imaging
on T1-weighted fat saturation MR images: Clini-
cal correlation. J. Magn. Reson. Imag. 5:267-
271.
Key References
Shellock, F.G. 1996.
Covers a number of important patient management
issues related to MR imaging, including recom-
mended safety procedures, a list of metallic implants
that have been tested for MR compatibility, and a
list of other sources on MR safety.
Contributed by Kathy Wilber and
Richard Semelka
University of North Carolina-Chapel Hill
Chapel Hill, North Carolina
Laurie Fisher
Siemens Uptime Service Center
Cary, North Carolina
Pancreas
A18.1.13
MRI of the Male Pelvis UNIT A19.1 A screening form is signed by each patient or legal guardian prior to bringing the patient
into the exam area.
MRI provides comprehensive information on many male pelvic diseases such as prostate The presence of ferromagnetic materials may be a health hazard to the patient while in the
cancer and bladder cancer as well as lymph nodes and bone metastases. The authors magnetic field and/or adversely affect image quality. To determine the safety of scanning,
employ a set protocol incorporating T1- and T2-weighted sequences including transverse such ferromagnetic materials, see Shellock (1996).
and sagittal data acquisitions, and the routine use of intravenous gadolinium. The presence of ferromagnetic materials in the globe of the eye is contraindicated for MRI.
Patients with prior metal exposure to the eye should have plain x-rays of the orbital area
to ensure that all metal has been removed prior to placing them in the magnetic field.
IMAGING THE MALE PELVIS BASIC
PROTOCOL 2. Request the patient to change into a gown and remove all personal effects such as,
The sequences described herein are based on the authors’ experience with a Siemens 1.5 T
jewelry, hearing aids, glasses, etc., prior to entering the MRI scan room. All personal
Vision scanner, but are expected to be equally applicable to machines from other
belongings should be secured during the examination. If the procedure is a research
manufacturers.
protocol, have the patient sign any necessary consent forms.
Scanning a patient or volunteer is a joint effort among technologists, nurses, and
3. Explain the procedure to the patient and record relevant clinical history. Ensure that
physicians, with the technologist normally responsible for following proper scanning
the patient understands what is expected and ask them if they have any questions;
protocols and techniques. Unless otherwise specified, in what follows the person to whom
answer appropriately.
directions are given is assumed to be the technologist. Table A19.1.1 lists the hardware
necessary to perform the procedure, along with appropriate parameters. 4. Fill a 20-ml syringe with normal saline and attach to saline filled extension tubing
(35 in.). Obtain i.v. access utilizing a 22-G angiocatheter and attach saline prepared
The following nine sequences comprise the imaging protocol for the male pelvis. This extension tubing and syringe. This will allow you to flush the extension tubing while
protocol employs multiple data acquisitions and serial post-gadolinium imaging. Most the patient waits to be imaged. In cases in which a power injector is not available,
sequences require the patient to be able to suspend respiration for ∼25 sec. It is imperative this will allow you to prepare for bolus injection. Secure the position of the angio-
that there is clear communication between the technologist and the patient throughout the catheter with tegraderm or tape.
exam. This protocol results in consistent, reproducible image quality that is effective for
evaluating the full spectrum of male pelvic diseases. Obtaining i.v. access prior to entering the scan room will promote patient throughput and
eliminate “dead” time of starting the i.v. while the patient is on the exam table. Follow
NOTE: Be sure that technologists and nurses have immediate access to any emergency power injector manufacturer guidelines with regard to appropriate gauge needle/angio-
equipment that may be relevant to a given study, or that may be needed for a particular catheter to be used; this will depend on chosen injection/flow rates.
patient, such as crash carts or oxygen. IMPORTANT NOTE: If there is no access to a power injector, you will still be able to
perform dynamic imaging as the extension tubing will allow the saline syringe to be placed
Materials at the foot of the patient table during pre-contrast imaging. In this case, you will need to
Normal saline (0.9% NaCl, sterile), 40 ml minimum draw-up contrast in another syringe. When you are ready to bolus inject the contrast agent,
Extravascular contrast agent (e.g., Magnevist, Omniscan, or Prohance), volume by simply disconnect the saline syringe and connect the syringe filled with contrast agent;
patient weight once you have injected the bolus of contrast agent, reconnect the saline syringe and bolus
an appropriate volume of flush, usually ∼10 ml. Alternatively, to eliminate the need of
Set up patient and equipment switching syringes, incorporate the use of a 3-way stopcock.
1. Interview the patient to assess for contraindications such as cardiac pacemaker, 5. Set up the exam room by securing the circularly polarized (CP)-body array coil onto
implanted mechanical devices, and/or ferromagnetic materials. Also, determine if the the table and providing a clean exam table.
patient will need sedation medication necessitating the use of appropriate monitoring
equipment. 6. Set-up the power injector as specified by the manufacturer. A minimum of 40 ml
normal saline should be drawn-up to ensure sufficient saline is available to keep the
vein open (KVO) throughout the exam. To determine the amount of contrast agent to
be used, reference the contrast agent packet insert and draw-up the amount indicated
Table A19.1.1 Equipment Specifications Needed to Perform the Following per killogram of patient weight. There is no need to double dose.
Imaging Sequences
7. Escort the patient to the MR examination room and ask them to lie down accordingly
Coil type Circularly polarized body phased array coil with respect to the exam to be performed. Connect the extension tubing secured to
Manufacturer and system type Siemens, Vision the syringe to the power injector extension tubing. Review the following items with
Field strength 1.5 T the patient:
Gradient coil strength 24 mT/m (or whatever the system permits)
a. Provide earplugs or headphones to the patient to minimize the loud knocking noise
Knee cushion Yes
that will be produced by the gradients but ensure them that they will still be able
Use of contrast agents Yes
to hear you.
Pulse oximeter If patient requires sedation
Power injector Yes b. Provide the patient with a safety squeeze-bulb and demonstrate how it works;
Normal saline Yes explain to the patient when to use the squeeze-bulb (i.e., if they need assistance
35" extension tubing Yes MRI of the during the exam).
Male Pelvis Male Pelvis
Contributed by Kathy Wilber, Richard Semelka, and Laurie Fisher A19.1.1 A19.1.2
Current Protocols in Magnetic Resonance Imaging (2001) A19.1.1-A19.1.11
Copyright © 2002 by John Wiley & Sons, Inc. Supplement 7 Supplement 7 Current Protocols in Magnetic Resonance Imaging
 c. Explain to the patient that you will be talking to them between imaging sequences Table A19.1.2 Imaging Parameters for Sequence 1 (Scout Sequence)
that will be when the loud knocking noise stops. Additionally, review breath-hold-
ing instructions with the patient. Patient position Supine
Scan type Gradient echo
d. Explain to the patient that it is imperative that they remain motionless during the
Imaging plane (orientation) Sagittal, transverse, and coronal
loud knocking noise to ensure good results; also explain that they should not
Central slice or volume center Center to pelvis
reposition their body between imaging sequences.
Echo time (TE) 6 msec
e. Position a support under the patient’s knees to enhance patient comfort. Repeat time (TR) 15 msec
f. Provide the patient with an approximate time that the examination will take. Flip angle (FA) 30°
Field of view (FOVx, FOVy) 450 mm, 450 mm
8. Secure the top portion of the CP-body array coil to prevent it from moving side-to-side Resolution (Δx,Δy) 1.76 mm, 3.52 mm
during breath-holding imaging sequences. Usually straps are provided by the manu- Number of data points collected (Nx, Ny) 256, 128
facturer that are directly attached to the coil. Display matrix (Dx, Dy) 256, 256
9. Using the laser light center the patient’s pelvis to the coil. Position the CP-body array Slice thickness (Δz) 10 mm
coil to include the entire pelvis and as much of the abdomen as possible. Do not Number of slices 3
sacrifice coil coverage at the bottom of the pelvis to include more of the abdomen. Slice gap Not applicable
10. Advance the patient table to isocenter. Number of acquisitions (Nacq) 1
Swap read and phase encoding No
11. Program the power injector for a contrast agent and saline injection rate of 2 ml/sec. Total Slice locations Not applicable
volume of saline following contrast agent injection should be programmed for 10 ml. Saturation pulses Not applicable
Program a scan delay of 25 sec (contrast agent is injected, scan is initiated 25 sec after Scan time 16 sec
contrast agent and saline are delivered). Arm the power injector and keep the vein open.
Do not inject the contrast agent at this time!
Sequence 1: Three-plane positioning scout (breath-hold) Table A19.1.3 Imaging Parameters for Sequence 2 (Gradient Echo)
12. To validate the patient’s position and to have a reference to prescribe successive
Patient position Supine
imaging sequences, acquire a three-plane orthogonal scout sequence. See Table
Scan type Gradient echo
A19.1.2 for specific parameters.
Imaging plane (orientation) Transverse
Most MR scanners can be programmed to acquire the scout automatically after coil tuning Central slice or volume center Slices posted on coronal; center to
or after the patient has been placed in isocenter (for systems that do not require tuning). mid-abdomen
13. Instruct the patient to take in a deep breath and exhale, take in another deep breath Echo time (TE) 4.5 msec
and hold it. Initiate the scan. Repeat time (TR) 140 msec
Flip angle (FA) 90°
Sequence 2: Transverse gradient echo (mid-abdomen) Field of view (FOVx, FOVy) 350 mm, 263 mm
14. Display both the coronal and transverse scout images in two separate quadrants on Resolution (Δx, Δy) 1.37 mm, 1.83 mm
the scan monitor. Change imaging parameters to those listed in Table A19.1.3. Number of data points collected (Nx, Ny) 256, 144
Position slices on the coronal scout ensuring that the mid-abdomen is covered. Display matrix (Dx, Dy) 256, 256
15. Instruct the patient to take in a deep breath and exhale, take in another deep breath Slice thickness (Δz) 7 mm
and hold it. Initiate the scan. Number of slices 18
Slice gap 1.4 mm
Sequence 3: Transverse gradient echo (pelvis) Number of acquisitions (Nacq) 1
16. Display both the coronal and transverse scout images in two separate quadrants on Swap read and phase encoding No
the scan monitor. Change imaging parameters to those listed in Table A19.1.4. Slice locations Centered to cover abdomen
Position slices on the coronal scout ensuring that the pelvis is covered. Saturation pulses No
17. Instruct the patient to take in a deep breath and exhale, take in another deep breath Slice series Interleaved
and hold it. Initiate the scan. Scan time 20 sec

Sequence 4: Transverse gradient echo with fat suppression

18. Display the coronal image and the transverse scout image in two separate quadrants
on the scan monitor. Change imaging parameters to those listed in Table A19.1.5.
Position slices to cover the pelvis.
19. Instruct the patient to take in a deep breath and exhale, take in another deep breath
MRI of the
and hold it. Male Pelvis Male Pelvis

A19.1.3 A19.1.4
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A19.1.4 Imaging Parameters for Sequence 3 (Gradient Echo)

Patient position Supine

Scan type Gradient echo
Imaging plane (orientation) Transverse
Central slice or volume center Slices posted on coronal; center to
pelvis
Echo time (TE) 4.5 msec
Repeat time (TR) 140 msec
Flip angle (FA) 90°
Field of view (FOVx, FOVy) 350 mm, 263 mm
Resolution (Δx, Δy) 1.37 mm, 1.83 mm
Number of data points collected (Nx, Ny) 256, 144
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 7 mm
Number of slices 18
Slice gap 1.4 mm
Number of acquisition (Nacq) 1
Swap read and phase encoding No
Slice location Centered to cover pelvis
Saturation pulses No
Slice series Interleaved
Scan time 20 sec

Table A19.1.5 Imaging Parameters for Sequence 4 (Gradient Echo with Fat
Suppression)

Patient position Supine

Scan type Gradient Echo
Imaging plane (orientation) Transverse
Central slice or volume center Slices posted on coronal; center to
pelvis Figure A19.1.1 High resolution sagittal T2-weighted turbo spin
Echo time (TE) 4.1 msec echo image.
Repeat time (TR) 147.2 msec
Flip angle (FA) 80° 20. Initiate the scan.
Field of view (FOVx, FOVy) 350 mm, 263 mm
Sequence 5: High resolution sagittal T2-weighted turbo spin echo (Fig. A19.1.1)
Resolution (Δx, Δy) 1.37 mm, 2.05 mm
21. Display both the coronal and sagittal scout images in two separate quadrants on the
Number of data points collected (Nx, Ny) 256, 128
scan monitor. Change imaging parameters to those listed in Table A19.1.6. Position
Display matrix (Dx, Dy) 256, 256
slices to cover the pelvis.
Slice thickness (Δz) 10 mm
Number of slices 20 22. Instruct the patient to remain motionless and to breathe normally as the scan will
Slice gap 2 mm begin and last for ∼3 min.
Number of acquisitions (Nacq) 1
Sequence 6: High resolution transverse T2-weighted turbo spin echo (Fig. A19.1.2)
Swap read and phase encoding No
23. Display both the coronal and transverse scout images in two separate quadrants on
Slice location Centered to cover pelvis
the scan monitor. Change imaging parameters to those listed in Table A19.1.7.
Saturation pulses No
Position slices to cover the pelvis.
Fat suppression Yes
Slice series Interleaved 24. Instruct the patient to remain motionless and to breathe normally as the scan will
Scan time 19 sec begin and last for ∼3 min.
Sequence 7: Transverse gradient echo with fat suppression (immediate post-contrast
scan; Fig. A19.1.3)
MRI of the
NOTE: See patient set up section for specific instructions on preparation for contrast agent
Male Pelvis Male Pelvis injection. This preparation can be done prior to placing the patient in the scanner.
A19.1.5 A19.1.6
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A19.1.6 Imaging Parameters for Sequence 5 (High Resolution
T2-Weighted Turbo Spin Echo)

Patient position Supine

Scan type Turbo spin echo
Imaging plane (orientation) Sagittal
Central slice or volume center Slices posted on coronal; center to
prostate
Echo time (TE) 132 msec
Echo train length (ETL) 15
Repeat time (TR) 4902 msec
Flip angle (FA) 180°a
Field of view (FOVx, FOVy) 350 mm, 263 mm
Resolution (Δx,Δy) 0.68 mm, 0.97 mm
Number of data points collected (Nx, Ny) 512, 270
Display matrix (Dx, Dy) 512, 512
Slice thickness (Δz) 5 mm Figure A19.1.2 High resolution transverse T2-weighted turbo spin echo image.
Number of slices 19
Slice gap 1.0 mm
Table A19.1.7 Imaging Parameters for Sequence 6 (High Resolution
Number of acquisitions (Nacq) 2 T2-Weighted Turbo Spin Echo)
Swap read and phase encoding No
Slice location Centered to cover prostate Patient position Supine
Saturation pulses No Scan type Turbo spin echo
Fat suppression No Imaging plane (orientation) Transverse
Slice series Interleaved Central slice or volume center Slices posted on coronal; center to
Scan time 3 min, 1 sec prostate
aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this Echo time (TE) 132 msec
sequence is 90°. Echo time length (ETL) 15
Repeat time (TR) 4902 msec
25. Display the midline slice of the scout coronal image and the transverse scout image Flip angle (FA) 180°a
in two separate quadrants on the scan monitor. Change imaging parameters to those Field of view (FOVx, FOVy) 350 mm, 263 mm
listed in Table A19.1.8. Position slices to cover the pelvis. Resolution (Δx, Δy) 0.68 mm, 0.97 mm
Number of data points collected (Nx, Ny) 512, 270
26. Explain to the patient that you will now be injecting the contrast agent and that he or Display matrix (Dx, Dy) 512, 512
she may feel a cool sensation in his or her arm. Initiate the injection. Do not begin Slice thickness (Δz) 5 mm
scanning until the 25-sec scan delay has expired. However, breathing instructions Number of slices 19
should be delivered when 10 sec of delay are remaining (see step 27). Slice gap 1.0 mm
If you do not have access to a power injector and are “hand” injecting follow step 26. Number of acquisitions (Nacq) 2
However, after you have completed the bolus contrast agent injection reattach the saline Swap read and phase encoding No
filled syringe and flush with 10 ml of saline. Begin breathing instructions. After all of the Slice location Centered to cover prostate
saline has been injected then proceed to initiate the scan. The process of switching syringes Saturation pulses No
must be completed as quickly as possible and thus, the suggestion of incorporating the use Fat suppression No
of a 3-way stopcock. Slice series Interleaved
27. When 15 sec after the contrast agent injection has passed, instruct the patient to take Scan time 3 min, 1 sec
aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this
in a deep breath and exhale, take in another deep breath and hold it.
sequence is 90°.
28. Initiate the scan.
Sequence 8: Sagittal gradient echo with fat suppression (post contrast)
29. Display the transverse image and the sagittal scout image in two separate quadrants
on the scan monitor. Change imaging parameters to those listed in Table A19.1.9.
Position slices to cover the pelvis.
30. Instruct the patient to take in a deep breath and exhale, take in another deep breath MRI of the
and hold it. Male Pelvis Male Pelvis

A19.1.7 A19.1.8
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A19.1.9 Imaging Parameters for Sequence 8 (Gradient Echo with Fat
Suppression)

Patient position Supine

Scan type Gradient echo
Imaging plane (orientation) Sagittal
Central slice or volume center Slices posted on transverse; center
to pelvis
Echo time (TE) 4.1 msec
Repeat time (TR) 147.2 msec
Flip angle (FA) 80°
Field of view (FOVx, FOVy) 350 mm, 300 mm
Resolution (Δx, Δy) 1.37 mm, 1.74 mm
Number of data points collected (Nx, Ny) 256, 172
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 8-10 mm
Number of slices 20
Figure A19.1.3 Transverse gradient echo with fat suppression immediate post contrast image. Slice gap 1.6-2 mm
Number of acquisitions (Nacq) 1
Swap read and phase encoding No
Table A19.1.8 Imaging Parameters for Sequence 7 (Gradient Echo with Fat Slice location Centered to cover pelvis
Suppression; Immediate Post-Contrast Scan)
Saturation pulses No
Patient position Supine Fat suppression Yes
Scan type Gradient echo Slice series Interleaved
Imaging plane (orientation) Transverse Scan time 25 sec
Central slice or volume center Slices posted on coronal; center to
pelvis
Table A19.1.10 Imaging Parameters for Sequence 9 (Gradient Echo with Fat
Echo time (TE) 4.1 msec Suppression)
Repeat time (TR) 147.2 msec
Flip angle (FA) 80° Patient position Supine
Field of view (FOVx, FOVy) 350 mm, 300 mm Scan type Gradient echo
Resolution (Δx, Δy) 1.37 mm, 1.74 mm Imaging plane (orientation) Transverse
Number of data points collected (Nx, Ny) 256, 172 Central slice or volume center Slices posted on coronal; center to
Display matrix (Dx, Dy) 256, 256 mid-abdomen
Slice thickness (Δz) 8 mm Echo time (TE) 4.1 msec
Number of slices 20 Repeat time (TR) 147.2 msec
Slice gap 1.6 mm Flip angle (FA) 80°
Number of acquisitions (Nacq) 1 Field of view (FOVx, FOVy) 350 mm, 263 mm
Swap read and phase encoding No Resolution (Δx, Δy) 1.37 mm, 2.74 mm
Slice location Centered to cover pelvis Number of data points collected (Nx, Ny) 256, 96
Saturation pulses No Display matrix (Dx, Dy) 256, 256
Fat suppression Yes Slice thickness (Δz) 8-10 mm
Slice series Interleaved Number of slices 20
Scan time 25 sec Slice gap 1.6-2 mm
Number of acquisitions (Nacq) 1
Swap read and phase encoding No
Slice location Centered to cover abdomen
Saturation pulses No
Fat suppression Yes
Slice series Interleaved
Scan time 14 sec

MRI of the
Male Pelvis Male Pelvis

A19.1.9 A19.1.10
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
31. Initiate the scan.
Sequence 9: Transverse gradient echo with fat suppression (post contrast;
mid-abdomen)
32. Display the coronal image and the transverse scout image in two separate quadrants
on the scan monitor. Change imaging parameters to those listed in Table A19.1.10.
Position slices to cover the mid-abdomen.
33. Instruct the patient to take in a deep breath and exhale, take in another deep breath
and hold it.
34. Initiate the scan
COMMENTARY
Background Information
The major indication for MRI of the male
pelvis is for the investigation of prostate cancer
(Mirowitz et al., 1994; Nunes et al., 1995;
Hricak et al., 1994; Schnall et al., 1991; Huch
Boni et al., 1996). The male pelvis MR protocol
is designed to evaluate the prostate gland and
the most common forms of metastases, lymph
nodes, and bone metastases.
Critical Parameters and
Troubleshooting
High-resolution T2-weighted images are
critical for the detection of prostate cancer (Mi-
rowitz et al., 1994; Nunes et al., 1995; Hricak
et al., 1994; Schnall et al., 1991; Huch Boni et
al., 1996). Using local receiver coils helps
achieve high spatial resolution imaging. The
above described protocol employs a phased
array multicoil. Other investigators have used
endorectal coils.
Anticipated Results
The non-contrast T1-weighted spoiled gra-
dient echo images of the upper abdomen and
pelvis are used to evaluate for lymph nodes.
Five millimeter thick transverse and sagittal
T2-weighted images are the principal technique
to evaluate the prostate and seminal vesicles.
The post-gadolinium T1-weighted fat sup-
pressed spoiled gradient echo sequence is the
most consistent sequence to detect bony metas-
tases. Using the above described sequences, the
detection and staging of prostate cancer is well
performed. This protocol is also effective at
evaluating other pelvis tumors such as bladder
cancer.
Literature Cited
Hricak, H., White, S., Vigneron, D., Kurhanewicz,
J., Kosco, A., Levin, D., Weiss, J., Narayan, P.,
and Carroll, P.R. 1994. Carcinoma of the prostate
Current Protocols in Magnetic Resonance Imaging
gland: MR imaging with pelvic phased-array
coils versus integrated endorectal-pelvic phased-
array coils. Radiology. 193:703-709.
Huch Boni, R., Meyenberger, C., Pok Lundquist, J.,
Trinkler, F., Lutolf, U., and Krestin, G. 1996.
Value of endorectal coil versus body coil MRI
for diagnosis of recurrent pelvic malignancies.
Abdom. Imaging. 21:345-352.
Mirowitz, S.A., Heiken, J.P., and Brown, J.J. 1994.
Evaluation of fat saturation technique for T2-
weighted endorectal coil MRI of the prostate.
Magn. Reson. Imaging. 12:743-747.
Nunes, L.W., Scheibler, M.S., Rauschning, W.,
Schnall, M.D., Tomaszewski, J.E., Pollack, H.,
and Kressel, H. 1995. The normal prostate and
periprostatic structures: Correlation between
MR images made with an endorectal coil and
cadaveric microtome sections. AJR Am. J.
Roentgenol. 164:923-927.
Schnall, M., Imai, Y., Tomaszewski, J., Pollack, H.,
Lenkinski, R., and Kressel, H. 1991. Prostate
cancer: Local staging with endorectal surface
coil MR imaging. Radiology. 178:797-802.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Key References
Shellock, 1996. See above.
Covers a number of important patient management
issues related to MR imaging, including recom-
mended safety procedures, a list of metallic implants
that have been tested for MR compatibility, and a
list of other sources on MR safety.
Contributed by Kathy Wilber and
Richard Semelka
University of North Carolina
Chapel Hill, North Carolina
Laurie Fisher
Siemens Uptime Service Center
Cary, North Carolina
Male Pelvis
A19.1.11
MRI of the Female Pelvis UNIT A20.1
MRI provides an effective means for evaluating the female pelvis. The benign female
pelvis protocol is designed to evaluate for the presence of fibroids or adenomyosis, uterine
anomalies, and endometriosis. This protocol also provides limited evaluation of adnexal
disease processes including: identification of ovaries, hemorrhagic ovarian cysts, and
dermoid tumors.
IMAGING THE FEMALE PELVIS BASIC
PROTOCOL
The benign disease protocol includes high-resolution turbo T2 and gradient echo se-
quences through the pelvis, and does not include gadolinium-enhanced images. The
sequences described herein are based on the authors’ experience with a Siemens 1.5 T
Vision scanner, but are expected to be equally applicable to machines from other
manufacturers.
Scanning a patient or volunteer is a joint effort among technologists, nurses, and
physicians, with the technologist normally responsible for following proper scanning
protocols and techniques. Unless otherwise specified, in what follows the person to whom
directions are given is assumed to be the technologist. Table A20.1.1 lists the hardware
necessary to perform the procedure, along with appropriate parameters.
The following seven sequences comprise the imaging protocol for the female pelvis.
Gradient echo sequences require the patient to be able to suspend respiration for ∼20 sec.
It is imperative that there is clear communication between the technologist and the patient
throughout the exam. This protocol results in consistent, reproducible image quality that
is effective for evaluating benign pelvic diseases.
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.
Set up patient and equipment
1. Interview (screen) the patient to assess for contraindications such as cardiac pace-
maker, implanted mechanical devices, and/or ferromagnetic materials. Also, deter-
mine if the patient will need sedation medication necessitating the use of appropriate
monitoring equipment.
A screening form is signed by each patient or legal guardian prior to bringing the patient
into the exam area.
Table A20.1.1 Equipment Specifications Needed to Perform the Following
Imaging Sequences
Coil type Circularly polarized body phased array coil
Manufacturer and system type Siemens, Vision
Field strength 1.5 T
Gradient coil strength 24 mT/m (or whatever the system permits, but
minimum of 24 mT/m for sequence 2)
Knee cushion Yes
Use of contrast agents No
Pulse oximeter If patient requires sedation
Female Pelvis
Contributed by Kathy Wilber, Richard Semelka, and Laurie Fisher A20.1.1
Current Protocols in Magnetic Resonance Imaging (2001) A20.1.1-A20.1.11
Copyright © 2001 by John Wiley & Sons, Inc.
 The presence of ferromagnetic materials may be a health hazard to the patient while in the Table A20.1.2 Imaging Parameters for Sequence 1 (Scout Sequence)
magnetic field and/or adversely affect image quality. To determine the safety of scanning
such ferromagnetic materials see Shellock (1996). Patient position Supine
The presence of ferromagnetic materials in the globe of the eye is contraindicated for MRI. Scan type Gradient echo
Patients with prior metal exposure to the eye should have plain x-rays of the orbital area Imaging plane (orientation) Sagittal, transverse, and coronal
to ensure that all metal has been removed prior to placing them in the magnetic field. Central slice or volume center Center to pelvis
Echo time (TE) 6 msec
2. Request the patient to change into a gown and remove all personal effects such as, Repeat time (TR) 15 msec
jewelry, hearing aids, glasses, etc., prior to entering the MRI scan room. All personal Flip angle (FA) 30°
belongings should be secured during the examination. Field of view (FOVx, FOVy) 450 mm, 450 mm
3. Explain the procedure to the patient and record relevant clinical history. Ensure that Resolution (Δx, Δy) 1.76 mm, 3.52 mm
the patient understands what is expected and ask them if they have any questions; Number of data points collected (Nx, Ny) 256, 128
answer appropriately. Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 10 mm
4. Set up the exam room by securing the circularly polarized (CP)-body array coil onto
Number of slices 3
the table and providing a clean exam table.
Slice gap Not applicable
5. Escort the patient to the MR examination room and ask them to lie down accordingly Number of acquisitions (Nacq) 1
with respect to the exam to be performed. Review the following items with the patient: Swap read and phase encoding No
a. Provide earplugs or headphones to the patient to minimize the loud knocking noise Slice location Not applicable
that will be produced by the gradients but ensure them that they will still be able Saturation pulses Not applicable
to hear you. Scan time 16 sec

b. Provide the patient with a safety squeeze-bulb and demonstrate how it works;
explain to the patient when to use the squeeze-bulb (i.e., if they need assistance
during the exam).
c. Explain to the patient that you will be talking to them between imaging sequences,
which will be when the loud knocking noise stops. Additionally, review breath-
holding instructions with the patient.
d. Explain to the patient that it is imperative that they remain motionless during the
loud knocking noise to ensure good results; also explain that they should not
reposition their body between imaging sequences.
e. Position a support under the patient’s knees to enhance patient comfort.
f. Provide the patient with an approximate time that the examination will take.
6. Secure the top portion of the CP-body array coil to prevent it from moving side-to-side
during breath-holding imaging sequences. Usually straps are provided that are
directly attached to the coil.
7. Using the laser light center the patient’s pelvis to the coil.
8. Advance the patient table to isocenter.

Sequence 1: Three-plane positioning scout

9. To validate the patient’s position and to have a reference to prescribe successive
imaging sequences, acquire a three-plane orthogonal scout sequence. See Table
A20.1.2 for specific parameters. Run the scan.
Most MR scanners can be programmed to acquire the scout automatically after coil tuning
or after the patient has been placed in isocenter (for systems that do not require tuning).

Sequence 2: Coronal half-acquisition (partial Fourier) turbo spin echo (Fig.

A20.1.1) Figure A20.1.1 Coronal half-acquisition turbo spin echo image.
10. Display both the transverse and coronal scout images in two separate quadrants on
MRI of the the scan monitor. Change imaging parameters to those listed in Table A20.1.3.
Female Pelvis Female Pelvis

A20.1.2 A20.1.3
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Position slices to center of the transverse scout ensuring that the pelvis and as much Table A20.1.4 Imaging Parameters for Sequence 3 (High Resolution
of the abdomen as possible is covered. T2-Weighted Turbo Spin Echo)

11. Instruct the patient to remain motionless and to breathe normally as the scan will Patient position Supine
begin and last for ∼40 sec. Run the scan. Scan type Turbo spin echo
Sequence 3: High resolution transverse T2-weighted turbo spin echo (Fig. A20.1.2) Imaging plane (orientation) Transverse
12. Display both the coronal and transverse scout images in two separate quadrants on Central slice or volume center Slices posted on coronal; center to
pelvis
the scan monitor. Change imaging parameters to those listed in Table A20.1.4.
Echo time (TE) 132 msec
Position slices to cover the pelvis.
Echo train length (ETL) 15
Repeat time (TR) 4902 msec
Table A20.1.3 Imaging Parameters for Sequence 2 (Half-Acquisition Turbo Flip angle (FA) 180°a
Spin Echo)
Field of view (FOVx, FOVy) 350 mm, 263 mm
Patient position Supine Resolution (Δx, Δy) 0.68 mm, 0.97 mm
Scan type Half acquisition turbo spin echo Number of data points collected (Nx, Ny) 512, 270
Imaging plane (orientation) Coronal Display matrix (Dx, Dy) 512, 512
Central slice or volume center Slices posted on transverse scout; Slice thickness (Δz) 5 mm
center to abdomen and pelvis Number of slices 19
Echo time (TE) 90 msec Slice gap 1.0 mm
Repeat time (TR) 4.4 msec (NOTE: The true TR is Numbers of acquisitions (Nacq) 2
infinite, 4.4 msec represents the Swap read and phase encoding No
echo spacing) Slice location Centered to cover pelvis
Delay time (TD) 1500 msec Saturation pulses No
Flip angle (FA) 150° Fat suppression No
Field of view (FOVx, FOVy) 400 mm, 400 mm Slice series Interleaved
Resolution (Δx, Δy) 1.56 mm, 2.08 mm Scan time 3 min, 1 sec
Number of data points collected (Nx, Ny) 256, 192 (using half Fourier) aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this
Display matrix (Dx, Dy) 256, 256 sequence is 90°.
Slice thickness (Δz) 8–10 mm
Number of slices 20 13. Instruct the patient to remain motionless and to breathe normally as the scan will
Slice gap 1.6–2 mm begin and last for ∼3 min. Run the scan.
Number of acquisitions (Nacq) 1
Swap read and phase encoding No Sequence 4: High resolution sagittal T2-weighted turbo spin echo (Fig. A20.1.3)
Slice location Centered to abdomen and pelvis 14. Display both the transverse and sagittal scout images in two separate quadrants on
Saturation pulses No the scan monitor. Change imaging parameters to those listed in Table A20.1.5.
Slice series Interleaved Position slices to cover the pelvis.
Scan time 40 sec
15. Instruct the patient to remain motionless and to breathe normally as the scan will
begin and last for ∼3 min. Run the scan.

Sequence 5: Transverse gradient echo

16. Display the midline slice of the coronal and the transverse scout image in two separate
quadrants on the scan monitor. Change imaging parameters to those listed in Table
A20.1.6. Position slices to cover the pelvis.
17. Instruct the patient to take in a deep breath and exhale, take in another deep breath
and hold it. Run the scan.

Sequence 6: Sagittal gradient echo with fat suppression

18. Display both the transverse and sagittal scout images in two separate quadrants on
the scan monitor. Change imaging parameters to those listed in Table A20.1.7.
Position slices to center of the transverse scout ensuring that the pelvis covered.
MRI of the
Female Pelvis Female Pelvis
Figure A20.1.2 High resolution transverse T2-weighted turbo spin echo image.
A20.1.4 A20.1.5
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A20.1.5 Imaging Parameters for Sequence 4 (High Resolution
T2-Weighted Turbo Spin Echo)

Patient position Supine

Scan type Turbo spin echo
Imaging plane (orientation) Sagittal
Central slice or volume center Slices posted on coronal; center to
pelvis
Echo time (TE) 132 msec
Echo train length (ETL) 15
Repeat time (TR) 4902 msec
Flip angle (FA) 180°a
Field of view (FOVx, FOVy) 350 mm, 263 mm
Resolution (Δx, Δy) 0.68 mm, 0.97 mm
Number of data points collected (Nx, Ny) 512, 270
Display matrix (Dx, Dy) 512, 512
Slice thickness (Δz) 5 mm
Number of slices 19
Slice gap 1.0 mm
Number of acquisitions (Nacq) 2
Swap read and phase encoding No
Slice locations Centered to cover pelvis
Saturation pulses No
Fat suppression No
Slice series Interleaved
Scan time 3 min, 1 sec
aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this

sequence is 90°.

Table A20.1.6 Imaging Parameters for Sequence 5 (Gradient Echo)

Patient position Supine

Figure A20.1.3 High resolution sagittal turbo spin echo-image. Scan type Gradient echo
Imaging plane (orientation) Transverse
Central slice or volume center Slices posted on coronal; center to
pelvis
19. Instruct the patient to take in a deep breath and exhale, take in another deep breath Echo time (TE) 4.5 msec
and hold it. Repeat time (TR) 140 msec
20. Initiate the scan. Flip angle (FA) 80°
Field of view (FOVx, FOVy) 350 mm, 263 mm
Sequence 7: Transverse gradient echo with fat suppression Resolution (Δx, Δy) 1.37 mm, 2.05 mm
21. Display both the coronal and transverse scout images in two separate quadrants on Number of data points collected (Nx, Ny) 256, 128
the scan monitor. Change imaging parameters to those listed in Table A20.1.8. Display matrix (Dx, Dy) 256, 256
Position slices to center of the coronal scout ensuring that the pelvis covered. Slice thickness (Δz) 8–10 mm
Number of slices 19
22. Instruct the patient to take in a deep breath and exhale, take in another deep breath Slice gap 1.6–2 mm
and hold it. Number of acquisitions (Nacq) 1
23. Initiate the scan. Swap read and phase encoding No
Slice location Centered to cover pelvis
Saturation pulses No
Slice series Interleaved
Scan time 18 sec
MRI of the
Female Pelvis Female Pelvis

A20.1.6 A20.1.7
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A20.1.7 Imaging Parameters for Sequence 6 (Gradient Echo with Fat COMMENTARY
Suppression)
Background Information and endometriosis (T1-weighted fat suppressed
Patient position Supine Noncontrast enhanced MR protocols em- spoiled gradient echo sequences). Limited
Scan type Gradient echo phasizing T2-weighted sequences (Hricak, evaluation of adnexal disease processes is also
1986; Smith et al., 1992; Scoutt et al., 1991; achieved including: identification of ovaries
Imaging plane (orientation) Sagittal
Pellerito et al., 1992; Carrington et al., 1990; (T2-weighted sequences), hemorrhagic ovarian
Central slice or volume center Slices posted on transverse; center
Reinhold et al., 1997; Zawin et al., 1990; We- cysts (T1-weighted fat suppressed spoiled gra-
to pelvis
inreb et al., 1990; Hricak et al., 1986; Mark et dient echo sequences), and dermoid tumors
Echo time (TE) 4.1 msec
al., 1987; Ascher et al., 1994; Togashi et al., (combined T1-weighted conventional and fat
Repeat time (TR) 147.2 msec 1989), and T1-weighted (short echo) sequences suppressed spoiled gradient echo sequences;
Flip angle (FA) 80° especially with fat suppression (Kier et al., Kier et al., 1992; Stevens et al., 1993; Sugimura
Field of view (FOVx, FOVy) 350 mm, 263 mm 1992; Stevens et al., 1993; Sugimura et al., et al., 1993; Outwater et al., 1996; Togashi et
Resolution (Δx, Δy) 1.37 mm, 2.05 mm 1993; Outwater et al., 1996; Togashi et al., al., 1991; Ascher et al., 1995; Ha et al., 1994;
Number of data points collected (Nx, Ny) 256, 128 1991; Ascher et al., 1995; Ha et al., 1994; Siegelman et al., 1994).
Display matrix (Dx, Dy) 256, 256 Siegelman et al., 1994) are effective at evaluat- The high soft tissue contrast resolution of
Slice thickness (Δz) 7 mm ing benign diseases in the female pelvis. The MRI of benign disease of the female pelvis has
Number of slices 20 benign female pelvis protocol is designed to been extensively documented (Hricak, 1986;
Slice gap 1.4 mm evaluate for the presence of fibroids or ade- Smith et al., 1992; Scoutt et al., 1991; Pellerito
Number of acquisitions (Nacq) 1 nomyosis (T2-weighted sequences; Hricak, et al., 1992; Carrington et al., 1990; Reinhold
Swap read and phase encoding No 1986; Smith et al., 1992; Scoutt et al., 1991; et al., 1997; Zawin et al., 1990; Weinreb et al.,
Zawin et al., 1990; Weinreb et al., 1990; Hricak 1990; Hricak et al., 1986; Mark et al., 1987;
Slice locations Centered to cover pelvis
et al., 1986; Mark et al., 1987; Ascher et al., Ascher et al., 1995; Togashi et al., 1989; Kier
Saturation pulses No
1995; Togashi et al., 1989), uterine anomalies et al., 1992; Stevens et al., 1993; Sugimura et
Fat suppression Yes (T2-weighted sequences; Pellerito et al., 1992; al., 1993; Outwater et al., 1993; Togashi et al.,
Slice series Interleaved Carrington et al., 1990; Reinhold et al., 1997) 1991; Ascher et al., 1995; Ha et al., 1994;
Scan time 19 sec

Table A20.1.9 Imaging Parameters for Transverse Half-Acquisition Turbo Spin

Echoa
Table A20.1.8 Imaging Parameters for Sequence 7 (Gradient Echo with Fat
Suppression)
Patient position Supine
Scan type Half acquisition turbo spin echo
Patient position Supine
Imaging plane (orientation) Transverse
Scan type Gradient echo
Central slice or volume center Slices posted on coronal; center to
Imaging plane (orientation) Transverse
pelvis
Central slice or volume center Slices posted on coronal; center to
Echo time (TE) 90 msec
pelvis
Repeat time (TR) 4.4 msec (NOTE: the true TR is
Echo time (TE) 4.1 msec
infinite, 4.4 msec represents the
Repeat time (TR) 147.2 msec
echo spacing)
Flip angle (FA) 80° Delay time (TD) 1500 msec
Field of view (FOVx, FOVy) 350 mm, 263 mm Flip angle (FA) 150°
Resolution (Δx, Δy) 1.37 mm, 2.05 mm
Field of view (FOVx, FOVy) 350 mm, 263 mm
Number of data points collected (Nx, Ny) 256, 128
Resolution (Δx, Δy) 1.37 mm, 1.37 mm
Display matrix (Dx, Dy) 256, 256 Number of data points collected (Nx, Ny) 256, 192 (using half Fourier)
Slice thickness (Δz) 8 mm Display matrix (Dx, Dy) 256, 256
Number of slices 20
Slice thickness (Δz) 8–10 mm
Slice gap 1.6 mm
Number of slices 20
Number of acquisitions (Nacq) 1 Slice gap 1.6–2 mm
Swap read and phase encoding No Number of acquisitions (Nacq) 1
Slice location Centered to cover pelvis
Swap read and phase encoding No
Saturation pulses No
Slice location Centered to cover the pelvis
Fat suppression Yes Saturation pulses Yes, superior and inferior to slices
Slice series Interleaved Fat suppression No
Scan time 19 sec
Slice series Interleaved
MRI of the Scan time 40 sec
Female Pelvis aUse instead of high resolution T -weighted turbo spin echo for patients who are unable to hold still.
Female Pelvis
2
A20.1.8 A20.1.9
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A20.1.10 Imaging Parameters for Sagittal Half-Acquisition Turbo Spin Literature Cited ment. Lippincott Williams and Wilkins, Phila-
Echoa Ascher, S.M., Agrawal, R., Bis, K.G., Brown, E.D., delphia.
Maximovich, A., Markham, S.M., Patt, R.H., Siegelman, E.S., Outwater, E.K., Wang, T., and
Patient position Supine and Semelka, R.C. 1995. Endometriosis: Ap- Mitchell, D.G. 1994. Solid pelvic masses caused
pearance and detection with conventional and by endometriosis: MR imaging features. AJR
Scan type Half acquisition turbo spin echo contrast-enhanced fat-suppressed spin-echo Am. J. Roentgenol. 163:357-361.
Imaging plane (orientation) Sagittal techniques. J. Magn. Reson. Imaging. 5:251-
257. Smith, R.C., Reinhold, C., McCauley, T.R., Lange,
Central slice or volume center Slices posted on coronal; center to R.C., Constable, R.T., Kier, R., McCarthy, S.
pelvis Ascher, S.M., Arnold, L.L., Patt R.H., Schruefer, 1992. Multicoil high-resolution fast spin-echo
Echo time (TE) 90 msec J.J., Bagley, A.S., Semelka, R.C., Zeman, R.K., MR imaging of the female pelvis. Radiolog y
and Simon, J.A. 1994. Adenomyosis: Prospec- 184:671-675.
Repeat time (TR) 4.4 msec (NOTE: the true TR is tive comparison of MR imaging and transvaginal
infinite, 4.4 msec represents the sonography. Radiology 190:803-806. Stevens, S.K., Hricak, H., and Campos, Z. 1993.
echo spacing) Teratomas versus cystic hemorrhagic adnexal
Carrington, B.M., Hricak, H., Nuruddin, R.N., Se- lesions: Differentiation with proton-selective
Delay time (TD) 1500 msec caf, E., Laros, R.K., and Hill, E.C. 1990. fat-saturation MR imaging. Radiology 186:481-
Flip angle (FA) 150° Mullerian duct anomalies. MR imaging evalu- 488.
Field of view (FOVx, FOVy) 350 mm, 263 mm ation. Radiology 176:715-720.
Sugimura, K., Okizuka, H., Imaoka, I., Kaji, Y.,
Resolution (Δx, Δy) 1.37 mm, 1.37 mm Ha, H.K., Lim, Y.T., Kim, H.S., Suh, T.S., Song, Takahashi, K., Kitao, M., and Ishida, T. 1993.
H.H., and Kim, S.J. 1994. Diagnosis of pelvic Pelvic endometriosis: Detection and diagnosis
Number of data points collected (Nx, Ny) 256, 192 (using half Fourier)
endometriosis: Fat-suppressed T1-weighted vs. with chemical shift MR imaging. Radiology
Display matrix (Dx, Dy) 256, 256 co nventional MR images. AJ R Am . J. 188:435-438.
Slice thickness (Δz) 8–10 mm Roentgenol. 163:127-131.
Togashi, K., Nishimura, K., Kimura, I., Tsuda, Y.,
Number of slices 20 Hricak, H. 1986. MRI of the female pelvis: A re- Yamashita, K., Shibata, T., Nakano, Y., Konishi,
Slice gap 1.6–2 mm view. AJR Am. J. Roentgenol. 146(6):1115- J., Konishi, I., and Mori, T. 1991. Endometrial
1122. cyst: Diagnosis with MR imaging. Radiology
Number of acquisitions (Nacq) 1
Hricak, H., Tscholakoff, D., Heinrichs, L., Fisher, 180:73-78.
Swap read and phase encoding No
M.R., Dooms, G.C., Reinhold, C., and Jaffe, Togashi, K., Ozasa, H., Konishi, I., Itoh, H.,
Slice locations Centered to cover the pelvis R.B. 1986. Uterine leiomyomas: Correlation of Nishimura, K., Fujisawa, I., Noma, S., Sagoh, T.,
Saturation pulses Yes, superior and inferior to slices MR histopathologic findings, and symptoms. Minami, S., Yamashita, K., et al. 1989. Enlarged
Fat suppression No Radiology 158:385-391. uterus: Differentiation between adenomyosis
Slice series Interleaved Kier, R., Smith, R.C., and McCarthy, S.M. 1992. and leiomyoma with MR imaging. Radiology
Value of lipid- and water-suppression MR im- 171:531-534.
Scan time 40 sec ages in distinguishing between blood and lipid Weinreb, J.C., Barkoff, N.D., Megibow, A., and
aUse instead of high resolution T -weighted turbo spin echo for patients who are unable to hold still.
2 with ovarian masses. AJR Am. J. Roentgenol. Demopoulos, R. 1990. The value of MR imaging
158:321-325. in distinguishing leiomyomas from other solid
Mark, A.S., Hricak, H., Heinrichs, L.W., Hendrick- pelvic masses when sonography is indetermi-
Siegelman et al., 1994). No other imaging mo- In the event that patients are unable to hold son, M.R., Winkler, M.L., Bachica, J.A., and nate. AJR Am. J. Roentgenol. 154:295-299.
dality can demonstrate the same clarity of find- relatively still, it has been found that obtaining Stickler, J.E. 1987. Adenomyosis and Zawin, M., McCarthy, S., Scoutt, L.M., and Comite,
ings in the evaluation of benign diseases of the transverse and sagittal projection images using leiomyoma: Differential diagnosis with MR im- F. 1990. High-field MRI and US evaluation of
female pelvis. the half-acquisition turbo spin echo sequence aging. Radiology 163:527-529. the pelvis in women with leiomyomas. Magn.
(see Tables A20.1.9 and A20.1.10) is adequate Outwater, E.K. and Mitchell, D.G. 1996. Normal Reson. Imaging. 8:371-376.
Critical Parameters and to define normal organs and the majority of ovaries and functional cysts: MR appearance.
Troubleshooting disease processes. Radiology 198:397-402. Key References
The critical sequences for the evaluation of The potential problem with fat suppression Pellerito, J.S., McCarthy, S.M., Doyle, M.B., Glick- Shellock and Kanal, 1996. See above.
benign diseases of the female pelvis are T2- is heterogeneous fat suppression that results in man, M.G., and DeCherney, A.H. 1992. Diagno- Covers a number of important patient management
sis of uterine anomalies: Relative accuracy of issues related to MR imaging, including recom-
weighted sequences and T1-weighted fat sup- difficulty interpreting the variations of tissue MR imaging, endovaginal sonography, and hys- mended safety procedures, a list of metallic implants
pressed sequences. It is also important to ac- signal intensity. Because of the relatively ho- terosalpingography. Radiology 183:795-800. that have been tested for MR compatibility, and a
quire images in both the transverse and sagittal mogeneous magnetic susceptibility and large Reinhold, C., Hricak, H., Forstner, R., Ascher, S.M., list of other sources on MR safety.
planes. The sagittal plane is crucial as the fe- volume of tissues in the pelvis, and relatively Bret, P.M., Meyer, W.R., and Semelka, R.C.
male gynecological organs all line up in this stationary tissue position through the respira- 1997. Primary amenorrhea: Evaluation with MR
projection and can be readily identified by fol- tory cycle, fat suppression is generally homo- imaging. Radiology 203: 383-390.
Contributed by Kathy Wilber and
lowing the course of the vagina to the cervix to geneous. Scoutt, L.M., Flynn, S.D., Luthringer, D.J., Richard Semelka
the uterus. Coronal plane images are used to McCauley, T.R., and McCarthy, S.M. 1991. University of North Carolina
supplement the transverse and sagittal plane Anticipated Results Junctional zone of the uterus: Correlation of MR Chapel Hill, North Carolina
imaging and histologic examination of hysterec-
images. The cervix is a useful landmark as it Using an MR protocol that employs image tomy specimens. Radiology 179:403-407.
can be readily identified as a dark tubular struc- acquisition in the transverse and sagittal planes, Laurie Fisher
Shellock, F.G. and Kanal, E. 1996. Magnetic Reso- Siemens Uptime Service Center
ture on T2-weighted sequences. T2-weighted images, and fat suppressed T1- nance Bioeffects, Safety, and Patient Manage-
In general, the authors attempt to acquire weighted images, a comprehensive range of Cary, North Carolina
high-resolution (512 points in the read direc- female pelvic diseases can be evaluated.
MRI of the tion) T2-weighted turbo spin echo sequences.
Female Pelvis Female Pelvis

A20.1.10 A20.1.11
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
Ovarian Cancer Diagnosis and Staging
The techniques used for detecting and staging ovarian cancer are similar to those used for
general imaging of the liver/abdomen and the pelvis for other applications. However, the
importance of imaging both the abdomen and pelvis, and the greater reliance on delayed
rather than early post-gadolinium images have led us to include a separate unit for imaging
patients with suspected or confirmed ovarian cancer. The authors’ experience has been
acquired primarily using GE equipment. However, the basic approach should be quite
similar, with most details relating to minor vendor-specific differences. Readers are
encouraged to read UNIT A15.2, the alternate liver unit for GE users, for general comments
on breath-holding techniques as well as the other chapters on abdominal and pelvic
imaging techniques. These, as well as the authors’ other protocols, are available on the
web at: http://www.mri.tju.edu/.
OVARIAN CANCER
Whenever possible, patients should be triaged to MRI machines that utilize phased-array
technology to improve signal-to-noise while maintaining temporal and spatial resolution, and
have gradient and radiofrequency (RF) subsystems that are adequate in imaging large portions
of the abdomen within a single breathhold. For initial diagnosis of ovarian cancer, T1-weighted
and T2-weighted thin-slice images should be acquired with small pixel size. In the lower
pelvis, breathholding is not essential, especially on unenhanced images, but the breathhold
technique will greatly improve image sharpness above the true pelvis, especially on gadolin-
ium-enhanced images. The dynamic technique is less essential for many other applications,
but should generally be included for the liver portion of the examination to maximize
diagnostic yield and to reduce the likelihood of ambiguous findings.
The most comprehensive method for detecting and staging ovarian carcinoma begins with
high resolution T1-weighted and T2-weighted imaging of the pelvis using the torso phased
array coil. Then, the coil is moved to the abdomen for imaging the liver, including
examination with dynamic gadolinium-enhanced technique. Finally, the coil is reposi-
tioned inferiorly to obtain post-contrast imaging of the pelvis. In post-operative patients
who cannot tolerate a long examination, thicker image slices can be used for the
pre-contrast portion of the pelvic examination. This examination should require ≤1 hr.
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen. Table A20.2.1 lists the hardware necessary to
perform the procedure, along with appropriate parameters.
Table A20.2.1 Equipment Specifications Needed to Perform the
Imaging Sequences
Coil type Torso phased array coil
Manufacturer and system GE Signa
type
Field strength 1.5 T
Gradient strength 25 mT/m (or whatever the system permits)
Knee cushion Yes
Pulse oximeter If patient requires sedation
Power injector Yes, if available
35-in. extension tubing Yes
Use of contrast agents Yes
Contributed by Donald G. Mitchell, Peter Natale, and George Holland
Current Protocols in Magnetic Resonance Imaging (2002) A20.2.1-A20.2.18
Copyright © 2002 by John Wiley & Sons, Inc.
UNIT A20.2 Materials
Normal saline (0.9% NaCl), sterile, 100 ml minimum
Extravascular contrast agent (e.g., Magnevist, Omniscan, or Prohance), 20 ml for
most patients
1 mg glucagon for intramuscular injection
22-G angiocatheter
Extension tubing (110-in.)
Power injector
Set up patient and equipment
1. Interview the patient to assess for contraindications such as a cardiac pacemaker,
implanted mechanical devices, and/or ferromagnetic materials. Also, determine if the
patient will need sedation medication necessitating the need for appropriate moni-
toring equipment.
BASIC Generally, standard screening forms are used for all patients scanned in a magnetic
PROTOCOL resonance system (see APPENDIX 1).
The presence of ferromagnetic materials may be a health hazard to the patient while in the
magnetic field and/or adversely affect image quality. To determine the safety of scanning
such ferromagnetic materials, see Shellock and Kanal (1996).
The presence of ferromagnetic materials in the globe of the eye is contraindicated for MRI.
Patients with prior metal exposure to the eye should have plain X-rays of the orbital area
to ensure that all metal has been removed prior to placing them in the magnetic field. See
Shellock and Kanal (1996) for discussion of what implants may be scanned safely using
magnetic resonance.
In patients with prior surgery, metallic clips may produce artifacts that are especially
deleterious for fat-suppressed techniques. If metallic clips are present, the shortest possible
TEs should be used, and unenhanced fat-suppressed images should be examined carefully
for areas with poor fat suppression, especially when accompanied by water suppression.
In areas such as these, a non-fat-suppressed technique should be used for the post-contrast
images.
2. If the procedure is a research protocol, have the patient sign any necessary consent
forms.
3. Instruct the patient to change into a gown and remove all personal effects, such as
jewelry, hearing aids, glasses, and other objects, prior to entering the MRI scan room.
All personal belongings should be secured during the examination.
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating
and image artifacts.
5. Explain the procedure to the patient and record relevant clinical history. Ensure that
the patient understands what is expected and ask them if they have any questions;
answer appropriately.
6. Obtain intravenous (i.v.) access utilizing a 22-G angiocatheter and attach to a
saline-filled extension tubing (110-in.), with saline set to slow infusion by gravity
drip or power injector to keep the line open. Secure position of angiocatheter with
sterile clear, occlusive dressing.
Obtaining i.v. access prior to entering the scan room will promote patient throughput and
eliminate “dead” time of starting the i.v. while the patient is on the exam table. Follow
power-injector manufacturer guidelines with regard to appropriate needle gauge/angio-
Ovarian Cancer catheter to be used; this will depend on chosen injection/flow rates.
Diagnosis and
Female Pelvis Staging
A20.2.1 A20.2.2
Supplement 6 Supplement 6 Current Protocols in Magnetic Resonance Imaging
 If a power injector is not accessible, dynamic imaging is still obtainable. In this case,
draw-up the contrast agent in one, 20-ml syringe and saline in two others, and flush the
timing and examination boluses. Alternatively, larger syringes can be used, drawing 20 ml
of contrast agent into one and 40 ml of saline into the other. To eliminate the need for
attaching and detaching syringes when switching from contrast to flush, incorporate the
use of a 3-way stopcock or Y-tubing.
7. Set-up the power injector as specified by the manufacturer.
A minimum of 40 ml normal saline should be drawn-up to ensure sufficient saline is
available to keep the vein open (KVO) throughout the exam. While there is no need to double
dose for hepatic MRI, a higher dose is useful for improving the conspicuity of enhanced
tissue throughout the abdomen and pelvis. The dose should be between 20 ml and 40 ml,
but we are not aware of any comparative study for this application to determine the
improvement of using >20 ml.
8. Escort the patient to the MR examination room and ask her to lie down accordingly
with respect to the exam to be performed. Connect the extension tubing secured to
the syringe to the power injector extension tubing. Review the following items with
the patient:
a. Provide earplugs or headphones to the patient to minimize the loud knocking noise
that will be produced by the gradients but ensure her that she will still be able to
hear you.
b. Provide the patient with a safety squeeze-bulb and demonstrate how it works;
explain to the patient when to use the squeeze-bulb (i.e., if she needs assistance
during the exam).
c. Explain to the patient that you will be talking to her between imaging sequences,
which will be when the loud knocking noise stops. Additionally, review breath-
holding instructions with the patient.
d. Explain to the patient that it is imperative that she remain motionless during the
imaging to ensure good results; also explain that she should not reposition her
body between imaging sequences.
e. Nevertheless, the patient may call out at any time if she feels it necessary.
f. Provide the patient with an approximate time that the examination will take.
g. Attach the power injector to the i.v., and set it to keep the vein open.
h. Inject 1 mg glucagon intramuscularly.
9. Secure the torso phased array coil to the patient, centering approximately four fingers
below the iliac crest in most patients.
While the pelvic phased array coil, if available, provides a slightly better signal-to-noise
ratio (SNR) of the pelvis than does the torso coil, the latter should be used since it can be
used for both the pelvic and abdominal portions.
Usually, straps are provided by the manufacturer that are directly attached to the coil. A
foam pad can be placed between the patient’s abdomen and the coil to reduce near-field
artifacts.
10. Place a support under the patient’s knees to enhance patient comfort.
11. Insert nasal canula for oxygen and set the flow rate to 2 liter/min, if available, to aid
in suspended respiration. A total of 2 liter of oxygen should be administered.
Hyperventilation before breath-holding is helpful. The administration of oxygen is most
helpful during the examination.
Current Protocols in Magnetic Resonance Imaging
12. Use the laser light to center the patient, at about three fingers below the iliac crests.
Advance the patient table to isocenter.
Once this step has been performed, so long as the patient does not move on the table, the
table itself can be moved and then replaced in the same position as before without
jeopardizing the positioning of one scan relative to another.
13. Program the power injector for a contrast agent and saline injection rate of 2 ml/sec.
Program the volume of saline following contrast agent injection for 20 ml. The scan
delay in healthy young patients averages 18 sec (when contrast agent is injected,
initiate scan 18 sec after contrast agent and saline are delivered), but this is highly
variable in the elderly or those with cardiovascular disease.
As indicated in the authors’ unit on liver MRI (UNIT A15.2), they strongly recommend
determination of the optimum delay in each patient. This can be done either with the
automated bolus detection (SMARTPREP) feature, or by use of a 2-ml timing bolus
sequence, as per Earls et al. (1997) and described later in this unit. Both of these methods
can be adapted to either hand or power injection.
14. Arm the power injector and keep the vein open. Do not inject the contrast agent.
Sequence 1: Coronal single-shot fast-spin echo (SSFSE)
Although some technologists precede this with a three-plane orthogonal scout sequence,
the authors find that this can be omitted safely by most experienced technologists. The
coronal single shot fast-spin echo (FSE) sequence of the pelvis is nearly always of high
quality and high information content (Fig. A20.2.1). These can be acquired using a 320
mm to 360 mm square field of view (FOV) and 5-mm image slices. The image should be
centered at or slightly below the iliac crests.
The average patient’s pelvis or abdomen can be covered in two breath-holdings, while
larger patients may require three. Adjust the number of acquisitions before pause to limit
times of suspended respiration to 20 to 25 sec. For example, if 28 slices are needed, this
can be accomplished by acquiring two sets of 14 images, with suspended respirations of
∼20 sec each. The pause between image sets should be ∼10 sec, to allow the patient to
rest between suspended respirations.
15. Instruct the patient to remain motionless, and to take a deep breath in, blow it out,
deep breath in, blow it out, and hold it. Run the sequence according to the parameters
listed in Table A20.2.2. Let the patient breathe for 10 sec, ask patient to hold her
breath, and finish running this sequence.
If a torso phased array coil is used, the kidneys are outside its optimal range. The pelvis
should be included at the bottom of the image, while at least the lower portion of the kidneys
is included at the top part of the image. Even though the kidneys are outside the optimal
range of the torso phased array coil, the images will be sufficient to verify their presence
and positions, and to detect or exclude urinary obstruction. If benign disease is diagnosed
on the pelvic examination, scanning of the abdomen may not be necessary.
16. Confirm positioning of the torso phased array coil by making certain that all parts of
the pelvis are included within the sensitive volume of the coil. Reposition and repeat
the sequence, if necessary.
Sequence 2: Transverse T2-weighted FSE for pelvis
17. Using the coronal images, position slices to ensure that the entire pelvis is covered
on transverse images, from bottom of symphysis pubis to iliac crests (Fig. A20.2.2).
Run the sequence according to the parameters listed in Table A20.2.3.
Ovarian Cancer
Diagnosis and The patient should be instructed to breath quietly, with minimal body movement.
Female Pelvis Staging
A20.2.3 A20.2.4
Supplement 6 Supplement 6 Current Protocols in Magnetic Resonance Imaging
Figure A20.2.1 Coronal T2-weighted single-shot FSE sequence used as a localizer, with field of Figure A20.2.2 Transverse T2-weighted FSE image. Large white arrow indicates left adnexal
view of 36 cm. White arrow indicates a cystic mass in the left adnexa. Black arrows indicate the cystic mass, which has dependent low signal. Thin white arrow indicates the right ovary, black arrow
lower poles of the kidneys; although they are out of ideal range of the torso phased array coil, image indicates the uterus. There is a small amount of free pelvic fluid.
quality is sufficient to verify their locations and lack of collecting system dilatation.
Table A20.2.3 Imaging Parameters for Sequence 2 (Fast-Spin Echo)
Table A20.2.2 Imaging Parameters for Sequence 1 (Single-Shot Fast-Spin Echo)
Patient position Supine
Patient position Supine Scan type Fast spin echo
Scan type Single-shot fast-spin echo Imaging plane (orientation) Transverse
Imaging plane (orientation) Coronal Variable bandwidth Yes
Variable bandwidth Yes Pulse sequence database (PSD) fse
Pulse sequence database (PSD) ssfse Central slice or volume center Center to pelvis
Central slice or volume center Center 3 fingers below iliac crests Echo time (TE) 100 msec (effective)
Echo time (TE) 100 msec (effective) Receiver bandwidth (RBW) ±31 kHz
Receiver bandwidth (RBW) ±31 kHz Echo train length (ETL) 12
Repeat time (TR) Infinite Repeat time (TR) 4000 msec
Flip angle (FA) Default (90°) Flip angle (FA) Default (90°)
Fields of view (FOVx, FOVy) 360 mm, 360 mm Fields of view (FOVx, FOVy) 200 mm, 200 mm
Resolution (Δx, Δy) 1.41 mm, 2.77 mm Resolution (Δx, Δy) 0.78 mm, 0.78 mm
Number of data points collected (Nx, Ny) 256, ∼130 Number of data points collected (Nx, Ny) 256, 256
Display matrix (Dx, Dy) 256, 256 Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 5 mm Slice thickness (Δz) 4 mm
Number of slices 14 for one set (28 in total)a Number of slices 32
Slice gap 0 Slice gap 0.5 mm
Number of acquisitions (Nacq) 0.5 (half Fourier) Number of acquisitions (Nacq) 2
Swap read and phase encoding No Swap read and phase encoding Yes
Slice locations A40–P40 to cover the entire pelvis Read direction Anterior–posterior
Saturation pulses None Ovarian Cancer Slice location Pelvis
Scan time 20 sec for one set (40 sec total)a Diagnosis and
Female Pelvis Staging Saturation pulses Superior, inferior
aSee text under sequence 1. Scan time 4 min
A20.2.5 A20.2.6
Current Protocols in Magnetic Resonance Imaging Supplement 6 Supplement 6 Current Protocols in Magnetic Resonance Imaging
Sequence 3: T1-weighted transverse spin echo
18. Obtain images as for sequence 3 according to the parameters listed in Table A20.2.4
(Fig. A20.2.3).
The patient should be instructed to breathe quietly, with minimal body movement.
19. Pull the patient out of the magnet. Reposition the torso phased array coil for imaging
the upper abdomen, centering it three fingers below the xyphoid (or switch to
abdominal portion of abdominal/pelvic coil). Send the patient into the magnet.
If abdomen is to be included, which is necessary for staging ovarian cancer, proceed to
sequence 4. If a full abdominal exam is not needed, as for a patient with a benign adnexal
mass, skip to sequence 11.
Sequence 4: Coronal single-shot FSE of abdomen
20. Run sequence 4 using a 360-mm square FOV and 5-mm slice thickness, according
to the parameters listed in Table A20.2.5.
Adjust the number of slices before running the sequence to limit times of suspended
respiration to 20 to 25 sec. For example, if 28 slices are needed, this can be accomplished
by acquiring two sets of 14 images, with suspended respirations of ∼20 sec each. The pause
between image sets should be ∼10 sec, to allow the patient to rest between suspended
respirations.
21. Instruct the patient to remain motionless, and to take a deep breath in, blow it out,
deep breath in, blow it out, and hold it. Begin acquisition. Let the patient breathe for
10 sec, ask the patient to hold her breath, and finish running the sequence.
22. Confirm positioning of the torso phased array coil by making certain that all parts of
the liver are included within the sensitive volume of the coil. Reposition and repeat Figure A20.2.3 Transverse T1-weighted spin echo image at the same level as the image in Figure
the sequence, if necessary. A20.2.2.

Table A20.2.4 Imaging Parameters for Sequence 3 (T1-Weighted Spin Echo) Table A20.2.5 Imaging Parameters for Sequence 4 (Single-Shot Fast-Spin Echo)

Patient position Supine Patient position Supine

Scan type Spin echo Scan type Single-shot fast-spin echo
Imaging plane (orientation) Transverse Imaging plane (orientation) Coronal
Variable bandwidth No Variable bandwidth Yes
Pulse sequence database (PSD) se Pulse sequence database (PSD) ssfse
Central slice or volume center Center to pelvis Central slice or volume center Center to xyphoid
Echo time (TE) 8 msec Echo time (TE) 100 msec (effective)
Receiver bandwidth (RBW) ±16 kHz Receiver bandwidth (RBW) ±31 kHz
Repeat time (TR) 500 msec Repeat time (TR) Infinite
Flip angle (FA) Default (90°) Flip angle (FA) Default (90°)
Fields of view (FOVx, FOVy) 200 mm, 200 mm Fields of view (FOVx, FOVy) 360 mm, 360 mm
Resolution (Δx, Δy) 0.78 mm, 1.04 mm Resolution (Δx, Δy) 1.41 mm, 1.88 mm
Number of data points collected (Nx, Ny) 256, 192 Number of data points collected (Nx, Ny) 256, ∼192
Display matrix (Dx, Dy) 256, 256 Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 5 mm Slice thickness (Δz) 5 mm
Number of slices 32 Number of slices 14 for one set (28 in total)a
Slice gap 0.5 mm Slice gap 0
Number of acquisitions (Nacq) 2 Number of acquisitions (Nacq) 0.5 (half Fourier)
Swap read and phase encoding Yes Swap read and phase encoding No
Read direction Anterior–posterior Slice locations Abdomen
Slice location Pelvis Saturation pulses None
Saturation pulses Superior, inferior Scan time 20 sec for one set (40 sec total)a
Fat suppression No Ovarian Cancer aSee annotation under step 20.
Diagnosis and
Scan time 4 min Female Pelvis Staging

A20.2.7 A20.2.8
Current Protocols in Magnetic Resonance Imaging Supplement 6 Supplement 6 Current Protocols in Magnetic Resonance Imaging
 Table A20.2.6 Imaging Parameters for Sequence 5 (Single-Shot Fast-Spin Echo) Table A20.2.7 Imaging Parameters for Sequence 6 (Fast-Spin Echo)

Patient position Supine Patient position Supine

Scan type Single-shot fast-spin echo Scan type Fast-spin echo
Imaging plane (orientation) Transverse Imaging plane (orientation) Transverse
Variable bandwidth Yes Variable bandwidth Yes
Pulse sequence database (PSD) ssfse Pulse sequence database (PSD) frfse-opt
Central slice or volume center Center to liver Central slice or volume center Center to liver
Echo time (TE) 180 msec (effective) Echo time (TE) 80 msec (effective)
Receiver bandwidth (RBW) ±31 kHz Receiver bandwidth (RBW) ±62 kHz
Repeat time (TR) Infinite Echo train length (ETL) 19
Flip angle (FA) Default (90°) Repeat time (TR) 2850 msec
Fields of view (FOVx, FOVy) 320 mm, 240 mm Flip angle (FA) Default (90°)
Resolution (Δx, Δy) 1.25 mm, 2.00 mm Fields of view (FOVx, FOVy) 320 mm, 240 mm
Number of data points collected (Nx, Ny) 256, ∼120 Resolution (Δx, Δy) 1.25 mm, 1.67 mm
Display matrix (Dx, Dy) 256, 256 Number of data points collected (Nx, Ny) 256, ∼144
Slice thickness (Δz) 5 mm Display matrix (Dx, Dy) 256, 256
Number of slices 24 Slice thickness (Δz) 7 mm
Slice gap 0 Number of slices 26
Number of acquisitions (Nacq) 0.5 (half Fourier) Slice gap 0.5 mm
Swap read and phase encoding No Number of acquisitions (Nacq) 0.5 (half Fourier)
Slice location Abdomen Swap read and phase encoding No
Saturation pulses None Slice location Cover the liver
Scan time 24 sec Saturation pulses Superior, inferior
Fat suppression Yes
For examples of abdominal imaging using sequences 4 through 9, please refer to UNIT A15.2, Scan time 46 sec
for liver imaging.
Table A20.2.8 Imaging Parameters for Sequence 7 (Gradient Echo)
Sequence 5: Heavily T2-weighted (TE = 180 msec) transverse single-shot FSE
23. Using the coronal images, position slices to ensure that the entire liver is covered on Patient position Supine
transverse images. Run sequence 5 according to the parameters listed in Table Scan type 2-D double-gradient echo
A20.2.6, during one or two suspended respirations to cover the entire abdomen. Imaging plane (orientation) Transverse
Variable bandwidth Yes
Sequence 6: Moderately T2-weighted FSE with fat suppression Pulse sequence database (PSD) fgre-dual
24. Prescribe the volume to cover the liver. Usually, this will require two separate, Central slice or volume center Center to liver
overlapping breath-holdings. Do not obtain these as interleaved images. “Pause Echo time (TE) 2.3 msec and 4.6 msec
before scan” must be set as “None”. Instruct the patient to remain motionless, and to Receiver bandwidth (RBW) ±62 kHz
take a deep breath in, blow it out, deep breath in, blow it out, and hold it. Run the Repeat time (TR) 200 msec
sequence according to the parameters listed in Table A20.2.7 with 13 slices. Allow Flip angle (FA) 90°
the patient to breathe for 10 sec. Instruct the patient to hold her breath and run the Fields of view (FOVx, FOVy) 320 mm, 240 mm
sequence with the remaining 13 slices. Resolution (Δx, Δy) 1.25 mm, 3.00 mm
Number of data points collected (Nx, Ny) 256, ∼80
25. Evaluate for image quality. This is the first motion-sensitive scan in the abdomen,
Display matrix (Dx, Dy) 256, 256
and will therefore be the first to be degraded if breath holding is not successful.
Slice thickness (Δz) 5 mm
If there is substantial motion artifact, remind the patient about the importance of breath Number of slices 24
holding, and repeat the acquisition. Consider reducing coverage or matrix, or increasing Slice gap 0
image thickness, to reduce acquisition time if necessary. Number of acquisitions (Nacq) 0.5 (half Fourier)
Swap read and phase encoding No
Sequence 7: In-phase and out-of-phase transverse gradient echo Slice location Cover the liver
If possible, these should be obtained as a dual-echo sequence, with TEs of ∼2.3 and 4.6 Saturation pulses Superior, inferior
msec. If this is not available, separating in-phase and opposed-phase acquisitions will be Scan time 24 sec
necessary. Although the TEs mentioned are ideal, software may be configured to yield Ovarian Cancer
slightly different TEs, with acceptable results. Diagnosis and
Female Pelvis Staging

A20.2.9 A20.2.10
Current Protocols in Magnetic Resonance Imaging Supplement 6 Supplement 6 Current Protocols in Magnetic Resonance Imaging
26. Position slices to cover entire liver and adjust levels if necessary. Instruct the patient Table A20.2.9 Imaging Parameters for Sequence 8 (Timing Bolus)
to remain motionless, and to take a deep breath in, blow it out, deep breath in, blow
it out, and hold it. Run sequence 7 according to the parameters listed in Table A20.2.8. Patient position Supine
Scan type Gradient echo
Sequence 8: Timing bolus sequence (if SMARTPREP is not used) Imaging plane (orientation) Sagittal
If SMARTPREP is used, skip steps 27 and 28; a timing bolus sequence is not needed. Variable bandwidth No
Pulse sequence database (PSD) fspgr
27. Inject 2 ml of gadolinium chelate of 2 ml/sec, followed immediately by a 20-ml flush Central slice or volume center Upper abdominal aorta
of saline (Earls et al., 1997). Echo time (TE) Minimum (∼2 msec)
The size of the syringe used should be the same as that used for injecting the bolus of Receiver bandwidth (RBW) ±16 kHz
gadolinium chelate, and the speed of injection should be as identical as possible. Repeat time (TR) 7 msec
Flip angle (FA) 70°
28. Begin image acquisition at the beginning of the flush, according to the parameters Fields of view (FOVx, FOVy) 320 mm, 240 mm
listed in Table A20.2.9. Breath-holding is not necessary. Resolution (Δx, Δy) 1.25 mm, 2.50 mm
The aorta will have low signal intensity on all of these images, until the 2-ml bolus of Number of data points collected (Nx, Ny) 256, 96
gadolinium arrives, at which point there will be a noticeable increase in aortic signal Display matrix (Dx, Dy) 256, 256
intensity. The time after the beginning of the acquisition is annotated in the upper left hand Slice thickness (Δz) 20 mm
corner of the image as the delay time (DT), and can also be determined by noting the image Number of slices 1
number and multiplying by 1 sec. Slice gap N/A
If the acquisition of the sequence begins when the flush begins, when signal intensity peaks Number of acquisitions (Nacq) 1
in the mid aorta, the delay time is the time it takes an injection of the contrast agent to Swap read and phase encoding No
travel from the i.v. tubing port to the abdominal aorta. This should correspond to the delay Slice location Upper abdominal aorta
time between beginning the major gadolinium bolus and beginning the acquisition. For Saturation pulses None
example, if it takes 25 sec for the gadolinium to travel from the i.v. tubing port to the Multi-phase 60
mid-abdominal aorta, breathholding instructions should be given so that the actual pulse Scan time 50 sec
sequence begins 25 sec following the beginning of the gadolinium injection.
29. Set up SMARTPREP region of interest, if SMARTPREP is used. This should cover Table A20.2.10 Imaging Parameters for Sequence 9 (Dynamic 3-D Gradient
the upper abdominal aorta on the coronal scout image obtained from sequence 1, at Echo)
the level of the upper poles of the kidneys. It should be 50 mm long by 20 mm wide.
The trigger should be set for a 20% increase in intensity, with a 7-sec delay to allow Patient position Supine
breathholding instructions. Scan type 3-D gradient echo
Imaging plane (orientation) Transverse
Sequence 9: Dynamic multi-phasic pre/post 3-D short TR gradient echo (liver) Variable bandwidth Yes
30. Prescribe the volume to cover the liver. TR and TE should be as short as possible, and Pulse sequence database (PSD) lufgre3d
flip angle should be 12° to 15°. Instruct the patient to remain motionless, and to take Central slice or volume center Center to liver
a deep breath in, blow it out, deep breath in, blow it out, and hold it. Run the sequence Echo time (TE) Minimum (∼1.5 msec)
according to the parameters listed in Table A20.2.10 but with a certain number of Receiver bandwidth (RBW) ±62 kHz
slices such that the scan time is ∼20 sec. Allow the patient to breathe and move the Repeat time (TR) 4 msec
region of interest to the next slab. Instruct the patient to hold her breath again and Flip angle (FA) 12°to 15°
continue to run the sequence. Repeat this procedure until the entire region of interest Fields of view (FOVx, FOVy) 320 mm, 240 mm
is covered. Resolution (Δx, Δy) 1.25 mm, 3.00 mm
Number of data points collected (Nx, Ny) 256, ∼80
There will be visible contrast in the kidneys and collecting systems if a timing bolus was Display matrix (Dx, Dy) 256, 256
used, but there will be no significant effect on any other tissues. Slice thickness (Δz) 5 mm
31. Explain to the patient that the contrast agent will be injected and that she may feel a Number of slices 192
cool sensation in her arm. Initiate the injection. Do not begin scanning until the scan Slice gap −2.5 mm (overlap)
delay has expired. However, breathing instructions should be initiated when ∼5 sec Number of acquisitions (Nacq) 0.5 (half Fourier)
of delay remain. Swap read and phase encoding No
Slice location Cover the liver
If access to a power injector is not available and are manual injecting, reattach the
ZIP2 Yes
saline-filled syringe and flush with 20 ml of saline. The process of switching syringes must
Saturation pulses None
be completed as quickly as possible and thus, the suggestion of incorporating the use of a
3-way stopcock. Ovarian Cancer Fat suppression Yes, spectral inversion at lipids
Diagnosis and (SPECIAL)
Female Pelvis Staging
Scan time 72 sec
A20.2.11 A20.2.12
Current Protocols in Magnetic Resonance Imaging Supplement 6 Supplement 6 Current Protocols in Magnetic Resonance Imaging
 A

B C
Figure A20.2.4 In a different patient, with history of ovarian cancer, 2-D fat-suppressed short TR
gradient echo images with TR/TE = 19/2 msec, and 30° flip angle, ∼5 min after gadolinium chelate
injection. Arrows indicate solid enhancing perihepatic tissue consistent with peritoneal metastases.

Table A20.2.11 Imaging Parameters for Sequence 10 (Fat-Suppressed Gradient

Echo)

Patient position Supine

Scan type 2-D gradient echo
Imaging plane (orientation) Transverse
Variable bandwidth No
Pulse sequence database (PSD) fspgr
Central slice or volume center Center to liver
Echo time (TE) Minimum (∼2 msec)
Receiver bandwidth (RBW) ±16 kHz
Repeat time (TR) 19 msec Figure A20.2.5 3-D fat-suppressed short TR gradient echo images before (A), immediately after (B), and 45 sec after (C)
Flip angle (FA) 30° gadolinium chelate injection. Through Fourier interpolation, images are reconstructed every 2.5 mm. Large arrow in A
indicates the left adnexal cyst, small arrows indicate dependent increased signal intensity. Arrows in C indicate uniform
Fields of view (FOVx, FOVy) 320 mm, 240 mm
thickness of the enhancing wall of the cyst, without solid or nodular enhancement to suggest tumor. Findings are most
Resolution (Δx, Δy) 1.25 mm, 2.0 mm compatible with a benign hemorrhagic cyst.
Number of data points collected (Nx, Ny) 256, 120
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 5 mm 32. Obtain the first set of post-contrast images as soon as possible according to Table
Number of slices ∼30 A20.2.10. Repeat step 30. Allow the patient ∼5 sec rest between the first and second
Slice gap 0 post-contrast acquisitions. Repeat step 30 again and obtain a second set of post-con-
Number of acquisitions (Nacq) 1 trast acquisitions. It is best to observe the patient’s respirations on the screen to verify
Swap read and phase encoding No that she is suspending respirations successfully when directed.
Slice location Entire liver
Fat suppression Yes Sequence 10: 2-D transverse gradient echo with fat suppression (Fig. A20.2.4)
Scan time ∼60 sec 33. Instruct the patient to remain motionless, and to take a deep breath in, blow it out,
deep breath in, blow it out, and hold it. Run sequence 10 according to the parameters
listed in Table A20.2.11 but with a certain number of slices such that the scan time
is ∼20 sec. Allow the patient to breathe and move the region of interest to the next
Ovarian Cancer slab. Instruct the patient to hold her breath again and continue to run the sequence.
Diagnosis and
Female Pelvis Staging Repeat this procedure until the entire region of interest is covered.
A20.2.13 A20.2.14
Current Protocols in Magnetic Resonance Imaging Supplement 7 Supplement 7 Current Protocols in Magnetic Resonance Imaging
 Table A20.2.12 Imaging Parameters for Sequence 11 (Dynamic 3-D Gradient
Echo)

Patient position Supine

Scan type 3-D gradient echo
Imaging plane (orientation) Transverse
Variable bandwidth Yes
Pulse sequence database (PSD) lufgre3d
Central slice or volume center Center to pelvis
Echo time (TE) Minimum (∼1.5 msec)
Receiver bandwidth (RBW) ±62 kHz
Repeat time (TR) 4 msec
Flip angle (FA) 12°to 15°
Fields of view (FOVx, FOVy) 240 mm, 240 mm
Resolution (Δx, Δy) 0.94 mm, 2.82 mm
Number of data points collected (Nx, Ny) 256, ∼85
Display matrix (Dx, Dy) 256, 256
Slice thickness (Δz) 5 mm
Number of slices 256
Slice gap −2.5 mm
Number of acquisitions (Nacq) 0.5 (half Fourier)
Swap read and phase encoding No
Slice location Pelvis
ZIP 2 Yes
Saturation pulses None Figure A20.2.6 In the same patient as in Figures A20.2.1 to A20.2.4, 2-D sagittal fat suppressed
Fat suppression Yes, spectral inversion at lipids short TR gradient echo images with TR and TE = 19/2 msec, and 30° flip angle, about five minutes
(SPECIAL) after gadolinium chelate injection. The cyst wall and surrounding tissues enhance, but there is no
Scan time 96 sec internal enhancement.

Table A20.2.13 Imaging Parameters for Sequence 12 (Fat-Suppressed Gradient

TR should be ∼20 msec. These images show delayed extracellular space contrast, and all Echo)
vessels are white because of the combined effects of gadolinium and time-of-flight. Motion
artifact is minimal because the short TR allows completion of each image in <2 sec. These Patient position Supine
should be obtained as several stacks of overlapping slices, to ensure that the fat suppression Scan type 2-D gradient echo
is adjusted properly for the volume of interest. The upper- and mid-abdomen should be
Imaging plane (orientation) Transverse or sagittal
imaged without significant gaps between image stacks.
Variable bandwidth No
Sequence 11: Dynamic 3-D short TR gradient echo (pelvis) Pulse sequence database (PSD) fspgr
If an abdominal examination is not included, such as if the pelvic images are not Central slice or volume center Center to pelvis
suggestive of ovarian cancer, the dynamic multi-phasic examination should be modified Echo time (TE) Minimal (∼2 msec)
for imaging of the pelvis (Fig. A20.2.5). Receiver bandwidth (RBW) ±16 kHz
Repeat time (TR) 19 msec
34. Pull the patient out of the magnet. Reposition torso phased array coil for pelvis (or Flip angle (FA) 30°
switch to pelvic portion of abdominal/pelvic coil). Center the patient and advance Fields of view (FOVx, FOVy) 240 mm, 240 mm
the patient into the magnet. Resolution (Δx, Δy) 0.94 mm, 1.88 mm
Number of data points collected (Nx, Ny) 256, 128
35. Repeat step 30 (with the coverage of pelvis) but run sequence 11 according to the Display matrix (Dx, Dy) 256, 256
parameters in Table A20.2.12. This involves the use of smaller field of view. If desired, Slice thickness (Δz) 5 mm
slice thickness can be reduced. A timing bolus is generally not used for pelvic Number of slices ∼30
imaging. Obtain three sets of post-contrast images. The procedures are similar to Slice gap 0
those in steps 31 and 32. Number of acquisitions (Nacq) 1
Although dynamic scanning is less important for imaging the adnexa than for imaging the Swap read and phase encoding No
abdomen, vessels and some tissues may be defined better. Next, sequence 12 should be Slice location Pelvis
obtained. Ovarian Cancer
Diagnosis and Fat suppression Yes
Female Pelvis Staging Scan time ∼60 sec
A20.2.15 A20.2.16
Current Protocols in Magnetic Resonance Imaging Supplement 6 Supplement 6 Current Protocols in Magnetic Resonance Imaging
Sequence 12: 2-D transverse or sagittal gradient echo with fat suppression (Fig. nosis and staging of ovarian cancer: Comparative Yamashita, Y., Torashima, M., Hatanaka, Y., Harada,
values of doppler and conventional US, CT, and M., Higashida, Y., Takahashi, M., Mizutani, H.,
A20.2.6) MR imaging correlated with surgery and his- Tashiro, H., Iwamasa, J., Miyazaki, K., et al.
36. Instruct the patient to remain motionless, and to take a deep breath in, blow it out, topathologic analysis—report of the radiology 1995. Adnexal masses: Accuracy of charac-
deep breath in, blow it out, and hold it. Cover the pelvis, and run sequence 12 diagnostic oncology group. Radiology 212:19- terization with transvaginal US and precontrast
according to the parameters in Table A20.2.13 but with a certain number of slices 27. and postcontrast MR imaging. Radiology
A comprehensive multimodality multi-institutional 194:557-565.
such that the scan time is ∼20 sec. Allow the patient to breathe and move the region
study Comparison of MRI to ultrasound
of interest to the next slab. Instruct the patient to hold her breath again and continue
to run the sequence. Repeat this procedure until the entire region of interest is covered. Outwater, E.K., Huang, A.B., Dunton, C.J., Taler-
man, A., and Capuzzi, D.M. 1997. Papillary pro- Internet Resources
jections in ovarian neoplasms: Appearance on http://www.mri.tju.edu/
COMMENTARY MRI. J. Magn. Reson. Imaging 7:689-695. A non-commercial site that lists all body MRI pro-
MRI-histogic correlation of ovarian cancer tocols, continually updated, used by updated GE
Background Information method such as SMARTPREP. If abdominal Signa scanners by the Thomas Jefferson University
A comprehensive protocol for imaging the images are not acquired, the timing of gadolin- Semelka, R.C., Lawrence, P.H., Shoenut, J.P., Hey- Department of Radiology. Additionally, there are
abdomen and pelvis in women with ovarian ium injection is not critical; the delayed post- wook, M., Kroeker, M.A., and Lotocki, R. 1993. descriptions and explanations of the various pulse
Primary ovarian cancer: Prospective comparison sequences, tips for problems solving, examples of
cancer that includes in-phase and out-of-phase gadolinium images are most important. clinical applications.
of contrast-enhanced CT and pre- and postcon-
T1-weighted images, moderately and heavily Another problem occurs in patients with trast, fat-suppressed MR imaging, with his-
T2-weighted, and early and late post-gadolin- implanted metal, such as from surgical clips. tologic correlation. J. MRI 3:99-106.
ium images is described in this unit. Each ab- Artifacts from metal on gradient echo images Comparison between MRI and CT Contributed by Donald G. Mitchell, Peter
dominal sequence is acquired in one or two can be minimized by using the shortest possible Natale, and George Holland
breath-holds, while un-enhanced T1-weighted TE, avoiding fat suppression, and increasing the Shellock and Kanal, 1996. See above.
Thomas Jefferson University
and T2-weighted images are non-breath-hold number of acquisitions (Nacq) or number of Discussion of safety issues Philadelphia, Pennsylvania
and are used in the pelvis. It is possible that excitations (NEX) from 0.5 to 1.0. TE on the
comparable or better image quality can be ob- 3-D-short TR gradient echo sequence is mini-
tained using an entirely breath-hold protocol, mized by increasing the sampling bandwidth
but the authors have retained the longer se- from ±31 to ±62 KHz. The acquisition time will
quences in the pelvis to ensure high spatial be slightly longer, so adjustments in number of
resolution and SNR. If the initial pelvic images image slices and/or acquisition matrix may be
and/or patient history are not strongly sugges- necessary.
tive of ovarian cancer, staging is unnecessary,
and the abdominal portion of the examination Anticipated Results
may be omitted, as described above. Most reports comparing MRI and computed
The most important pulse sequence for di- tomography (CT) indicate that MRI is superior
agnosing and staging ovarian carcinoma is the for depicting normal and abnormal gyne-
delayed fat-suppressed post-gadolinium se- cologic anatomy, and for distinguishing be-
quence. In cooperative patients, thinner over- tween benign and malignant adnexal and he-
lapping slices with higher SNR can be obtained patic masses. Importantly, proper use of gad-
by using the fast 3-D short TR gradient echo olinium and fat suppression leads to a higher
sequence (efgre3d or lufgre3d in the GE ma- sensitivity of MRI for detecting small perito-
chine). However, these images are more mo- neal implants from ovarian carcinoma.
tion-sensitive, and do not depict blood vessels
as well, compared to the single-slice 2-D se- Literature Cited
quences described here. Earls, J.P., Rofsky, N.M., Lee, V.S., DeCorato, D.R.,
Krinsky, G.A., and Weinreb, J.C. 1997. Hepatic
arterial-phase dynamic gadolinium-enhanced
Critical Parameters and MR imaging: Optimization with a test examina-
Troubleshooting tion and a power injector. Radiology 202:268-
The two major potential problems with this 273.
protocol are that, for the abdominal portion, Shellock, F.G. and Kanal, E. 1996. Magnetic Reso-
patients may not be able to suspend respiration nance Bioeffects, Safety, and Patient Manage-
or timing of gadolinium administration may be ment. Lippincott, Philadelphia.
incorrect. The first problem can be addressed
by substituting short TR (≤10 msec) single-slice Key References
2-D gradient echo T1-weighted sequences with Kurtz, A.B., Tsimikas, J.V., Tempany, C.M., Ham-
per, U.M., Arger, P.H., Bree, R.L., Wechsler, R.J.,
a flip angle of ∼30°. The second problem is
Francis, I.R., Kuhlman, J.E., Siegelman, E.S.,
prevented in most cases by using a test dose of Mitchell, D.G., Silverman, S.G., Brown, D.L.,
gadolinium to determine timing (Earls et al., Sheth, S., Coleman, B.G., Ellis, J.H., Kurman, Ovarian Cancer
R.J., Caudry, D.J., and McNeil, B.J. 1999. Diag- Diagnosis and
1997), or by using an automated detection Female Pelvis Staging

A20.2.17 A20.2.18
Current Protocols in Magnetic Resonance Imaging Supplement 12 Supplement 12 Current Protocols in Magnetic Resonance Imaging
Pelvic Floor Relaxation UNIT A20.3 Set up patient and equipment
1. Interview the patient to assess for contraindications such as cardiac pacemakers,
implanted mechanical devices, and/or ferromagnetic materials. Also, determine the
Pelvic floor relaxation–abnormal descent of the bladder (cystocele), uterus and/or vagina
use of appropriate monitoring equipment.
(uterine or vaginal vault prolapse), small bowel (enterocele), or rectum (rectocele), is a
significant women’s health issue affecting primarily parous women >50 years of age. Up A screening form (see APPENDIX 1) is signed by each patient or legal guardian prior to
to 50% of such women have some degree of genital prolapse. Symptoms are present in bringing the patient into the exam area.
10% to 20% of this group and most commonly include pelvic pressure, protrusion of The presence of ferromagnetic materials may be a health hazard to the patient while in the
tissue through the pelvic floor and urinary incontinence. In addition to age, risk factors magnetic field and/or adversely affect image quality. If in doubt as to the exact composition
include multiparity, menopause, and obesity. The purpose of this work is to describe a of the items, it is best to exclude patients with any metal implants; see Shellock (2001) for
technique for production and review of MR images that can be used to quickly and discussion of what implants may be safely scanned using magnetic resonance.
accurately identify significant pelvic floor defects as an aid to treatment planning (Gold Patients may be accompanied into the magnet room by a friend or family member, who can
et al., 1999; Fielding et al., 2000). sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
IMAGING THE PELVIC FLOOR BASIC 2. Instruct the patient to change into a gown and remove all personal effects, such as
PROTOCOL
In order to achieve an adequate signal-to-noise ratio using the ultra-fast pulse se- jewelry, hearing aids, glasses, etc., prior to entering the MRI scanning room. Have
quences required, MR imaging of the pelvic floor is best done on a high-field strength the patient wash off any mascara and other makeup to avoid local tissue heating and
magnet (1.0 T or 1.5 T). The sequences described below are based on the author’s image artifacts. Have the patient void to completion. If the procedure is a research
experience with a Siemens 1.5 T vision scanner, but have been applied with success protocol, have the patient sign any necessary consent forms.
to machines made by other manufacturers. The following 3 sequences comprise the
3. Explain the procedure to the patient and record relevant clinical history. Ensure that
routine protocol for evaluation of pelvic floor descent. When urinary incontinence or
the patient understands what is expected and ask if she has any questions; answer
uterine prolapse is the primary clinical presentation, no rectal contrast material needs
them appropriately.
be administered. In the case of global pelvic floor relaxation or suspected rectal
prolapse, 60 ml of ultrasound gel should be administered rectally prior to imaging 4. Escort the patient to the MR examination room and review the following items with
using a catheter-tip syringe. The entire Basic Protocol can be completed in 15 min. the patient:
There is no need to sedate the patient.
a. Provide earplugs or headphones to the patient to minimize the loud knocking noise
Table A20.3.1 lists the hardware necessary to perform the procedure, along with appro- that will be produced by the gradient coils but ensure that she will still be able to
priate parameters. The available gradient strength will depend on the scanner, and the hear you.
echo times given in other tables below may be varied accordingly (the smaller the gradient b. Provide the patient with a safety squeeze-bulb and demonstrate how it works;
strength, the longer the echo time for a particular scan). explain to the patient when to use the squeeze-bulb (i.e., if assistance is needed
during the exam).
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular c. Explain to the patient that you will be talking to her between imaging sequences
patient, such as crash carts or oxygen. (when the loud knocking noise stops). Additionally, review breath-holding instruc-
tions with the patient.
d. Explain to the patient that it is imperative that she remain motionless during the
loud knocking noise to ensure good results; also explain that she should not
Table A20.3.1 Equipment Parameters for Pelvic Floor Relaxation
reposition between imaging sequences.
Coil type Multi-coil array such as torso or pelvis e. Nevertheless, the patient may call out at any time if he or she feels it necessary.
phased array coil f. Inform the patient with an approximate length of time for the entire examination.
Field strength 1.0 or 1.5 T (preferred)
Gradient strength 25 mT/m (or whatever system permits) 5. Install the bottom portion of the pelvis phased array or torso phased array coil on the
Cardiac gating No table. Instruct the patient to lie down on the table with head toward the scanner, and
Peripheral gating No place a pillow under the patient’s knees to reduce discomfort. Secure the top portion
Respiratory gating No of the phased array coil to eliminate it from moving side-to-side during breath-hold-
Respirator If required by patient ing imaging sequences.
Oxygen If required by patient
Usually, straps are provided by the manufacturer that are directly attached to the coil.
Motion cushions Useful
Use of contrast agents Only for evaluation of rectal prolapse (see 6. Using the laser light, center the patient’s pelvis in the coil. Include the entire pelvis
text) and uppermost portion of the thighs. Do not sacrifice coil coverage at the level of the
upper thighs in order to cover the uppermost portion of the pelvis.
Pelvic Floor
Female Pelvis Relaxation
Contributed by Julia R. Fielding A20.3.1 A20.3.2
Current Protocols in Magnetic Resonance Imaging (2003) A20.3.1-A20.3.7
Copyright © 2003 by John Wiley & Sons, Inc. Supplement 8 Supplement 8 Current Protocols in Magnetic Resonance Imaging
 Table A20.3.2 Primary Clinical Imaging Parameters for Sequence 1 (Pilot Scan) Table A20.3.3 Primary Clinical Imaging Parameters for Sequence 2 (SSFSE or
HASTE)
Patient position Supine
Scan type 3-D gradient echo Patient position Supine
Imaging plane (orientation) Sagittal, coronal, and transverse Scan type Single-shot fast-spin echo
Central slice or volume center Laser light centered on mid pelvis Imaging plane (orientation) Sagittal
Echo time (TE) As short as possible (e.g., 5 msec) Central slice or volume center Centered on bladder base
Repeat time (TR) As short as possible (e.g., 15 msec) Echo time (TE) 90 msec
Flip angle (FA) 10° Repeat time (TR) 4.4 msec (the true TR is infinite, 4.4
Fields of view (FOVx, FOVy) 350 mm, 350 mm msec represents the echo spacing)
Resolution (Δx, Δy) 1.37 mm, 2.73 mm Flip angle (FA) 180°a
Number of data points collected (Nx, Ny) 256, 128 Fields of view (FOVx, FOVy) 350 mm, 350 mm
Slice thickness (Δz) 10 mm Resolution (Δx, Δy) 1.37 mm, 2.06 mm
Number of slices 3 per orientation Number of data points collected (Nx, Ny) 256, 170
Slice gap Not applicable Display matrix (Dx, Dy) 512, 512
Number of acquisitions (Nacq) 1 Slice thickness (Δz) 10 mm
Swap read and phase encoding Standard default Number of slices 1
Scan time ∼15 sec Slice gap Not applicable
Number of acquisitions (Nacq) 0.5 (half Fourier)
Read direction Anterior–posterior
7. Advance the patient table to isocenter. Slice locations Midline pelvis
Saturation pulses No
Sequence 1: Rapid three-plane positioning pilot Scan time 1 sec (the total scan time is 2 sec, 1 at
8. Instruct the patient to take in a deep breath and exhale, take in another deep breath, rest and 1 at strain, see text)
and then hold it. Initiate the scan as in step 9. aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this sequence

is 90°.
9. Run the system’s pilot (or scout) scan to ensure correct location of the pelvis in three
directions using the imaging sequence given in Table A20.3.2. Table A20.3.4 Primary Clinical Imaging Parameters for Sequence 3 (FSE or TSE)
Please refer to manufacturer’s specifications for standard default setting and swap read
and phase encoding directions as necessary. Patient position Supine
Scan type FSE
Sequence 2: Sagittal T2-weighted midline pelvis images at rest and at strain Imagine plane (orientation) Transverse
In general, a woman can only hold maximal downward pelvic strain for 7 sec. For this Central slice or volume center Centered on midpoint of
reason, ultra-fast pulse sequences must be employed. Common acronyms are single-shot pubococcygeal line
fast-spin echo (SS-FSE; GE) and half acquisition single-shot turbo-spin echo (HASTE; Echo time (TE) 132 msec
Siemens). These pulse sequences perform phase encoding of all data simultaneously and Echo train length (ETL) 8
Repeat time (TR) 5000 msec
sample only half of k-space.
Flip angle (FA) 180°a
10. Use the transverse and sagittal images from sequence 1 to choose a midline image Fields of view (FOVx, FOVy) 240 mm, 180 mm
that shows the symphysis, bladder neck, cervix, and coccyx. Resolution (Δx, Δy) 0.67 mm, 0.47 mm
Number of data points collected (Nx, Ny) 360, 384
In order to determine which of the three main compartments of the pelvic floor descend
Display matrix (Dx, Dy) 512, 512
abnormally, it is imperative that all be encompassed on a single slice.
Slice thickness (Δz) 3 mm
11. Instruct the patient to take in a deep breath and exhale, take in another deep breath, Number of slices 19
and then hold it. Slab thickness 57 mm
Slice gap 0 mm
12. Run sequence 2 according to Table A20.3.3 while the patient’s pelvic floor is at rest.
Number of acquisitions (Nacq) 2
13. Instruct the patient to take in a deep breath and exhale, take in another deep breath, Read direction Anterior–posterior
and hold it. Slice locations Pelvic floor
Saturation pulses None
14. Coach the patient verbally to strain her pelvis maximally downward and re-run Fat supression No
sequence 2 according to Table A20.3.3. Slice series Interleaved
If an adequate strain is not obtained, repeat steps 13 and 14 with further coaching. Scan time ∼7 min
Pelvic Floor aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this sequence
Female Pelvis Relaxation
is 90°.
A20.3.3 A20.3.4
Current Protocols in Magnetic Resonance Imaging Supplement 8 Supplement 8 Current Protocols in Magnetic Resonance Imaging
Sequence 3: Transverse fast spin echo (FSE or TSE, turbo spin echo) images of the
pelvic floor
15. After completion of sequence 2, review the resting image and identify the pelvic floor. A B
The pelvic floor usually extends from the last joint of the coccyx to the inferiormost aspect
of the symphysis.
16. Use console calipers, or visually estimate a line drawn between the aforementioned
two points (the pubococcygeal line).
L L
PC PC
17. Instruct the patient that she may breathe normally for the next series. M line
M line
18. Use sequence 3 to identify signal changes and tears within the supporting muscles H line
H line
of the pelvic floor. Run the sequence according to Table A20.3.4.

COMMENTARY
Background Information
For the majority of patients with mild to
moderate stress incontinence and pelvic floor
relaxation, the combination of physical exami-
nation findings and urodynamic pressure read-
ings is diagnostic and no further imaging is
required. For those patients with severe urinary
or fecal incontinence thought to be multifacto-
rial, multiple compartment involvement, or
have failed prior surgery, imaging can be ex-
tremely valuable.
There are several techniques that can be used
to evaluate the pelvic organs. These include
voiding cystourethrography (with or without
concurrent video urodynamic tracings), ultra-
sound of the bladder neck and anal sphincter,
colpocystodefecography, and magnetic reso-
nance imaging.
Voiding cystourethrography
Voiding cystourethrography (VCUG) is
often performed during an incontinence work-
up to exclude anatomic abnormalities such as
duplicated collecting systems and ureters, blad-
der and urethral diverticula, and vesicoureteral
reflux. Even when these findings are not pre-
sent, examination of the bladder can provide
useful information. Trabeculation is a sign of
urge incontinence, or detrusor overactivity
(Fielding, 2002). The hallmark of this disease
is the sudden onset and imminent need to void
due to a bladder contraction. This is the most
common type of incontinence in the elderly,
affecting up to 30% of those living at home and
50% in long-term care institutions. Identifica-
tion of an unsuspected neurogenic bladder is
another significant finding. These patients
often have a history of spinal injury. The blad-
der remains contracted and severely trabecu-
lated during the entirety of the exam and may
have a markedly tapered dome. While VCUG
is rarely required to diagnose stress inconti-
nence, it can be useful in severe cases when the
bladder extends so far inferiorly that bladder
neck kinks, masking symptoms.
Ultrasound
Ultrasound of the bladder neck or perineum is
an alternative way to diagnose stress inconti-
nence. Again, it is usually only of clinical use for
assessment of bladder neck mobility. Trans rectal
ultrasound (TRUS) has been shown to be reliable
in the identification of rectal sphincter tears and
atrophy (Gold et al., 1999). It is often used prior
to surgery to gauge the size and depth of tear and
identify adjacent hypoechoic scar tissue. A small
probe providing circumferential images is placed
in the rectal canal. Images are obtained at 5-mm
intervals along the length of the sphincter, ~2.5
cm. The normal internal sphincter is a circumfer-
ential black band of uniform thickness. The ex-
ternal sphincter is harder to see because it is
echogenic. Tears are defined as areas of disconti-
nuity. Large tears respond poorly even to aggres-
sive surgical intervention.
Defecography
Colpocystodefecography is a method of ob-
serving all three compartments of the pelvic
floor during rest, upward contraction, and Val-
salva. Because it is done in the sitting position,
it most closely mimics the physiologic state
(Kelvin, 1998). In most institutions, it is used
primarily to identify posterior compartment
abnormalities, therefore the bladder is not
opacified. One hour prior to the study, the
patient is given a barium meal to coat the small
bowel loops. Barium paste is placed into the
rectum, usually with the aid of a caulking gun,
and the patient places a tampon soaked in con-
trast material in the vagina. In the upright po-
sition, multiple spot fluoroscopic images are
Figure A20.3.1 Sagittal MR images of a 55-year-old woman with stress incontinence and incomplete
bladder emptying. Sagittal T2-weighted image at rest (A) shows pelvic organs and floor to have normal
appearance. On strain (B), there is development of a cystocele and rotation of the urethra into the
horizontal plane (arrow) indicating damage to the internal sphincter. Reproduced from Fielding (2002)
with permission from Radiographics.
obtained during pelvic floor maneuvers, con-
cluding with defecation. With a pelvic floor
contraction, a sharp anorectal angle is an indi-
cator of good muscular support, however, spe-
cific measurements have not been found useful.
Retained barium within a portion of the anterior
rectum that bulges >2 cm into the rectovaginal
space is defined as an anterior rectocele. Pa-
tients will often be able to manually empty the
rectocele. Descent of small bowel loops into the
rectovaginal space is defined as an enterocele
and indicates a tear in the rectovaginal fascia.
Many gastroenterologists believe that entero-
cele gives a patient the feeling of incomplete
defecation despite the presence of an empty
rectum leading to persistent and ineffectual
straining. Intusussception usually involves
only the rectum or rectosigmoid and resolves
with cessation of Valsalva.
Magnetic resonance imaging
During the past 10 years, magnetic reso-
nance imaging (MRI) has emerged as a com-
petitor to other imaging modalities for evalu-
ation of the female pelvic floor. The main ad-
vantages of MRI are the ability to evaluate the
three compartments of the pelvic floor simul-
taneously during rest and strain and the direct
visualization of supporting structures (Fielding
et al., 1998; Unterweger et al., 2001). Disad-
vantages include the requirement that the exam
be performed in the supine or left lateral decu-
bitus position, although one group working
with an open configuration magnet reported no
Pelvic Floor significant difference between upright and su-
Female Pelvis Relaxation pine findings (Bertschinger et al., 2002).
Critical Parameters and
Troubleshooting
The two most common reasons for poor
quality MR images of the pelvic floor are (1) a
patient’s inability to perform adequate pelvic
strain, and (2) incorrect patient positioning for
obtaining the critical midline sagittal images.
Before beginning the exam and after the
patient is situated in the magnet, it is helpful to
practice pelvic strain. Instruct the patient to take
a slow deep breath in and then bear down as if
she is having a bowel movement. Ask her to
keep her buttocks and thighs on the table while
she strains. During the exam, if the operator
suspects that an adequate strain is not obtained,
simply repeat the sagittal scan following further
coaching.
Reinforce with the technologists the neces-
sity of absolute midline positioning. If the scout
does not demonstrate all of the midline pelvic
structures, the patient must be repositioned so
that she rests absolutely flat on the table. If the
scout is difficult to read, the author often finds
it helpful to expand the first sagittal HASTE
series (sequence 2) to 3 or 4 slices in order to
find the best midline image.
Anticipated Results
On sagittal images, the pubococcygeal
(PCL) line should be drawn between the last
joint of the coccyx and the inferiormost aspect
of the symphysis. Urologists and gynecologists
use this line as an indicator of the pelvic floor.
In an early work, Yang et al. (1991) used gra-
dient echo images to define maximal normal
descent of the bladder base (1.0 cm below),

A20.3.5 A20.3.6
Current Protocols in Magnetic Resonance Imaging Supplement 8 Supplement 8 Current Protocols in Magnetic Resonance Imaging
vagina (1.0 cm above), and rectum (2.5 cm
below) with respect to the pubococcygeal line.
In practical terms, descent of the bladder or
vagina >1 cm below the PCL indicates some
degree of laxity while descent >2 cm in asymp-
tomatic patient often requires surgical therapy
(Fig. A20.3.1). Rectal abnormalities such as
anterior rectocele and intussusception as well
as enterocele are identified in the same fashion
as with defecography. There are other impor-
tant findings on sagittal images. The levator
plate should remain parallel to the pubococcy-
geal line at all times. Caudal angulation of the
levator plate >10° indicates loss of pelvic floor
support (Ozasa et al., 1992; Hoyte et al., 2001).
Measurements of the H and M lines are
useful ways to quantify loss of pelvic floor
support (Comiter et al., 1999). The H line is
drawn from the inferior aspect of the symphysis
pubis to the posterior wall of the rectum and
measures the anterior-posterior dimension of
the pelvic hiatus. The M line is drawn as a
perpendicular from the PCL to the posterior-
most aspect of the H line. It measures the height
of the hiatus. In healthy women, the H line
should not exceed 5 cm and the M line should
not exceed 2 cm. Values greater than these
indicate loss of pelvic floor support.
Transverse images should be reviewed for
muscle integrity and signal intensity and for the
vaginal shape and location. The puborectalis
should extend from the parasymphysial inser-
tion posterior to the rectum. It should be of
similar width along its entire course without
evidence of gaps or fraying. The width of the
levator hiatus at the level of the symphysis
rarely exceeds 4.5 cm in healthy volunteers,
however, there is some overlap with incontinent
patients. The vagina should normally be of H
or butterfly shape and centered in the pelvis
(Klutke et al., 1990).
Literature Cited
Bertschinger, K.M., Hetzer, F.H., Roos, J.E.,
Treiber, K., Marincek, B., and Hilfiker, P.R.
2002. Dynamic MR imaging of the pelvic floor
performed with patient sitting in an open-magnet
unit versus with patient supine in a closed-mag-
net unit. Radiology 223:501-508.
Comiter, C.V., Vasavada, S.P., Barbaric, Z.L.,
Gousse, A.E., and Raz, S. 1999. Grading pelvic
prolapse and pelvic floor relaxation using dy-
namic magnetic resonance imaging. Urology
3:454-457.
Current Protocols in Magnetic Resonance Imaging
Fielding, J.R. 2002. Practical MR imaging of female
pelvic floor weakness. Radiographics 22:295-
304.
Fielding, J.R., Griffiths, D.J., Versi, E., Mulkern,
R.V., Lee, M.L., and Jolesz, F.A. 1998. MR
imaging of pelvic floor continence mechanisms
in the supine and sitting positions. A.J.R.
171:1607-1610.
Fielding, J.R., Lee, J.H., Dubeau, C.E., Zou, K.H.,
and Resnick, N.M. 2000. Voiding cystoure-
thrography findings in elderly women with urge
incontinence. J. Urol. 163:1216-1218.
Gold, D.G., Halligan, S., Kmiot, W., and Bartram,
C.I. 1999. Anal endosonography: Inter- and in-
tra-observer agreement. Br. J. Surg. 86:371-375.
Hoyte, L., Schierlitz, L., Zou, K., Flesh, G., and
Fielding, J.R. 2001. Two- and 3-dimensional
MRI comparison of levator ani structure, vol-
ume, and integrity in women with stress inconti-
nence and prolapse. Am. J. Obstet. Gynecol.
185:11-19.
Kelvin, F.M., Maglinte, D.D., Hornback, J.A., and
Benson, J.T. 1998. Pelvic prolapse: Assessment
with evacuation proctography (defecography).
Radiology 184:547-551.
Klutke, C., Golomb, J., Barbaric, Z., and Raz, S.
1990. The anatomy of stress incontinence: Mag-
netic resonance imaging of the female bladder
neck and urethra. J. Urol. 143:563-566.
Ozasa, H., Mori, T., and Togashi, K. 1992. Study of
uterine prolapse by magnetic resonance imag-
ing: Topographical changes involving the levator
ani muscle and the vagina. Gynecol. Obstet. In-
vest. 24:43-48.
Shellock, F.G. 2001. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Unterweger, M., Marincek, B., Gottstein-Aalame,
N., Debatin, J.F., Seifert, B., Ochsenbein-Imhof,
N., Perucchini, D., Kubik-Huch, R.A. 2001. Ul-
trafast MR imaging of the pelvic floor.
A.J.R.176:959-963.
Yang, A., Mostwin, J.L., Rosenshein, N.B., and
Zerhouni, E.A. 1991. Pelvic floor descent in
women: Dynamic evaluation with fast MR im-
aging and cinematic display. Radiology 179:25-
33.
Internet Resource
http://www.mrisafety.com
This is the definitive site for MR safety including
medical devices. It is free to all users following
registration.
Contributed by Julia R. Fielding
University of North Carolina at Chapel Hill
Chapel Hill, North Carolina
Female Pelvis
A20.3.7
Supplement 8
CHAPTER A21
Breast
INTRODUCTION
ue to the high prevalence, the high mortality, and the operative therapy which is
D currently required, breast cancer represents a great problem from the medical, psy-
chological, and health policy point of view. In the past 20 years, breast cancer has been the
most frequently occurring cancer of women in the western hemisphere. It is the most fre-
quent cause of death of women below 50 years in age and, as a whole, the most frequent
cause of cancer mortality of women. Despite the constant improvement of diagnostic
procedures, the introduction of novel mammography techniques and ultrasonic methods,
in principle, nothing has changed. This dilemma is to be regretted, as the relatively slow
growth rates of breast cancer, with tumor doubling times of ∼100 to 300 days, give suf-
ficient time to detect the tumor in a curable stage, i.e., before the formation of external
metastases. It is known that a tumor <1 cm in size has an excellent prognosis irrespective
of its internal histological structure. It is estimated that a breast cancer needs a period of
∼7 to 10 years for reaching a size of ∼1 cm on an average.
By the introduction of magnetic resonance mammography (MRM), diagnosis of breast
cancer has improved significantly. In the early years, there was no agreement as to the
value of this method, because of the multitude of different measurement possibilities.
The considerably worse spatial resolution compared to X-ray mammography, the inabil-
ity to detect microcalcifications, the need to use contrast agents, and the high costs of the
method were all considered major disadvantages, causing many experienced mammog-
raphy experts to regard this method with great skepticism; however, after a clinical test
phase of ∼17 years, several studies have shown that MRM yields the highest sensitivity
in the diagnosis of small breast cancers and that the multifocality/multicentricity of breast
cancers can be recognized adequately only with this method.
The utility of contrast media–enhanced MRM can be explained through the generation
of early tumor angiogenesis, which probably represents a reliable sign of a breast tumor
≥3 mm in size. To grow uncontrollably, such a tumor needs an increased blood supply
for nourishment and removal of metabolic waste materials. Following the injection of
a contrast medium, breast cancer exhibits a rapid and often typical quantitative and
morphological increase in signal intensity with an early plateau formation or washout
effect, whereas the most benign tumors and normal parenchyma usually show a slower
and more steady increase in signal intensity.
Meanwhile, there appears to be consensus that for an exact diagnosis of breast cancer,
the selection of high spatial and temporal resolution in a so-called dynamic technique
is of utmost importance. In these dynamic sequences, i.e., the repetitive imaging and
measurement of the same slices before and in short time intervals after the injection of
contrast medium, the best differentiation between benign and malignant lesions occurs
within the first 2 min. Therefore, it is imperative to maintain optimal study conditions.
The evaluation of dynamic MRM is simple in some cases but exceptionally delicate in
others, and demands a high level of experience on the part of the examiner.
Breast
Contributed by Werner A. Kaiser A21.0.1
Current Protocols in Magnetic Resonance Imaging (2005) A21.0.1-A21.0.2
C 2005 by John Wiley & Sons, Inc.
Copyright  Supplement 13
 False positive results can occur due to biological or technical reasons. Biologically false
positives can be achieved in the examination of myxoid, fibroadenoma, proliferative
displeasures, and acute mastitis; however, most of these cases do not show a plateau
phenomenon or wash-out effect, which is the decrease of signal intensity after the initial
striking increase observed within the first 2 min. This decrease can be explained by arte-
riovenous shunts within the tumorangiogenetic network, which induce a sudden washout
of contrast medium and therefore a drop in signal intensity. Only a few benign cases in the
literature show these so-called wash-out effects. Furthermore, the type of enhancement
(centrifugally, centripetally), the presence membranes or septations within in the lesion,
as well as the signal-intensity in T2 -weighted scans can be used for the differentiation
between malignant and benign enhancing lesions.
Besides false-positive cases due to pathophysiological reasons, there exists an enor-
mous variety of other reasons for different types of enhancement. These include the
wrong dosage of contrast medium, the different types of physiological enhancement, and
hormone effects. In addition, there exists an enormous variety of technical reasons for
various enhancements—for example, wrong coil adjustments, inhomogenous coil ad-
justments, injection technique, out-of-phase-echo times, and comparison within different
MR-sequences without any objective standardization or calibration. The number of false-
positive diagnoses due to biological or technical reasons effects the level of specificity. As
long as there are no definite standardizations or optimizations, MRM must be considered
as a research technique in demand of further development.
False negatives, on the other hand, are fairly rare, and this is meanwhile accepted in a large
series of cases revealing levels of sensitivity above 98%. However, different in situ cancers,
especially low-grade ductal carcinoma in situ (DCIS) cancers, need not show a tumor
angiogenesis and therefore could be left undetected in MRM. Especially for the detection
of DCIS, the combined analysis of kinetic and morphological pattern of enhancement
(e.g., segmental, unilateral) is important. Other causes for false negatives include rare
cases of special lobular cancers and situations after previous bleeding following a core,
fine-needle, or open biopsy.
The current approaches to imaging breast cancer using MRM are presented in UNIT A21.1,
which describes dynamic MRM techniques (multislice 2- and 3-D), as well as fat satura-
tion techniques, which are used mainly in the U.S.
Breast implants are well known to develop problems over time. The ability to detect
ruptures is critical to determining the status of the implant. Thus, UNIT A21.2 describes
examination of implants, focusing on the detection of internal or external implant ruptures.
Sequences have been developed to focus on the implant and any leakage into surrounding
tissue, including fat saturation methods and methods to generally enhance the contrast
between the implant (silicon) and surrounding tissue.
Werner A. Kaiser
Introduction
A21.0.2
Supplement 13 Current Protocols in Magnetic Resonance Imaging
MRI of Breast Lesions UNIT A21.1
Magnetic Resonance Mammography (MRM) has evolved in the last 17 years as a
powerful tool for diagnosing breast lesions. The tomographic aspect of imaging and a
high soft tissue contrast without radiation proved to have the highest sensitivity in
detecting especially small invasive breast lesions. Some major advantages of MRM are
the detection of multifocality/multicentricity; differentiation between scar and tumor;
delineation of breast implants; analysis of “surprising” lesions in the contralateral breast;
search for primary tumor in a CUP syndrome (Cancer of Unknown Primary) after having
detected malignant lymph nodes in the axillary region; and early analysis of response to
chemotherapy. Several unanswered questions remain, including: the varying specificities
reported in the literature, depending on the technical and methodological variations and
problems; MRM’s high price; the lack of detection of microcalcification and some cases
of DCIS (Ductal Carcinoma in situ); the need for contrast and sophisticated data
evaluation as well as delineation of hormonal aspects; and a broad variety of technical
pitfalls.
For differentiating between malignant and benign lesions, a high spatial as well as
temporal resolution is of utmost importance. Therefore, a so-called dynamic technique
(i.e., the repetitive imaging of the same slices before and in short time intervals after the
injection of contrast medium), is essential to detect the differences in initial enhancements
between malignant and benign lesions which are reflected by the tumorangiogenetic
vascular network of malignant lesions. This network results not only in an increased
number of vessels, but also and especially in a changed vascular architecture consisting
of primitive membranes, including defects in basal membranes, which are responsible for
an increased interstitial space and AV-shunts, which might explain the so-called wash-out
effect, i.e., the decrease of signal intensity after the initial rise in the first 1 to 2 min after
contrast agent injection. The protocols given here focus on dynamic techniques at high
field strength (1.5 T), where the majority of experience has occurred in MR Mammogra-
phy in the last decade.
DYNAMIC MR MAMMOGRAPHY BASIC
PROTOCOL
Magnetic Resonance Mammography was mainly evaluated at field strengths of 1.0 T and
1.5 T. There is only little experience at mid-field (0.5 T) or low field (0.2 T and lower).
At low field strength some compromises have to be made, for example, increase of
contrast medium, decrease of repeat time, and increase of flip angle in order to compensate
the relatively low decrease of T1 relaxation time induced by gadolinium based contrast
agents compared with high fields. The sequences described herein are based on the
author’s experience with a Philips 1.5 T ACS II and a Siemens 1.5 T Vision Scanner, but
are expected to be equally applicable to devices from other manufacturers.
The following technical advice is given to the technologist and radiologist, who, espe-
cially in the field of dynamic MR Mammography, have to work together as an effectively
coordinated team in order to make the examination successful.
Dynamic MR Mammography can be performed either as a multi-slice 2-D or a dynamic
3-D sequence. The 3-D sequence offers higher through-plane resolution than the 2-D
sequence; however, some machines and coils still lack sufficient signal homogeneity in
3-D techniques. In these cases, the critical subtraction images can be deteriorated by
artifacts. Most machines can perform multi-slice 2-D gradient echo sequences in a
sufficient homogeneity over a wider range of field of view, which is especially important
Breast
Contributed by Werner A. Kaiser A21.1.1
Current Protocols in Magnetic Resonance Imaging (2001) A21.1.1-A21.1.21
Copyright © 2001 by John Wiley & Sons, Inc.
 Table A21.1.1 Equipment Parameters for Dynamic MR Mammography There are two major disadvantages of low field strength. These are, first, the reduced
signal-to-noise, and, second, a smaller decrease in T1 induced by gadolinium based
Coil type Bilateral circularly polarized breast coil (or contrast agents as a result of the initially low T1 values of the tissue (for example, 600
phased array double breast coil, if available) msec at 0.5 T, 900 msec at 1.5 T). However, these disadvantages can be compensated at
Gradient coil strength 15-25 mT/m (or whatever the system permits) mid field by additional use of strong gradients (≥15 mT/m) and high slew rates (≥15 to
Cardiac gating No 25 T/m/sec). We also recommend increasing sensitivity towards Gd at low field by making
Peripheral gating No the sequence more T1 dependent (shortening of repeat time, increase of contrast medium
Respiratory gating No dose, etc.).
Oxygen Usually not
Motion cushions Very useful Besides the exclusion criteria listed above, we advise that one rigorously select for strong
Subtraction software Very useful indications for MR Mammography. Today, one should perform this relatively expensive
Automatic tuning Yes, but no additional tuning after the first and time-consuming technique only if high-quality X-ray mammography and ultrasound
pre-contrast scan (“switched off” or “offline” examinations still leave some doubts about the presence or absence, the malignancy or
technique) benignancy of a lesion, and/or the multicentricity or multifocality of lesions.
Fast dynamic technique Yes
Automatic image subtraction Yes Sequence overview in dynamic MRM (1.5 T)
Signal intensity versus time Very useful Table A21.1.2 gives an overview of the sequences used in the complete dynamic MR
curves examination at 1.5 T (for example, a Philips 1.5 T ACS II machine). After a short scout
Use of contrast agent Yes sequence, a coronal T1-weighted gradient-echo and a transverse T2-weighted turbo-spin-
echo sequence follow. The key part of this dynamic procedure is a T1-weighted gradient-
in imaging of both breasts in transverse or coronal orientation and for a later analysis of echo sequence before and 7 times after injection of contrast agent. Finally, the T1-weighted
lesions in subtracted images. coronal sequence is repeated so that a complete pre- and post-contrast image data set is
available in two orientations. In special patients (see Critical Parameters and Trou-
Table A21.1.1 lists the hardware necessary to perform the procedure along with the major bleshooting) a dynamic examination in coronal orientation is advised; in these cases the
appropriate parameters. The double breast coil is essential in order to evaluate both breasts orientations of the sequences outlined in Table A21.1.2 are listed in parentheses.
simultaneously. This is especially important to compare enhancement of both breasts, i.e.,
in cases of in-flow phenomena, DCIS, inhomogeneities, and detection of lesions in either Materials
one or both breasts simultaneously. Once the contrast medium is injected, the simultane- Normal saline (0.9% NaCl), sterile
ous examination of both breasts is strongly recommended. Extravascular contrast agent (e.g., Magnevist, Omniscan, or Prohance)
Patients having any ferromagnetic metal should be excluded from the examination, e.g., NOTE: Be sure that technicians and nurses have immediate access to any emergency
ferromagnetic particles, pacemakers, etc. This should be clarified by phone, before the equipment that may be relevant to a given study, or that may be needed for a particular
MR examination is scheduled. patient, such as crash carts or oxygen.
The currently available MR breast coils and MR machines have relatively limited space Set up equipment and patient
because of the space taken up by the doublebreast coil. In our experience, patients 1. Interview (screen) the patient to ensure that she has no contraindications such as
weighing more than 95 kg cannot be put in today’s machines, and so patients should be cardiac pacemakers or other implants containing ferromagnetic materials. Also,
screened by phone in advance. Patients taking hormones (i.e., hormone replacement determine if the patient has any health conditions that may require the presence of
therapy after menopause) should be advised to stop taking the hormones for ∼3 months
before MRM examination. Patients in the menstrual cycle can normally be examined.
However, the progesterone effect in the second half of the menstrual cycle usually induces Table A21.1.2 Dynamic MRM Examination at 1.5 Ta
a slightly increased contrast uptake, which might be a problem for inexperienced readers. Sequence Time
Therefore, an MR examination in the first half of the menstrual cycle, at least between
day 7 and 20, is advisable but not necessary. 1. Scout (transverse + coronal) 37 sec
2. Coronal (transverse) T1-weighted gradient echo 1 min, 47 sec
Besides pacemakers or ferromagnetic parts, other exclusion criteria are: previous allergies 3. Transverse (coronal) T2-weighted turbo spin echo 3 min, 12 sec
to gadolinium (Gd); unclear implants; clip implantations (especially in arteries or in the 4. Transverse (coronal) T1-weighted gradient echo 1 min, 0 sec
breast); psychiatric conditions, or claustrophobia. If there is any doubt about the possi- Injection of contrast agent: 0.1 mmol/kg Gd-DTPA intravenous (or 0.2 ml/kg per dose)
bility of some unclear implants, please contact the website http://www.mrisafety.com, for 5. Transverse (coronal) T1-weighted gradient echo 1 min, 0 sec per scan (total 7
further information. (7 repetitions) min)
Since many technical and methodological pitfalls can occur, it is strongly advised to Automatic image subtraction (between sequences 5 and 4)
perform an MRM examination only if sufficient experience in analyzing the images is 6. Coronal (transverse) T1-weighted gradient echo 1 min, 47 sec
MRI of Breast available. aTotal scanning time of these sequences is 15 min and 23 sec.
Lesions Breast

A21.1.2 A21.1.3
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 special emergency equipment during the scanning procedure, or if she will need Be sure that each breast is not deformed within the coil, since a later description of the
sedation medication necessitating the use of appropriate monitoring equipment. Use exact position of a lesion in the breast is performed by describing the breast as a
sedation only if it cannot at all be avoided. If you are unsure about the presence of half-spherical mass with a nipple at the top and the lesion being positioned as xyz mm
medially, laterally, cranially, caudally or dorsally in relation to the nipple.
any ferromagnetic materials, please contact the referring physician or the web site
http://www.mrisafety.com. 10. Be sure that the patient has stretched arms along the body and not bent arms crossed
The presence of any ferromagnetic metals may be a health hazard to the patient when she
over the head, in order to guarantee constant influx of the contrast medium.
is inside the magnet, and will also affect the imaging. If in doubt as to the exact composition Injection of contrast medium is performed in the magnet without moving the patient out of
of the items, it is best to exclude patients with any metal implants; see also Shellock (1996) the magnet.
for discussion of what implants may be safely scanned using magnetic resonance.
11. If needed, place a pillow or other support under the knees to make the examination
Patients may be accompanied into the magnet room by a friend or family member, who can more comfortable for the patient. Have the patient cushioned as comfortably as
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
possible in order to let her relax and to obtain images without motion artifacts.
It is strongly advised that this accompanying person does not talk to the patient during the 12. Use the centering light to position the middle of the breast coil (this should coincide
examination to avoid any movements. with the middle of the breast at the level of the nipple) and to put her into the isocenter
2. Clarify that the patient has had no previous allergy to or contraindications against of the magnet.
Gd. Explain the examination in detail and have the patient sign any necessary consent 13. If the patient is very nervous, try to calm her or bring her out of the magnet after a
form. few seconds and talk to her. In most cases this is sufficient.
3. Have the patient remove all jewelry and change into a gown to eliminate any metal Try to avoid any medical sedation, as that reduces the possibility of exchanging information
that might be found in clothing. or of detecting an allergy, and it may induce a paradoxic reaction.
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating
and image artifacts. Especially look for metallic rings in any body parts. Sequence 1: Rapid two-plane positioning pilot
14. To validate the patient’s position, run the system’s pilot (or scout) scan to ensure
5. Inform the patient about what will occur during the procedure, what she will correct location of the breast in two dimensions, using the imaging sequence given
experience while in the magnet, and how to behave, including the following: in Table A21.1.3 or similar parameters. This sequence usually consists of two
orthogonal planes to allow localization.
a. If earphones or headphones are used to protect the ears from the loud sounds
produced by the gradients, the patient will be asked to wear these, but will be able
Sequence 2: T1-weighted gradient echo
to communicate with you at any time during the imaging. 15. Run sequence 2 according to Table A21.1.4.
b. The patient will be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this works). This coronal T1-weighted sequence allows the complete examination of the axillary region.
The slice thickness in this and all following sequences depends on breast size and varies
c. Explain to the patient that movement of the head or coughing deteriorates image
quality. For good results the patient should not talk, and should avoid or minimize
swallowing or other movement during each scan—i.e., as long as the banging Table A21.1.3 Primary Clinical Imaging Parameters for Pilot Scan (Sequence 1)
sounds continue. Between scans, talking and swallowing are allowed in most
Patient position Prone
cases, but should be avoided when comparative positional studies are being
Scan type Spin echo (multi-slice,
performed; the patient will be informed when this is the case. half-Fourier)
d. Nevertheless, the patient may call out any time if she feels it necessary. Imaging plane (orientation) Transverse and coronal
Central slice or volume center Middle of the breast (usually at
6. Help the patient mount onto the table and stand beside the patient during positioning the nipple)
in the magnet. Either before or right after the patient lies down, set up any triggering Echo time (TE) 13 msec
devices or other monitoring equipment that is to be used. Repeat time (TR) 121 msec
7. If the patient is nervous, calm her and talk to her as much as possible before Flip angle (FA) 90°
positioning her into the machine. After having positioned her, explain that you will Fields of view (FOVx, FOVy) 450 mm, 450 mm
not talk to her in order to avoid any motion, unless communication is necessary. Resolution (Δx, Δy) 2.53 mm, 1.76 mm
Number of data points collected (Nx, Ny) 178, 256
8. Insert the plastic cannula into the cubital vein, fix it and connect a plastic tube before Slice thickness (Δz) 5 mm
putting the patient into the machine. After positioning the patient in the machine, Number of slices 5 in each orientation
connect the plastic tube with the automatic injector filled with a sufficient dose of Slice gap 3 mm
contrast medium and a bolus of sodium chloride (physiological concentration). Number of acquisitions (Nacq) 1
9. Place patient’s breasts in the coil. Scan time 37 sec
MRI of Breast
Lesions Breast

A21.1.4 A21.1.5
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 between 3 and 5 mm; correspondingly, the gap between the slices varies between 0.1 and Sequence 4: Transverse T1-weighted gradient echo (pre-constrast scan)
1.0 mm, depending on breast size. Slice thickness and slice gaps are kept constant Since a variety of pitfalls can occur, this dynamic scan procedure has to be carefully
throughout the whole examination of the patient. checked at three different checkpoints (see Fig. A21.1.1). Checkpoint I is just before the
precontrast scan. At this point, define the exact position of the breast in the coil by
Sequence 3: T2-weighted turbo spin echo (TSE) analyzing the scout scan and correct any deformation or malpositioning of the nipple. The
16. Run sequence 3 according to Table A21.1.5. orientation of the dynamic slices is checked (see below), the phase encoding direction is
The additional evaluation of signal intensity in T2-weighted images is important for the optimized (see below), and the scan parameters are finally controlled.
delineation of fluid cysts, myxoid fibroadenomas, edema, blood, abscesses, septations in
the lesion, etc. 17. Perform checkpoint I.
18. Run sequence 4 according to Table A21.1.6.
Table A21.1.4 Primary Clinical Imaging Parameters for T1-Weighted Gradient
Echo (Sequences 2 and 6) 19. After running sequence 4, checkpoint II of the dynamic scan is reached. At this point,
i.e., after the precontrast scan and before the injection of contrast medium (see Fig.
Patient position Prone A21.1.1), evaluate the homogeneity of the coil by observing how constant the fat
Scan type T1-weighted 2-D gradient echo signal is across both breasts. This is to ensure that the placement of the coil is
Imaging plane (orientation) Coronal
Central slice or volume center Middle of the breast
Echo time (TE) 5 msec
Repeat time (TR)a 100 msec checkpoints
dynamic MRM
Flip angle (FA) 80°
Fields of view (FOVx, FOVy) 350 mm, 315 mm contrast + NaCI
Resolution (Δx, Δy) 1.37 mm, 1.53 mm check I check II waiting check III
Number of data points collected (Nx, Ny) 256, 206 time
Slice thickness (Δz) 3–5 mm (depending on breast size)
Number of slicesb 24
Slice gap 0.1–1.0 mm (depending on breast X X X
size) pre post 1 ..2 ..3 ..4 ..5 ..6 ..7
Number of acquisitions (Nacq) 2
Scan time 1 min, 47 sec
aT is 100 msec for the Philips ACSII-1.5 T-machine or 240 msec for the Siemens Vision 1.5T machine. Figure A21.1.1 Special sequential performance of the scans in the dynamic MRM (sequences 4
R to 6). There are 3 “checkpoints” where the examiner has to control the performance: before the
b3 slabs of 8 slices each for the Phillips machine or 1 slab of 24 slices for the Siemens machine.
pre-contrast scan (checkpoint I), after the pre-contrast scan (checkpoint II), and after the last
post-contrast scan (checkpoint III; see text).
Table A21.1.5 Primary Clinical Imaging Parameters for T2-Weighted Turbo
Spin Echo (TSE; Sequence 3) Table A21.1.6 Primary Clinical Imaging Parameters for T1-Weighted Gradient
Echo (Pre-Contrast) Scan (Sequence 4)
Patient position Prone
Scan type T2-weighted turbo spin echo Patient position Prone
Imaging plane (orientation) Transverse Scan type T1-weighted 2-D gradient echo
Central slice or volume center Middle of the breast Imaging plane (orientation) Transverse
Echo time (TE) 300 msec Central slice or volume center Middle of the breast
Echo train length (ETL) 35 Echo time (TE) 5 msec
Repeat time (TR) 4000 msec Repeat time (TR)a 100 msec
Flip angle (FA) 90° Flip angle (FA) 80°
Fields of view (FOVx, FOVy) 315 mm, 350 mm Fields of view (FOVx, FOVy) 315 mm, 350 mm
Resolution (Δx, Δy) 1.64 mm, 1.37 mm Resolution (Δx, Δy) 1.23 mm, 1.37 mm
Number of data points collected (Nx, Ny) 192, 256 Number of data points collected (Nx, Ny) 256, 256
Slice thickness (Δz) 3–5 mm Slice thickness (Δz) 3–5 mm
Number of slices 24 Number of slicesb 24
Slice gap 0.1–1.0 mm Slice gap 0.1–1.0 mm
Number of acquisitions (Nacq) 1 Number of acquisitions (Nacq) 1
Scan time 3 min, 12 sec Scan time 1 min, 0 sec
aT is 100 msec for the Philips ACSII-1.5 T-machine or 240 msec for the Siemens Vision 1.5T machine.
R
MRI of Breast b3 slabs of 8 slices each for the Phillips machine or 1 slab of 24 slices for the Siemens machine.
Lesions Breast

A21.1.6 A21.1.7
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 appropriate and no tuning artifacts occurred and that the complete breast parenchyma
is included in the dynamic scans. Be sure that the automatic adjustment is switched
off for the following post-contrast scans in order to ensure an identical adjustment
position in pre- and post-contrast scans. Look for and analyze possible artifacts (e.g.,
metals, clips). If the patient had shown any motion, advise her not to move and in
that case, the pre-contrast scan is repeated once again.

Sequence 5: Transverse T1-weighted gradient echo (post-contrast scan)

20. Leaving the patient in the magnet, inject a dose of 0.1 mmol/kg (= 0.2 ml/kg)
Gd-DTPA at a bolus of 3 ml/sec followed by a sodium chloride bolus (physiological
concentration) of 20 ml (if the position is in the cubital vein) or 30 ml (if the position
is at the lower arm or hand vein).
21. After the end of the sodium chloride injection, wait 20 sec before beginning the first
post-contrast scan, in order to let the contrast medium flow in and to be able to
compare the results with previously published dynamic procedures. This waiting time
also makes it possible to avoid artifacts from interleaved data acquisitions of different
slices during the post-contrast scans.
Figure A21.1.2 Four images of the same slice before
22. After that waiting time, repeat without interruption seven post-contrast acquisitions
injection of contrast medium (upper left) and 1 min (upper
with identical measurements and sequence parameters as the pre-contrast scan. The right), 2 min (lower left) and 7 min (lower right) after
total scan time is 7 min (Table A21.1.7). injection of contrast medium. Strong initial signal in-
creases in a focal lesion (ductal invasive cancer) as well
Automatic image subtraction as segmental, dotted increases in the left segmental area
23. The software will perform an automatic image subtraction as soon as data points have oriented towards the nipple (neighbored DCIS). Signal
been measured; this automatic image subtraction is finished shortly after the end of decreases (washout) in the focal lesion in the late scan.
sequence 5. These signal changes can be better detected in the
corresponding subtractions in Figure A21.1.3.
24. Data evaluation begins at the monitor as soon as images are displayed on the screen
(see Table A21.1.2). As soon as the dynamic scan is finished, an automatic subtraction
allows for a complete fat subtraction and a quick detection of enhancing lesions
among these “innumerable” images. The subtraction images also allow the detection
of possible motions or inhomogeneities (Fig. A21.1.2 and Fig. A21.1.3). The auto-

Table A21.1.7 Primary Clinical Imaging Parameters for T1-Weighted Gradient

Echo (Post-Contrast) Scan (Sequence 5)

Patient position Prone

Scan type T1-weighted 2-D gradient echo
Imaging plane (orientation) Transverse
Central slice or volume center Middle of the breast
Echo time (TE) 5 msec
Repeat time (TR) 100 msec
Flip angle (FA) 80°
Fields of view (FOVx, FOVy) 315 mm, 350 mm
Resolution (Δx, Δy) 1.23 mm, 1.37 mm
Number of data points collected (Nx, Ny) 256, 256
Slice thickness (Δz) 3–5 mm
Number of slices 24 Figure A21.1.3 Here the image subtractions (post-mi-
Slice gap 0.1–1.0 mm nus pre-contrast scan) are seen after 1 min (upper left),
Number of acquisitions (Nacq) 1 2 min (upper right), 3 min (lower left), and 7 min (lower
right) post-contrast injection. The focal lesion including
Number of repetitions 7
the washout effect as well as the corresponding segmen-
Scan time 1 min, 0 sec (after 7 repetitions tal enhancement in the DCIS are documented.
MRI of Breast the scan time adds to 7 min)
Lesions Breast

A21.1.8 A21.1.9
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 matic data evaluation enables a real time navigation through all slices in four Table A21.1.9 Primary Clinical Imaging Parameters for Dynamic 3D MR
dimensions (three spatial dimensions and one time dimension). Mammography at 1.5 T (Sequence 8)

25. However, pitfalls concerning automatic data evaluation should not be overlooked (see Patient position Prone
Critical Parameters and Troubleshooting). Some examples are an inadequate or Scan type 3-D short TR gradient echo
wrong injection technique; an inappropriate receiver/transmitter adjustment; a mis- Imaging plane (orientation) Transverse
interpretation of inflow phenomena; hormone effects, and vessels as lesions. Central slice or volume center Middle of the breast
Echo time (TE) 4 msec
Sequence 6: coronal T1-weighted gradient echo (post-contrast scan) Repeat time (TR) 8.1 msec
The final examination is a repetition of the T1-weighted gradient echo sequence in the Flip angle (FA) 20°
coronal orientation. Fields of view (FOVx, FOVy) 320 mm, 240 mm
Resolution (Δx, Δy) 1.25 mm, 1.25 mm
26. Run sequence 6 according to Table A21.1.4.
Number of data points collected (Nx, Ny) 256, 192
27. After the last post-contrast scan at checkpoint III (see Fig. A21.1.1), the contrast Slice thickness (Δz) 1–2 mm
uptake is tested according to enhancement in breast parenchyma, vessels, muscles, Number of slices 48
etc. Slice gap 0
The constancy of tuning parameters during the complete time of the dynamic scan has to Number of acquisitions (Nacq) 1
be checked by a constant time-intensity curve in the fatty tissue, because that does not Fat suppression Temporal subtraction
enhance. If the fat signal is inhomogeneous and/or varying over time, the signal-intensity Scan time 75 sec
versus time curve is not helpful or will be misleading (see Critical Parameters and
Troubleshooting).
Set up equipment and patient
1. Use the same equipment and perform patient set-up as for the previous method (see
ALTERNATE MR IMAGING BY CONTRAST-ENHANCED DYNAMIC 3-D TECHNIQUE Basic Protocol, steps 1 to 13). This is also valid for the establishment of an intervenous
PROTOCOL 1
In the past years, different dynamic techniques have been evaluated, including multi-slice line as well as the information and informed consent obtained from the patient.
2-D (see Basic Protocol) or 3-D techniques. 3-D scans allow a better spatial resolution
with thinner slices. However, the homogeneity of 3-D sequences on some machines has Sequence 7: Scout scan
been limited, so that inhomogeneity artifacts deteriorated the kinetic data evaluation. 2. Run a rapid 3-D pilot scan to demonstrate the position of the breast in the coil (Table
Today both multi-slice 2-D and 3-D sequences can be applied with equal results in most A21.1.8).
machines, and therefore this alternate protocol is given in detail here. One can also run sequence 1 (Table A21.1.3) instead.

Table A21.1.8 Primary Clinical Imaging Parameters for Scout Scan (Sequence 7)
Sequence 8: 3-D short TR gradient echo scans
3. Run a 3-D short TR gradient echo sequence according to Table A21.1.9.
Patient position Prone
4. Repeat steps 20 and 21 of the Basic Protocol (to inject the contrast agent).
Scan type 3-D gradient echo
Imaging plane (orientation) Transverse or coronal (but not 5. After a 20 sec wait time, run the 3-D gradient echo sequence 5 times according to
sagittal) Table A21.1.9 to get 6 data points for a signal-intensity time curve.
Central slice or volume center Centered at the nipple level
This is not the same as setting the number of acquisitions to be 5.
Echo time (TE) As short as possible (but in a
“in-phase” echo time, i.e.,
multiple of 4.8 msec) FAT-SATURATED 3-D SCAN: HIGH SPATIAL RESOLUTION ALTERNATE
Repeat time (TR) As short as possible (e.g., 10 msec) PROTOCOL 2
In the United States fat saturation techniques have been evaluated which focus especially
Flip angle (FA) 10° to 15°
on the high spatial resolution used to describe morphological features of lesions. These
Fields of view (FOVx, FOVy) 350 mm, 350 mm
techniques are also mentioned in this unit.
Resolution (Δx, Δy) 1.37 mm, 1.37 mm
Number of data points collected (Nx, Ny) 256, 256 Set up equipment and patient
Slice thickness (Δz) 1–2 mm (depending on breast size) 1. Use the same equipment and perform the same patient set up as in the Basic Protocol
Number of slices 48 (steps 1 to 14).
Slice gap No gap
Number of acquisitions (Nacq) 1 Sequence 9: 3-D short TR gradient echo scans: fat suppression
Scan time <90 sec (depending on TR) 2. Run a 3-D short TR gradient echo sequence according to Table A21.1.10.
3. Repeat steps 20 and 21 of the Basic Protocol (to inject the contrast agent).
MRI of Breast
Lesions Breast

A21.1.10 A21.1.11
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A21.1.10 Primary Clinical Imaging Parameters for Focusing High Spatial Sequence 10: 3-D short TR gradient echo scans: fat suppression
Resolution at the Cost of High Temporal Resolution (Sequence 9) 2. Run a 3-D short TR gradient echo sequence according to Table A21.1.11.
Patient position Prone 3. Repeat steps 20 and 21 of the Basic Protocol (to inject the contrast agent).
Scan type 3-D short TR gradient echo
4. After a 20-sec wait time, run the 3-D gradient echo sequence according to Table
Imaging plane (orientation) Sagittal (dual volume) with lateral
compression A21.1.11.
Central slice or volume center Middle of the breast
Echo time (TE) 2.2 msec COMMENTARY
Repeat time (TR) 22 msec
Background Information formed in a prone position to minimize move-
Flip angle (FA) 30° The development of MR Mammography ment and artifacts caused by breathing. How-
Fields of view (FOVx, FOVy) 160 mm, 160 mm has included several phases up to now. At first ever, at that time the available spin-echo and
Resolution (Δx, Δy) 0.31 mm, 0.625 mm MR tomography in general had to be devel- inversion-recovery sequences did not allow a
Number of data points collected (Nx, Ny) 512, 256 oped, then special breast coils (single breast definite detection and differentiation of small
Slice thickness (Δz) 1.875 mm coils, followed by double breast coils and later lesions in all cases, yet the advantage of imag-
Number of slices 32 combined imaging/interventional coils) had to ing in thin slices in any orientation in a variable
Slice gap 0 be built and optimized. Contrast media were soft tissue contrast without X-rays was already
Number of acquisitions (Nacq) 1 used at first in spin-echo sequences and later in clear (Heywang et al., 1985; Kaiser and Zeitler,
Fat suppression Yes sophisticated so-called “dynamic” techniques 1986a,b). Further progress was the develop-
Scan time 180 sec using gradient echo sequences for a better dif- ment and introduction of the MR contrast me-
ferential diagnosis. The main topics today are dium Gadolinium-DTPA (Weinmann et al.,
technological optimization, evaluation of diag- 1984), because experiments were possible in
4. After a 20-sec wait time, run the 3-D gradient echo sequence twice according to Table nostic criteria, and development of interven- analogy to computer tomographic results using
A21.1.10. tional procedures. ionized X-ray contrast medium and radioactive
Already in 1971 different relaxation times iodine uptake from the mid 1970s (Eskin et al.,
This is not the same as setting the number of acquisitions to be 2.
in tumor tissue as compared with normal tissue 1974; Chang et al., 1978). The first group,
were measured using in vitro experiments which got and used contrast medium, was Hey-
ALTERNATE FAT-SATURATED 3-D SCAN: RODEO (ROTATING DELIVERY OF (Damadian, 1971). However, MR imaging of wang et al. (1986). Initially spin-echo se-
PROTOCOL 3 EXCITATION OFF-RESONANCE) the human body was possible only after the quences in relatively long examination times
As mentioned in Alternate Protocol 2, this is another alternate protocol to describe application of local gradient fields (Lauterbur, and high contrast dosages were applied and
morphological features of lesions. 1973). The first tissue samples of the breast reported; however, the uptake of contrast me-
were published by Mansfield (1979). These dium of cancers, normal tissue, and prolifera-
Set up equipment and patient tissue samples were taken ∼90 min after opera- tive changes could not be differentiated suffi-
1. Use the same equipment and perform the same patient setup as in the Basic Protocol tion. The first in vivo images of whole-body ciently. After the introduction of fast gradient
(steps 1 to 14). MR machines from 128 breasts of 65 patients echo sequences (Haase et al., 1986), the first
including 7 cancers were taken in 1982 (Ross dynamic examinations were established that
et al., 1982). In these reports a whole-body used repetitive measurements of the same slices
Table A21.1.11 Primary Clinical Imaging Parameters for RODEO (see Harms
et al., 1993; Sequence 10) scanner in a field strength of 0.045 T was used before and in short- time intervals after contrast
and each patient was imaged in a supine posi- medium injection (Kaiser and Oppelt, 1987;
Patient position Prone tion, providing 2 proton density–weighted im- Kaiser and Zeitler, 1989). However how to
Scan type 3-D short TR gradient echo ages and 6 T1 relaxation time measurements. In better differentiate between benign and malig-
Imaging plane (orientation) Sagittal 1983, El-Yousef published results of 2 patients nant lesions was the subject of scientific dis-
using an experimental surface coil in a field cussion for a long time (see subsection in Lit-
Central slice or volume center Middle of the breast
strength of 0.03 T and reported a reduced signal erature Cited listing References on Differenti-
Echo time (TE) 5 msec
intensity of both cancers. In a following publi- ating Benign and Malignant Breast Lesions).
Repeat time (TR) 20 msec
cation, El-Yousef (1984) reported images of 10 The following development of a double breast
Flip angle (FA) 45° volunteers and 45 patients describing a reduced coil (Kaiser and Kess, 1989) allowed routine
Fields of view (FOVx, FOVy) 120 mm, 120 mm signal intensity of breast lesions in spin echo measurements of both breasts of a patient in a
Resolution (Δx, Δy) 0.47 mm, 0.94 mm and inversion-recovery sequences. single examination with a good signal-to-noise
Number of data points collected (Nx, Ny) 256, 128 Starting in 1983, special breast coils were ratio.
Slice thickness (Δz) 1.2 mm developed which were used only for breast After these initial steps, a period of evalu-
Number of slices 128 imaging and later sold commercially. Until ation of dynamic techniques using different
Slice gap 0 1986 only single breast coils were available measurement sequences and dosages followed,
Number of acquisitions (Nacq) 1 (Fritschy et al., 1984; Kaiser, 1985; Stelling et mainly in Europe. Since 1991 American groups
Fat suppression Yes al., 1985). After first disappointing trials in a reported contrast-enhanced MR Mammogra-
MRI of Breast supine position, the examinations were per- phy, too, using mainly fat saturation sequences
Lesions Scan time 320 sec Breast

A21.1.12 A21.1.13
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 (Harms et al., 1993; Orel et al., 1995) in single
breast coils. The results of the enormous variety
of measurement techniques were confusing and
resulted in a wide variation of opinion about
the usefulness of MR Mammography (Kaiser,
1996).
By using MR Mammography, doctors could
detect small lesions on the size of a few milli-
meters. These lesions were often not seen by
X-ray or ultrasound. Therefore, the need for the
development of interventional MR procedures
increased, at first for positioning of markers,
later for biopsy and therapeutic removal (Huss-
man et al., 1993; de Souza et al., 1995; Silver-
man et al., 1995; Doler et al., 1996; Mahfouz
et al., 1996; Sittek et al., 1996; Kuhl et al., 1997;
Wurdinger et al., 1997; Thiele et al., 1998). Up
to now these techniques required additional
MR imaging on another day, i.e., a second
measurement in a repetitive use of the MR
device and contrast medium injection. At pre-
sent, single breast biopsy coils in so-called
“closed” as well as in “open” MR machines are
tested. A bilateral coil for combined imaging
and simultaneous intervention has been clini-
cally tested since 1997 (Wurdinger et al., 1997).
Today modern high-field MR machines
(above 1.0 T) do not allow direct access to the
breasts during data acquisition. The patient has
to be in the isocenter of the main magnetic field
during imaging and must be moved outside the
machine for the following intervention. The
position of a wire marking, a core biopsy, or a
laser fiber has to be checked after the position-
ing of the interventional device with further
imaging in the MR machine, therefore increas-
ing measurement time, artifacts and pitfalls. An
MR compatible robotic system for simultane-
ous imaging and immediate biopsy/interven-
tion at high field strength is in clinical testing
(Kaiser et al., 2000).
Critical Parameters and
Troubleshooting
MR Mammography has to be performed in
pre- and post-contrast scans to detect the tumor-
specific vessel structure which is called tumo-
rangiogenesis. Up to now a non-contrast scan
alone is insufficient in detecting very small
lesions. In addition to morphological evalu-
ations, the time intensity curve is very impor-
tant in making the diagnosis. The signal inten-
sity in the first post-contrast images is abso-
lutely crucial for a correct interpretation of the
data. This first contrast signal intensity image
depends critically on exact measurement pa-
rameters. Injection time, injection site, sodium
chloride bolus, length of plastic tube, arm po-
sition, and tuning parameters are major factors
of this critical image signal intensity. Figure A21.1.5 Nine images of the same slice before (upper left) and every minute after
According to our experience, an increase of injection of contrast medium: a malignant lesion shows a striking initial signal increase
>90% of signal intensity in the first 90 sec after especially in the tumor periphery and a constancy of signal intensity during the following
the start of contrast medium injection (“90-90 dynamic examination. Note the slower signal change of the normal breast parenchyma.
rule”) is a threshold for a malignant enhance-
ment using 2-D dynamic imaging at 1.5 T (Fig.
A21.1.4 and Fig. A21.1.5). To reach this critical
internal filling
threshold, the measurement using a region of dynamic MRM
interest (ROI) procedure has to be performed
carefully. Only the “vital” tumor areas (Fig.
A21.1.5 and Fig. A.21.1.6) show a critical in-
itial enhancement followed by a plateau phe-
nomenon or a washout effect. An inclusion of
necrotic areas of the tumor or surrounding nor-
centripetal centrifugal
mal glandular or fatty tissue will deteriorate and

Figure A21.1.6 Different types of enhancement after injection of contrast medium: centrifugal
(from inside to outside) and centripetal (from outside to inside).
kinetic description
MRM
falsify the effect of kinetic data evaluation. It is a pretty high spatial resolution, but have a
initial necessary to look for the fastest enhancing limited value in the kinetic data evaluation.
delayed
portion of the lesion. Since a pre-pulse in a low bandwidth is neces-
It is also important not to include just any sary, the adjustment parameters in pre- and
enhancing spot and describe it as malignant. post-contrast scans are not identical. Present fat
Vessels are cut in different orientations in the saturation techniques usually have a relatively
slice and are therefore displayed in a round, high signal inhomogeneity, so that a quantita-

Signal intensity (%)

oval or comet shaped manner. The observance
of inflow phenomena, the delineation of vessels
in other slices or other orientations or, if you
have still any doubts, an additional MR angiog-
Time (min) raphy sequence (see Chapter A1 for MRA se-
quence) all help clarify whether the tiny en-
hancing area is a vessel or not.
Figure A21.1.4 Signal intensity versus time curve showing relative (percentage) changes of
MRI of Breast signal after injection of contrast agent: the early (“initial”) and later (“delayed”) phases are described
Fat saturation techniques are in a relatively
Lesions widespread use in the United States. They allow
separately.
A21.1.14
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
tive evaluation is difficult and subtractions are
difficult, nevertheless often necessary. The
measurement time in most cases is longer than
in non fat saturation techniques and the “diag-
nostic window” for the detection of differences
between benign and malignant enhancement
criteria is restricted.
Fat suppression methods may either be by
temporal subtraction (which is ideal for fat Breast
A21.1.15
 suppression, since only enhancing structures
are delineated as high signal intensity lesions),
or by pre-pulses with the above-mentioned
limitations present.
The injection of contrast medium is made in
a bolus (3 ml/sec) either manually or by an
automatic injector followed by a post-contrast
bolus of physiological sodium chloride of 20
ml (cubital vein) or 30 ml (lower arm/hand). It
is preferable not to move or talk to the patient
during injection, but rather to explain the pro-
cedure before positioning the patient in the
machine. According to our experience, the dose
of contrast medium should not be higher than
0.1 mmol/kg. A higher dose shortens the al-
ready short “diagnostic window”, i.e., the time
difference of only 1 to 2 min where the contrast
uptake between malignant and benign lesions
is sufficiently different in order to enable a
differential diagnosis.
The data evaluation begins by looking at
fatty tissue on both breasts and to control if the
signal intensity of fatty tissue is constant eve-
rywhere in the breast. If this signal intensity of
fatty tissue as an “internal standard” varies,
consider that it might be due to a changing
receiver adjustment or field inhomogeneity ef-
fects or artifacts. In these cases a quantitative
evaluation is difficult or impossible, since the
technical problems overlap any morphological
information.
Dynamic MRM should be performed using
transverse scans (sequence 5), because in this
orientation the identification of vessels, nip-
ples, posterior borders of implants, fatty layer
between parenchyma and muscle, signal homo-
geneity of the coil and signal correlation with
the aorta is much easier. However, if the lateral
border of the parenchyma towards the axillary
region and/or lymph nodes is especially impor-
tant, a dynamic MRM in coronal scans is ad-
vantageous. A sagittal orientation should be
avoided because of doubled measurement time
and the lack of correlation with the other breast.
Furthermore, it is difficult to detect inflow phe-
nomena and hormone effects, etc., in this ori-
entation
If only the axillary region is important, the
rotation of phase encoding direction towards a
dorsoventral orientation is recommended to
keep the lateral axillary region free of phase
encoding artifacts.
It is of crucial importance to use echo times
where fat and water protons are in “in-phase”
conditions, because only in these echo times
MRI of Breast will Gd have a signal-increasing effect on the
Lesions water protons in the voxel. Especially tiny re-
A21.1.16
ticular structures like DCIS or lobular cancers
can be examined only in this “in-phase” image.
“In-phase” situations for 1.5 T are even multi-
ple numbers of 2.4 msec, i.e., 4.8 msec, 9.6
msec, etc. Odd multiple numbers of 2.4 msec,
for example, 7.2 msec, are “forbidden” because
of the “opposed effect” of water and fat protons.
This effect is field strength dependent and
should be adjusted according to the field
strength used.
During the dynamic examination it is essen-
tial not to change tuning parameters between
scans in order to guarantee identical pulses in
pre- and post-contrast scans. Only under these
conditions is a correct signal intensity versus
time evaluation possible.
The pre-contrast signal intensity of paren-
chyma should be in a typical range for this
sequence. If it is too low, look for an inappro-
priate receiver adjustment or inappropriate coil
as the problem and try to change the receiver
adjustment. If this pre-contrast signal intensity
is too high, it could also suggest an inappropri-
ate receiver adjustment or some other medi-
cal/biological reason for the problem, such as
bleeding after puncture/biopsy, hormone ef-
fect, pregnancy, previous operation and/or ra-
diation.
Anticipated Results
Morphological and kinetic information
Diagnosis of breast lesions in MR Mam-
mography is always made by a combined mor-
phological and kinetic analysis. It is the main
feature of MRM that tumortypic “tumorangio-
genesis” can be detected by signal intensity
versus time curves. Various curves are listed in
Figure A21.1.4. A typical malignant lesion
shows the striking initial increase (wash-in phe-
nomenon) within the first 1 to 2 min after the
injection of contrast medium followed by either
a constant signal (plateau phenomenon) or—
more specifically—a decrease in signal inten-
sity (washout effect). The sudden increase and
the following constancy or washout effect make
the so-called “cancer-corner” of the signal-time
curve. Benign lesions normally show a slower
initial increase which is continuouly rising over
the complete dynamic examination. Morpho-
logical analysis describes the type of enhance-
ment as shown schematically in Figure
A21.1.6. The distribution of a lesion is de-
scribed as either regional or patchy or diffuse
or symmetric, and the morphology of the lesion
itself is described as a focal mass with sharp or
non-sharp margins, linear, linear-branched, or
Current Protocols in Magnetic Resonance Imaging
Table A21.1.12 Morphological Analysis of Breast
Lesions
Shape Margin
Round Smooth
Oval Scalloped
Lobulated Irregular
Irregular Spiculated
Stellate
segmental aspects (Harms, 1999). The initial
and late enhancement is described as being
homogeneous, heterogeneous, rim enhance-
ment, bright or dark septations after contrast
medium injection and centripetal or centrifugal
filling effect over time. A centrifugal filling
(from inside to outside) directs mainly towards
a benign lesion, a centripetal filling (from out-
side to inside) towards a malignant lesion (Fig.
A21.1.6). An overview of the morphological
analysis according to shape, margins and en-
hancement is listed in Table A21.1.12.
Acknowledgement
I would like to thank my secretary Dorit
Dirlam for typing this unit.
Literature Cited
Chang, C.H.J., Sibala, J.L., Lin, F., Fritz, S.L., Gal-
lagher, J.H., Dwyerill, S.J., and Templeton, A.W.
1978. Computed tomographic evaluation of the
breast. Am. J. Roentgenol. 131:459-464.
Damadian, R. 1971. Tumor detection by nuclear
magnetic resonance. Science 171:1151-1153.
de Souza, N.N., Kormos, D.W., Krausz, T., Coutts,
G.A., Hall, A.S., Burl, M., Schwieso, J.E., Puni,
R., and Vernon, C. 1995. MR-guided biopsy of
the breast after lumpectomy and radiation ther-
apy using two methods of immobilization in the
lateral decubitus position. J. Magn. Reson. Im-
aging 5:525-528.
Doler, W., Fischer, U., Metzger, I., Harder, D., and
Grabbe, E. 1996. Stereotaxic add-on device for
MR-guided biopsy of breast lesions. Radiology
200:863-864.
El Yousef, S.J., Alfidi, R.J., and Duchesneau, R.H.
1983. Initial experience with nuclear magnetic
resonance (NMR) imaging of the human breast.
J. Comput. Assist. Tomogr. 7:215-218.
El Yousef, S.J., Duchesneau, R.H., Alfidi, R.J.,
Haaga, J.R., Bryan, P.J., and LiPuma, J.P. 1984.
Magnetic resonance imaging of the breast. Radi-
ology 150:761-766.
Eskin, B.A., Parker, J.A., Bassett, J.G., and George,
D.L. 1974. Human breast uptake of radioactive
iodine. Obstet. Gynecol. 44:398-492.
Current Protocols in Magnetic Resonance Imaging
In the mass
Homogeneous
Heterogeneous
Rim
Septations
Enhancing
Non-enhancing
Fritschy, P., Müller, E., Sauter, R., and Kaiser, W.
1984. MR-imaging with special coils for visual-
isation of superficial organs and structures.
RSNA 1984, Radiology 153(P):243-244.
Haase, A., Frahm, J., Matthaei, D., et al. 1986.
FLASH-imaging: Rapid NMR imaging using
low flip angle pulses. J. Magn. Reson. 67:258-
268.
Harms, S.E. (ed.) 1999. Technical Report of the
International Working Group on Breast MRI. J.
Magn. Reson. Imaging 10:979-1015.
Harms, S.E., Flamig, D.P., Hesley, K.L., Meiches,
M.D., Jensen, R.A., Evans, W.P., Savino, D.A.,
and Wells, R.V. 1993. MR imaging of the breast
with rotating delivery of excitation off reso-
nance: Clinical experience with pathologic cor-
relation. Radiology 187:493-501.
Heywang, S.H., Fenzl, G., Edmaier, M. et al. 1985.
Kernspintomographie in der Mammadiagnostik.
Fortschr. Röntgenstr. 143:207-212.
Heywang, S.H., Hahn, D., Schmidt, H., Krischke, I.,
Eiermann, W., Bassermann, R.J., and Lissner, J.
1986. MR imaging of the breast using Gadolin-
ium-DTPA. J. Comput. Assist. Tomogr. 10:199-
204.
Hussman, K., Renslo, R., Phillips, J.J., Fischer, H.J.,
Khalkhali, I., Braslau, D.L., and Sinow, R.M.
1993. MR mammographic localization. Work in
progress. Radiology 189:915-917.
Kaiser, W. 1985. MRI of the female breast. First
clinical results. Arch. Int. Physiol. Biochim.
93:67-76.
Kaiser, W.A. and Kess, H. 1989. Prototyp-Doppel-
spule für die Mamma-MR-Messung. Fortschr.
Röntgenstr. 151:103-105.
Kaiser, W.A. 1996. Gegenwärtiger Stand der MR-
Mammographie (Present state of MR mammog-
r aph y, ed itorial). Fortschr. Röntgenstr.
165(5):425-427.
Kaiser, W.A. and Oppelt, A. 1987. KST der Mamma
mit Schnellbildverfahren. MR’87, 2. Interna-
tionales Kernspintomographie Symposium,
29.1.-1.2.87, Garmisch-Partenkirchen, Refer-
ateband, 303-310. Schnetztor-Verlag, Konstanz.
Kaiser, W. and Zeitler, E. 1986a. Die Kernspinto-
mographie der Mamma—Diagnose, Differenti-
aldiagnose, Probleme und Lösungsmöglichkei-
Breast
A21.1.17
 ten. Teil I: Untersuchungsverfahren. Fortschr.
Röntgenstr. 144(5):459-465.
Kaiser, W. and Zeitler, E. 1986b. Die Kernspinto-
mographie der Mamma—Diagnose, Differen-
tialdiagnose, Probleme und Lösungsmöglichkei-
ten. Teil II: Diagnostik. Fortschr. Röntgenstr.
144(5):572-579.
Kaiser, W.A. and Zeitler, E. 1989. MR-imaging of
the breast: Fast imaging sequences with and
without Gd-DTPA. Radiology 170:681-686.
Kaiser, W.A., Fischer, H., Vagner, J., and Selig, M.
2000. Robotic system for biopsy and therapy of
breast lesions in a high-field whole-body mag-
netic resonance tomograph. Invest. Radiol.
35(8):513-519.
Kuhl, C.K., Elevelt, A., Leutner, C.C., Gieseke, J.,
Pakos, E., and Schild, H.H. 1997. Interventional
breast MR imaging: Clinical use of a stereotactic
localization and biopsy device. Radiology
204:667-675.
Lauterbur, P.C. 1973. Image formation by induced
local interactions: Examples employing nuclear
magnetic resonance. Nature 242:190-191.
Mahfouz, A.E., Rahmouni, A., Zylbersztejn, C., and
Mathieu, D. 1996. MR-guided biopsy using ul-
trafast T1- and T2-weighted reordered turbo fast
low-angle shot sequences: Feasibility and pre-
liminary clinical applications. AJR Am. J.
Roentgenol. 167:167-169.
Mansfield, P., Morris, P.G., Ordidge, R., et al. 1979.
Carcinoma of the breast imaged by nuclear mag-
netic resonance (NMR). Br. J. Radiol. 52:242-
243.
Orel, S.G., Schnall, M.D., Powell, C.M., Hochman,
M.G., Solin, L.J., Fowble, B.L., Torosian, M.H.,
and Rosato, E.F. 1995. Staging of suspected
breast cancer: Effect of MR imaging and MR-
guided biopsy. Radiology 196(1):115-122.
Ross, R.J., Thompson, J.S., Kim, K., and Bailey,
R.A. 1982. Nuclear magnetic resonance imaging
and evaluation of human breast tissue: Prelimi-
nary clinical trials. Radiology 143:195-205.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Silverman, S.G., Collick, B.D., Figueira, M.R.,
Khorasani, R., Adams, D.F., Newman, R.W., and
Topulos, G.P. 1995. Interactive MR-guided bi-
opsy in an open-configuration MR imaging sys-
tem. Radiology 197:175-182.
Sittek, H., Kessler, M., Muller, U., Lisse, U., Untch,
M., Bohmert, H., and Reiser, M. 1996. Tech-
niques for the preoperative marking of nonpal-
pable br east lesions in MRT. Fortschr.
Röntgenstr. 165:84-87.
Stelling, C.B., Wang, P.C., Lieber, A., Mattingly,
S.S., Griffen, W.O., and Powell, D.E. 1985. Pro-
totype coil for magnetic resonance imaging of
the female breast: Work in progress. Radiology
154:457-462.
Thiele, J., Schneider, J.P., Franke, P., Lieberenz, S.,
MRI of Breast and Schmidt, F. 1998. New method of MR-
Lesions
A21.1.18
guided mammary biopsy. Fortschr. Röntgenstr.
168:374-379.
Weinmann, H.J., Laniado, M., and Mützel, W. 1984.
Pharmacokinetics of gadolinium-DTPA/di-
meglumine after intravenous injection into
healthy volunteers. Physiol. Chem. Phys. Med.
NRM 16:167-172.
Wurdinger, S., Noras, H., Straube, K., Michaelsen,
S., and Kaiser, W.A. 1997. New MRI breast coil
for imaging and doublesided intervention: Pre-
liminary results. ECR ’97, European Congress
of Radiology, 2-7 March 1997, Eur. Radiol.
7:243.
References on differentiating benign and
malignant breast lesions
Boetes, C., Strijk, S.P., Holland, R., Barentsz, J.O.,
Van Der Sluis, R.F., and Ruijs, J.H. 1997. False-
negative MR imaging of malignant breast tu-
mors. Eur. Radiol. 7:1231-1234.
Bone, B., Aspelin, P., Bronge, L., Isberg, B., Per-
beck, L., and Veress, B. 1996. Sensitivity and
specificity of MR mammography with his-
topathological correlation in 250 breasts. Acta
Radiol. 37:208-213.
Brenner, R.J. Breast MR imaging. An analysis of its
role with respect to other imaging and interven-
tional modalities. Magn. Reson. Imaging Clin. N.
Am. 2(4):705-723.
Brookes, J.A., Redpath, T.W., Gilbert, F.J., Need-
ham, G., Murray, A.D. 1996. Measurement of
spin-lattice relaxation times with FLASH for
dynamic MRI of the breast. Br. J. Radiol. 69:206-
214.
Buadu, L.D., Murakami, J., Murayama, S.,
Hashiguchi, N., Sakai, S., Masuda, K.,
Toyoshima, S., Kuroki, S., and Ohno, S. 1996.
Breast lesions: correlation of contrast medium
enhancement patterns on MR images with his-
topathologic findings and tumor angiogenesis.
Radiology 200:639-49.
Buttner, A., Bauer, M., Schulz-Wendtland, R.,
Grebmeier, H., Tulusan, H.A., and Lang, N.
1993. Significance of MR mammography for
evaluation of the tumor and adjacent tissue. Gy-
nakol. Geburtshilfliche Rundsch. 33(Suppl
1):229-230.
Daldrup, H.E., Roberts, T.P., Muhler, A., Gossmann,
A., Roberts, H.C., Wendland, M., Rosenau, W.,
and Brasch, R.C. 1997. Macromolecular contrast
media for MR mammography. A new approach
to characterizing breast tumors. Radiologe
37:733-740.
Drew, P.J., Kerin, M.J., Turnbull, L.W., Imrie, M.,
Carleton, P.J., Fox, J.N., and Monson, J.R. 1998.
Routine screening for local recurrence following
breast-conserving therapy for cancer with dy-
namic contrast-enhanced magnetic resonance
imaging of the breast. Ann. Surg. Oncol.
5(3):265-270.
Ercolani P., Giovagnoni A., Giuseppetti G., Baldas-
sarre, S., De Nigris, E., and Amici, F. 1991.
Imaging dinamico con resonanza magnetica
Current Protocols in Magnetic Resonance Imaging
nella senologia diagnostica. Radiol. Med. Torino
82:422-426.
Farria, D.M., Gorczyca, D.P., Barsky, S.H., Sinha,
S., and Bassett, L.W. 1996. Benign phyllodes
tumor of the breast: MR imaging features. Am.
J. Roentgenol. 167:187-189.
1st International Congress on MR Mammography.
1997. Jena, Germany, 29-31 May 1997. Ab-
stracts. Eur. Radiol. 7(4):605-624.
Fischer, U., Kopka, L., Brinck, U., Korabiowska,
M., Schauer, A., and Grabbe E. 1997. Prognostic
value of contrast-enhanced MR mammography
in patients with breast cancer. Eur. Radiol.
7:1002-1005.
Fischer, U., von Heyden D, Vosshenrich, R., Vieweg,
I., and Grabbe, E. 1993. Signal characteristics of
malignant and benign lesions in dynamic 2D-
MRT of the breast. Rofo. Fortschr. Geb.
Rontgenstr. Neuen Bildgeb. Verfahr. 158:287-
292.
Fobben, E.S., Rubin, C.Z., Kalisher, L., Dembner,
A.G., Seltzer, M.H., and Santoro, E.J. 1995.
Breast MR imaging with commercially available
techniques: radiologic-pathologic correlation.
Radiology 196:143-152.
Frankel, S.D. and Sickles, E.A. 1997. Morphologic
criteria for interpreting abnormalities seen at
breast MR imaging. Radiology 202:633-634.
Friedman, E.P., Hall-Craggs, M.A., Mumtaz, H.,
and Schneidau, A. 1997. Breast MR and the
appearance of the normal and abnormal nipple.
Clin. Radiol. 52:854-861.
Gilles, R., Guinebretiere, J.M., Shapeero, L.G., Les-
nik, A., Contesso, G., Sarrazin, D., Masselot, J.,
and Vanel, D. 1993. Assessment of breast cancer
recurrence with contrast-enhanced subtraction
MR imaging: Preliminary results in 26 patients.
Radiology 188:473-478.
Gilles, R., Guinebretiere, J.M., Toussaint, C., Spiel-
man, M., Rietjens, M., Petit, J.Y., Contesso, G.,
Masselot, J., Vanel, D. 1994. Locally advanced
breast cancer: contrast-enhanced subtraction
MR imaging of response to preoperative chemo-
therapy. Radiology 191:633-638.
Gorczyca, D.P. 1993. Bilateral surface coils in MR
breast imaging: Potential hazards. Radiology
186:906-907.
Gorczyca, D.P., DeBruhl, N.D., Ahn, C.Y., Hoyt, A.,
Sayre, J.W., Nudell, P., McCombs, M., Shaw,
W.W., and Bassett, L.W. 1994. Silicone breast
implant ruptures in an animal model: Compari-
son of mammography, MR imaging, US, and CT.
Radiology 190:227-232.
Graham, R.A., Homer, M.J., Sigler, C.J., Safaii, H.,
Schmid, C.H., Marchant, D.J., and Smith, T.J.
1994. The efficacy of specimen radiography in
evaluating the surgical margins of impalpable
breast carcinoma [see comment in Am. J.
Roentgenol. 1995, 164:766]. Am. J. Roentgenol.
162:33-36.
Gribbestad, I.S., Nilsen, G., Fjosne, H.E.,
Kvinnsland, S., Haugen, O.A., and Rinck, P.A.
1997. Comparative signal intensity measure-
Current Protocols in Magnetic Resonance Imaging
ments in dynamic gadolonium-enhanced MR
mammography. J. Magn. Reson. Imaging 4:477-
480.
Guillemaud, R., Brady, M. 1997. Estimating the bias
field of MR images. IEEE Trans. Med. Imaging
16:238-251.
Hachiya, J., Seki, T., Okada, M., Nitatori, T.,
Korenaga, T., and Furuya, Y. 1991. MR imaging
of the breast with Gd-DTPA enhancement: Com-
parison with mammography and ultrasonogra-
phy. Radiat Med. 9:232-240.
Harms, S.E. 1996. MRI in breast cancer diagnosis
and treatment. Curr. Probl. Diagn. Radiol.
25:193-215.
Harms, S.E. and Flamig, D.P. 1994. Staging of
breast cancer with MR imaging. Magn. Reson.
Imaging Clin. N. Am. 2:573-584.
Harms, S.E., Flamig, D.P., Hesley, K.L., Meiches,
M.D., Jensen, R.A., Evans, W.P., Savino, D.A.,
and Wells, R.V. 1993. MR imaging of the breast
with rotating delivery of excitation off reso-
nance: clinical experience with pathologic cor-
relation. Radiology 187:493-501.
Hess, T., Knopp, M.V., Brix, G., et al. 1993. Phar-
makokinetische Analyse der Kontrastmittelan-
reicherung in der MRT beim Mammakarzinom.
Zentralbl. Radiol. 147:969.
Heywang, S.H., Wolf, A., Pruss, E., Hilbertz, T.,
Eiermann, W., and Permanetter, W. 1989. MR
Imaging of the breast with Gd-DTPA: Use and
limitations. Radiology 171:95-103.
Hochman, M.G., Orel, S.G., Powell, C.M., Schnall,
M.D., Reynolds, C.A., and White, L.N. 1997.
Fibroadenomas: MR imaging appearances with
radiologic-histopathologic correlation. Radiol-
ogy 204:123-129.
Hoffmann, U., Brix, G., Knopp, M.V., Hess, T., and
Lorenz, W.J. 1995. Pharmacokinetic mapping of
the breast: a new method for dynamic MR mam-
mography. Magn. Reson. Med. 33:506-514.
Hylton, N.M. and Frankel, S.D. 1994. Imaging tech-
niques for breast MR imaging. Magn. Reson.
Imaging Clin. N. Am. 2:511-525.
Kaiser, W.A. 1990. Dynamic MR-Breast Imaging
using a double breast coil: A step towards routine
MR examination of the breast. Eur. Frontiers
Radiol. 7:39-68.
Knopp, M.V., Brix, G., Junkermann, H.J., and Sinn,
H.P. 1994. MR mammography with pharmacok-
inetic mapping for monitoring of breast cancer
treatment during neoadjuvant therapy. Magn.
Reson. Imaging Clin. N. Am. 2:633-658.
Knopp, M.V., Hoffmann, U., Brix, G., Hawighorst,
H., Junkermann, H.J., and van Kaick, G. 1995.
Fast MRI contrast medium dynamics for charac-
terization of tumors. Experiences with func-
tional MR-mammography. Radiologe 35:964-
972.
Kuhl, C.K., Kreft, B.P., Hauswirth, A., Elevelt, A.,
Steudel, A., Reiser, M., and Schild, H.H. 1995.
MR mammography at 0.5 and 1.5 T. I. Compari-
son of image quality and sensitivity of MR mam-
mography at 0.5 and 1.5 T. Rofo. Fortschr. Geb. Breast
A21.1.19
 Rontgenstr. Neuen Bildgeb. Verfahr. 162:381-
389.
Kuhl, C.K., Kreft, B.P., Hauswirth, A., Gieseke, J.,
Elevelt, A., Reiser, M., and Schild, H.H. 1995.
MR mammography at 0.5 tesla. II. The capacity
to differentiate malignant and benign lesions in
MR mammography at 0.5 and 1.5 T. Rofo.
Fortschr. Geb. Rontgenstr. Neuen Bildgeb. Ver-
fahr. 162:482-491.
Kuhl, C.K., Seibert, C., Sommer, T., Kreft, B., Gie-
seke, J., and Schild, H.H. 1995. Focal and diffuse
lesions in dynamic MR-mammography of
h ealthy probands. Rofo. Fortschr. Geb.
Rontgenstr. Neuen Bildgeb. Verfahr. 163:219-
224.
Kurtz, B., Achten, C., Audretsch, W., Rezai, M., and
Zocholl, G. 1996. MR mammography of fatty
tissue necrosis. Rofo. Fortschr. Geb. Rontgenstr.
Neuen Bildgeb. Verfahr. 165:359-363.
Melani, E., Sardanelli, F., Ottonello, C., Parodi,
R.C., Castaldi, A., and Succi, S. 1995. Magnetic
resonance mammography in suspected tumor
recurrences. Radiol. Med. Torino 89:219-224.
Morris, E.A., Schwartz, L.H., Dershaw, D.D., van
Zee, K.J., Abramson, A.F., and Liberman, L.
1997. MR imaging of the breast in patients with
occult primary breast carcinoma. Radiology 205
(2):437-440.
Muller, R.D., Barkhausen, J., Sauerwein, W., and
Langer, R. 1998. Assesment of local re currence
after breast-conserving therapy with MRI. J.
Comput. Assist. Tomogr. 22:408-412.
Muller-Schimpfle, M., Ohmenhauser, K., Sand, J.,
Stoll, P., and Claussen, C.D. 1997. Dynamic
3D-MR mammography: is there a benefit of
sophisticated evaluation of enhancement curves
for clinical routine? J. Magn. Reson. Imaging
7:236-240.
Muller-Schimpfle, M., Ohmenhauser, K., Stoll, P.,
Dietz, K., and Claussen, C.D. 1997. Menstrual
cycle and age: UInfluence on parenchymal con-
trast medium enhancement in MR imaging of the
breast. Radiology 203:145-149.
Mumtaz, H., Davidson, T., Spittle, M., Tobias, J.,
Hall-Craggs, M.A., Cowley, G., and Taylor, I.
1996. Breast surgery after neoadjuvant treat-
ment. Is it necessary? Eur. J. Surg. Oncol.
22:335-341.
Mussurakis, S., Buckley, D.L., and Horsman, A.
1997. Prediction of axillary lymph node status in
invasive breast cancer with dynamic contrast-en-
hanced MR imaging. Radiology 203:317-321.
Mussurakis, S., Buckley, D.L., Drew, P.J., Fox, J.N.,
Carleton, P.J., Turnbull, L.W., and Horsman, A.
1997. Dynamic MR imaging of the breast com-
bined with analysis of contrast agent kinetics in
the differentiation of primary breast tumours.
Clin. Radiol. 52:516-526.
Nunes, L.W., Schnall, M.D., Siegelman, E.S., Lan-
glotz, C.P., Orel, S.G., Sullivan, D., Muenz, L.A.,
Reynolds, C.A., and Torosian, M.H. 1997. Diag-
nostic performance characteristics of architec-
MRI of Breast tural features revealed by high spatial-resolution
Lesions
A21.1.20
MR imaging of the breast. Am. J. Roentgenol. Flamig, D.P. 1996. Three-dimensional RODEO rate for benign breast disease. J. Magn. Reson.
169:409-415. breast MR imaging of lesions containing ductal Imaging 4:563-568.
Orel, S.G., Hochman, M.G., Schnall, M.D., carcinoma in situ. Radiology 201:427-432. Van Goethem, M., Van Breusegem, L., Ceulemans,
Reynolds, C., and Sullivan, D.C. 1996. High- Soo, M.S., Kornguth, P.J., Walsh, R., Elenberger, C., R., De Schepper, A., and de Moor, J. 1994.
resolution MR imaging of the breast: Clinical Georgiade, G.S., DeLong, D., and Spritzer, C.E. Infraclinical breast carcinoma. J. Belge Radiol.
context. Radiographics 16:1385-1401. 1997. Intracapsular implant rupture: MR find- 78:6-10.
Orel, S.G., Schnall, M.D., LiVolsi, V.A., and ings of incomplete shell collapse. J. Magn. Weinreb, J.C. and Newstead G. 1995. MR imaging
Troupin, R.H. 1994. Suspicious breast lesions: Reson. Imaging 7:724-730. of the breast. Radiology 196:593-610.
MR imaging with radiologic-pathologic correla- Soo, M.S., Kornguth, P.J., Walsh, R., Elenberger, Weinreb, J.C. and Newstead, G. Controversies in
tion. Radiology 190:485-493. C.D., and Georgiade, G.S. 1996. Complex radial breast MRI. Magn. Reson. Q. 10:67-83.
Orel, S.G., Schnall, M.D., Powell, C.M., Hochman, folds versus subtle signs of intracapsular rupture
of breast implants: MR findings with surgical Zuo, C.S., Jiang, A., Buff, B.L., Mahon, T.G., and
M.G., Solin, L.J., Fowble, B.L., Torosian, M.H., Wong, T.Z. 1996. Automatic motion correction
and Rosato, E.F. 1995. Staging of suspected correlation. Am. J. Roentgenol. 166:1421-1427.
for breast MR imaging. Radiology 198:903-906.
breast cancer: effect of MR imaging and MR- Stack, J.P., Redmond, O.M., Codd, M.B., Dervan,
guided biopsy. Radiology 196:115-122. P.A., and Ennis, J.T. 1990. Breast disease: Tissue
characterization with Gd-DTPA enhancement Internet Resources
Pedevilla, M. 1997. Enhancement velocity not nec- http://www.mediteach.de
essarily reflective of contrast material uptake in profiles. Radiology 174:491-494.
breast MR imaging. Radiology 205:580-581. Stelling, C.B. 1995. Breast cancer staging with con- List of detailed courses in MR Mammography, his-
trast material-enhanced MR imaging: Should it tory of MR Mammography, MRM publications of the
Peller, M., Stehling, M.K., Sittek, H., Kessler, M., author.
and Reiser, M. 1996. Effects of partial volume change patient treatment? Radiology 196:16-18.
and phase shift between fat and water in gradi- Stelling, C.B. 1995. MR imaging of the breast for
http://www.mrisafety.com
ent-echo magnetic resonance mammography. cancer evaluation. Current status and future di-
MAGMA 4:105-113. rections. Radiol. Clin. North Am. 33:1187-1204. List of possible safe or unsafe materials for a quick
overview.
Piccoli, C.W. 1994. The specificity of contrast-en- Tukel, S., Erden, I., Ciftci, E., Kocak, S. 1997.
hanced breast MR imaging. Magn. Reson. Imag- Hydatid cyst of the breast: MR imaging findings.
ing Clin. N. Am. 2:557-571. Am. J. Roentgenol. 168:1386-1387.
Contributed by Werner A. Kaiser
Pierce, W.B., Harms, S.E., Flamig, D.P., Griffey, Turkat, T.J., Klein, B.D., Polan, R.L., and Richman,
R.H., Evans, W.P., and Hagans, J.E. 1991. Three- R.H. 1994. Dynamic MR mammography: A Hospital of the Friedrich-Schiller-University
dimensional gadolinium-enhanced MR imaging technique for potentially reducing the biopsy Jena, Germany
of the breast: pulse sequence with fat suppression
and magnetization transfer contrast. Work in pro-
gress. Radiology 181:757-763.
Rieber, A., Merkle, E., Bohm, W., Brambs, H.J., and
Tomczak, R. 1997. MRI of histologically con-
firmed carcinoma: Clinical relevance of diagnos-
tic procedures for detection of multifocal or con-
tralateral secondary carcinoma. J. Comput. As-
sist. Tomogr. 21:773-779.
Rieber, A., Tomczak, R.J., Mergo, P.J., Wenzel, V.,
Zeitler, H., Brambs, H.J. 1997. MRI of the breast
in the differential diagnosis of mastitis versus
inflammatory carcinoma and follow-up. J. Com-
put. Assist. Tomogr. 21:128-132.
Rieber, A., Zeitler, H., Rosenthal, H., Gorich, J.,
Kreienberg, R., Brambs, H.J., and Tomczak, R.
1997. MRI of breast cancer: influence of chemo-
therapy on sensitivity. Br. J. Radiol. 70:452-458.
Rodenko, G.N., Harms, S.E., Pruneda, J.M., Farrell,
R.S. Jr., Evans, W.P., Copit, D.S., Krakos, P.A.,
and Flamig, D.P. 1996. MR imaging in the man-
agement before surgery of lobular carcinoma of
the breast: Correlation with pathology. Am. J.
Roentgenol. 167:1415-1419.
Santyr, G.E. 1994. MR imaging of the breast. Imag-
ing and tissue characterization without intrave-
nous contrast. Magn. Reson. Imaging Clin. N.
Am. 2:673-690.
Seki, T., Hachiya, J., Nitatori, T., Yokoyama, K.,
Fukushima, H., and Uchigasaki S. 1996. MR
imaging for breast cancer. Nippon Geka Gakkai
Zasshi 97:347-356.
Soderstrom, C.E., Harms, S.E., Copit, D.S., Evans,
W.P., Savino, D.A., Krakos, P.A., Farrell, R.S. Jr, Breast
A21.1.21
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
Silicone Breast Implants
Magnetic resonance imaging is a powerful noninvasive means by which to study the
integrity of silicone breast implants. Silicone has a unique MR resonance frequency and
long T1 and T2 relaxation times that allow for several MR sequences to provide excellent
diagnostic images. The most commonly used sequences include T2-weighted and STIR
(short tau inversion recovery) imaging. The T2-weighted and STIR sequences are com-
monly used in conjunction with water suppression. The following are sequences that can
be used to investigate normal and ruptured silicone breast implants.
IMAGING OF SILICONE BREAST IMPLANTS
An understanding of the MR characteristics of silicone will be helpful to understand why
certain MR sequences are used to differentiate silicone from surrounding breast paren-
chyma. As previously stated, silicone has a unique MR resonance frequency and long T1
and T2 relaxation times that allow for several MR sequences to provide diagnostic images.
The MR characteristics of silicone will be more thoroughly discussed in the Critical
Parameters section.
MR imaging of silicone breast implants is perfomed with a dedicated bilateral breast coil
using a 1.5 T superconducting magnet to help separate the MR signals from silicone, fat,
and water. To image patients with silicone breast implants, 4 MR sequences are used,
which require ∼30 min to complete the examination. Low field strength magnets can be
utilized to image the silicone breast implants, however, the signal separation between
silicone, fat, and water is poor on certain MR sequences.
Table A21.2.1 lists the necessary hardware to perform the procedure. The available
gradient strength will vary from scanner to scanner. The echo times will depend on the
MR system used. A dedicated bilateral breast coil is optimal for scanning, however, a
shoulder coil, round coil, or flex coil can be used on one breast at a time in the supine
position.
Set up patient and equipment
1. Interview (screen) the patient to ensure that he or she has no counterindications such
as cardiac pacemakers or other implants containing ferromagnetic materials. Also,
be sure to find out if the patient has any health conditions that may require the
presence of special emergency equipment during the scanning procedure, or neces-
sitate any other precautions.
Generally, standard screening forms are used for all patients scanned in a magnetic
resonance system. The presence of any ferromagnetic metals may be a health hazard to the
Table A21.2.1 Equipment Parameters Necessary for Imaging of Silicone Breast
Implants
Coil type Dedicated bilateral breast coil (phased-array if
available)
Gradient coil strength Depends on MR scanner
Cardiac gating No
Peripheral gating For safety only
Respiratory gating No
Respirator Patient dependent
Oxygen Patient dependent
Use of contrast agent No
Contributed by David P. Gorczyca
Current Protocols in Magnetic Resonance Imaging (2001) A21.2.1-A21.2.15
Copyright © 2001 by John Wiley & Sons, Inc.
UNIT A21.2 patient when he or she is inside the magnet, and will also affect the imaging. If in doubt as
to the exact composition of the items, it is best to exclude patients with any metal implants;
see Shellock and Kanal (1996) for discussion of what implants may be safely scanned using
magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing.
2. If this is a research protocol, the patient should sign any necessary forms.
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
BASIC
PROTOCOL 4. Have the patient wash off any mascara and other makeup to avoid local tissue heating
and image artifacts.
5. Inform the patient about what will occur during the procedure, what he or she will
experience while in the magnet, and how to behave, including the following:
a. If earphones or headphones are used to protect the ears from the loud sounds
produced by the gradients, the patient will be asked to wear these, but will be able
to communicate with you at any time during the imaging.
b. The patient will be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this works).
c. For good results, the patient should not talk, and should avoid or minimize
swallowing or other movement, during each scan—i.e., as long as the banging
sounds continue.
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
6. The patient should be told to remove all clothes from the waist up and change into a
gown with the opening in front.
No contrast agent will be used for this MR examination so intraveneous access is not
required.
7. Have the patient mount onto the table. Either before or right after the patient lies
down, set up any triggering devices or other monitoring equipment that is to be used.
8. Then ask the patient to lie prone on the dedicated bilateral breast coil. The gown
should be opened in the front so the breast can be accurately placed in the breast coil.
9. Several pillows should be placed under the patient’s head. Have the patient’s arms
placed parallel to their body. If needed, place a pillow or other support under the knees
to make the patient more comfortable.
10. Use the centering light and center the patient’s nipples in the coil.
If the MR technologist is male, a female employee should be present at all times during set
up.
Once this step has been performed, so long as the patient does not move on the table, the
table itself can be moved and then replaced in the same position as before without
jeopardizing the positioning of one scan relative to another.
11. If the patient is unable to hold still, provide an appropriate sedative.
12. The following 4 sequences consist of the preferred protocol, (1) a transverse scout
sequence, (2) a sagittal T2-weighted sequence with water suppression, (3) a transverse
Silicone Breast T2-weighted sequence, and (4) a transverse STIR sequence with water suppression.
Breast Implants A GE signa 1.5 T scanner was used for this protocol.
A21.2.1 A21.2.2
Current Protocols in Magnetic Resonance Imaging
Sequence 1: Rapid transverse positioning pilot 15. Use the pilot images to locate the breast.
13. To validate the patient’s position, run the system’s pilot (or scout) scan to ensure
The image slices for this sequence should only cover the silicone implants in order to obtain
correct location of the breast, using the imaging sequence given in Table A21.2.2 or thinner slices in less time. This sequence is not designed to evaluate the entire breast
similar parameters. parenchyma for breast tumors.
Sequence 2: Sagittal T2-weighted fast spin echo with water suppression 16. Run the scan.
14. Bring the sequence for the sagittal T2-weighted fast spin echo with water suppression
up onto the console. Set the imaging parameters as shown in Table A21.2.3. Sequence 3: Transverse T2 weighted fast spin echo
17. Bring the sequence for the sagittal T2-weighted fast spin echo up onto the console.
Table A21.2.2 Primary Clinical Imaging Parameters for Scout Sequence 1
Set the imaging parameters as shown in Table A21.2.4.
(Pilot Scan)
18. Use the pilot images to locate the breast.
Patient position Prone The image slices for this sequence should only cover the silicone implants in order to obtain
Scan type Gradient echo thinner slices in less time. This sequence is not designed to evaluate the entire breast
Imaging plane (orientation) Transverse parenchyma for breast tumors.
Central slice or volume center Laser light centered on nipple of
19. Run the scan.
breast
Echo time (TE) As short as possible
Sequence 4: Transverse inversion recovery fast spin echo sequence with water
Repeat time (TR) As short as possible
suppression
Flip angle (FA) 15°
20. Bring the sequence for the transverse T2-weighted fast spin echo up onto the console.
Fields of view (FOVx, FOVy) 360 mm, 360 mm
Set the imaging parameters as shown in Table A21.2.5.
Resolution (Δx, Δy) 1.41 mm, 2.81 mm
Number of data points collected (Nx, Ny) 256, 128 21. Use the pilot images to locate the breast.
Slice thickness (Δz) 10 mm
The image slices for this sequence should only cover the silicone implants in order to obtain
Number of slices 10 thinner slices in less time. This sequence is not designed to evaluate the entire breast
Slice gap 5 mm parenchyma for breast tumors.
Number of acquisitions (Nacq) 1
Swap read and phase encoding No This sequence provides more robust signal separation between silicone, fat, and water. The
fat and water signal will be suppressed making the silicone signal relatively high (silicone
Scan time ∼30 sec

Table A21.2.3 Primary Clinical Imaging Parameters for Sequence 2 Table A21.2.4 Primary Clinical Imaging Parameters for Sequence 3
(T2-Weighted Fast Spin Echo With Water Suppression) (T2-Weighted Fast Spin Echo)

Patient position Prone Patient position Prone

Scan type Fast spin echo Scan type Fast spin echo
Imaging plane (orientation) Sagittal Imaging plane (orientation) Transverse
Central slice or volume center Image silicone implants only Central slice or volume center Image silicone implants only
Echo time (TE) 190–220 msec Echo time (TE) 190–220 msec
Echo train length (ETL) 8–16 Echo train length (ETL) 8–16
Repeat time (TR) 3000–5000 msec Repeat time (TR) 3000–5000 msec
Flip angle (FA) 90° Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 200 mm, 320 mm (depends on Fields of view (FOVx, FOVy) 320 mm, 320 mm (depends on
patient) patient)
Resolution (Δx, Δy) 0.78 mm, 1.25 mm Resolution (Δx, Δy) 1.25 mm, 1.25 mm
Number of data points collected (Nx, Ny) 256, 256 Number of data points collected (Nx, Ny) 256, 256
Slice thickness (Δz) 3 mm Slice thickness (Δz) 3 mm
Number of slices 30–38 (depends on patient) Number of slices 30–38 (depends on patient)
Slice gap 1 mm Slice gap 1 mm
Number of excitations (NEX) 2 Number of excitations (NEX) 2
Number of acquisitions (Nacq) 2 Number of acquisitions (Nacq) 2
Swap read and phase encoding No Swap read and phase encoding No
Water suppression Yes Water suppression No
Scan time ∼4 min Silicone Breast Scan time ∼4 min
Breast Implants

A21.2.3 A21.2.4
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A21.2.5 Primary Clinical Imaging Parameters for Sequence 4 (Inversion sults (Letterman and Schurter, 1989; Steinbach brous or calcific contracture, localized pain,
Recovery Fast Spin Echo Sequence) et al., 1993). pareshtesias, and possibly even generalized
autoimmune disorders. In 1992, the U.S. Food
Patient position Prone Injectable materials and Drug Administration (FDA) held hearings
Scan type Inversion recovery fast spin echo During the last century, many different ma- about potential complications resulting from
Imaging plane (orientation) Transverse terials have been injected into the breast in the use of silicone gel implants. As a result, the
Central slice or volume center Image silicone implants only attempts to augment the breast. One of the first FDA announced it would allow the use of
Echo time (TE) 190–220 msec materials injected was paraffin. Several other silicone gel implants only under special condi-
Echo train length (ETL) 8–16 materials were also injected into the breast, but tions (Kessler, 1992).
Repeat time (TR) 3000–5000 msec most, if not all of the materials injected even- Because of concerns about the potential dan-
Inversion time (TI) 150 msec tually resulted in a poor cosmetic outcome gers of rupture and leakage of silicone-gel im-
secondary to complications. Complications in- plants, radiologists are often requested to evalu-
Flip angle (FA) 180°
cluded granulomatous reactions, inflammatory ate the integrity of breast prostheses. Mammog-
Fields of view (FOVx, FOVy) 320 mm, 320 mm (depends on
reactions, necrosis, pulmonary embolism, and raphy, sonography, and MR imaging have been
patient)
death. Direct silicone injections started in the used to evaluate the integrity of breast implants
Resolution (Δx, Δy) 1.25 mm, 1.25 mm 1950’s, again most with eventual complications (Ahn et al., 1993; Gorczyca et al., 1992; De-
Number of data points collected (Nx, Ny) 256, 256 similar to those identified with paraffin injec- Bruhl et al., 1993; Destouet et al., 1992; Eklund
Slice thickness (Δz) 3 mm tions (Letterman and Schurter, 1989; Steinbach et al., 1988; Harris et al., 1993; Sinha et al.,
Number of slices 30–38 (depends on patient) et al., 1993). 1993). When compared with other imaging
Slice gap 1 mm techniques, MR imaging appears to be the most
Number of excitations (NEX) 2 Implantable prostheses accurate method for evaluating the integrity of
Number of acquisitions (Nacq) 2 In the 1950’s, synthetic sponge prostheses, breast implants (Gorczyca et al., 1994a,b;
Swap read and phase encoding No composed of polyvinyl alcohol or Ivalon, were Mund et al., 1993; Gorczyca and Brenner,
Water suppression Yes implanted in the breast. Initial reports were 1997).
Scan time ∼5 min promising, however, scar tissue quickly made
these prostheses hard and caused them to Critical Parameters
shrink. Other synthetic materials were used An understanding of the MR characteristics
such as etheron, polyether polyurethane, of silicone will be helpful to understand the
only sequence). This is a good sequence if one is looking for free silicone in the breast polypropylene, and even polytef (Teflon), all different MR sequences that can be used to
parenchyma. This sequence can only be performed consistently on a 1.5 T machine. resulting in complications or poor cosmetic differentiate silicone from surrounding breast
results. A new approach with promising cos- parenchyma. The chemical composition of
22. Run the scan. metic results and less complications was re- most medical grade silicones is dimethyl
ported by Cronin and Gerow in 1963 when they polysiloxane with varying degrees of polym-
Process and view the data described their use of a silicone gel prosthesis erization (Fig. A21.2.1; Habal, 1984). The MR
23. To better view the silicone implant images, adjust window level settings to see the (Cronin and Gerow, 1964). signal is derived from the protons of the methyl
interior characteristics of the implants. Several hundred different variations of sili- groups. The implant shell (envelope) is also
cone gel prostheses have been made commer- composed of silicone but differs from the gel
These MR sequences are optimized to evaluate silicone implants, not breast parenchyma
cially available over the past 40 years. Breast because of the many additional cross linkages
abnormalities. Therefore, the window level settings should be set to optimize the visuali-
zation of the silicone breast implants even at the expense of visualizing contrast in the breast implants have been placed in 1 to 2 million between the methyl groups that result in an
parenchyma. For the transverse breast images, have the technologist flip the images 180° women for augmentation mammoplasty or re- elastic solid. Although the implant shell is
prior to printing in order to have the images in radiologic anatomic position. construction after mastectomies. Despite the composed of silicone, only minimal MR signal
initial enthusiasm, complications related to the is produced from the silicone shell because of
silicone-gel implants were reported. These the many additional methyl group cross link-
COMMENTARY complications include rupture and leakage, fi- ages.
Background Information Autogenous tissue transplantation
One of the first surgical uses of autogenous
History of breast augmentation tissue to correct a breast defect was performed
During the last century, many different by Czerny in 1895. He removed a lipoma from CH3 CH3 CH3
methods have been tried to augment or recon- a patient’s thigh to fill out a breast defect. In the
struct the breast. Unfortunately, the majority of mid 1920’s, free fat grafts were used for breast Si O Si O Si O
approaches were disappointing with many as- augmentation; however, with time, liquefaction
sociated complications. The methods used to or absorption of the free fat graft diminished
augment or reconstruct the breast can be placed the cosmetic results. Abdominal fat flaps were CH3 CH3 CH3
into one of three categories, autogenous tissue first described in the mid 1950’s and are still x
transplantation, injectable materials, and im- used today, many with excellent cosmetic re-
plantable prostheses. Silicone Breast
Breast Implants Figure A21.2.1 Chemical composition of silicone (dimethyl polysiloxane).

A21.2.5 A21.2.6
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 fat silicone
A
water

220 Hz 100 Mz

Figure A21.2.2 Relative resonance frequency differences between water, fat, and silicone at 1.5
T. The resonance frequency of silicone is ∼320 Hz lower than that of water and 100 Hz lower than
that of fat.

Table A.21.2.6 Relative Signal Intensities of Silicone, Fat, and Water

MR Sequence Silicone Fat Water

Fast spin echo (FSE) T2-weighted (TR ≅ 5000 High Medium Very high
msec, TE ≅ 200 msec)
FSE with water suppression High Medium Low
Inversion recovery FSE (IRFSE) (TR ≅ 5000 High Low Very high
msec, TE ≅ 200 msec, TI = 150 msec)
IRFSE with water suppression High Low Low

The selection of MR pulse sequences used Troubleshooting

to image breast implants is determined by the
relative Larmor precessional frequencies, as
well as the T1 and T2 properties of the tissues
(fat, muscle, and silicone). The relative reso-
nance frequency of silicone is ∼100 Hz lower
than fat and 320 Hz lower than water at 1.5 T
(Fig. A21.2.2). Since the resonance frequency
of silicone is close to fat, when chemical sup-
pression techniques (fat or water suppression)
are used, the MR signal from silicone behaves
similar to fat. As a result, the silicone signal is
high when water suppression is used (Table
A21.2.6) and the silicone signal is low when fat
suppression is used.
The relative relaxation times of silicone, fat,
and water can also be used to obtain MR images
that selectively emphasize the signal from sili-
cone. The relaxation times of fat are shorter
than those of silicone; therefore, one can use
the relaxation time properties of silicone and
fat to suppress the fat while maintaining a
strong signal from silicone. The use of inver-
sion recovery with a short TI (STIR) will sup-
press the signal from fat while maintaining
signal from silicone. To obtain a more selective
silicone image, a water suppression pulse can
be used in conjunction with an STIR sequence
to produce a silicone selective sequence.
Failure of water suppression
Occasionally, auto prescan will fail when a
water suppression sequence is being per-
formed. To correct this, go into manual prescan
B
(one should see a wave form similar to Fig.
A21.2.2—occasionally the fat and silicone
peak will merge into one flat peak), adjust to
the fat peak, save the frequency, and scan. One
helpful suggestion to individuals just starting
to image silicone implants is to place a small
bag of saline and a small silicone implant in the
breast coil when scanning the patient. The sa-
line and silicone bag act as controls so that the
radiologist can easily see on the images how
the saline and silicone are behaving on the MR
sequence being used.
Figure A21.2.3 A 34-year-old woman with normal subpectoral single-lumen silicone implants. (A)
Motion artifact Sagittal T2-weighted FSE with water suppression. Pectoralis major muscle = curved arrow. (B)
Do not confuse motion, cardiac, or respira- Transverse T2-weighted FSE demonstrates a normal signal-lumen subpectoral silicone implant with
tory artifacts with the collapsed silicone im- normal radial folds (arrows) of the silicone shell extending to the periphery of the implant. These
plant shell. Most artifacts will extend beyond folds are not indicative of rupture or leak.
the confines of the breast implants.
Window level
It is extremely important to adjust the win-
dow level setting to clearly see inside the sili-
cone breast implants. By adjusting the window Silicone Breast
level, the radiologist will be able to see the Breast Implants

A21.2.7 A21.2.8
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 A A

Figure A21.2.4 A 48-year-old woman with normal subglandular double-lumen silicone implants.
(A) Sagittal T2-weighted FSE with water suppression. Pectoralis major muscle = curved arrow. (B)
Transverse T2-weighted FSE. The very high signal intensity (arrows) surrounding the lower signal
silicone on the transverse image (B) represents the saline-filled outer lumen. The saline outer lumen
is of low-signal-intensity (straight arrows) on the sagittal T2-weighted FSE with water suppression
(A).
Silicone Breast
Breast
A21.2.9
Current Protocols in Magnetic Resonance Imaging
Figure A21.2.5 Intracapsular rupture. (A) Unlike an intact implant (left), in early intracapsular
ruptures (center), silicone gel surrounds implant shell, but is contained by the fibrous capsule. Later,
the collapsed implant shell floats within the silicone gel (linguine sign, right). Light gray line = fibrous
capsule, black line = implant shell. (B) Patient presented with pain in both breasts. Sagittal
Implants T2-weighted image shows multiple curvilinear low-signal-intensity lines within the left implant known
as the linguine sign (arrows). Intracapsular rupture of the left implant was found at surgery.
A21.2.10
Current Protocols in Magnetic Resonance Imaging
 collapsed implant shell or normal radial folds.
These MR sequences are optimized to evaluate
A silicone implants, not breast parenchymal ab-
normalities, therefore, concentrate on window-
ing on the silicone implants, not the breast
tissue.
Saline implants
Radiologists new to MR imaging of silicone
implants can become easily confused by saline
implants if they are imaged by MRI. If a breast
implant is completely dark when a water sup-
pression sequence is performed, the implant is
not a silicone implant—it is a saline implant. If
the radiologist is not sure, a bag of saline can
be placed in the coil and compared with the
breast implant. Furthermore, saline implants do
not have to be imaged by MR imaging because
when saline implants rupture, the saline is ab-
sorbed by the body. Ruptured saline implants
can be easily detected by mammography.
B figuration commonly known as stacked im-
Anticipated Results
Appearance of normal implants
Several hundred different types of breast
implants were commercially available over the
past 40 years (Middleton and McNamara,
2000). Most silicone implants are composed of
an outer silastic elastomer shell (envelope),
filled with viscous silicone gel. The implants
are usually oval and have a smooth or textured
surface. After surgical placement, a thin fibrous
capsule (scar tissue) normally forms around the
prosthesis. MR imaging often shows radial
folds, normal infoldings of the silastic elas-
tomer shell (Fig. A21.2.3). These folds may
appear prominent, but should not be confused
with implant rupture or leak. Radial folds, even
when prominent, extend to the periphery of the
implant and the folds are relatively few in
number.
Although hundreds of different types of sili-
cone implants have been produced, single-lu-
men (Fig. A21.2.3) and double-lumen silicone
implants (Fig. A21.2.4) are the most commonly
encountered. A single-lumen silicone implant
has an outer silastic shell containing the viscous
silicone gel. A double-lumen silicone implant
typically has an inner lumen that contains the
thick viscous silicone gel surrounded by a
smaller outer lumen that contains saline. Breast
implants may be surgically placed in a subglan-
dular location, that is, anterior to the pectoralis
major muscle, or subpectoral, posterior to the
pectoralis major muscle (Fig. A21.2.3 and Fig.
A21.2.4).
A variety of other types of implantable pros-
theses are occasionally encountered, including
expander implants (reverse double lumens—
saline in the inner lumen, silicone in the outer
lumen), multicompartmental implants, foam
implants, and single lumen silicone implants
with saline directly injected into the silicone at
the time of surgery. Occasionally, two or even
more implants are placed in one breast, a con-
plants. Some implants have a coating of poly-
urethane covering the surface of the silicone
envelope. These implants typically have a mod-
erate to large amount of reactive fluid surround-
ing the implant. The many different implant-
able prosthesis can be easily differentiated from
direct silicone injections into the breast. Saline
implants, where the lumen is filled with saline
have become more popular since the FDA lim-
ited the use of silicone gel implants in 1992.
Appearance of ruptured breast implants
Implant ruptures can be divided into two
major categories, intracapsular and extracapsu-
lar rupture.
Intracapsular implant rupture, the most
common type of rupture, is defined as rupture
of the implant shell (elastomer envelope) with
silicone leakage that does not macroscopically
extend beyond the fibrous capsule that com-
monly forms around silicone implants (Gor-
czyca et al., 1992). The most reliable MR cri-
terion for intracapsular rupture is the presence

Figure A21.2.6 Extracapsular rupture. (A) Intact implant has uninterrupted shell and fibrous
capsule (left). Disruption of shell and fibrous capsule (center) will allow silicone to extravasate into
surrounding breast tissue (right). Light gray line = fibrous capsule, black line = implant shell. (B)
Patient presented with right breast pain. The sagittal T2-weighted image demonstrates extracapsu-
lar silicone (straight white arrow) and collapse of the silicone shell (straight black arrow), M =
pectoralis muscle.
Silicone Breast
Breast
Figure A21.2.7 (at right) Focal rupture without complete collapse (uncollapsed implant rupture)
or extensive gel bleed of implant shell. (A) Normal infolding of implant shell causes radial folds (left).
Gel bleed is microscopic silicone leakage through an intact implant shell (center). Silicone may
enter a radial fold, resulting in an inverted teardrop sign (right). A focal tear in the implant shell
without complete collapse of the implant shell (uncollapsed implant rupture) can have an identical
appearance. Light gray line = fibrous capsule, black line = implant shell. (B) Focal rupture of implant
shell without complete collapse of implant shell (uncollapsed implant rutpure). Sagittal T2-weighted
FSE image shows inverted teardrop signs (straight arrow), indicating silicone within a radial fold
and outside of the implant lumen itself. At surgery, a small tear within the implant shell was found.
The inverted teardrop sign is not specific, this sign has been seen with both focal ruptures and
Implants extensive gel bleeds with intact implant shells.

A21.2.11 A21.2.12
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 of multiple curvilinear low-signal-intensity
lines seen within the high-signal-intensity sili-
A cone gel, the so-called linguine sign (Fig.
A21.2.5). These curvilinear lines represent the
collapsed implant shell floating within the sili-
cone gel. Rarely, intracapsular rupture will
show multiple hyperintense foci on T2-
weighted images or multiple hypointense foci
on water-suppression images within the im-
plant lumen. When less than six foci of water
droplets are identified within a silicone implant
without other evidence of rupture, one must be
careful not to definitely diagnose a ruptured
implant. The authors have three cases that all
showed several small water droplets within the
implants (no other MR findings were noted to
suggest rupture) and all proved to be intact at
the time of surgery.
Extracapsular silicone implant ruptures are
defined as ruptures of both the implant shell
and the fibrous capsule with macroscopic sili-
B cone leakage that extends beyond the fibrous
capsule into surrounding tissues. Focal areas of
high signal intensity, representing free silicone,
can be identified on MR images (Fig. A21.2.6).
In addition to free silicone in the surrounding
breast parenchyma, the linguine sign is often
present with extracapsular ruptures. The mul-
tiplanar capabilities of MR imaging allow pre-
cise localization of free silicone.
Unlike ruptures, gel bleed is microscopic
silicone leakage through an intact implant shell
(Brody, 1977). Most if not all implants will
eventually have gel bleed, however, the major-
ity of gel bleeds cannot be detected by MR
imaging. Only when gel bleed is extensive can
silicone gel be detected outside the silicone
shell. A focal or early intracapsular rupture can
have a similar appearance to a large gel bleed
and it can be difficult if not impossible to
differentiate these two entities on MR images
(Fig. A21.2.7).
In the authors’ experience, other signs, such
as focal or diffuse irregularity of the contour of
the implant or reactive fluid surrounding the
implant, are not reliable signs of silicone im-
plant rupture.
Literature Cited
Ahn, C.Y., Shaw, W.W., Narayanan, K., Gorczyca,
D.P., Sinha, S., DeBruhl, N.D., and Bassett, L.W.
1993. Definitive diagnosis of breast implant rup-
Figure A21.2.7 (See legend on facing page.) ture using magnetic resonance imaging. Plast.
Reconstr. Surg. 92:681-691.
Brody, G.S. 1977. Fact and fiction about breast
implant “bleed.” Plast. Reconstr. Surg. 60:615-
Silicone Breast 616.
Breast Implants
A21.2.13 A21.2.14
Current Protocols in Magnetic Resonance Imaging
Cronin, T.D. and Gerow, F. 1964. Augmentation
mammoplasty: A new “natural feel” prosthesis.
In: Transactions of the Third International Con-
gress of Plastic Surgeons. Exerpta Medica, Am-
sterdam.
DeBruhl, N.D., Gorczyca, D.P., Ahn, C.Y., Shaw,
W.W., and Bassett, L.W. 1993. Silicone breast
implants: US evaluation.Radiology 189:95-98.
Destouet, J.M., Monsees, B.S., Oser, R.F., Neme-
cek, J.R., Young, V.L., and Pilgram, T.K. 1992.
Screening mammography in 350 women with
breast implants: Prevelence and findings of im-
plant complications. Am. J. Roentgenol.
159:973-978.
Eklund, G.W., Busby, R.C., Miller, S.H., and Job,
J.S. 1988. Improved imaging of the augmented
breast. Am. J. Roentgenol. 151:469-473.
Gorczyca, D.P. and Brenner, R.J., 1997. The Aug-
mented Breast: Radiologic & Clinical Perspec-
tives. Stuttgart, Thieme, New York.
Gorczyca, D.P., Sinha, S., Ahn, C.Y., DeBruhl, N.D.,
Hayes, M.K., Gausche, V.R., Shaw, W.W., and
Bassett, L.W. 1992. Silicone breast implants in
vivo: MR Imaging. Radiology 185:407-410.
Gorczyca, D.P., DeBruhl, N.D., Ahn, C.Y., Hoyt, A.,
Sayre, J.W., Nudell, P., McCombs, M., Shaw,
W.W., and Bassett, L.W. 1994a. Silicone breast
implant ruptures in an animal model: Compari-
son of mammography, MR imaging, US, and CT.
Radiology 190:227-232.
Gorczyca, D.P., Schneider, E., DeBruhl, N.D., Foo,
T.K., Ahn, C.Y., Sayre, J.W., Shaw, W.W., and
Bassett, L.W. 1994b. Silicone breast implant rup-
ture: Comparison between three-point Dixon
and fast spin-echo MR imaging. Am. J.
Roentgenol. 162:305-310.
Habal, M.B. 1984. The biologic basis for the clinical
application of the silicones. Arch. Surg. 119:843-
848.
Harris, K.M., Ganott, M.A., Shestak, K.C., Losken,
H.W., and Tobon, H. 1993. Silicone implant rup-
ture: Detection with US. Radiology 187:761-
768.
Kessler, D.A. 1992. The basis of the FDA’s decision
on breast implants. N. Engl. J. Med. 326:1713-
1715.
Letterman, G. and Schurter, M. 1989. History of
aesthetic breast surgery. In: The Art of Aesthetic
Plastic Surgery, Vol 1. (J.R. Lewis, ed.) pp. 21-
27. Little, Brown, Boston, Massachusettes.
Middleton, M.S. and McNamara, M.P., Jr. 2000.
Breast Implant Classification with MR imaging
correlation. Radiographics 20:E1.
Mund, D.F., Farria, D.M., Gorczyca, D.P., DeBruhl,
N.D., Ahn, C.Y., Shaw, W.W., and Bassett, L.W.
1993. MR imaging of the breast in patients with
silicone-gel implants: Spectrum of findings. Am.
J. Roentgenol. 161:773-778.
Shellock, F.G. and Kanal, E. 1996. Magnetic Reso-
nance Bioeffects, Safety and Patient Manage-
ment. Lippincott Raven, Philadelphia.
Current Protocols in Magnetic Resonance Imaging
Sinha, S., Gorczyca, D.P., DeBruhl, N.D., Shellock,
F.G., Gausche, V.R., and Bassett, L.W. 1993. MR
imaging of silicone breast implants: Comparison
of different coil arrays. Radiology 187:284-286
Steinbach, B.G., Hardt, N.S., Abbitt, P.L., Lanier, L.,
and Caffee, H.H. 1993. Breast implants, com-
mon complications, and concurrent breast dis-
ease. Radiographics 13:95-118.
Key References
Gorczyca and Brenner, 1997. See above.
A comprehensive reference text written for the clini-
cian and radiologist covering the history of silicone
breast implants, classification, radiologic imaging,
Current Protocols in Magnetic Resonance Imaging
pathology, immunologic, and medical legal aspects
of silicone breast implants.
Shellock and Kanal, 1996. See above.
Covers a number of important patient management
issues related to MR imaging, including recom-
mended safety procedures, a list of metallic implants
that have been tested for MR compatibility, and a
list of other sources on MR safety.
Contributed by David P. Gorczyca
Sunrise Mountain View Hospital
Las Vegas, Nevada
Breast
A21.2.15
CHAPTER A22
Shoulder
INTRODUCTION
he second most frequently and widely performed musculoskeletal extremity MRI
T examination involves the shoulder. Patients presenting with shoulder pain and dys-
function may have many derangements which can potentially coexist and can frequently
overlap in clinical presentation. When physical examination is insufficient for diagnosis,
or when additional information defining structural lesions is desired to assist the treating
physician in planning medical or operative management, MRI can present a comprehen-
sive examination of osseous and soft tissue structures about the shoulder. The two most
frequently suspected sources of shoulder derangement involve rotator cuff disease and
instability. The standard shoulder MRI protocol should evaluate both of these possibilities
because of the very real possibility and frequency of overlap and coexistence of these
problems in any given patient. In UNIT A22.1, the authors present their standard shoulder
MRI examination, written to address possible cuff disease; however, this protocol also
presents a very comprehensive look at possible instability lesions as well as other sources
of shoulder problems.
MR arthrography of the shoulder has been fairly widely presented in the literature, and
has been proposed to increase the sensitivity and accuracy of examiner evaluation of the
shoulder MRI exam. This modification of shoulder MRI has not only been reported to
improve evaluation of possible instability lesions, but also for cuff undersurface or artic-
ular margin partial tears, although not all users or authors believe MR arthrography is
always necessary for evaluating these possible lesions. The procedure does have its draw-
backs, including the transition to an invasive procedure and necessitating more complex,
time-consuming, and possibly disruptive logistics of scheduling between the MR and
fluoroscopy suites. Furthermore, patient acceptance may not be universal. However, it is
an additional tool for shoulder MRI that all users may find helpful in specific settings. In
UNIT A22.2, the authors present their modified sequence protocol for MR arthrography, writ-
ten to address possible instability; however, the resulting examination is not and should
not be limited to evaluating this single possible entity.
Charles P. Ho
Shoulder
Contributed by Charles P. Ho A22.0.1
Current Protocols in Magnetic Resonance Imaging (2005) A22.0.1
C 2005 by John Wiley & Sons, Inc.
Copyright  Supplement 13
Rotator Cuff Disease UNIT A22.1 Table A22.1.1 Equipment Parameters for MRI of the Rotator Cuff

Coil type Phased array shoulder or standard

MR imaging has revolutionized the evaluation of the glenohumeral joint and has emerged shoulder coil
as an accurate, noninvasive means to assess disorders of the shoulder. Its multiplanar Gradient coil strength 22 mT/m
imaging capability and superior soft tissue contrast offer information not readily provided Cardiac gating No
by other imaging techniques allowing both detection and characterization of pathology. Pheripheral gating No
Respitatory gating No
Shoulder pain is a common clinical presentation. Though the spectrum of disorders
Oxygen No
affecting the shoulder is quite diverse; emphasis is often placed on abnormalities of the
Motion cushions Useful
rotator cuff and glenohumeral joint instability. The rotator cuff plays an important role in
Use of contrast agents No
both the function and stability of the shoulder. Moreover, its pathology, with early and
accurate diagnosis, is amenable to treatment, which can markedly affect patient outcome.
excellent fat-suppression and are extremely sensitive to free water protons. The entire
This unit will present an MR imaging protocol for evaluation of the glenohumeral joint. protocol can be performed in 45 min.
Though its specific emphasis is on the rotator cuff, it can certainly be used for a global
assessment of the shoulder. Table A22.1.1 lists the hardware and parameters needed to perform the examination.
NOTE: Be sure that technologists and nurses have immediate access to any emergency
IMAGING OF THE ROTATOR CUFF BASIC equipment that may be relevant to a given study, or that may be needed for a particular
PROTOCOL patient, such as crash carts or oxygen.
MR imaging scans of the glenohumeral joint can be attained at a variety of field strengths
with a variety of magnet types. Suffice it to say that the primary limitation of low-field-
strength MR imaging is a decreased signal-to-noise ratio which can be increased by using Set up patient and equipment
lower spatial resolution and longer acquisition times. The sequences described here are 1. Initial screening of the patient is achieved by completion of a standardized form
based on a 1.5 T magnet, and rely on fat saturation in many cases. In low- and mid-field designed to ensure that he or she has no internal ferromagnetic materials. In our
strength scanners, these sequences can be replaced by inversion recovery or T2-weighted screening process, specific reference is made to the following items, which can prove
sequences. to be a health hazard to the patient or interfere with image acquisition.

Sequence options for musculoskeletal applications include conventional spin-echo, fast a. Cardiac pacemaker
spin-echo (FSE), gradient echo, short tau inversion recovery (STIR), and fast STIR. This b. Retained metal fragments in eyes
imaging protocol will emphasize the use of fast spin-echo imaging techniques, also known c. Heart valve replacement, venous umbrella
as turbo spin-echo. This adaptation of the rapid acquisition with relaxation enhancement d. Vascular clips
(RARE) technique (Hennig et al., 1986) revolutionized MR imaging, allowing shortened
e. Prosthetic devices in the eyes and joints
acquisition times, increased spatial resolution, or improved signal-to-noise ratio. Though
contrast is similar between conventional-spin echo and fast-spin echo, there are important f. Hearing aid, neurostimulator, insulin pump
differences. One such difference is the bright fat signal seen with fast-spin echo imaging, g. Intrauterine device (I.U.D.)
which can obscure areas of pathology adjacent to fat on T2-weighted sequences. For this h. Shunts/stents (ventricular, spinal, biliary), metal mesh/coil implant
reason, frequency-selective fat saturation is often used with fast-spin echo proton density
i. Orthopedic hardware
and T2-weighted sequences. Fat-saturated fast-spin echo T2-weighted imaging for detec-
tion of rotator cuff pathology has proven quite successful (Quinn et al., 1995; Reinus et Questions regarding safe scanning of implants can be researched in Shellock (1996).
al., 1995; Singson et al., 1996) and will also form an important part of the imaging protocol 2. In addition, the screening form includes general inquiries regarding health issues that
presented in this unit. are pertinent to the performance of the MR imaging study, the need for any emergency
equipment (crash cart, oxygen) and its interpretation.
Routine MR imaging of the shoulder is usually performed in three different planes.
Though there is some degree of redundancy in the anatomic information provided, certain a. Pregnancy status of patient
features are optimally evaluated in specific planes and it is often desirable to evaluate b. Respiratory difficulties or nausea when lying in supine position
complex structures in more than one plane. The imaging protocol reflects this, keeping
c. Past medical history and current medications
in mind that the hallmark of rotator cuff tears is abnormally high signal intensity in the
tendon on T2-weighted images, which should be constant between imaging planes. The d. History of claustrophobia
sequences that comprise the protocol include: transverse fat-suppressed FSE proton 3. In conjunction with the standardized screening form, ask the patient to complete a
density weighted, sagittal oblique FSE T1-weighted, sagittal fat-suppressed FSE T2- background information form. The questionnaire includes:
weighted, coronal oblique FSE T1-weighted, coronal oblique fat-suppressed FSE proton
density weighted MR images, and coronal oblique STIR MR images, which provide a. Requesting physician and subspecialty
Rotator Cuff b. Origin and progression of symptoms
Shoulder Disease
Contributed by Leopoldo M. Gigena, Christine B. Chung, and Donald Resnick A22.1.1 A22.1.2
Current Protocols in Magnetic Resonance Imaging (2001) A22.1.1-A22.1.11
Copyright © 2001 by John Wiley & Sons, Inc. Current Protocols in Magnetic Resonance Imaging
 c. Request for diagram of the symptomatic area Sequence 2: Transverse fat-suppressed fast spin echo (FSE) proton density
d. Activities that exacerbate symptoms 13. After the grid lines are set, changes in the sequence prescription can still be made.
This proves to be an invaluable source of history and clinical information given the paucity
Check Table A22.1.3 to see that the parameters have been correctly entered.
of information often provided on the requisition forms. The information that the patient 14. Let the patient know that you will begin the scan. Run sequence 2.
provides may help the technician to do a more specific study and the radiologist to render
the most accurate interpretation of the exam. 15. After the images have been acquired, if they are satisfactory in appearance, plan both
the sagittal and coronal oblique imaging planes from these transverse images.
4. Review data sheets and interview the patient to ensure he or she understands the nature
of all questions and that the data sheets have been filled out as completely and
accurately as possible.
5. Have the patient sign any necessary consent forms.
6. Instruct the patient to remove all jewelry or metal from their body, and change into
a gown.
7. Explain what will happen during the course of the procedure and what the patient
will experience while in the magnet, including the following.
a. The patient will be given earphones to wear during the examination to protect from
the loud knocking sound produced during the study. This knocking will occur
when images are being acquired, take place 6 to 7 times, and last a few minutes
each time. The technologist and patient may communicate at any time during the
study by simply speaking out, though it is preferable to wait until the knocking
sound has stopped.
b. The patient is asked to remain quiet and avoid any movement during the time
images are being acquired—i.e., when the knocking sound is occurring.
c. The patient will be provided a safety squeeze-bulb. Demonstrate how it works and
explain to the patient when to use the squeeze-bulb (i.e., if they need assistance
during the exam). Figure A22.1.1 The arm is extended at the patient’s side and in neutral
position, with the thumb pointed toward the ceiling. The coil is placed over the
8. Have the patient mount onto the table in a supine position. Place a wedge under the glenohumeral joint.
patient’s knees to facilitate comfort and lessen the likelihood of motion. Cover the
patient with a sheet to maintain personal privacy.
9. Fully extend the arm at the patient’s side and in neutral position. This is achieved by
having the thumb pointed toward the ceiling (see Fig. A22.1.1). When the thumb
points away from the body, the humerus is in external rotation; when it points toward
the body, internal rotation.
10. Place a dedicated shoulder coil over the region of the glenohumeral joint (see Fig.
22.1.1).
11. Use the centering light to assign a reference point of the patient’s position (middle
of humeral head) on the table relative to the bore of the magnet and localize the region
of the shoulder to be imaged. Then place the patient into the bore of the magnet.
In order to preserve a true frame of reference from one scan relative to another, the patient
must not move after this point.

Sequence 1: Spoiled gradient-recalled (SPGR) coronal localizer

12. This large field of view localizer will result in an image of the upper extremity, axilla,
and thorax. From this, plan the transverse images. Next, set the grid lines (graphic
prescription) to begin in the soft tissues just above the AC joint and extend to a point
below the inferior most aspect of the glenoid (see Fig. A22.1.2). Let the patient know
that you are ready. Run sequence 1 (Pilot scan) following the parameters in Table Figure A22.1.2 SPGR coronal localizer. The grid lines are set from above
A22.1.2. Rotator Cuff the acromioclavicular joint and extend to a point below the inferior aspect of
Shoulder Disease
the glenohumeral joint.
A22.1.3 A22.1.4
Current Protocols in Magnetic Resonance Imaging Supplement 3 Supplement 3 Current Protocols in Magnetic Resonance Imaging
Sequence 3: Sagittal oblique fast spin echo (FSE) T1-weighted 17. Let the patient know that you will begin to scan. Run sequence 3 following the
16. Review the transverse images and select an image at the level of the mid-scapula in parameters in Table A22.1.4.
which the scapular spine is visualized. Place the localizer line parallel to the articular
Sequence 4: Sagittal oblique fat-suppressed fast spin echo (FSE) T2-weighted
surface of the glenoid and perpendicular to the spine of the scapula. The grid lines
18. Use the same localizer for this sequence. Run this sequence according to the
should begin 2 cm medial to the glenoid and extend beyond the lateral margin of the
parameters in Table A22.1.5.
humeral head. (see Fig. A22.1.3).

Table A22.1.2 Primary Clinical Imaging Parameters for Sequence 1 (Pilot Scan):
Spoiled Gradient-Recalled (SPGR) Coronal Localizer

Patient position Supine

Scan type 2-D spoiled gradient-recalled, fast
multiplanar
Imaging plane (orientation) Coronal
Central slice or volume center Center anatomy
Echo time (TE) 4.2 msec
Repeat time (TR) 175 msec
Flip angle (FA) 60°
Fields of view (FOVx, FOVy) 320 mm, 320 mm (off set to side)
Resolution (Δx, Δy) 1.25 mm, 2.5 mm
Number of data points collected (Nx, Ny) 256, 128
Display matrix (Dx, Dy) 512, 512
Slice thickness (Δz) 7 mm
Number of slices 15
Slice gap 2 mm
Number of acquisitions (Nacq) 2
Swap read and phase encoding No
Figure A22.1.3 The grid lines are placed perpendicular to
Saturation pulses None
the spine of the scapula to perform the sagittal oblique
Scan time 46 sec images.

Table A22.1.3 Primary Clinical Imaging Parameters for Sequence 2: Transverse Table A22.1.4 Primary Clinical Imaging Parameters for Sequence 3: Sagittal
Fat-Suppressed Fast Spin Echo (FSE) Proton Density Oblique Fast Spin Echo (FSE) T1-Weighted

Patient position Supine Patient position Supine

Scan type Fast spin echo Scan type Fast spin echo
Imaging plane (orientation) Transaxial Imaging plane (orientation) Sagittal oblique
Central slice or volume center Superior AC joint to inferior glenoid Central slice or volume center Perpendicular to the spine of the
Echo time (TE) 10 msec scapula
Echo train length (ETL) 4–8 Echo time (TE) 13.6 msec
Repeat time (TR) 2400 msec Echo train length (ETL) 2
Flip angle (FA) 90° Repeat time (TR) 600 msec
Fields of view (FOVx, FOVy) 160 mm, 160 mm Flip angle (FA) 90°
Resolution (Δx, Δy) 0.63 mm, 0.63 mm Fields of view (FOVx, FOVy) 140 mm, 140 mm
Number of data points collected (Nx, Ny) 256, 256 Resolution (Δx, Δy) 0.55 mm, 0.55 mm
Display matrix (Dx, Dy) 512, 512 Number of data points collected (Nx, Ny) 256, 256
Slice thickness (Δz) 4 mm Display matrix (Dx, Dy) 512, 512
Number of slices 22 Slice thickness (Δz) 4 mm
Slice gap 1 mm Number of slices 20
Number of acquisition (Nacq) 3 Slice gap 1 mm
Swap read and phase encoding No Number of acquisitions (Nacq) 2
Saturation pulses Inferior Swap read and phase encoding No
Rotator Cuff Saturation pulses Inferior
Scan time 4 min, 46 sec Shoulder Disease
Scan time 4 min, 34 sec
A22.1.5 A22.1.6
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
Sequence 5: Coronal oblique fast spin echo (FSE) T1-weighted
19. The coronal oblique images are also planned from the transverse images in one of
two ways. Either choose a high transverse section at the level of the supraspinatus
tendon allowing direct parallel alignment of the graphic prescription with the tendon.
The lines should extend to the soft tissue margins both anteriorly and posteriorly.
Alternatively, perform a mid transverse localizer. In this case, place the localizer lines
along the long axis of the visualized scapular spine, and perpendicular to the glenoid
articular surface (see Fig. A22.1.4). Extend the lines 2 cm anterior and posterior to
the humeral head. The latter method underestimates the degree of obliquity needed
to section the supraspinatus tendon along its true long axis (see Fig. A22.1.5). Run
sequence 5 following the parameters in Table A22.1.6.

Table A22.1.5 Primary Clinical Imaging Parameters for Sequence 4: Sagittal

Oblique Fat-Suppressed Fast Spin Echo (FSE) T2-Weighted

Patient position Supine

Scan type Fast spin echo
Imaging plane (orientation) Sagittal oblique
Central slice or volume center Perpendicular to the spine of the
scapula
Echo time (TE) 75 msec
Echo train length (ETL) 4–8
Repeat time (TR) 2500 msec
Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 140 mm, 140 mm
Resolution (Δx, Δy) 0.55 mm, 0.73 mm
Number of data points collected (Nx, Ny) 256, 192
Display matrix (Dx, Dy) 512, 512
Slice thickness (Δz) 4 mm
Number of slices 20
Slice gap 1 mm
Number of acquisitions (Nacq) 2
Figure A22.1.5 The grid lines are located parallel to the su-
Swap read and phase encoding No praspinatus tendon to perform the coronal oblique images.
Saturation pulses Inferior
Scan time 3 min, 40 sec
Table A22.1.6 Primary Clinical Imaging Parameters for Sequence 5: Coronal
Oblique Fast Spin Echo (FSE) T1-Weighted

Patient position Supine

Scan type Fast spin echo
Imaging plane (orientation) Coronal oblique
Central slice or volume center Parallel to the long axis of the spine of
the scapula
Echo time (TE) 13 msec
Echo train length (ETL) 2
Repeat time (TR) 625 msec
Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 140 mm, 140 mm
Resolution (Δx, Δy) 0.55 mm, 0.55 mm
Number of data points collected (Nx, Ny) 256, 256
Display matrix (Dx, Dy) 512, 512
Slice thickness (Δz) 4 mm
Number of slices 22
Slice gap 1 mm
Number of acquisitions (Nacq) 2
Swap read and phase encoding Yes
Saturation pulses Inferior
Figure A22.1.4 The grid lines are placed parallel to the spine Rotator Cuff Scan time 5 min, 21 sec
of the scapula to perform the coronal oblique images. Shoulder Disease

A22.1.7 A22.1.8
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A22.1.7 Primary Clinical Imaging Parameters for Sequence 6: Coronal Sequence 6: Coronal oblique fat-suppressed FSE proton density
Oblique Fat-Suppressed FSE Proton Density 20. Use the same localizer for this sequence. Once again, follow the appropriate sequence
prescribed and name. Run sequence 6 according to the parameters in Table A22.1.7.
Patient position Supine
Scan type Fast spin echo Sequence 7: Coronal oblique short inversion time (tau) inversion recovery (STIR)
Imaging plane (orientation) Coronal oblique 21. The set of imaging sequences presented in the protocol above is an effective means
Central slice or volume center Parallel to the long axis of the spine of to evaluate the glenohumeral joint. The addition of an STIR sequence, as previously
the scapula
mentioned, confers the added advantage of excellent fat suppression with exquisite
Echo time (TE) 12 msec
sensitivity to free water protons. This can prove extremely helpful for evaluation of
Echo train length (ETL) 4–8
signal abnormalities in the rotator cuff. It can be performed in any imaging plane, but
Repeat time (TR) 2500 msec
usually proves most helpful in the coronal oblique plane. Follow Table A22.1.8 for
Flip angle (FA) 90°
parameters to run sequence 7.
Fields of view (FOVx, FOVy) 140 mm, 140 mm
Resolution (Δx, Δy) 0.55 mm, 0.55 mm
Number of data points collected (Nx, Ny) 256, 256 COMMENTARY
Display matrix (Dx, Dy) 512, 512 Background Information of hyperintense signal on short TE sequences
Slice thickness (Δz) 4 mm The rotator cuff is a set of muscles and and makes close scrutiny of the T2-weighted
Number of slices 22 tendons composed of the supraspinatus, in- sequences imperative before an abnormality is
Slice gap 1 mm fraspinatus, subscapularis, and teres minor diagnosed.
Number of acquisitions (Nacq) 2 groups. These muscles arise from the scapula Iatrogenic changes can prove misleading in
Swap read and phase encoding Yes to insert on the tuberosities of the humeral head the MR evaluation of the rotator cuff. Steroid
Saturation pulses Inferior fat saturated and serve to both facilitate motion of the shoul- or local anesthetic injections can cause in-
Scan time 4 min, 25 sec der girdle and stabilize this cup and saucer type creased signal intensity in and around the rota-
articulation. MR imaging of the glenohumeral tor cuff. Post-surgical changes can result in a
joint is second only to the knee, reflecting the host of findings that could easily be misinter-
common nature of pathology in this articula- preted as primary abnormalities in the absence
Table A22.1.8 Primary Clinical Imaging Parameters for Sequence 7: Coronal
tion. It is considered the imaging study of of thorough clinical history.
Oblique STIR
choice for evaluation of the rotator cuff, largely
Patient position Supine due to its multiplanar imaging capabilities and Anticipated Results
superior soft tissue contrast, accounting for its Injury to the rotator cuff leads to shoulder
Scan type Inversion recovery fast spin echo
high sensitivity, specificity, and accuracy in this pain and dysfunction, a common clinical com-
Imaging plane (orientation) Coronal oblique
realm. plaint. The spectrum of pathology affecting this
Central slice or volume center Parallel to the long axis of the spine of Other methods for evaluation of the gleno- articulation is quite variable and ranges from
the scapula
humeral joint and rotator cuff include conven- the acute partial or full thickness tear to chronic
Echo time (TE) 34 msec tional radiographs, standard arthrography, ul- intrinsic or extrinsic tendon changes. The goal
Echo train length (ETL) 3–8 trasound, and computed tomography with and of MR imaging in evaluation of the rotator cuff
Repeat time (TR) 2600 msec without arthrography. Though all of these stud- is to identify and characterize the extent of
Inversion time (TI) 150 msec ies serve a purpose in evaluation of the shoulder disease in an attempt to guide treatment and
Flip angle (FA) 90° and offer useful information in diagnosis of its treat patients. MR imaging has emerged as the
Fields of view (FOVx, FOVy) 140 mm, 140 mm pathology, none approaches the level of com- imaging method of choice for evaluation of the
Resolution (Δx, Δy) 0.55 mm, 0.73 mm pleteness achieved with MR imaging. rotator cuff and its pathology, primarily due to
Number of data points collected (Nx, Ny) 256, 192 its superior soft tissue contrast.
Display matrix (Dx, Dy) 512, 512 Critical Parameters and
Slice thickness (Δz) 4 mm Troubleshooting Acknowledgement
Number of slices 22 The application of the protocol and interpre- The authors would like to thank Dexter
tation of the images acquired from this unit is Walpole, R.T., for reviewing this material and
Slice gap 1 mm
generally done with ease. Perhaps the most providing helpful comments.
Number of acquisitions (Nacq) 2
troublesome factor to consider for the radiolo-
Swap read and phase encoding Yes gist is that of the magic angle phenomenon Editor’s Note
Saturation pulses Inferior (Timins et al., 1995). When collagen fibers are This unit has been written by the authors
Scan time 6 min, 25 sec oriented at ∼55° with respect to the main mag- clearly for high field scanners with fat satura-
netic field (as in the case of the inserting fibers tion sequences. For those users who do not have
of the portions of the rotator cuff), they may fat saturation capability, such as on mid- or
display artifactually increased signal intensity low-field scanners, a reasonable alternative
Rotator Cuff on T1 and proton density weighted sequences. protocol might be: 1) oblique coronal proton
Shoulder Disease This creates some confusion with the presence density weighted and T2-weighted fast/turbo

A22.1.9 A22.1.10
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
spin echo images; 2) oblique sagittal proton
density weighted and T2-weighted fast/turbo
spin echo images; 3) transverse T2* gradient
echo images; and 4) transverse STIR (short tau
inversion recovery) images. In the future, we
expect to have a separate unit discussing low
field protocols.
Also, for a given field strength, the field of
view should be adjusted to the size of the
patient, within the constraints of the scanner.
On a high field scanner, 12 to 14 cm or smaller
fields of view are easily achievable and can
cover the entire shoulder in a average sized
patient to give the best resolution. The field of
view should be adjusted down for smaller pa-
tients and up for larger patients as necessary.
On mid- and low-field scanners, these smaller
fields of view may not be achievable in reason-
able scan times and with sufficient signal to
noise. In that case, the smallest field of view
should be used (sacrificing some resolution as
needed) to maintain sufficient signal to noise
and the images then magnified (zoomed).
Literature Cited
Hennig, J., Nauerth, A., and Friedburg, H. 1986.
RARE imaging: A fast imaging method for clini-
cal MR. Magn. Reson. Med. 3:823-833.
Quinn, S.F., Sheley, R.C., Demlow, T., and Szu-
mowski, J. 1995. Rotator cuff tendon tears:
Current Protocols in Magnetic Resonance Imaging
Evaluation with fat-suppressed MR imaging
with arthroscopic correlation in 100 patients.
Radiology 195:497-501.
Reinus, W.R, Shady, K.L., Mirowitz, S.A., and
Totty, W.G. 1995. MR diagnosis of rotator cuff
tears of the shoulder: Value of using T2-weighted
fat-saturated images. Am. J. Roentgenol.
164:1451-1455.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Singson, R.D., Hoang, T., Dan, S., and Fridman, M.
1996. MR evaluation of rotator cuff pathology
using T2-weighted fast spin-echo technique with
and without fat suppression. Am. J. Roentgenol.
166:1061-1065.
Timins, M.E., Erickson, S.J., Estkowski, L.D., Car-
rera, G.F., and Komorowski, R.A. 1995. In-
creased signal in the normal supraspinatus ten-
don on MR imaging: Diagnostic pitfall caused
by the magic-angle effect. Am. J. Roentgenol.
165:109-114.
Contributed by Leopoldo M. Gigena,
Christine B. Chung, and Donald Resnick
Veterans Affairs San Diego Healthcare
System
University of California San Diego
La Jolla, California
Shoulder
A22.1.11
MR Arthrography of Shoulder UNIT A22.2
The glenohumeral joint boasts the greatest range of motion of any peripheral joint in the
body, but not without cost; it is also the most frequently dislocated joint in the body.
Stability of this articulation is limited for two major reasons. The articulating surface of
the glenoid is significantly smaller than that of the humeral head, and the joint capsule is
redundant and provides little support.
Generally speaking, the term instability when applied to the glenohumeral joint, refers to
a recurrence of dislocation or subluxation. This is not an unexpected complication, given
the complex biomechanics of this articulation, the sophisticated movement achieved, and
the high performance demanded from the body. There is much debate and controversy
surrounding the mechanism and imaging evaluation of shoulder instability.
This unit will focus on an MR arthrography protocol for evaluation of glenohumeral joint
instability.
ROLE OF EVALUATION IN SHOULDER INSTABILITY BASIC
PROTOCOL
As previously mentioned, MR imaging has revolutionized the evaluation of the gleno-
humeral joint, in particular, internal derangements. It offers excellent visualization of soft
tissue structures and unparalleled soft tissue contrast. Although conventional MR imaging
allows direct visualization of the major anatomic structures of the glenohumeral joint,
many investigators feel smaller intraarticular structures and their pathology are difficult
to evaluate without distention of the joint capsule (Flanagan et al., 1990; Chandnani et
al., 1993; Chung et al., in press). The disadvantages of MR arthrography include the
necessity for a more complex scheduling system as the arthrography portion of the study
is usually performed fluoroscopically, and the transformation of a noninvasive study to
an invasive one. Though this remains a topic of debate, the authors feel MR arthrography
is the imaging method of choice for evaluation of glenohumeral joint instability and will
continue to gain acceptance in the community as more sophisticated study interpretation
is demanded from referring physicians.
This unit will present an MR arthrography protocol for evaluation of glenohumeral joint
instability. It will not cover the arthrogram technique as this is performed as a completely
separate procedure. It will not cover the arthrogram technique as this is performed as a
completely separate procedure.
Although the arthrogram technique will not be discussed in detail, the operator has the
choice of media with which to distend the glenohumeral joint. In the early literature, it
was felt that dilute gadolinium solutions offered optimal contrast for delineation of
intraarticular structures as compared to several other agents including saline (Hajek et al.,
1987). Although this has been validated in the recent literature where superior contrast-
to-noise ratios have been demonstrated with dilute gadolinium (4 mmol/l gadoteridate)
MR arthrography as compared with saline MR arthrography, the differences in diagnostic
efficacy have not proven significant (Zanetti et al., 1997). The primary technical differ-
ences resulting from the use of saline as the contrast medium would be the emphasis on
fluid sensitive sequences such as proton density fat-suppressed fast spin echo (FSE),
T2-weighted fat-suppressed fast spin echo, or inversion recovery imaging to optimally
evaluate the distribution of saline in the joint. The authors feel the advantages of using a
dilute gadolinium solution as the contrast agent in conjunction with T1- and T2-weighted
images offer the significant advantage of distinguishing between existing native fluid
collections and those resulting from the intraarticular distention. Table A22.2.1 lists
Shoulder
Contributed by Christine B. Chung, Leopoldo M. Gigena, and Donald Resnick A22.2.1
Current Protocols in Magnetic Resonance Imaging (2001) A22.2.1-A22.2.13
Copyright © 2001 by John Wiley & Sons, Inc.
 Table A22.2.1 Equipment Parameters for MR Arthrography 5. In conjunction with the standardized screening form, have the patient complete a
background information form.
Coil type Shoulder phased array or standard shoulder coil
Gradient coil strength 22 mT/m This proves to be an invaluable source of history and clinical information given the paucity
of information often provided on the requisition forms. The information that the patient
Cardiac gating No
provides may help the technician to do a more specific study and the radiologist to render
Peripheral gating No the most accurate interpretation of the exam.
Respiratory gating No
Oxygen No The questionnaire includes:
Motion cushions Useful a. Requesting physician and subspecialty
Use of contrast agents Yes (intraarticular)
b. Origin and progression of symptoms
c. Request for diagram of the symptomatic area
hardware and parameters needed to perform the examination. This protocol will take ∼45
min to perform. d. Activities that exacerbate symptoms
6. Review the data sheets and interview the patient to ensure he or she understands the
Set up patient and equipment
nature of all questions and that the data sheets have been filled out as completely and
1. The patient will be scheduled for a two-part examination. The first part is the standard
accurately as possible.
arthrogram (where contrast is administered into the articulation), which is scheduled
30 min prior to the MR imaging portion of the study. The coordination of the timing 7. Have the patient sign any necessary consent forms.
of these two components is of the utmost importance. After appropriate informed
consent has been obtained and the arthrogram is completed, transport the patient to 8. Instruct the patient to remove all jewelry or metal from their body, and change into
the MR imaging suite with minimal movement of the shoulder being imaged to a gown.
decrease the possibility of rupturing the joint capsule. It is best to have them 9. Explain what will happen during the course of the procedure and what the patient
transported by wheelchair. will experience while in the magnet.
2. Begin MR imaging within 15 to 30 min of the arthrogram procedure to lessen the a. The patient will be given earphones to wear during the examination to protect from
possibility of resorption of the contrast material. the loud knocking sound produced during the study. This knocking will occur
3. Screen the patient by having him/her complete a standardized form designed to ensure when images are being acquired, take place six to seven times and last a few
that he or she has no internal ferromagnetic materials. For questions regarding safe minutes each time. The technologist and patient may communicate at any time
scanning of implants, see Shellock (1996). In the authors’ screening process, specific during the study by simply speaking out, though it is preferable to wait until the
reference is made to the following items, which can prove to be a health hazard to knocking sound has stopped.
the patient or interfere with image acquisition. These include: b. The patient is asked to remain quiet and avoid any movement during the time
images are being acquired—i.e., when the knocking sound is occurring.
a. Cardiac pacemaker
c. The patient will be provided with a safety squeeze-bulb. Demonstrate how it works
b. Retained metal fragments in eyes and explain to the patient when to use the squeeze-bulb (i.e., if they need assistance
c. Heart valve replacement, venous umbrella during the exam).
d. Vascular clips
10. Have the patient mount onto the table and assume a supine position. Place a wedge
e. Prosthetic devices in the eyes and joints under the patient’s knees to facilitate their comfort and lessen the likelihood of
f. Hearing aid, neurostimulator, insulin pump motion. Cover the patient with a sheet to maintain personal privacy.
g. Intrauterine device (I.U.D.; birth control device) 11. The arm should be fully extended at the patient’s side and in neutral position. This is
h. Shunts/stents (ventricular, spinal, biliary), metal mesh/coil implant achieved by having the thumb pointed towards the ceiling (see Fig. A22.2.1). When
i. Orthopedic hardware the thumb points away from the body, the humerus is in external rotation, when it
points towards the body, it is in internal rotation.
4. In addition, the screening form includes general inquiries to health issues that are
pertinent to the performance of the MR imaging study, the need for any emergency 12. Place a dedicated shoulder coil over the region of the glenohumeral joint (see Fig.
equipment (e.g., crash cart, oxygen), and its interpretation. These include: A22.2.1).

a. Pregnancy status of patient 13. The centering light is used to assign a reference point of the patient’s position on the
table relative to the bore of the magnet and localize the region of the shoulder to be
b. Respiratory difficulties or nausea when lying in supine position
imaged. Then place the patient into the bore of the magnet. In order to preserve a true
c. Past medical history and current medications frame of reference from one scan relative to another, the patient must not move after
MR d. History of claustrophobia this point.
Arthrography of
Shoulder Shoulder

A22.2.2 A22.2.3
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 3
 Sequence 1: Spoiled gradient-recalled (SPGR) coronal localizer (pilot scan) Table A22.2.2 Clinical Parameters for Sequence 1: Spoiled Gradient-Recalled
14. This large field of view localizer will result in an image of the upper extremity, axilla, (SPGR) Coronal Localizer (Pilot Scan)
and thorax. From this, the transverse images are planned from the beginning of the
Patient position Supine
soft tissues just above the AC joint to a point below the inferior most aspect of the
Scan type 2-D spoiled gradient-recalled
glenoid (see Fig. A22.2.2). Let patient know that you are ready. Run sequence 1
Imaging plane (orientation) Coronal localizer
according to the parameters in Table A22.2.2.
Central slice or volume center Center of anatomy
Echo time (TE) 4.2 msec
Repeat time (TR) 150 msec
Flip angle (FA) 60°
Fields of view (FOVx, FOVy) 360 mm, 360 mm
Resolution (Δx, Δy) 1.41 mm, 2.82 mm
Number of data points collected (Nx, Ny) 256, 128
Display matrix (Dx, Dy) 512, 512
Slice thickness (Δz) 7 mm
Number of slices 13
Slice gap 2 mm
Number of acquisitions (Nacq) 2
Swap read and phase encoding No
Saturation pulses None
Scan time 40 sec

Table A22.2.3 Clinical Parameters for Sequence 2: Transverse

Fat-Suppressed Fast Spin Echo (FSE) T1-Weighted

Patient position Supine

Scan type Fast spin echo
Figure A22.2.1 The arm is extended at the patient’s side and in
neutral position, with the thumb pointed towards the ceiling. The coil is Imaging plane (orientation) Transverse
placed over the glenohumeral joint. Pulse sequence database (PSD) FSE-XLC
Central slice or volume center Acromioclavicular joint to inferior
glenoid
Echo time (TE) 12.2 msec
Receiver bandwidth (RBW) 17.86 KHz
Echo train length (ETL) 2
Repeat time (TR) 650 msec
Flip Angle (FA) 90°
Fields of view (FOVx, FOVy) 160 mm, 160 mm
Resolution (Δx, Δy) 0.63 mm, 0.83 mm
Number of data points collected (Nx, Ny) 256, 192
Slice thickness (Δz) 3 mm
Number of slices 27
Slice gap 1 mm
Number of acquisitions (Nacq) 2
Swap read and phase encoding No
Saturation pulses Inferior
Fat suppression Yes
Scan time 6 min, 23 sec

Figure A22.2.2 SPGR coronal localizer. The grid lines are set from
MR above the acromioclavicular joint and extend to a point below the inferior
Arthrography of
Shoulder aspect of the glenohumeral joint. Shoulder

A22.2.4 A22.2.5
Supplement 3 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Sequence 2: Transverse fat-suppressed fast spin echo (FSE) T1-weighted
15. After the grid lines (graphic prescriptions) are set, changes in the sequence prescrip-
tion can still be made. Check to see that all parameters have been correctly entered
according to Table A22.2.3 before running this sequence.
16. Let the patient know that you will begin the scan and run it.
17. After the images have been acquired, if they are satisfactory in appearance, both the
sagittal and coronal oblique imaging planes can be planned from the transverse
images.
Sequence 3: Sagittal oblique fat-suppressed fast spin echo (FSE) T1-weighted
18. Review the transverse images and select an image at the level of the mid-scapula in
which the scapular spine is visualized. Place the localizer line parallel to the articular
surface of the glenoid and perpendicular to the spine of the scapula (see Fig. A22.2.3).

Figure A22.2.3 The grid lines are placed perpendicular to the

spine of the scapula to perform the sagittal oblique images.

Table A22.2.4 Clinical Parameters for Sequence 3: Sagittal Oblique

Fat-Suppressed Fast Spin Echo (FSE) T1-Weighted Figure A22.2.4 The grid lines are placed parallel to the
spine of the scapula to perform the coronal oblique images.
Patient position Supine
Scan type Fast spin echo
Imaging plane (orientation) Sagittal oblique
Pulse sequence database (PSD) FSE-XLC
Central slice or volume center Perpendicular to the long axis of
the spine of the scapula
Echo time (TE) 12.4 msec
Receiver bandwidth (RBW) 17.86 KHz
Echo train length (ETL) 2
Repeat time (TR) 550 msec
Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 140 mm, 140 mm
Resolution (Δx, Δy) 0.55 mm, 0.73 mm
Number of data points collected (Nx, Ny) 256, 192
Slice thickness (Δz) 3 mm
Number of slices 24
Slice gap 1 mm
Number of acquisitions (Nacq) 2
Swap read and phase encoding No
Saturation pulses Inferior
MR Fat suppression Yes
Arthrography of Scan time 5 min, 24 sec Figure A22.2.5 The grid lines are located parallel to the su-
Shoulder praspinatus tendon to perform the coronal oblique images. Shoulder

A22.2.6 A22.2.7
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 Table A22.2.5 Clinical Parameters for Sequence 4: Coronal Oblique The grid lines should begin 2 cm medial to the glenoid and extend beyond the lateral
Fat-Suppressed Fast Spin Echo (FSE) T1-Weighted margin of the humeral head. Run sequence 3 following the parameters in Table
A22.2.4.
Patient position Supine
Scan type Fast spin echo 19. Let the patient know that you will begin and scan.
Imaging plane (orientation) Coronal oblique Sequence 4: Coronal oblique fat-suppressed fast spin echo (FSE) T1-weighted
Pulse sequence database (PSD) FSE-XLC 20. The coronal oblique images are also planned from the transverse images in one of
Central slice or volume center Parallel to the long axis of the two ways. One can choose a high transverse section at the level of the supraspinatus
spine of the scapula
tendon allowing direct alignment of the localizer with the tendon. The localizer
Echo time (TE) 12.4 msec
should extend to the soft tissue margins both anteriorly and posteriorly. In this case,
Receiver bandwidth (RBW) 17.86 KHz
Echo train length (ETL) 2
Repeat time (TR) 600 msec
Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 140 mm, 140 mm
Resolution (Δx, Δy) 0.55 mm, 0.73 mm
Number of data points collected (Nx, Ny) 256, 192
Slice thickness (Δz) 3 mm
Number of slices 24
Slice gap 1 mm
Number of acquisitions (Nacq) 2
Swap read and phase encoding Yes
Saturation pulses Inferior
Fat suppression Yes
Scan time 5 min, 53 sec

Table A22.2.6 Clinical Parameters for Sequence 5: Coronal Oblique

T2-Weighted
Figure A22.2.6 ABER position with coil placement. See the abducted
Patient position Supine arm and the coil behind the axilla.
Scan type Fast spin echo
Imaging plane (orientation) Coronal
Pulse sequence database (PSD) FSE-XL
Central slice or volume center Parallel to the long axis of the
spine of the scapula
Echo time (TE) 102 msec
Echo train length (ETL) 9
Repeat time (TR) 3375 msec
Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 140 mm, 140 mm
Resolution (Δx, Δy) 0.55 mm, 0.73 mm
Number of data points collected (Nx, Ny) 256, 192
Slice thickness (Δz) 3 mm
Number of slices 24
Slice gap 1 mm
Number of acquisitions (Nacq) 2
Swap read and phase encoding Yes
Saturation pulses Inferior
Fat suppression Yes
Scan time 2 min, 35 sec

MR
Arthrography of Figure A22.2.7 Localizer images parallel to the long axis of
Shoulder the humerus. Shoulder

A22.2.8 A22.2.9
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 3
 the localizer lines are placed along the long axis of the visualized scapular spine, and
perpendicular to the glenoid articular surface (see Fig. A22.2.4). The lines should
extend 2 cm anterior and posterior to the humeral head. The latter method underes-
timates the degree of obliquity needed to section the supraspinatus tendon along its
true long axis (see Fig. A22.2.5). Run sequence 4 according to the parameters in Table
22.2.5.
Sequence 5: Coronal oblique fast spin echo (FSE) T2-weighted
21. The same localizer is used for this sequence 5, run this one according to the
parameters in Table A22.2.6.
Change position and coil
22. Move the scanning table out of the magnet and remove the phased array shoulder
coil. Ask the patient to place his or her arm in an abducted and externally rotated
position (ABER), i.e., hand behind the head or neck while remaining supine (see Fig.
A22.2.6). This can be an uncomfortable position for patients with shoulder pain. The
importance of the sequence should be emphasized to the patient to encourage full
cooperation with this potentially difficult position. A general purpose surface coil is
placed between the shoulder and the scanning table. The scanning table is centered
localizing the region of the shoulder and the patient is returned to the magnet.

Sequence 6: Spoiled gradient-recalled (SPGR) coronal localizer for ABER position

23. Image the upper extremity, axilla and thorax with this large field of view localizer.
From this, the transverse oblique images are planned. An image should be chosen
that shows the humeral shaft. Place the localizer lines along the longitudinal axis of
the humeral shaft, with the center point over the glenohumeral joint (see Fig.
A22.2.7). This results in an transverse oblique set of images (see Fig. A22.2.8). Run
sequence 6 according to the parameters in Table A22.2.7.

Sequence 7: ABER fat-suppressed fast spin echo T1-weighted

24. Check to be sure the parameters in Table A22.2.8 have been entered correctly. Assure
the patient that this is the last sequence to be performed and again request full
cooperation with respect to lack of motion. Run this sequence following the parame-
ters in Table A22.2.8.

Figure A22.2.8 Example of the transverse oblique in the ABER

position. This picture also shows a Perthes lesion (designation for
an avulsion of the anteroinferior portion of the glenoid labrum Table A22.2.8 Clinical Parameters for Sequence 7: ABER-Transverse
Fat-Suppressed Fast Spin Echo (FSE) T1-Weighted
without displacement and with stripping of the periosteal mem-
brane).
Patient position Supine
Scan type Fast spin echo
Table A22.2.7 Clinical Parameters for Sequence 6: Spoiled Gradient-Recalled
(SPGR) Coronal Localizer for ABER Position (Pilot Scan) Imaging plane (orientation) Transverse oblique
Pulse sequence database (PSD) FSE-XLC
Patient position Supine Central slice or volume center Center to the glenoid
Scan type 2-D spoiled gradient-recalled Echo time (TE) 12.2 msec
Imaging plane (orientation) Coronal Receiver bandwidth (RBW) 17.86 KHz
Central slice or volume center Center of anatomy Echo train length (ETL) 2
Echo time (TE) 4.2 msec Repeat time (TR) 600 msec
Repeat time (TR) 150 msec Flip angle (FA) 90°
Flip angle (FA) 60° Fields of view (FOVx, FOVy) 160 mm, 160 mm
Fields of view (FOVx, FOVy) 360 mm, 360 mm Resolution (Δx, Δy) 0.63 mm, 0.83 mm
Resolution (Δx, Δy) 1.41 mm, 2.82 mm Number of data points collected (Nx, Ny) 256, 192
Number of data points collected (Nx, Ny) 256, 128 Slice thickness (Δz) 3 mm
Slice thickness (Δz) 7 mm Number of slices 19
Number of slices 13 Slice gap 1 mm
Slice gap 2 mm Number of acquisitions (Nacq) 2
Number of acquisitions (Nacq) 2 Swap read and phase encoding No
Swap read and phase encoding No Saturation pulses Inferior
MR Saturation pulses None Fat suppression Yes
Arthrography of
Shoulder Scan time 40 sec Scan time 4 min, 58 sec Shoulder

A22.2.10 A22.2.11
Supplement 3 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
 COMMENTARY
Background Information
The shoulder is clearly one of the most
biomechanically complex articulations in the
body. As previously mentioned, it has the great-
est range of motion of any peripheral joint, but
is the most frequently dislocated. The body has
countered by providing an intricate, layered
stabilizing system comprised of the osseous
infrastructure, the glenoid labrum, the gleno-
humeral ligaments (capsular thickenings), the
joint capsule, the rotator cuff, and supporting
muscles. Despite the excellent soft tissue con-
trast afforded by MR imaging, identifying the
redundant, delicate supporting layers of the
glenohumeral joint can prove challenging. The
distention of the joint capsule with contrast
material in conjunction with specialized arm
positioning offers a welcome advantage for
identification of the structures associated with
shoulder stability (Tirman et al., 1994).
Common indications for MR arthrography
of the shoulder include any history of instability
(especially previous dislocation), a suggestion
of undersurface tear of the rotator cuff, or rota-
tor cuff interval abnormality (Flanagan et al.,
1990; Chung et al., in press).
Other methods for evaluation of gleno-
humeral joint instability have included conven-
tional MR imaging (Gusmer et al., 1996), and
indirect MR arthrography (Vahlensieck et al.,
1996). Though the optimal means of evaluation
of this entity remains controversial, the detailed
anatomic information offered by well per-
formed MR arthrography cannot be denied.
Critical Parameters and
Troubleshooting
Clearly, anytime an invasive procedure is
involved, great attention to technique must be
provided. It is desirable for the injection to be
entirely intraarticular. Though the shoulder is
an accessible articulation, errors with contrast
administration can occur and familiarity of the
normal appearance of a distended joint may
salvage a study. Care should be taken to avoid
the introduction of air into the joint. This may
result in susceptibility artifact. In addition,
background work has been performed to opti-
mize the millimolar concentration of gadolin-
ium solution used for MR arthrography (Hajek
et al., 1990). If undiluted gadolinium is placed
in the joint, the result is complete loss of signal,
i.e., a black appearance in the surrounding joint
MR space.
Arthrography of
Shoulder
A22.2.12
Literature Cited MR arthrography: In vitro evaluation and prac-
Chandnani, V.P., Yeager, T.D., DeBernardino, T., tical observations. Am. J. Roentgenol. 149:97-
Though fat-suppressed imaging techniques Christensen, K., Gagliardi, J.A., Heitz, D.R., and 104.
further augment image contrast, inhomogene- Hansen, M.F. 1993. Glenoid labral tears: Pro- Shellock, F.G. 1996. Pocket Guide to MR Proce-
ous suppression can prove problematic. spective evaluation with MRI imaging. MR dures and Metallic Objects. Lippincott-Raven,
STIR (short tau inversion recovery) imaging arthrography, and CT arthrography. Am. J. Philadelphia.
Roentgenol. 161:1229-1235.
is not an appropriate means of fat suppression Tirman, P.F., Bost, F.W., Steinbach, L.S., Mall, J.C.,
in this setting because T1 shortening by gad- Chung, C.B., Dwek, J.D., Cho, G.J., Lektrakul, N., Peterfy, C.G., Sampson, T.G., Sheehan, W.E.,
Trudell, D., and Resnick, D. Rotator cuff inter- Forbes, J.R., and Genant, H.K. 1994. MR
olinium-containing contrast results in de- val: Evaluation with MR imaging and MR arthrographic depiction of tears of the rotator
creased image contrast. Arthrography of the shoulder in 32 cadavers. J. cuff: Benefit of abduction and external rotation
Comput. Assisted Tomogr. In press. of the arm. Radiology 192:851-856.
Anticipated Results Flannigan, B., Kursunoglu-Brahme, S., Snyder, S., Vahlensieck, M., Peterfy, C.G., Wischer, T., Som-
Glenohumeral joint instability is a compli- Karzel, R., Del Pizzo, W., and Resnick, D. 1990. mer, T., Lang, P., Schlippert, U., Genant, H.K.,
cated entity paralleling the complex structure MR arthrography of the shoulder: Comparison and Schild, H.H. 1996. Indirect MR arthrogra-
and function of the articulation. Though the with conventional MR imaging. Am . J. phy: Optimization and clinical applications. Ra-
Roentgenol. 155:829-832. diology 200:249-254.
optimal means of imaging for evaluation of
instability remains a controversial subject, the Gu smer, P.B., Potter, H.G., Schatz, J.A., Zanetti, M. and Hodler, J. 1997. Contrast media in
Wickiewicz, T.L., Altchek, D.W., O’Brien, S.J., MR arthrography of the glenohumeral joint: In-
goal is accurate identification and charac- and Warren, R.F. 1996. Labral injuries: Accuracy tra-articular gadopentetate vs saline: Preliminary
terization of abnormalities of the stabilizing of detection with unenhanced MR imaging of the results. Eur. Radiol. 7:498-502.
structures of the shoulder. shoulder. Radiology 200:519-524.
Hajek, P.C., Sartoris, D.J., Gylys-Morin, V.,
Acknowledgments Haghighi, P., Engel, A., Kramer, F., Neumann,
C.H., and Resnick, D. 1990. The effect of intra-
Contributed by Christine B. Chung,
The authors would like to thank R.T. Dexter Leopoldo M. Gigena, and Donald Resnick
Walpole for reviewing this material and provid- articular gadolinium-DTPA on synovial mem-
brane and cartilage. Investig. Radiol. 25:179- Veterans Affairs San Diego Healthcare
ing helpful comments. 183. System
University of California San Diego
Hajek, P.C., Sartoris, D.J., Neumann, C.H., and
Editor’s Note Resnick, D. 1987. Potential contrast agents for La Jolla, California
This unit has been written by the authors
clearly for high field scanners with fat satura-
tion sequences. For those users who do not have
fat saturation capability, such as on mid- or
low-field scanners, a reasonable alternative
protocol might be: (1) oblique coronal proton
density weighted and T2-weighted fast/turbo
spin echo images; (2) oblique sagittal proton
density weighted and T2-weighted fast/turbo
spin echo images; (3) transverse T2* gradient
echo images; and (4) transverse STIR (short tau
inversion recovery) images. In the future, the
authors expect to have a separate unit discuss-
ing low field protocols.
Also, for a given field strength, the field of
view should be adjusted to the size of the
patient, within the constraints of the scanner.
On a high field scanner, 12 to 14 cm or smaller
fields of view are easily achievable and can
cover the entire shoulder in an average sized
patient to give the best resolution. The field of
view should be adjusted down for smaller pa-
tients and up for larger patients, as necessary.
On mid- and low-field scanners, these smaller
fields of view may not be achievable in reason-
able scan times and with sufficient signal to
noise. In that case, the smallest field of view
should be used (sacrificing some resolution as
needed) to maintain sufficient signal to noise
and the images then magnified (zoomed).
Shoulder
A22.2.13
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging
MRI of the Acute Injured Knee
Magnetic resonance imaging (MRI) has become a mainstay in the assessment of internal
derangement of the knee (Carmichael et al., 1997). MR imaging prior to surgery increases
diagnostic confidence (Maurer et al., 1997). It also influences clinical practice by
identifying alternative diagnoses, e.g., for an osteochondral injury that clinically may
mimic meniscal tears, different surgical approaches exist for many cases. MRI helps avoid
unnecessary arthroscopy (Ruwe et al., 1992; Spiers et al., 1993; Mackenzie et al., 1996;
Bui-Mansfield et al., 1997; Carmichael et al., 1997).
This unit presents the standard protocol for imaging injured knees in a clinical setting.
The parameters given in this unit are derived from a 1.5 T machine and may need to be
altered slightly depending on the main magnetic field strength and the equipment
manufacturer.
IMAGING THE KNEE
Magnetic resonance imaging scans can be run at a range of different field strengths. The
lower signal-to-noise ratio (SNR) inherent in low field systems means that trade-offs must
be made in the field of view, number of excitations (or number of acquisitions), slice
thickness, and acquisition matrix size to maintain adequate signal. Whenever the spatial
resolution is kept constant, the lower SNR of a low field system generally results in the
need to reduce the receiver bandwidth and to increase the number of excitations in order
to increase the signal. These alterations in protocol result in a longer acquisition time. On
high field systems, a general protocol of 4 to 5 sequences will result in a comprehensive
evaluation of the injured knee generally in <25 min.
In general, protocols utilizing a short TE sequence (T1, proton density-weighted or gradient
echo) and a long TE sequence (T2-weighted), especially with fat saturation, are useful in
the musculoskeletal system.
Regardless of the primary disease process suspected clinically, the end result, anatomi-
cally and pathologically, of acute injuries is a T2-weighted prolongation (edema and/or
hemorrhage). Therefore, a sensitive T2-weighted examination is needed and it is generally
felt that a fast spin echo T2-weighted sequence with fat saturation or short tau inversion
recovery (STIR) sequence works best. In the authors’ experience, the short TE sequence
of choice for the evaluation of meniscal disorders is a standard (conventional) spin echo
or gradient echo sequence as fast spin echo (FSE) sequences tend to be blurry and are
slightly less sensitive. Recently, improvements in gradients and coils have made the use
of high resolution fast spin echo proton density sequences more practical.
The following five sequences (including one optional sequence) encompass the authors’
preferred Basic Protocol.
Table A23.1.1 provides a list of the hardware necessary to perform the procedure, along
with appropriate parameters. Intravenous or intra-articular contrast agent administration
is neither necessary nor recommended in an acute trauma setting.
NOTE: Be sure that technicians and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as a crash cart or oxygen.
Contributed by Rolf W. Huegli and Phillip F.J. Tirman
Current Protocols in Magnetic Resonance Imaging (2001) A23.1.1-A23.1.11
Copyright © 2001 by John Wiley & Sons, Inc.
UNIT A23.1 Table A23.1.1 Equipment Parameters for Standard Knee Imaging
Coil type Circumferential extremity coil with a
transmit, receive, quadrature, or phase
array design
Gradient coil strength 25 mT/m (or whatever the system permits)
Motion cushions Helpful
Use of contrast agents Not necessary
Set up patient and equipment
1. Interview the patient to ensure that there are no contraindications for the MRI exam
such as cardiac pacemakers or other ferromagnetic materials. Find out if the patient
has any health conditions that may require the presence of any special emergency
equipment during the scanning procedure, or necessitate any other precautions.
Standard screening forms (APPENDIX 1) are generally used for all patients scanned in MRI
BASIC systems.
PROTOCOL
Any ferromagnetic metals may be a health hazard to the patient inside the magnet, and may
also affect the imaging quality. If in doubt as to the exact composition of the devices, it is
best to exclude patients with any metal implants; see Shellock (1996) for a discussion of
which implants may be safely scanned using magnetic resonance.
The patient (or volunteer) may be accompanied into the magnet, by someone who can sit
through the scan and comfort the patient as needed. This accompanying person must also
be screened to ensure the absence of loose metal objects on the body or clothing.
2. If the scan is a research protocol, have the patient sign any necessary consent forms.
3. Ask the patient to remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
4. Have the patient wash off any mascara and other makeup in order to avoid local tissue
heating and image artifacts.
5. Inform the patient of what will happen during the procedure, what he or she will
experience while in the magnet, and how to behave including the following:
a. If earphones or headphones are used to protect the ears from the loud sounds
produced by the magnet, the patient will be asked to wear these, but will be able
to communicate at any time during the examination.
b. The patient will be given a safety bell or similar equipment to request assistance
at any time (demonstrate how it works).
c. In order to obtain good results, the patient should not move or talk during each
scan—i.e., as long as the banging sound continues. Between the scans, talking is
allowed in most cases, but should be avoided when comparative positional studies
are being performed; the patient will be informed when this is the case.
d. Nevertheless, the patient may call out at any time if he or she feels it is necessary.
6. Have the patient positioned on the table with feet toward the magnet. Either before
or right after the patient lies down, set up any monitoring equipment that is to be used.
7. Center the patient in a knee coil at the region where the key information is needed.
Make sure that the knee is constrained to prevent motion, especially if high resolution
scans are to be run.
MRI of the Acute Generally, the patient’s knee is fixed in a straight horizontal neutral position. The comfort-
Knee Injured Knee able installation of the patient at the beginning of the study is important to limit motion
A23.1.1 A23.1.2
Supplement 3 Supplement 3 Current Protocols in Magnetic Resonance Imaging
 Table A23.1.2 Primary Clinical Imaging Parameters for Sequence 1 (Scout Scan)

Patient position Supine

Scan type Fast spin echo
Imaging plane (orientation) Transverse
Central slice or volume center Laser light centered on the lower third
of the patella
Echo time (TE) 42 msec
Receiver bandwidth (RBW) ±11.36 kHz
Echo train length (ETL) 8
Repeat time (TR) 3075 msec
Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 120–140 mm, 120–140 mm
Resolution (Δx, Δy) 0.47–0.55 mm, 0.47–0.55 mm
Number of data points collected (Nx, Ny) 256, 256
Slice thickness (Δz) 5 mm
Number of slices 17
Slice gap 2 mm
Number of acquisitions (Nacq) 2
No phase wrap (NPW)a Yes
Saturation pulses Frequency selective fat saturation
Scan time 3 min, 23 sec
Figure A23.1.1 Fast spin echo proton density image with fat saturation. Note the bone
aSee Index of Terms in UNIT A7.1.
trabecular injury posterolaterally (straight arrow), the torn ACL in the notch of the knee (curved
arrow), and the extensive soft tissue damage (arrowheads).

artifacts. Care should be exercised in positioning the cushions and pads around the knee
in the extremity coil to make the examination as comfortable as possible.
8. If needed, place a pillow under the patient’s head to make him or her more comfort-
able.
9. Use the centering (laser) light to position the injured knee to the lower third of the
patella and put the patient into the center of the bore.
10. If the patient is not able to hold still, provide an appropriate sedative.

Sequence 1: Transverse fast spin echo proton density to T2-weighted image with fat
saturation (localizer; Fig. A23.1.1)
11. To validate the patient’s position, run the localizer (scout scan) to ensure the correct
location of the knee according to Table A23.1.2.
In the following sequences, when necessary, choose the perpendicular (sagittal and
coronal) planes using these transverse images.

Sequence 2: Sagittal dual spin echo proton density weighted/T2-weighted sequence

(Figs. A23.1.2 and A23.1.3)
12. Load the sequence for a sagittal dual spin echo proton density weighted/T2-weighted
sequence. Set the imaging parameters according to Table A23.1.3. Figure A23.1.2 Proton density weighted sagittal image. Note the normal lateral meniscus
(straight arrow) and the osteochondral injury of the posterior-most aspect of the lateral femoral
Sagittal and coronal images can be obliqued to become orthogonal to the knee joint by
condyle (curved arrow).
using the femoral condylar surface as an orienting internal landmark to compensate for
variability in patient positioning.
13. Alert the patient, and begin the scan.
MRI of the Acute
Knee Injured Knee

A23.1.3 A23.1.4
Current Protocols in Magnetic Resonance Imaging Supplement 11 Supplement 11 Current Protocols in Magnetic Resonance Imaging
 Figure A23.1.3 Standard spin echo T2-weighted sagittal image. Note the torn ACL (straight
arrow) and the relatively poor signal-to-noise ratio compared to fast spin echo images. The
standard T2-weighted image is chosen because the sequence is a double echo sequence
including a standard spin echo proton density weighted echo, which is more sensitive for
meniscal tears.

Table A23.1.3 Primary Clinical Imaging Parameters for Sequence 2 Figure A23.1.4 Fast spin echo proton density weighted sagittal image. This is a high
resolution image utilizing the ZIP 512 feature.
Patient position Supine
Scan type Dual spin echo proton density Table A23.1.4 Primary Clinical Imaging Parameters for Sequence 3 (Optional)a
weighted/T2-weighted
Imaging plane (orientation) Sagittal Patient position Supine
Central slice of volume center Laser light centered on the lower third Scan type Fast spin echo proton density weighted
of the patella Imaging plane (orientation) Sagittal
Echo time (TE) 20 msec and 80 msec Central slice or volume center Laser light centered on the lower third
Repeat time (TR) 2000 msec of the patella
Flip angle (FA) 90° Echo time (TE) 27 msec
Fields of view (FOVx, FOVy) 120–140 mm, 120–140 mm Echo train length (ETL) 6
Resolution (Δx, Δy) 0.47–0.55 mm, 0.63–0.73 mm Repeat time (TR) 2200 msec
Number of data points collected (Nx, Ny) 256, 192 Flip angle (FA) 90°
Slice thickness (Δz) 4 mm Fields of view (FOVx, FOVy) 120 mm, 120 mm
Number of slices 16 Resolution (Δx, Δy) 0.38 mm, 0.47 mm
Slice gap 1 mm Number of data points collected (Nx, Ny) 320, 256
Number of acquisitions (Nacq) 1 Slice thickness (Δz) 4 mm
No phase wrap (NPW) Yes Number of slices 17
Saturation pulses No Slice gap 0.5 mm
Scan time 6 min, 39 sec Number of acquisitions (Nacq) 2
No phase wrap (NPW) Yes
ZIP 512 Yes
Saturation pulses No
Scan time 3 min, 11 sec
MRI of the Acute aUse autoshimming to shim the field and choose “phase correct” to remove the field inhomogeneity-caused
Knee Injured Knee
artifact automatically.
A23.1.5 A23.1.6
Current Protocols in Magnetic Resonance Imaging Supplement 11 Supplement 11 Current Protocols in Magnetic Resonance Imaging
 Figure A23.1.5 T1-weighted coronal image. Note intact menisci (straight arrows) and subtle Figure A23.1.6 Fast spin echo proton density weighted coronal image with fat saturation.
edematous changes in the lateral femoral condyle (curved arrow). The edema within the bone Note the extensive bone trabecular injury of the lateral femoral condyle (straight arrow). Note
is shown to a much greater degree on the coronal proton density weighted fat saturated the normal hyaline articular cartilage of the edial femoral condyle (arrowheads).
images.

Table A23.1.5 Primary Clinical Imaging Parameters for Sequence 4 Table A23.1.6 Primary Clinical Imaging Parameters for Sequence 5

Patient position Supine Patient position Supine

Scan type T1-weighted conventional spin echo Scan type T2-weighted fast spin echo
Imaging plane (orientation) Coronal Imaging plane (orientation) Coronal
Central slice or volume center Laser light centered on the lower third Central slice or volume center Laser light centered on the lower third
of the patella of the patella
Echo time (TE) 20 msec Echo time (TE) 54 msec
Receiver bandwidth (RBW) ±15.63 kHz Echo train length (ETL) 10
Repeat time (TR) 600 msec Repeat time (TR) 4200 msec
Flip angle (FA) 90° Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 120–140 mm, 120–140 mm Fields of view (FOVx, FOVy) 120–140 mm, 120–140 mm
Resolution (Δx, Δy) 0.47–0.55 mm, 0.47–0.55 mm Resolution (Δx, Δy) 0.47–0.55 mm, 0.47–0.55 mm
Number of data points collected (Nx, Ny) 256, 256 Number of data points collected (Nx, Ny) 256, 256
Slice thickness (Δz) 5 mm Slice thickness (Δz) 5 mm
Number of slices 19 Number of slices 18
Slice gap 1 mm Slice gap 1 mm
Number of acquisitions (Nacq) 2 Number of acquisitions (Nacq) 2
No phase wrap (NPW) Yes No phase wrap (NPW) Yes
Saturation pulses No Saturation pulses Frequency selective fat saturation
Scan time 5 min, 15 sec Scan time 3 min, 42 sec

MRI of the Acute

Knee Injured Knee

A23.1.7 A23.1.8
Current Protocols in Magnetic Resonance Imaging Supplement 11 Supplement 11 Current Protocols in Magnetic Resonance Imaging
Sequence 3: Sagittal fast spin echo proton density weighted sequence (optional; Fig. sensitivity for detecting edema. T1-weighted Pulsation artifact
A23.1.4) images alone are not adequate to appreciate Popliteal artery pulsation artifacts lead to
An alternative pulse sequence is the fast spin echo proton density weighted sequence areas of edema and hemorrhage in a disrupted streaks in the MR image. They can be mini-
utilizing a ZIP 512 feature that acquires the data at a 320 by 256 acquisition matrix and ligament or tendon. mized by swapping the phase encoding and
then manipulates the information to obtain a high resolution 512 by 512 display matrix read directions.
image. Critical Parameters and
Troubleshooting Anticipated Results
14. Run sequence 3 according to Table A23.1.4. Magnetic resonance imaging of the injured The goal in studying the acute injured knee
knee is a potent, noninvasive tool. The sug- is the detection of soft tissue and bone abnor-
Sequence 4: Coronal spin echo T1-weighted sequence (Fig. A23.1.5) gested imaging method provides an approach malities that could lead to worsening by repeat
to acquire a standard set of images that allows trauma or chronic instability and joint degen-
15. Load the sequence for a coronal spin echo T1-weighted sequence. Set the imaging
a thorough analysis. Even choosing a reliable eration if not treated.
parameters according to Table A23.1.5. protocol allows some potential artifacts to oc- Magnetic resonance imaging offers a highly
16. Alert the patient, and begin the scan. cur during the acquisition. The most common sensitive, specific, and accurate diagnosis of
and important artifacts that may pose diagnos- meniscal tears. In a comparison of 12 studies
Sequence 5: Coronal fast spin echo intermediate to T2-weighted sequence with fat tic difficulties are outlined in the following with a study volume of ≥200 subjects done by
saturation (Fig. A23.1.6) paragraphs together with solutions for reme- Rubin and Paletta (2000), the sensitivity for
dies. diagnosing a medial meniscal tear was 86% to
17. Load the sequence for a coronal spin echo T2-weighted sequence with fat saturation.
96% with a specificity of 84% to 94%. For
Set the imaging parameters according to Table A23.1.6. Magic angle artifact lateral meniscal tears, the sensitivity was 68%
18. Alert the patient, and begin the scan. Near the notch of the knee, the posterior horn to 86% and the specificity was 92% to 98%.
of the lateral meniscus slopes upward. This The negative predictive value of these studies
approximates 55° of the external magnetic field was 91%. This means that, in an arthroscopy in
COMMENTARY and can lead to the magic angle artifact or >9 of 10 cases, a meniscal tear will be con-
Background Information FSE techniques have been recommended as diffuse increased signal intensity (Peterfy et al., firmed. The protocol described in this unit sup-
The knee is one of the most frequently in- some studies have shown no significant de- 1994). To overcome the problem, the knee can ports these results.
jured regions of the body. Substantial progress crease in sensitivity for meniscal tears (Ander- be imaged in a slightly different position (the Anterior cruciate ligament (ACL) ruptures
has been made in magnetic resonance imaging son et al., 1995; Escobedo et al., 1996). To authors choose abduction) in order to exclude are accompanied by medial meniscal tears in
since its initial application in 1984 for evaluat- prevent blurring with FSE and to allow an meniscal tears confidently. 43% and lateral meniscal tears in 66% with an
ing the meniscus. MRI is established as the accurate diagnosis, the echo train length (ETL) overall incidence of meniscal tears of 65% to
diagnostic procedure of choice, supporting the must be relatively short (e.g., <4; Anderson et Truncation artifact 80% (Stoller and Anderson, 1997).
physical examination and plain X-ray studies al., 1995; Escobedo et al., 1996). The imaging The artifact results from under-sampling of Isolated ligamentous lesions can be detected
for virtually all suspected disorders of the knee. time needed for a fast spin echo sequence, data so that interfaces of high and low signal reliably in clinical evaluation and MR imaging,
Moreover, MRI may serve as the “second” compared to a conventional spin echo se- are represented incorrectly on the image. If but physical examinations becomes less reli-
opinion before any surgical intervention. quence, is reduced by, approximately, the echo utilizing a small acquisition matrix, the trunca- able when multiple lesions exist (Rubin et al.,
The technique chosen in the Basic Protocol train length. In the authors’ experience, fast spin tion artifact can result in an increased signal 1998). Rubin et al. (1998) found an overall
allows a reliable detection of common soft echo techniques are slightly less sensitive for intensity through the meniscus that may appear sensitivity and specificity for diagnosing liga-
tissue and bony knee injuries. the detection of meniscal tears, which is why as a tear. Using a higher acquisition matrix can ment tears to be 94% and 99%, respectively,
Gradient echo imaging is an alternative to standard spin echo is chosen, although, at some minimize the artifact. when one or more ligament was torn and 88%
spin echo imaging for the meniscus. However, centers where the primary interpretations are and 84%, respectively, when two or more sup-
some studies have shown that the images are supplied by the authors, fast spin echo tech- Motion artifact porting structures were torn. Clinical examina-
more limited than spin echo imaging in their niques are employed. Alternating increased and decreased signal tion alone gave significantly less accurate re-
ability to show ligament, muscle, tendon, and In order to diagnose abnormalities of me- lines occur with motion and can mimic menis- sults compared to the MR imaging. Hodler et
bone marrow abnormalities (Reeder et al., nisci accurately, short TE sequences should be cal tears (Mirowitz, 1994). It is highly recom- al. (1993) showed a sensitivity, specificity, and
1989; Solomon et al., 1989; Heron and Calvert, applied. Long TE and TR sequences (T2- mended that this portion of the examination be accuracy for anterior cruciate ligament tears of
1992). In addition, higher signal appears within weighted images) are specific but not sensitive. repeated. 89%, 97%, and 95%, respectively. The imaging
a normal meniscus on gradient echo sequences, In other words, visualization of fluid within the protocol provided in this unit offers comparable
which can lead to an overestimation of meniscal meniscal substance is highly specific for a me- Vacuum phenomenon and ferromagnetic results to the cited studies and presents a non-
tears and decrease the specificity and negative niscal tear on a T2-weighted image. However, substances invasive and accurate method for detecting li-
predictive value (Guckel et al., 1995). Gradient if one does not observe fluid in the meniscus Magnetic susceptibility of intra-articular gamentous lesions of the knee.
echo sequences are also vulnerable to suscep- on a T2-weighted image, a tear is not ruled out. gas and ferromagnetic substances may produce Osteochondral defects in the knee can be
tibility artifacts that occur in the vicinity of The authors’ imaging routine protocols for the a low signal intensity void or blooming, espe- detected very reliably and graded with fast spin
ferromagnetic substances (e.g., microscopic knee include T2-weighting in three imaging cially on gradient echo images. This artifact echo T2-weighted sequences with frequency
metallic particles after shavings from prior sur- planes to maximize the sensitivity and specific- may be mistaken for a meniscal tear or articular selective fat saturation as described in se-
gery) or gas (e.g., vacuum phenomenon in the ity for detecting ligamentous and tendinous injury. Normal spin echo or fast spin echo quences 1 and 5. Applying these sequences,
knee joint). pathology. Fat saturation further increases the MRI of the Acute sequences are much less vulnerable for these Bredella et al. (1999) demonstrated that in com-
Knee Injured Knee artifacts because of the 180° refocusing pulse. bination of the axial and coronal fast spin echo
A23.1.9 A23.1.10
Current Protocols in Magnetic Resonance Imaging Supplement 3 Supplement 3 Current Protocols in Magnetic Resonance Imaging
T2-weighted sequences with fat saturation the
sensitivity for the diagnosis of cartilage defects
was 94%, specificity was 99%, and accuracy
was 98%. Of these lesions, 90% were within
one grade using MR imaging and arthroscopy
applying a standard arthroscopic grading
scheme adapted to MR imaging.
MR imaging affects the management and
diagnosis of acute knee injuries by decreasing
the numbers of arthroscopic procedures, im-
proving diagnostic confidence, and supporting
in management decisions (Maurer et al., 1997).
LITERATURE CITED
Anderson, M.W., Raghavan, N., Seidenwurm, D.J.,
Greenspan, A., and Drake, C. 1995. Evaluation
of meniscal tears: Fast spin-echo versus conven-
tional spin-echo magnetic resonance imaging.
Acad. Radiol. 2:209-214.
Bredella, M.A., Tirman, P.F., Peterfy, C.G., Zar-
lingo, M., Feller, J.F., Bost, F.W., Belzer, J.P.,
Wischer, T.K., and Genant, H.K. 1999. Accuracy
of T2-weighted fast spin-echo MR imaging with
fat saturation in detecting cartilage defects in the
knee: Comparison with arthroscopy in 130 pa-
tients. A.J.R. 172:1073-1080.
Bui-Mansfield, L.T., Youngberg, R.A., Warme, W.,
Pitche, J.D., and Nguyen, P.L. 1997. Potential
cost savings of MR imaging obtained before
arthroscopy of the knee: Evaluation of 50 con-
secutive patients. A.J.R. 168:913-918.
Carmichael, I.W., MacLeod, A.M., and Travlos, J.
1997. MRI can prevent unnecessary arthroscopy.
J. Bone Joint Surg. Br. 79:624-625.
Escobedo, E.M., Hunter, J.C., Zink-Brody, G.C.,
Wilson, A.J., Harrison, S.D., and Fisher, D.J.
1996. Usefulness of turbo spin-echo MR imag-
ing in the evaluation of meniscal tears: Compari-
son with a conventional spin-echo sequence.
A.J.R. 167:1223-1227.
Guckel, C., Jundt, G., Schnabel, K., and Gachter, A.
1995. Spin-echo and 3D gradient-echo imaging
of the knee joint: A clinical and histopathological
comparison. Eur. J. Radiol. 21:25-33.
Heron, C.W. and Calvert, P.T. 1992. Three-dimen-
sional gradient-echo MR imaging of the knee:
comparison with arthroscopy in 100 patients.
Radiology 183:839-844.
Hodler, J., Buess, E., Rodriguez, M., and Imhoff, A.
1993. [Magnetic resonance tomography (MRT)
of the knee joint: meniscus, cruciate ligaments
and hyaline cartilage]. Rofo. Fortschr. Geb.
Rontgenstr. Neuen. Bildgeb. Verfahr. 159:107-
112.
Mackenzie, R., Dixon, A.K., Keene, G.S., Holling-
worth, W., Lomas, D.J., and Villar, R.N. 1996.
Magnetic resonance imaging of the knee: As-
Current Protocols in Magnetic Resonance Imaging
sessment of effectiveness. Clin. Radiol. 51:245-
250.
Maurer, E.J., Kaplan, P.A., Dussault, R.G., Diduch,
D.R., Schuett, A., McCue, F.C., Hornsby, P.P.,
and Hillman, B.J. 1997. Acutely injured knee:
Effect of MR imaging on diagnostic and thera-
peutic decisions. Radiology 204:799-805.
Mirowitz, S.A. 1994. Motion artifact as a pitfall in
diagnosis of meniscal tear on gradient reoriented
MRI of the knee. J. Comput. Assist. Tomogr.
18:279-282.
Peterfy, C.G., Janzen, D.L., Tirman, P.F., van Dijke,
C.F., Pollack, M., and Genant, H.K. 1994.
“Magic-angle” phenomenon: A cause of in-
creased signal in the normal lateral meniscus on
short-TE MR images of the knee. A.J.R. 163:149-
154.
Reeder, J.D., Matz, S.O., Becker, L., and Andelman,
S.M. 1989. MR imaging of the knee in the sagit-
tal projection: Comparison of three-dimensional
gradient-echo and spin-echo sequences. A.J.R.
153:537-540.
Rubin, D.A. and Paletta, G.A. Jr. 2000. Current
concepts and controversies in meniscal imaging.
Magn. Reson. Imaging Clin. N. Am. 8:243-270.
Rubin, D.A., Kettering, J.M., Towers, J.D., and Brit-
ton, C.A. 1998. MR imaging of knees having
isolated and combined ligament injuries. A.J.R.
170:1207-1213.
Ruwe, P.A., Wright, J., Randall, R.L., Lynch, J.K.,
Jokl, P., and McCarthy, S. 1992. Can MR imag-
ing effectively replace diagnostic arthroscopy?
Radiology 183:335-339.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Solomon, S.L., Totty, W.G., and Lee, J.K. 1989. MR
imaging of the knee: Comparison of three-di-
mensional FISP and two- dimensional spin-echo
pulse sequences. Radiology 173:739-742.
Spiers, A.S., Meagher, T., Ostlere, S.J., Wilson, D.J.,
and Dodd, C.A. 1993. Can MRI of the knee
affect arthroscopic practice? A prospective study
of 58 patients. J. Bone Joint Surg. Br. 75:49-52.
Stoller, D.W. and Anderson, L.J. 1997. Magnetic
Resonance Imaging in Orthopaedics and Sports
Medicine. 2nd edition. Lippincott-Raven, Phila-
delphia.
Contributed by Rolf W. Huegli
University of California at San Francisco
San Francisco, California
Phillip F.J. Tirman
National Orthopaedic Imaging Associates
Greenbrae, California
Knee
A23.1.11
Supplement 3
MRI of the Post-Operative Meniscus and UNIT A23.2
ACL Graft
Conventional MR imaging has proven valuable in the diagnosis of knee injuries (Car- BASIC
michael et al., 1997; Maurer et al., 1997). Standard imaging protocols have been less PROTOCOL
reliable in imaging the postoperative knee than in an unoperated knee, especially for the
diagnosis of meniscal tears and the postoperative anterior cruciate ligament (ACL).
Postsurgically, intra-articular injection of contrast media allows improved visualization
of recurrent meniscal tears and of meniscal repairs (Applegate et al., 1993). In situations
where an ACL graft is not optimally visualized in a conventional MR imaging study and
its radiological evaluation is important for the clinical follow-up, an intravenous gadopen-
tate dimeglumine administration is sometimes helpful. A post-contrast sagittal T1-
weighted sequence with fat saturation (Sequence 2) may help to distinguish an ACL graft
from the enhancing periligamentous soft tissue synovium. These aforementioned cases
represent two indications for application of either intra-articular or intravenous contrast
media in a postoperative diagnostic setting.
In the postoperative setting where a metal artifact distorts the image, an artifact-reducing
sequence may be helpful (generally a fast spin echo technique with multiple refocusing
pulses, i.e., a large-number echo train length).
The parameters provided in this unit apply to a 1.5-T machine and may need to be altered
slightly depending on the main magnetic field strength and the equipment manufacturer.
Magnetic resonance imaging scans can be run at a range of different field strengths. The
lower signal-to-noise ratio (SNR) inherent in low-field systems means that trade-offs must
be made in the field of view, number of excitations (or number of acquisitions), slice
thickness, and acquisition matrix size to maintain adequate signal. Whenever the spatial
resolution is kept constant, the lower SNR of a low-field system generally results in the
need to increase the receiver bandwidth (or increase the number of data points, Nx) and
to increase the number of excitations in order to increase the signal. These alterations in
protocol result in a longer acquisition time. On high-field systems, a general protocol of
four sequences will result in a comprehensive evaluation of the postoperative knee
generally in <30 min.
In general, protocols utilizing a short TE sequence (T1 or proton density–weighted) and a
long TE sequence (T2-weighted), especially with fat saturation, are useful in the muscu-
loskeletal system as described below.
Spin-echo imaging techniques (especially fast spin echo with the multiple refocusing
pulses) are less prone to artifacts, because the refocusing pulse(s) decreases the field
inhomogeneity produced by the metal in and about the postoperative knee. Gradient-echo
images are the most sensitive to metallic artifacts, and therefore are infrequently used to
evaluate the postoperative knee unless one is specifically searching for loose or foreign
bodies where the “blooming”artifact caused by the bodies helps in their detection.
Regardless of the primary disease process suspected clinically, the end result, anatomi-
cally and pathologically, of acute injuries is a T2-weighted prolongation (edema and/or
hemorrhage). Therefore, a sensitive T2-weighted examination is needed, and it is generally
felt that a fast spin echo T2-weighted sequence with fat saturation or short tau inversion
recovery (STIR) sequence works best. It should be noted the STIR images are not
recommended in the arthrography setting, as a variable degree of contrast saturation can
occur.
Knee
Contributed by Rolf W. Huegli, Phillip F.J. Tirman, and Margie Garbarino A23.2.1
Current Protocols in Magnetic Resonance Imaging (2003) A23.2.1-A23.2.12
Copyright © 2003 by John Wiley & Sons, Inc. Supplement 10
 Table A23.2.1 Equipment Parameters for Postoperative Knee Imaging
Coil type Circumferential extremity coil with a
send-receive, quadrature, or phase array
design
Gradient coil strength 25 mT/m (or whatever the system permits)
Motion cushions Helpful
Use of contrast agents Yes
For the diagnosis of postoperative pathology such as recurrent meniscal tear, MR imaging
can be performed after the intra-articular injection of paramagnetic contrast (gadopentate
dimeglumine–saline mixture as described below) to obtain direct MR arthrography. In
the knee after meniscal repair or meniscectomy, this procedure can better verify the
presence of recurrent meniscal tears or the presence of mensical healing. Fat saturation
techniques are applied because they increase the conspicuity of the contrast in the joint
and potential extravasations into relevant locations. These images provide a high signal-
to-noise ratio, and offer an excellent separation of the intra-articular gadolinium mixture
from the menisci, cartilage, and surrounding soft tissue. Some radiologists obtain standard
images before injecting contrast agents, but many, including the authors, do not.
In the diagnostic setting of an ACL reconstruction, the intravenous application of
gadopentate dimeglumine along with fat suppression techniques may help to better
distinguish an ACL graft from the enhancing periligamentous soft tissue. The intravenous
injection adds ∼5 min to the time of the procedure. The entire protocol including patient
setup takes ∼35 to 40 min.
The following four sequences encompass the authors’ preferred Basic Protocol.
Table A23.2.1 provides a list of the hardware necessary to perform the procedure, along
with appropriate parameters.
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as a crash cart or oxygen.
Materials
Normal saline (0.9% NaCl), sterile
Paramagnetic contrast agent: gadopentate dimeglumine
Set up patient and equipment
1. Interview the patient to ensure that there are no contraindications for the MRI exam
such as contrast media allergy, cardiac pacemakers, or other ferromagnetic materials.
Find out if the patient has any health conditions that may require the presence of any
special emergency equipment during the scanning procedure, or necessitate any other
precautions.
Standard screening forms (see APPENDIX 1) are generally used for all patients scanned in
MRI systems.
Any ferromagnetic metals may be a health hazard to the patient inside the magnet, and may
also affect the imaging. If in doubt as to the exact composition of the devices, it is the best
to exclude patients with any metal implants; see Shellock (1996) for discussion of what
implants may be safely scanned using magnetic resonance.
MRI of the
Post-Operative
Meniscus and
ACL Graft
A23.2.2
Supplement 10 Current Protocols in Magnetic Resonance Imaging
The patient or volunteer may be accompanied into the magnet by another person, who can
sit during the scan and comfort the patient as needed. This accompanying person must be
screened as well to ensure the absence of loose metal objects on the body or clothing.
2. If the scan is a research protocol, have the patient sign any necessary consent forms.
3. Ask the patient to remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
4. Have the patient wash off any mascara and other makeup in order to avoid local tissue
heating.
5. Inform the patient of what will happen during the procedure, what he or she will
experience while in the magnet, and how to behave, including the following:
a. If earphones or headphones are used to protect the ears from the loud sounds
produced by the magnet, the patient will be asked to wear these, but will be able
to communicate with you at any time during the examination.
b. The patient will be given a safety bell or similar equipment to request assistance
at any time (demonstrate how it works).
c. In order to get good results, the patient should not move or talk during each
scan—i.e., as long as the banging sound continues. Between the scans, talking is
allowed in most cases, but should be avoided when comparative positional studies
are being performed; the patient will be informed when this is the case.
d. Nevertheless, the patient may call out at any time if he or she feels it is necessary.
6. Have the patient positioned comfortably on the table with feet toward the scanner.
Inject the appropriate contrast agent:
a. For postoperative meniscus imaging (direct, intra-articular): Dilute paramagnetic
contrast agent (gadopentate dimeglumine) with saline in a ratio of 1:100. Inject
the mixture into the knee until mild joint distention is produced (usually 40 to 50
ml).
b. For intravenous administration in postoperative ACL imaging: Administer
gadopentate dimeglumine intravenously in a dosage of 0.1 mmol per kg of body
weight.
Flush with 10 ml saline after the injection. Either before or right after the patient lies
down, set up any monitoring equipment that is to be used.
7. Center the patient in a knee coil at the region where the key information is needed.
Make sure that the knee is constrained to prevent motion, especially if high-resolution
scans are to be run.
Generally the patient’s knee is fixed in a straight, horizontal, neutral position. For a better
visualization of the ACL, some physicians recommend 10° to 15° of outer rotation (Lee et
al., 1988). This is less critical as slice thickness decreases (≤4mm).
8. If needed, place a pillow under the patient’s head to make him or her more comfort-
able.
The comfortable installation of the patient at the beginning of the study is important to
limiting motion artifacts. Care should be exercised in positioning the cushions and pads
around the knee in the extremity coil to make the examination as comfortable as possible.
9. Use the centering (laser) light to position the injured knee to the lower third of the
patella and put the patient into the center of the magnet.
10. If the patient is not able to hold still, provide an appropriate sedative. Knee
A23.2.3
Current Protocols in Magnetic Resonance Imaging Supplement 10
 Table A23.2.2 Primary Clinical Imaging Parameters for Sequence 1 Table A23.2.3 Primary Clinical Imaging Parameters for Sequence 2

Patient position Supine Patient position Supine

Scan type Fast spin echo Scan type Spin echo
Imaging plane (orientation) Transverse Imaging plane (orientation) Sagittal
Variable bandwidth Yes Variable bandwidth Yes
Central slice or volume center Laser light centered on the lower third Central slice or volume center Laser light centered on the lower third
of the patella of the patella
Echo time (TE) 42 msec Echo time (TE) Minimum (e.g., 11 msec)
Receiver bandwidth (RBW) ±13.89 kHz Receiver bandwidth (RBW) ±17.86 kHz
Echo train length (ETL) 8 Repeat time (TR) 650 msec
Repeat time (TR) 3000 msec Flip angle (FA) 90°
Flip angle (FA) 90° Field of view (FOVx, FOVy) 140 mm, 140 mm
Field of view (FOVx, FOVy) 130 mm, 130 mm Resolution (Δx, Δy) 0.55 mm, 0.55 mm
Resolution (Δx, (Δy) 0.51 mm, 0.51 mm Number of data points collected (Nx, Ny) 256, 256
Number of data points collected (Nx, Ny) 256, 256 Slice thickness (Δz) 4 mm
Slice thickness (Δz) 4 mm Number of slices 20–23
Number of slices 20 Slice gap 0 mm
Slice gap 1 mm Number of acquisitions (Nacq) 2
Number of acquisitions (Nacq) 2 ZIP 512 Yes
Tailored RF Yes Tailored RF Yes
No phase wrap (NPW) Yes No phase wrap (NPW) Yes
Fat suppression Yes; frequency-selective fat saturation, Fat suppression Yes; frequency-selective fat saturation
superior and inferior Scan time 5 min, 38 sec
Scan time 3 min, 23 sec

Figure A23.2.2 Sagittal spin-echo T1-weighted image with fat saturation, with intra-articular
Figure A23.2.1 Fast spin-echo T2-weighted image with fat saturation (localizer). gadolinium.

MRI of the
Post-Operative
Meniscus and
ACL Graft Knee

A23.2.4 A23.2.5
Supplement 10 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 10
 Figure A23.2.3 Sagittal spin-echo T1-weighted image with fat saturation, with intravenous gad- Figure A23.2.4 Coronal spin-echo T1-weighted image with fat saturation, with intra-articular
olinium. gadolinium.

Table A23.2.4 Primary Clinical Imaging Parameters for Sequence 3

Patient position Supine

Scan type Spin echo
Imaging plane (orientation) Coronal
Variable bandwidth Yes
Central slice or volume center Laser light centered on the lower third
of the patella
Echo time (TE) Minimium (e.g., 14.2 msec)
Receiver bandwidth (RBW) ±15.63 kHz
Repeat time (TR) 500 msec
Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 130 mm, 130 mm
Resolution (Δx, Δy) 0.41 mm, 0.51 mm
Number of data points collected (Nx, Ny) 320, 256
Slice thickness (Δz) 4 mm
Number of slices 14
Slice gap 1 mm
Number of acquisitions (Nacq) 2
Tailored RF Yes Figure A23.2.5 Coronal spin-echo T1-weighted image with fat saturation, with intravenous gad-
No phase wrap (NPW) Yes olinium.
Fat suppression Yes; frequency selective fat saturation
Scan time 4 min, 20 sec

MRI of the
Post-Operative
Meniscus and
ACL Graft Knee

A23.2.6 A23.2.7
Supplement 10 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 10
 Table A23.2.5 Primary Clinical Imaging Parameters for Sequence 4 Sequence 1: Transverse fast spin-echo T2-weighted sequence with fat saturation
(localizer)
Patient position Supine 11. To validate the patient’s position, run the localizer (scout scan) to ensure the correct
Scan type Fast spin echo location of the knee and plan the perpendicular (sagittal and coronal) orientations
Imaging plane (orientation) Coronal using the imaging sequence given in Table A23.2.2, or similar parameters.
Variable bandwidth Yes
Central slice or volume center Laser light centered on the lower third An example image is shown in Figure A23.2.1.
of the patella
Echo time (TE) 42 msec Sequence 2: Sagittal spin-echo T1-weighted sequence with fat saturation
Receiver bandwidth (RBW) ±11.36 kHz 12. Bring the sagittal spin-echo T1-weighted sequence with fat saturation up onto the
Echo train length (ETL) 10 console. Set the imaging parameters as shown in Table A23.2.3.
Repeat time (TR) 3000 msec 13. Let the patient know you are ready, and begin the scan.
Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 130 mm, 130 mm Example images are shown in Figures A23.2.2 and A23.2.3.
Resolution (Δx, Δy) 0.51 mm, 0.51 mm
Number of data points collected (Nx, Ny) 256, 256
Sequence 3: Coronal spin-echo T1-weighted sequence with fat saturation
Slice thickness (Δz) 4 mm
14. Bring the coronal spin-echo T1-weighted sequence with fat saturation up onto the
Number of slices 14
console. Set the imaging parameters as shown in Table A23.2.4.
Slice gap 1 mm 15. Let the patient know you are ready, and begin the scan.
Number of acquisitions (Nacq) 2
Tailored RF Yes Example images are shown in Figures A23.2.4 and A23.2.5.
No phase wrap (NPW) Yes
Sequence 4: Coronal fast spin-echo T2-weighted sequence with fat saturation
Fat suppression Yes; frequency selective fat saturation
16. Bring the coronal fast spin-echo T2-weighted sequence with fat saturation up onto
Scan time 2 min, 42 sec
the console. Set the imaging parameters as shown in Table A23.2.5.
17. Let the patient know you are ready, and begin the scan.
An example image is shown in Figure A23.2.6.

COMMENTARY
Background Information
The knee is one of the most frequently in-
jured regions of the body. Injury to the knee
menisci and ligaments are common especially
in the younger athletic population. Substantial
progress has been made in magnetic resonance
imaging since its initial application in 1984.
MRI is established as the diagnostic procedure
of choice, supporting the physical examination
and plain X-ray studies for virtually all sus-
pected disorders of the knee. Moreover MRI
may serve as the “second” opinion before any
surgical intervention.
Imaging of the postoperative menisci
Figure A23.2.6 Coronal fast spin echo T2-weighted image with fat saturation. Difficulty in the image evaluation of the
postoperative meniscus comes from multiple
factors. The assessment is made more difficult
by the spectrum of possible postoperative find-
ings in menisci.
The meniscal morphology is distorted after
a partial meniscectomy. Shape irregularity may
MRI of the
Post-Operative be misinterpreted as a recurrent meniscal tear
Meniscus and because, after an injury, the meniscus may
ACL Graft
never gain its original preinjury signal intensity
on the T1- and proton density–weighted images
(Arnoczky et al., 1994).
Furthermore a healing area within the me-
niscus may appear as a high signal alteration
line reaching the articular surface, and may
therefore be misinterpreted as a new meniscal
tear (Arnoczky et al., 1994). For that reason one
of the most reliable parameters in an acute
clinical setting—increased signal extending to
the meniscal surface on T1-weighted or proton
density–weighted images—is not as valid in the
postoperative knee. An approach to overcome
these difficulties after meniscus surgery is to
administer intra-articular gadolinium in the
knee. (Sciulli et al., 1999), which offers several
advantages compared to the imaging without
contrast media. On conventional T2-weighted
MR images, joint fluid usually has the same
signal intensity as granulation tissue and fi-
brovascular scar. For that reason, healing tears
are sometimes indistinguishable from recurrent
tears in the meniscal remnant. On T1-weighted
direct-arthrographic images, contrast solution
Knee

A23.2.8 A23.2.9
Supplement 10 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 10
 is not mistaken for scar tissue, because its signal
intensity is significantly greater as shown in a
study by Palmer (1996).
Gadolinium-based contrast material has a
lower viscosity compared to synovial fluid and
is therefore more likely to be imbibed into a
small cleft and highlight the presence of a
meniscal tear. Additionally, the use of gadolin-
ium-based contrast material provides imaging
with T1-weighted pulse sequences, with their
inherently favorable signal-to-noise ratio. The
increased intra-articular pressure that results
from applying the contrast medium provides a
distension of normally apposed structures, such
as edges of a nondisplaced meniscal tear (Sci-
ulli et al., 1999).
Imaging of the postoperative anterior
cruciate ligament (ACL)
The appearance of an intact ACL graft has
been the subject of multiple studies (Autz et al.,
1991; Howell et al., 1991; Rak et al., 1991;
Cheung et al., 1992; Yamato et al., 1992; May-
wood et al., 1993; Sanchis-Alfonso et al., 1993;
Stockle et al., 1998). In these studies, it is
commonly accepted that in the immediate post-
operative period the graft is of low signal inten-
sity, comparable to that of the patellar tendon.
Another consensus is that intact grafts which
are more than two years old are low in signal
intensity. However, in the perioperative period
the appearance of an intact graft is less well
defined. Some studies have stated that these
grafts could appear low in signal intensity or
attenuated with increased signal intensity. The
increased signal intensity was considered to be
the result of a graft vascularization. Another
study applying intravenous contrast material in
nonimpinged ACL grafts did not show any
revascularization of the graft within the first
two postoperative years, whereas other studies
showed increased signal intensity within the
graft in the first two postoperative years.
Despite the discrepant reports of the appear-
ance on an intact, nonimpinged ACL, one can
draw several conclusions. A homogenous ACL
graft of low signal intensity in the first two
postoperative years is intact. After the periop-
erative period, increased signal within the graft
may be seen within an unimpinged healthy
graft, an intact impinged graft, and disrupted
grafts.
The indication to give gadopentate dimeglu-
mine in this context is to help distinguish an
MRI of the ACL graft from the enhancing periligamentous
Post-Operative
Meniscus and tissue on T1-weighted images with fat satura-
ACL Graft tion.
A23.2.10
Supplement 10
Critical Parameters and
Troubleshooting
Magnetic resonance imaging of the injured
knee is a potent and noninvasive tool. The
suggested imaging method provides an ap-
proach for acquiring a standard set of images
that allows a thorough analysis. Even though a
reliable protocol has been chosen, some arti-
facts may potentially occur during the acquisi-
tion. The most common and important artifacts
that may pose diagnostic difficulties are out-
lined in the following paragraphs, together with
possible solutions.
Magic angle artifact
Near the notch of the knee, the posterior horn
of the lateral meniscus slopes upward. This
approximates 55° of the external magnetic field
and can lead to the magic angle artifact or
diffuse increased signal intensity (Peterfy et al.,
1994). To overcome the problem, the knee can
be imaged in a slightly different position (the
authors choose abduction) in order to confi-
dently exclude meniscal tears.
Truncation artifact
This artifact results from undersampling of
data, so that interfaces of high and low signal
are incorrectly represented on the image. If
utilizing a small matrix, the truncation artifact
can result in an increased signal intensity
through the meniscus which may appear as a
tear. Using a larger acquisition matrix (number
of data points collected) can minimize the arti-
fact.
Motion artifact
Alternating increased and decreased signal
lines occur with motion and can mimic menis-
cal tear (Mirowitz, 1994). If this is observed it
is highly recommended that the affected por-
tion of the examination be repeated and the
patient be asked to hold still. Alternatively, mild
sedation and anxiolysis with benzodiazepine
(e.g., Lorasepam) is possible.
Vacuum phenomenon and ferromagnetic
substances
Magnetic susceptibility of intra-articular
gas and ferromagnetic substances may produce
a low-signal-intensity void, or blooming, espe-
cially on gradient echo images. This artifact
may be mistaken for a meniscal tear or articular
injury. Normal spin-echo or fast spin-echo se-
quences are much less vulnerable to these arti-
facts because of the 180° refocusing pulse.
Current Protocols in Magnetic Resonance Imaging
Pulsation artifact
Popliteal artery pulsation artifacts lead to
streaks in the MR image. They can be mini-
mized by swapping the phase-encoding and
read directions.
Anticipated Results
The goal in studying the postoperative knee
is the detection of clinically suspected abnor-
malities in the follow-up that could lead to an
exacerbation by repeat trauma and/or delay a
surgical revision.
Routine MR imaging is not reliable for de-
tecting meniscal re-tears. Applegate et al.
(1993) showed in a series study of 37 patients,
who had either meniscectomy or meniscal re-
pair and were evaluated with both standard MR
imaging and with MR arthrography, that the
overall accuracy in the postoperative meniscus
was 66% when conventional MR imaging was
used and 88% with MR arthrography. In pa-
tients with an extensive meniscal resection,
accuracy was 65% with conventional MRI and
87% with MR arthrography. In patients with
only minimal meniscal resection, both methods
provided an accuracy of 89%. When only a
small meniscal remnant remained, the accuracy
was 50% with routine MR imaging and 100%
with MR arthrography.
Sciulli et al. (1999) studied a series of 33
patients after meniscal surgery with four differ-
ent imaging modalities: conventional
arthrography, conventional MR imaging, MR
arthrography with iodinated contrast material,
and MR arthrography with gadolinium-based
contrast material. Twelve patients underwent
repeat arthroscopy in their follow-up. The gad-
olinium-enhanced arthrograms provided the
most accurate diagnosis when compared with
the arthroscopic images. The evaluation of the
postoperative meniscus compared to the gold
standard (i.e, arthroscopy) was correct in 92%,
using MR arthrography with gadolinium-based
contrast media, 77% by conventional MR ex-
amination, 75% by MR arthrography, and 58%
by conventional arthrography.
These studies support the experiences
gained with the protocol described in this unit
and emphasize the importance of using MR
arthrography with gadolinium-based contrast
material in patients who have had significant
meniscal resections.
As described above (see Background Infor-
mation), the MRI evaluation of the postopera-
tive ACL was the subject of several different
studies revealing discrepant results with regard
to the MRI appearance of the graft during
Current Protocols in Magnetic Resonance Imaging
follow-up. The reason for giving intravenous
gadopentate dimeglumine is that it helps dis-
tinguish an ACL graft from the enhancing per-
iligamentous tissue on T1-weighted images
with fat saturation, thus allowing confirmation
of its intactness. This approach is supported by
a study from Stockle et al. (1998), who showed
that in 50% of their patients an accurate assess-
ment of a reconstructed ligament was only
possible in the post-contrast images.
Literature Cited
Applegate, G.R., Flannigan, B.D., Tolin, B.S., Fox ,
J.M., and Del Pizzo, W. 1993. MR diagnosis of
recurrent tears in the knee: Value of intraarticular
contrast material. AJR Am. J. Roentgenol.
161:821-825.
Arnoczky, S.P., Cooper, T.G., Stadelmaier, D.M.,
and Hannafin, J.A. 1994. Magnetic resonance
signals in healing menisci: An experimental
study in dogs. Arthroscopy 10:552-557.
Autz, G., Goodwin, C., and Singson, R.D. 1991.
Magnetic resonance evaluation of anterior cruci-
ate ligament repair using the patellar tendon
double bone block technique. Skeletal Radiol.
20:585-588.
Carmichael, I.W., MacLeod, A.M., and Travlos, J.
1997. MRI can prevent unnecessary arthroscopy.
J. Bone Joint Surg. Br. 79:624-625.
Cheung, Y., Magee, T.H., Rosenberg, Z.S., and
Rose, D.J. 1992. MRI of anterior cruciate liga-
ment reconstruction. J. Comput. Assist. Tomogr.
16:134-137.
Howell, S.M., Berns, G.S., and Farley, T.E. 1991.
Unimpinged and impinged anterior cruciate liga-
ment grafts: MR signal intensity measurements.
Radiology 179:639-643.
Lee, J.K., Yao, L., Phelps, C.T., Wirth, C.R., Czajka,
J., and Lozman, J. 1988. Anterior cruciate liga-
ment tears: MR imaging compared with arthro-
scopy and clinical tests. Radiology 166:861-864.
Maurer, E.J., Kaplan, P.A., Dussault, R.G., Diduch,
D.R., Schuett, A., McCue, F.C., Hornsby, P.P.,
and Hillman, B.J. 1997. Acutely injured knee:
Effect of MR imaging on diagnostic and thera-
peutic decisions. Radiology 204:799-805.
Maywood, R.M., Murphy, B.J., Uribe, J.W., and
Hechtman, K.S. 1993. Evaluation of arthro-
scopic anterior cruciate ligament reconstruction
using magnetic resonance imaging. Am. J. Sports
Med. 21:523-527.
Mirowitz, S.A. 1994. Motion artifact as a pitfall in
diagnosis of meniscal tear on gradient reoriented
MRI of the knee. J. Comput. Assist. Tomogr.
18:279-282.
Palmer, W.E. 1996. MR arthrography: Is it worth-
while? Top. Magn. Reson. Imaging 8:24-43.
Peterfy, C.G., Janzen, D.L., Tirman, P.F., van Dijke,
C.F., Pollack, M., and Genant, H.K. 1994.
“Magic-angle” phenomenon: A cause of in-
creased signal in the normal lateral meniscus on
Knee
A23.2.11
Supplement 10
 short-TE MR images of the knee. AJR Am. J.
Roentgenol. 163:149-154.
Rak, K.M., Gillogly, S.D., Schaefer, R.A., Yakes,
W.F., and Liljedahl, R.R. 1991. Anterior cruciate
ligament reconstruction: Evaluation with MR
imaging. Radiology 178:553-556.
Sanchis-Alfonso, V., Martinez-Sanjuan, V., and
Gastaldi-Orquin, E. 1993. The value of MRI in
the evaluation of the ACL deficient knee and in
the post-operative evaluation after ACL recon-
struction [published erratum appears in Eur. J.
Radiol. 16:255]. Eur. J. Radiol. 16:126-130.
Sciulli, R.L., Boutin, R.D., Brown, R.R., Nguyen,
K.D., Muhle, C., Lektrakul, N., Pathria, M.N.,
Pedowitz, R., and Resnick, D. 1999. Evaluation
of the postoperative meniscus of the knee: A
study comparing conventional arthrography,
conventional MR imaging, MR arthrography
with iodinated contrast material, and MR
arthrography with gadolinium-based contrast
material. Skeletal Radiol. 28:508-514.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
MRI of the
Post-Operative
Meniscus and
ACL Graft
A23.2.12
Supplement 10
Stockle, U., Hoffmann, R., Schwedke, J., Lubrich,
J., Vogl, T., Sudkamp, N.P., and Haas, N. 1998.
Anterior cruciate ligament reconstruction: The
diagnostic value of MRI. Int. Orthop. 22:288-
292.
Yamato, M. and Yamagishi, T. 1992. MRI of patellar
tendon anterior cruciate ligament autografts. J.
Comput. Assist. Tomogr. 16:604-607.
Contributed by Rolf W. Huegli
Osteoporosis and Arthritis Research Group
University of California San Francisco
San Francisco, California
Phillip F.J. Tirman
National Orthopaedic Imaging Associates
Greenbrae, California
Margie Garbarino
California Pacific Medical Center
San Francisco, California
Current Protocols in Magnetic Resonance Imaging
Bone Stress Injuries UNIT A24.1
Because of their insidious onset and overlap of clinical signs and symptoms, stress
fractures and other bone stress injuries present a diagnostic challenge for clinicians.
Magnetic resonance imaging (MRI) can be used to efficiently detect and grade stress
fractures and other bone stress injuries, allowing appropriate and timely treatment
(Fredericson et al., 1995). This unit outlines basic MR protocols for evaluating the most
common bone stress injuries including those of the tibia, femoral neck, femoral shaft,
metatarsal, and navicular bone. The protocols concentrate on the lower extremity, as the
majority of stress injures occur in this region (Hulkko and Orava, 1987; Matheson et al.,
1987). These protocols were developed using a 1.5 T system (Signa, General Electric
Medical Systems). Detection of subtle changes in bone marrow or periosteal signal
(edema) is essential in the evaluation of stress injury. By experience, the authors have
found high field strength systems to be excellent in this regard, therefore, are preferred
over low and mid field strength magnets for this application. However, the sequences
described here could be adapted to low or mid field scanners using STIR (short tau
inversion recovery) sequences to acquire the T2-weighted images. This technique would
require longer scan times and lower resolution matrices.
TIBIAL STRESS INJURIES BASIC
PROTOCOL 1
Stress injuries are attributed to running more often than any other athletic activity, and
the tibia is the most common site of stress injury among runners (Hulkko and Orava, 1987;
Matheson et al., 1987). Patients usually complain of pain with activity, which worsens
over time if the activity is continued. Eventually, pain may occur with routine ambulation
or at rest (Fredericson et al., 1995). MRI evaluation is focused on the area of signs and
symptoms.
Four primary sequences are utilized for this protocol: (1) coronal T1-weighted spin echo;
(2) coronal T2-weighted fast spin echo (FSE); (3) transverse T1-weighted spin echo; and
(4) transverse T2-weighted FSE. The approximate amount of time needed to complete this
protocol is 30 min.
Table A24.1.1 outlines the equipment needed to perform this examination. Cardiac and
respiratory gating are not used and contrast agents are not needed.
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.
Set up patient and equipment
1. Interview the patient to ensure that he or she has no contraindications such as cardiac
pacemakers or other implants containing ferromagnetic materials. Also, be sure to
find out if the patient has any health conditions that may require the presence of
Table A24.1.1 Equipment Parameters for Stress Injuries
Coil type Extremity or torso phased array
Gradient coil strength 22 mT/m (or whatever the system permits)
Cardiac gating No
Respiratory gating No
Use of contrast agents No Musculoskeletal
Stress Injuries
Contributed by Gabrielle Bergman and Timothy R. Jones A24.1.1
Current Protocols in Magnetic Resonance Imaging (2001) A24.1.1-A24.1.21
Copyright © 2001 by John Wiley & Sons, Inc. Supplement 3
 special emergency equipment during the scanning procedure, or necessitate any other Sequence 1: Coronal T1-weighted spin echo
precautions. 11. Access the spin echo sequence on the system console and set the imaging parameters
as shown in Table A24.1.2.
Generally, standard screening forms (APPENDIX 1) are used for all patients scanned in a
magnetic resonance system. 12. Inform the patient that the scan is about to begin.
The presence of any ferromagnetic metals may be a health hazard to the patient when he 13. Start the scan.
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock This sequence also serves as the scout scan. A separate scout sequence is not necessary.
(1996) for discussion of what implants may be safely scanned using magnetic resonance. Care should be taken that imaging is performed through the entire antero-posterior
dimension of the calf.
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing. Table A24.1.2 Tibial Stress Injuries: Imaging Parameters for Sequence 1
2. If the procedure is a research protocol, have the patient sign any necessary consent
Patient position Supine
forms.
Scan type Spin echo
3. Have the patient remove all jewelry and change into a gown to eliminate any metal Imaging plane (orientation) Coronal
that might be found in clothing. Central slice or volume center Centered on antero-posterior center of
the calf
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating Echo time (TE) “Minimum full” (usually ∼15 msec)
and image artifacts. Repeat time (TR) 600 msec
Flip angle (FA) 90°
5. Inform the patient about what will occur during the procedure, what he or she will Fields of view (FOVx, FOVy) 160–200 mm, 160–200 mm
experience while in the magnet, and how to behave, including the following: Resolution (Δx, Δy) 0.31–0.39 mm, 0.83–1.04 mm
Number of data points collected (Nx, Ny) 512, 192
a. If earphones or headphones are used to protect the ears from the loud sounds
Display matrix (Dx, Dy) 512, 512
produced by the gradients, the patient will be asked to wear these, but will be able Slice thickness (Δz) 4 mm
to communicate with you at any time during the imaging. Number of slices Variable (enough to cover the entire
b. The patient will be given a safety squeeze-bulb or similar equipment to request tibia from anterior to posterior)
assistance at any time (demonstrate how this works). Slice gap 1 mm
Number of acquisitions (Nacq) 2
c. For good results, the patient may breathe in a normal, relaxed manner but should
Saturation pulses None
avoid or minimize any other movement during each scan, i.e., as long as the Scan time 3 min, 54 sec
banging sounds continue. Between scans, the patient should not change the
position of the region of the body being imaged.
d. Nevertheless, the patient may call out at any time if he or she feels it necessary. Table A24.1.3 Tibial Stress Injuries: Imaging Parameters for Sequence 2

6. Have the patient lie on the table in the supine position. Patient position Supine
Scan type Fast spin echo
7. Position the extremity coil around the symptomatic calf. Imaging plane (orientation) Coronal
Have the patient point with one finger to the area of most intense pain and place a marker Central slice or volume center Centered on antero-posterior center of
on this point. Center the coil about the marker. If the point of pain is extremely superior or the calf
inferior, the coil may be placed such that the superior or inferior half of the tibia is covered. Echo time (TE) 70 msec
Echo train length (ETL) 8
8. If needed, place a pillow or other support under the knees to make the patient more Repeat time (TR) 4000 msec
comfortable. Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 160–200 mm, 160–200 mm
9. Use the laser light to point the center of the coil and put the patient into the center of Resolution (Δx, Δy) 0.31–0.39 mm, 0.83–1.04 mm
the magnet. Number of data points collected (Nx, Ny) 512, 192
Once this step has been performed, so long as the patient does not move on the table, the Display matrix (Dx, Dy) 512, 512
table itself can be moved and then replaced in the same position as before without Slice thickness (Δz) 4 mm
jeopardizing the positioning of one scan relative to another. Number of slices Variable (enough to cover the entire
tibia from anterior to posterior)
10. If the patient is unable to hold still, provide an appropriate sedative. Slice gap 1 mm
Number of acquisitions (Nacq) 2
Saturation pulses Frequency-selective fat saturation
Bone Stress Scan time 3 min, 28 sec Musculoskeletal
Injuries Stress Injuries

A24.1.2 A24.1.3
Supplement 3 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 3
 Table A24.1.4 Tibial Stress Injuries: Imaging Parameters for Sequence 3 Sequence 2: Coronal T2-weighted fast spin echo (FSE)
14. Call up the menu for the fast spin echo sequence on the console. Use the parameters
Patient position Supine seen in Table A24.1.3.
Scan type Spin echo
Imaging plane (orientation) Transverse 15. Run the sequence.
Central slice or volume center Centered on the marker
Echo time (TE) “Minimum full” (usually ∼15 msec) Sequence 3: Transverse T1-weighted spin echo
Repeat time (TR) 600 msec 16. Once again go to the spin echo sequence on the system console and set the imaging
Flip angle (FA) 90° parameters as shown in Table A24.1.4.
Fields of view (FOVx, FOVy) 140–180 mm, 140–180 mm (as small 17. Instruct the patient to remain still and run scan.
as possible but including the entire calf
diameter)
Sequence 4: Transverse T2-weighted fast spin echo
Resolution (Δx, Δy) 0.27–0.35 mm, 0.73–0.94 mm
18. Set up for a fast spin echo scan using the parameters outlined in Table A24.1.5.
Number of data points collected (Nx, Ny) 512, 192
Display matrix (Dx, Dy) 512, 512 19. Run the sequence.
Slice thickness (Δz) 4 mm
Number of slices ∼32–40 (image 8–10 cm FEMORAL NECK AND FEMORAL SHAFT STRESS INJURIES BASIC
approximately and same distance PROTOCOL 2
distal to the point of maximum Stress fractures of the femur are relatively common injuries in which MR imaging can be
tenderness as indicated by the point) extremely useful. Femoral neck stress injuries deserve special attention. They can
Slice gap 1 mm progress to displaced fractures with associated complications such as avascular necrosis
Number of acquisitions (Nacq) 2 and irreversible joint damage if not promptly diagnosed and properly treated. MRI is more
Saturation pulses None sensitive than radiography and more specific than radionuclide bone scan in evaluation
Scan time 3 min, 54 sec of proximal femoral abnormalities (Shin et al., 1996). It allows for differentiation between
stress fracture and other causes of hip pain such as avascular necrosis, symptomatic
synovial herniation pit, iliopsoas bursitis, iliopsoas or obturator externus tendinitis, or
Table A24.1.5 Tibial Stress Injuries: Imaging Parameters for Sequence 4 iliopsoas muscle tear (Bergman and Fredericson, 1999).
Patient position Supine MR imaging is also useful in the evaluation of possible femoral diaphyseal injuries. Stress
Scan type Fast spin echo fractures in this location may commonly be misdiagnosed as muscle injuries (Bergman
Imaging plane (orientation) Transverse and Fredericson, 1999). MRI, by evaluating both the bone and soft tissues, is effective in
Central slice or volume center Centered on the marker differentiating these abnormalities.
Echo time (TE) 70 msec
Echo train length (ETL) 8 Optimal evaluation of the femur includes T1- and T2-weighted imaging in the coronal
Repeat time (TR) 4150 msec (Figures A24.1.1 through A24.1.4) and transverse planes. This protocol requires ∼30 min
Flip angle (FA) 90° to complete.
Fields of view (FOVx, FOVy) 140–180 mm, 140–180 mm (as small
as possible but including the entire calf Sequence 5: Coronal T1-weighted spin echo
diameter) 1. Screen and prepare the patient as described in Basic Protocol 1, steps 1 to 6.
Resolution (Δx, Δy) 0.27–0.35 mm, 0.73–0.94 mm
2. With the patient in the supine position, center the phased array or torso (not extremity)
Number of data points collected (Nx, Ny) 512, 192
coil at the level of the greater trochanter if femoral neck fracture suspected or at the
Display matrix (Dx, Dy) 512, 512
level of maximum pain if the femoral shaft is the area of interest. For the coronal
Slice thickness (Δz) 4 mm
sequences, include both femurs in the field of view. Place a marker at the area of
Number of slices ∼32–40 (image 8–10 cm
approximately and same distance maximum pain. Repeat Basic Protocol 1, steps 8 to 10.
distal to the point of maximum 3. Run the scan according to the imaging parameters listed in Table A24.1.6.
tenderness as indicated by the point)
Slice gap 1 mm This sequence also serves as the scout scan. A separate scout sequence is not necessary.
Number of acquisitions (Nacq) 2 Care should be taken that imaging is performed through the entire antero-posterior
Saturation pulses Frequency-selective fat saturation dimension of the thigh.
Scan time 3 min, 36 sec
Sequence 6: Coronal T2-weighted fast spin echo
4. Scan the patient according to the parameters in Table A24.1.7.
Bone Stress Musculoskeletal
Injuries Stress Injuries

A24.1.4 A24.1.5
Supplement 3 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 3
 Figure A24.1.3 T1-weighted coronal image of the pelvis and proximal femurs in a 37-year-old
Figure A24.1.1 T1-weighted coronal image of the pelvis and proximal femurs in a 17-year-old distance runner with a femoral diaphyseal stress fracture.
distance runner with a femoral neck stress fracture.

Figure A24.1.4 T2-weighted coronal image of the pelvis and proximal femurs in a 37-year-old
distance runner with a femoral diaphyseal stress fracture.
Bone Stress Figure A24.1.2 T2-weighted coronal image of the pelvis and proximal femurs in a 17-year-old Musculoskeletal
Injuries distance runner with a femoral neck stress fracture. Stress Injuries

A24.1.6 A24.1.7
Supplement 3 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 3
 Table A24.1.6 Femoral Stress Injuries: Imaging Parameters for Sequence 5 Table A24.1.8 Femoral Stress Injuries: Imaging Parameters for Sequence 7

Patient position Supine Patient position Supine

Scan type Spin echo Scan type Spin echo
Imaging plane (orientation) Coronal Imaging plane (orientation) Transverse
Central slice or volume center Centered antero-posteriorly on the Central slice or volume center Centered on the greater trochanter or
greater trochanter (femoral neck the marker
fracture) or femoral shaft (diaphyseal Echo time (TE) “Minimum full” (usually ∼15 msec)
fracture) Repeat time (TR) 600 msec
Echo time (TE) “Minimum full” (usually ∼15 msec) Flip angle (FA) 90°
Repeat time (TR) 600 msec Fields of view (FOVx, FOVy) 200–240 mm, 200–240 mm (as small
Flip angle (FA) 90° as possible)
Fields of view (FOVx, FOVy) 320 mm, 320 mm Resolution (Δx, Δy) 0.39–0.47 mm, 1.04–1.25 mm
Resolution (Δx, Δy) 0.63 mm, 1.67 mm Number of data points collected (Nx, Ny) 512, 192
Number of data points collected (Nx, Ny) 512, 192 Display matrix (Dx, Dy) 512, 512
Display matrix (Dx, Dy) 512, 512 Slice thickness (Δz) 4 mm
Slice thickness (Δz) 4 mm Number of slices ∼32–40 (see Table A24.1.4)
Number of slices Variable (enough to cover the entire Slice gap 1 mm
thigh from anterior to posterior) Number of acquisitions (Nacq) 2
Slice gap 1 mm Saturation pulses None
Number of acquisitions (Nacq) 2 Scan time 7 min, 48 sec
Saturation pulses None
Scan time 3 min, 54 sec

Table A24.1.9 Femoral Stress Injuries: Imaging Parameters for Sequence 8

Table A24.1.7 Femoral Stress Injuries: Imaging Parameters for Sequence 6
Patient position Supine
Patient position Supine Scan type Fast spin echo
Scan type Fast spin echo Imaging plane (orientation) Transverse
Imaging plane (orientation) Coronal Central slice or volume center Centered on the greater trochanter or
Central slice or volume center Centered antero-posteriorly on the the marker
greater trochanter (femoral neck Echo time (TE) 70 msec
fracture) or femoral shaft (diaphyseal Echo train length (ETL) 8
fracture) Repeat time (TR) 4000 msec
Echo time (TE) 70 msec Flip angle (FA) 90°
Echo train length (ETL) 8 Fields of view (FOVx, FOVy) 200–240 mm, 200–240 mm (as small
Repeat time (TR) 4000 msec as possible)
Flip angle (FA) 90° Resolution (Δx, Δy) 0.39–0.47 mm, 1.04–1.25 mm
Fields of view (FOVx, FOVy) 320 mm, 320 mm Number of data points collected (Nx, Ny) 512, 192
Resolution (Δx, Δy) 0.63 mm, 1.67 mm Display matrix (Dx, Dy) 512, 512
Number of data points collected (Nx, Ny) 512, 192 Slice thickness (Δz) 4 mm
Display matrix (Dx, Dy) 512, 512 Number of slices ∼32–40 (see Table A24.1.4)
Slice thickness (Δz) 4 mm Slice gap 1 mm
Number of slices Variable (enough to cover the entire Number of acquisitions (Nacq) 2
thigh from anterior to posterior) Saturation pulses Frequency-selective fat saturation
Slice gap 1 mm Scan time 7 min, 12 sec
Number of acquisitions (Nacq) 2
Saturation pulses Frequency-selective fat saturation
Scan time 3 min, 44 sec

Bone Stress Musculoskeletal

Injuries Stress Injuries

A24.1.8 A24.1.9
Supplement 3 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 3
 Sequence 7: Transverse T1-weighted spin echo
5. Perform scan in the transverse plane according to the parameters in Table A24.1.8.

Sequence 8: Transverse T2-weighted fast spin echo

6. Run the scan in the transverse plane according to the parameters in Table A24.1.9.

BASIC IMAGING OF METATARSAL STRESS INJURIES

PROTOCOL 3
Metatarsal stress fractures or “march fractures” are commonly seen in military recruits.
They most often occur in the distal diaphyses of the second and third metatarsals. Although
radiography and bone scanning can be useful in these patients, MR imaging demonstrates
evidence of stress injury and stress fracture well. These are not critical fractures and
activity can usually be resumed as soon as there is no pain with ambulation.
Imaging of these injuries is performed with T1- and T2-weighted sequences the coronal,
sagittal, and transverse planes (Figures A24.1.5 to A24.1.7). The authors prefer to use an
inversion recovery (IR) FSE technique as the T2-weighted sequence in this region.
Suppression of fat signal with this sequence is superior to the standard T2-weighted FSE
with fat saturation because it is not affected by the local magnetic field inhomogeneity in
the region of the foot. This protocol requires ∼40 min to complete.

Sequence 9: Coronal T1-weighted spin echo

1. Screen and prepare the patient as described in Basic Protocol 1, steps 1 to 6.
2. Position the extremity coil with the center over the symptomatic forefoot. Repeat
Basic Protocol 1, steps 8 to 10. Figure A24.1.5 T1-weighted coronal image of the foot in a patient with a metatarsal stress fracture.
3. Run the scan according to the parameters in Table A24.1.10.
This sequence also serves as the scout scan. A separate scout sequence is not necessary.
Care should be taken that imaging is performed through the entire antero-posterior
dimension of the foot.
Note that the foot is usually best positioned with a right angle at the ankle, as if standing
up with the full planar surface on the ground (not with a pointed foot). With the foot in that
position, the imaging planes (transverse, coronal, and sagittal) of the foot will be the same
as in the body, projected down to the foot (see Figure A24.1.8).

Sequence 10: Coronal T2-weighted inversion recovery fast spin echo

4. Run the scan according to the parameters in Table A24.1.11.

Sequence 11: Sagittal T1-weighted spin echo

5. Scan the patient in the sagittal plane according to the parameters in Table A24.1.12.
Sequence 12: Sagittal T2-weighted inversion recovery fast spin echo
6. Scan in the sagittal plane according to the parameters in Table A24.1.13.
Sequence 13: Transverse T1-weighted spin echo
7. Perform scan in the transverse plane according to the parameters in Table A24.1.14.
Sequence 14: Transverse T2-weighted inversion recovery fast spin echo
8. Perform another scan in the transverse plane according to the parameters in Table
A24.1.15.

Bone Stress Figure A24.1.6 T2-weighted coronal image of the foot in a patient with a metatarsal stress fracture. Musculoskeletal
Injuries Stress Injuries

A24.1.10 A24.1.11
Supplement 3 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 3
 transverse transverse plane

coronal plane

table

Figure A24.1.8 With the foot positioned at right angle to the body, the transverse and coronal
planes as used in this text are in the same orientation as in the body.

Table A24.1.11 Metatarsal Stress Injuries: Imaging Parameters for Sequence 10

Patient position Supine

Scan type Inversion recovery fast spin echo
Imaging plane (orientation) Coronal
Central slice or volume center Centered on the antero-posterior
midpoint of the forefoot
Echo time (TE) 60 msec
Echo train length (ETL) 8
Repeat time (TR) 6200 msec
Figure A24.1.7 T1-weighted sagittal image of the foot in a patient with a metatarsal stress fracture.
Inversion time (TI) 160 msec
Flip angle (FA) 90°
Table A24.1.10 Metatarsal Stress Injuries: Imaging Parameters for Sequence 9 Fields of view (FOVx, FOVy) 160–200 mm, 160–200 mm
Resolution (Δx, Δy) 0.31–0.39 mm, 0.83–1.04 mm
Patient position Supine Number of data points collected (Nx, Ny) 512, 192
Scan type Spin echo Display matrix (Dx, Dy) 512, 512
Imaging plane (orientation) Coronal Slice thickness (Δz) 4 mm
Number of slices ∼20–30
Central slice or volume center Centered on the antero-posterior
midpoint of the forefoot Slice gap 1 mm
Number of acquisitions (Nacq) 1
Echo time (TE) “Minimum full” (usually ∼15 msec)
Saturation pulses Inversion recovery fat suppression
Repeat time (TR) 600 msec
Scan time 5 min, 23 sec
Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 160–200 mm, 160–200 mm
Table A24.1.12 Metatarsal Stress Injuries: Imaging Parameters for Sequence 11
Resolution (Δx, Δy) 0.31–0.39 mm, 0.83–1.04 mm
Number of data points collected (Nx, Ny) 512, 192 Patient position Supine
Display matrix (Dx, Dy) 512, 512 Scan type Spin echo
Slice thickness (Δz) 4 mm Imaging plane (orientation) Sagittal
Number of slices Variable Central slice or volume center Centered on the third metatarsal
Slice gap 1 mm Echo time (TE) “Minimum full” (usually ∼15 msec)
Number of acquisitions (Nacq) 2 Repeat time (TR) 650 msec
Saturation pulses None Flip angle (FA) 90°
Scan time 3 min, 54 sec Fields of view (FOVx, FOVy) 160–200 mm, 160–200 mm
Resolution (Δx, Δy) 0.31–0.39 mm, 0.83–1.04 mm
Number of data points collected (Nx, Ny) 512, 192
Display matrix (Dx, Dy) 512, 512
Slice thickness (Δz) 3 mm
Number of slices Variable
Slice gap 1 mm
Number of acquisitions (Nacq) 2
Saturation pulses None
Scan time 4 min, 14 sec
Bone Stress Musculoskeletal
Injuries Stress Injuries

A24.1.12 A24.1.13
Supplement 3 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 3
 Table A24.1.13 Metatarsal Stress Injuries: Imaging Parameters for Sequence 12 Table A24.1.15 Metatarsal Stress Injuries: Imaging Parameters for Sequence 14

Patient position Supine Patient position Supine

Scan type Inversion recovery fast spin echo Scan type Inversion recovery fast spin echo
Imaging plane (orientation) Sagittal Imaging plane (orientation) Transverse (parallel to metatarsal long
Central slice or volume center Centered on the third metatarsal axis)
Echo time (TE) 60 msec Central slice or volume center Centered on the supero-inferior center
Echo train length (ETL) 8 of the metatarsals
Repeat time (TR) 6200 msec Echo time (TE) 60 msec
Inversion time (TI) 160 msec Echo train length (ETL) 8
Flip angle (FA) 90° Repeat time (TR) 5000 msec
Fields of view (FOVx, FOVy) 160–200 mm, 160–200 mm Inversion time (TI) 160 msec
Resolution (Δx, Δy) 0.31–0.39 mm, 0.83–1.04 mm Flip angle (FA) 90°
Number of data points collected (Nx, Ny) 512, 192 Fields of view (FOVx, FOVy) 160–200 mm, 160–200 mm
Display matrix (Dx, Dy) 512, 512 Resolution (Δx, Δy) 0.31–0.39 mm, 0.83–1.04 mm
Slice thickness (Δz) 3 mm Number of data points collected (Nx, Ny) 512, 192
Number of slices ∼20–25 Display matrix (Dx, Dy) 512, 512
Slice gap 1 mm Slice thickness (Δz) 3 mm
Number of acquisitions (Nacq) 1 Number of slices Variable
Saturation pulses Inversion recovery fat suppression Slice gap 1 mm
Scan time 8 min, 40 sec Number of acquisitions (Nacq) 2
Saturation pulses Inversion recovery fat suppression
Scan time 4 min, 20 sec
Table A24.1.14 Metatarsal Stress Injuries: Imaging Parameters for Sequence 13

Patient position Supine

Scan type Spin echo IMAGING OF NAVICULAR STRESS INJURIES BASIC
Imaging plane (orientation) Transverse (parallel to metatarsal long PROTOCOL 4
axis) MR imaging is often essential in the identification of navicular stress fractures. Radio-
Central slice or volume center Centered on the supero-inferior center graphs are not as sensitive as MRI for the early detection of this injury, which can be
of the metatarsals debilitating if not diagnosed promptly (Khan et al., 1994; Pavlov et al., 1983). They may
Echo time (TE) “Minimum full” (usually ∼15 msec) be treated surgically or nonsurgically depending on the degree of displacement.
Repeat time (TR) 600 msec
T1- and T2-weighted images are obtained in three planes (Figures A24.1.9 and A24.1.10).
Flip angle (FA) 90°
IR FSE technique is again preferred for the T2-weighted sequences because of superior
Fields of view (FOVx, FOVy) 160–200 mm, 160–200 mm
suppression of fat signal in this area of local magnetic field inhomogeneity. This protocol
Resolution (Δx, Δy) 0.31–0.39 mm, 0.83–1.04 mm
may require ∼40 min to complete.
Number of data points collected (Nx, Ny) 512, 192
Display matrix (Dx, Dy) 512, 512
Sequence 15: Coronal T1-weighted spin echo
Slice thickness (Δz) 3 mm
1. Screen and prepare the patient as described in Basic Protocol 1, steps 1 to 6.
Number of slices Variable
Slice gap 1 mm 2. Place the extremity coil on the symptomatic foot with the center over the hindfoot.
Number of acquisitions (Nacq) 2 Repeat Basic Protocol 1, steps 8 to 10.
Saturation pulses None
3. Perform a coronal scan according to the parameters in Table A24.1.16.
Scan time 3 min, 54 sec
This sequence also serves as the scout scan. A separate scout sequence is not necessary.
Care should be taken that imaging is performed through the entire antero-posterior
dimension of the foot.

Sequence 16: Coronal T2-weighted inversion recovery fast spin echo

4. Perform a coronal scan according to the parameters in Table A24.1.17.

Sequence 17: Sagittal T1-weighted spin echo

5. Scan the patient in the sagittal plane according to the parameters in Table A24.1.18.
Bone Stress Musculoskeletal
Injuries Stress Injuries

A24.1.14 A24.1.15
Supplement 3 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 3
 Table A24.1.16 Navicular Stress Injuries: Imaging Parameters for Sequence 15

Patient position Supine

Scan type Spin echo
Imaging plane (orientation) Coronal
Central slice or volume center Centered on the antero-posterior
midpoint of the hindfoot
Echo time (TE) “Minimum full” (usually ∼15 msec)
Repeat time (TR) 500 msec
Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 160–200 mm, 160–200 mm
Resolution (Δx, Δy) 0.31–0.39 mm, 0.83–1.04 mm
Number of data points collected (Nx, Ny) 512, 192
Display matrix (Dx, Dy) 512, 512
Slice thickness (Δz) 3 mm
Number of slices ∼25
Slice gap 1 mm
Number of acquisitions (Nacq) 2
Saturation pulses None
Scan time 6 min, 32 sec

Table A24.1.17 Navicular Stress Injuries: Imaging Parameters for Sequence 16

Patient position Supine

Figure A24.1.9 T1-weighted coronal image in a 19-year-old distance runner with a navicular Scan type Inversion recovery fast spin echo
stress fracture. Imaging plane (orientation) Coronal
Central slice or volume center Centered on the antero-posterior
midpoint of the hindfoot
Echo time (TE) 60 msec
Echo train length (ETL) 8
Repeat time (TR) 5000 msec
Inversion time (TI) 160 msec
Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 160–200 mm, 160–200 mm
Resolution (Δx, Δy) 0.31–0.39 mm, 0.83–1.04 mm
Number of data points collected (Nx, Ny) 512, 192
Display matrix (Dx, Dy) 512, 512
Slice thickness (Δz) 3 mm
Number of slices ∼25
Slice gap 1 mm
Number of acquisitions (Nacq) 2
Saturation pulses Inversion recovery fat suppression
Scan time 8 min, 40 sec

Bone Stress Figure A24.1.10 T2-weighted coronal image in a 19-year-old distance runner with a navicular Musculoskeletal
Injuries stress fracture. Stress Injuries

A24.1.16 A24.1.17
Supplement 3 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 3
 Table A24.1.18 Navicular Stress Injuries: Imaging Parameters for Sequence 17 Table A24.1.20 Navicular Stress Injuries: Imaging Parameters for Sequence 19

Patient position Supine Patient position Supine

Scan type Spin echo Scan type Spin echo
Imaging plane (orientation) Sagittal Imaging plane (orientation) Transverse
Central slice or volume center Centered on the medio-lateral center of Central slice or volume center Centered on the supero-inferior center
the hindfoot of the hindfoot
Echo time (TE) “Minimum full” (usually ∼15 msec) Echo time (TE) “Minimum full” (usually ∼15 msec)
Repeat time (TR) 650 msec Repeat time (TR) 650 msec
Flip angle (FA) 90° Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 160–200 mm, 160–200 mm Fields of view (FOVx, FOVy) 160–200 mm, 160–200 mm
Resolution (Δx, Δy) 0.31–0.39 mm, 0.83–1.04 mm Resolution (Δx, Δy) 0.31–0.39 mm, 0.83–1.04 mm
Number of data points collected (Nx, Ny) 512, 192 Number of data points collected (Nx, Ny) 512, 192
Display matrix (Dx, Dy) 512, 512 Display matrix (Dx, Dy) 512, 512
Slice thickness (Δz) 3 mm Slice thickness (Δz) 3 mm
Number of slices Variable Number of slices Variable
Slice gap 1 mm Slice gap 1 mm
Number of acquisitions (Nacq) 2 Number of acquisitions (Nacq) 2
Saturation pulses None Saturation pulses None
Scan time 4 min, 14 sec Scan time 4 min, 14 sec

Table A24.1.19 Navicular Stress Injuries: Imaging Parameters for Sequence 18 Table A24.1.21 Navicular Stress Injuries: Imaging Parameters for Sequence 20

Patient position Supine Patient position Supine

Scan type Inversion recovery fast spin echo Scan type Inversion recovery fast spin echo
Imaging plane (orientation) Sagittal Imaging plane (orientation) Transverse
Central slice or volume center Centered on the medio-lateral center of Central slice or volume center Centered on the supero-inferior center
the hindfoot of the hindfoot
Echo time (TE) 60 msec Echo time (TE) 60 msec
Echo train length (ETL) 8 Echo train length (ETL) 8
Repeat time (TR) 5000 msec Repeat time (TR) 5000 msec
Inversion time (TI) 160 msec Inversion time (TI) 160 msec
Flip angle (FA) 90° Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 160–200 mm, 160–200 mm Fields of view (FOVx, FOVy) 160–200 mm, 160–200 mm
Resolution (Δx, Δy) 0.31–0.39 mm, 0.83–1.04 mm Resolution (Δx, Δy) 0.31–0.39 mm, 0.83–1.04 mm
Number of data points collected (Nx, Ny) 512, 192 Number of data points collected (Nx, Ny) 512, 192
Display matrix (Dx, Dy) 512, 512 Display matrix (Dx, Dy) 512, 512
Slice thickness (Δz) 3 mm Slice thickness (Δz) 3 mm
Number of slices ∼20–25 Number of slices ∼20–25
Slice gap 1 mm Slice gap 1 mm
Number of acquisitions (Nacq) 2 Number of acquisitions (Nacq) 2
Saturation pulses Inversion recovery fat suppression Saturation pulses Inversion recovery fat suppression
Scan time 8 min, 40 sec Scan time 8 min, 40 sec

Sequence 18: Sagittal T2-weighted inversion recovery fast spin echo

6. Run the scan according to the parameters in Table A24.1.19.
Sequence 19: Transverse T1-weighted spin echo
7. Scan the patient according to the parameters in Table A24.1.20.

Sequence 20: Transverse T2-weighted inversion recovery fast spin echo

Bone Stress
8. Perform the scan in the transverse plane according to the parameters in Table Musculoskeletal
Injuries A24.1.21. Stress Injuries

A24.1.18 A24.1.19
Supplement 3 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 3
 COMMENTARY
Background Information
Musculoskeletal stress injuries are among
the most commonly encountered sports medi-
cine conditions, with stress fractures account-
ing for ∼10% of visits to sports medicine clinics
(Matheson et al., 1987). Bone stress injuries
represent a spectrum of abnormalities ranging
from early stress reaction to frank cortical frac-
ture. They often present a diagnostic challenge
for clinicians as differentiation between the
more benign stress reactions (such as shin splint
syndrome in the tibia) and more advanced stress
injury or fracture can be difficult. As this point
of distinction is essential to the proper manage-
ment of these patients, the ability of MR grad-
ing of stress lesion severity has proven to be
clinically useful (Lee and Yao, 1988; Bergman
and Fredericson, 1999).
Bony stress lesions develop when an area of
accelerated bone remodeling occurs in re-
sponse to repetitive stress. Bone resorption is
greater than new bone formation in this region,
leading to bone weakening and eventually to
trabecular microfractures (stress injury). With
further activity, a full cortical break or stress
fracture can occur (Daffner and Pavlov, 1992).
The location of injury depends on the activ-
ity undertaken by the patient as well as other
biomechanical factors. For example, runners
often develop injuries in the tibia, the femoral
neck or diaphysis, or the navicular bone
(Daffner and Pavlov, 1992). Much overlap ex-
ists however.
Stress injuries usually present as focal pain
that begins during the end of an exercise session
or soon afterwards. Eventually, the pain occurs
throughout the exercise period and then during
other daily activities as well. On physical ex-
amination, tenderness and swelling are noted
over the affected area on direct palpation.
While radiography remains the initial test in
screening individuals with possible stress inju-
ries, it is negative in two thirds of patients at the
onset of symptoms and remains negative in
approximately one half of these patients (Floyd
et al., 1987). Radionuclide bone scanning (bone
scintigraphy) is an alternative test that is more
effective than radiography at detecting early
stress injury. A progression of bone scan find-
ings has been associated with corresponding
evolution of stress injury. Bone scans demon-
strate diffuse uptake with an early stress reac-
tion, and more focal and intense uptake with a
stress fractures (Roub et al., 1979).
Bone Stress
Injuries
A24.1.20
Supplement 3
However, MR imaging is superior to bone
scintigraphy in several respects. Unlike MRI,
bone scans do not effectively evaluate the sur-
rounding soft tissues. As a result, unsuspected
soft tissue injuries can be missed. In addition,
a positive bone scan finding is not specific for
stress injury (Shin et al., 1996). Because a bone
scan simply indicates areas of increased new
bone formation, other conditions can mimic
fractures. Finally, Milgrom et al. (1984) suggest
that a negative bone scan does not always ex-
clude a stress injury.
Critical Parameters and
Troubleshooting
Although MR imaging can detect stress in-
juries causing frank cortical fractures, it is most
useful in diagnosing more subtle injuries. For
this, accurate detection of even very subtle
periosteal and bone marrow edema is essential
for this function. Thus, the authors propose that
high magnetic field strength systems should be
used for this examination, as they are superior
in this regard.
In addition, use of T2-weighted sequences
that provide optimal fat signal suppression is
also important so that false positive results are
not obtained. In the femurs and tibias, T2-
weighted FSE sequences with a fat saturation
pulse perform well and have greater signal-to-
noise than IR FSE sequences. However, in
regions with significant local magnetic field
inhomogeneity such as the ankle and foot, the
superiority of the fat signal suppression with
the IR FSE sequence outweighs its signal-to-
noise deficiency.
Maximum resolution is desired so that small
areas of edema are not missed. An extremity
coil is almost always preferred for this reason.
If the patient has bilateral symptoms, a torso
phased array or body coil may be used initially,
then with additional scans using an extremity
coil to follow, if an area of abnormality is
identified. It is also important to center the
imaging on the site of maximal pain by having
the patient point to the area with one finger and
to position him or her accordingly. To further
maximize the resolution, the acquisition matrix
(Nx, Ny) should be set to Nx = 512 in the
frequency dimension, if practical.
Anticipated Results
Magnetic resonance imaging (MRI) is con-
sidered by many to be the imaging modality of
Current Protocols in Magnetic Resonance Imaging
choice for the evaluation of bone stress injuries.
It provides a means of making an early diagno-
sis, and allowing appropriate treatment to be
initiated by accurately staging the lesions. T2-
weighted imaging is most sensitive for the de-
tection of bone marrow and periosteal edema,
which are early signs of bone stress injury.
T1-weighted images are also helpful, in grading
stress injuries and in visualization of cortical
fracture lines. A grading system for tibial stress
injuries has been outlined by Fredericson et al.
(1995). Early stress reaction (Grade 1) is rep-
resented by abnormal periosteal signal on T2-
weighted images (periosteal edema). As the
injury progresses (Grade 2), bone marrow
edema is demonstrated on T2-weighted images.
Grade 3 images are seen as marrow and pe-
riosteal edema on both T2-weighted and T1-
weighted imaging. Finally, if a low signal frac-
ture line is present on both images, the injury
is classified as Grade 4. Similar findings can be
seen in stress injuries of other bones.
Classification of stress injuries is important
so that appropriate steps to treatment may be
taken. For example, tibial Grade 1 and 2 injuries
may be treated with a rest period varying from
2 to 6 weeks. Grade 3 injuries often require
longer periods of rest, and Grade 4 injuries may
need nonweightbearing, casting, or sometimes
surgery depending on location. All stages may
progress if improperly or incompletely treated,
sometimes with catastrophic long-term results.
Literature Cited
Bergman, A.G. and Fredericson, M. 1999. MR im-
aging of stress reactions, muscle injuries, and
other overuse injuries in runners. MRI Clinics of
North America. 7:151-174.
Daffner, R.H. and Pavlov H. 1992. Stress fractures:
Current concepts. A.J.R.. Am. J. Roentgenol.
159:245-252.
Floyd, W.N., Butler J.E., Clanton T., Kim, E.E., and
Pjura, G. 1987. Roentgenologic diagnosis of
Current Protocols in Magnetic Resonance Imaging
stress fractures and stress reactions. South. Med.
J. 80:433-439.
Fredericson, M., Bergman, A.G., Hoffman, K.L.,
and Dillingham, M.S. 1995. Tibial stress reac-
tion in runners: Correlation of clinical symptoms
and scintigraphy with a new magnetic resonance
imaging grading system. Am. J. Sports Med.
23:472-481.
Hulkko, A. and Orava, S. 1987. Stress fractures in
athletes. Int. J. Sports Med. 8:221-226.
Khan, K.M., Brukner, P.D., and Kearney, C. 1994.
Tarsal navicular stress fracture in athletes. Sports
Med. 17:65-76.
Lee, J.K. and Yao, L. 1988. Stress fractures: MR
imaging. Radiology 169:217-220.
Matheson, G.O., Clement, D.B., McKenzie, D.C.,
Taunton, J.E., Lloyd-Smith, D.R., and MacIn-
tyre, J.G. 1987. Stress fractures in athletes: a
study of 320 cases. Am. J. Sports Med. 15:46-58.
Milgrom, C., Chisin, R., Giladi, M., Stein, M., Kash-
tan, H., Margulies, J., and Atlan, H. 1984. Nega-
tive bone scans in impending tibial stress frac-
tures: A report of three cases. Am. J. Sports Med.
12: 488-491.
Pavlov, H., Torg, J.S., and Freiberger, R.H. 1983.
Tarsal navicular stress fractures: Radiographic
evaluation. Radiology 148:641-645.
Roub, L.W., Gumerman, L.W., and Hanley, E.N.
1979. Bone stress: A radionuclide imaging per-
spective. Radiology 132: 431-438.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Shin A.Y., Morin W.D., Gorman J.D., Jones, S.B.,
and Lapinsky, A.S.1996. The superiority of mag-
netic resonance imaging in differentiating the
cause of hip pain in endurance athletes. Am. J.
Sports Med. 24:168-176.
Contributed by Gabrielle Bergman and
Timothy R. Jones
Stanford University Medical Center
Stanford, California
Musculoskeletal
Stress Injuries
A24.1.21
Supplement 3
MRI of the Ankle and Hindfoot UNIT A25.1 Table A25.1.1 Ankle and Hindfoot Protocol

Type of weighting and sequences Imaging plane

MR imaging is the diagnostic modality of choice for the evaluation of traumatic ligamen- 1. Gradient echo three-plane positioning scout Sagittal, coronal, and transverse
tous and tendinous injuries of the ankle and hindfoot. MR imaging is also valuable in the 2. T1-weighted spin echo Transverse
detection of occult bony trauma, osteochondral injuries, avascular necrosis, osteomyelitis, 3. T2-weighted fast spin echo (FSE) Transverse
and a variety of other osseous conditions. Exquisite contrast resolution, noninvasiveness, 4. T1-weighted spin echo Coronal
and multiplanar capabilities are unique features that make MR imaging a powerful 5. Fat-suppressed proton density weighted fast spin echo Coronal
diagnostic technique. The authors employ the Basic Protocol using T1- and T2-weighted 6. Turbo STIR Sagittal
sequences and short tau inversion recovery (STIR) imaging. Additional imaging following
Optional sequence
intravenous gadolinium is indicated in selected cases.
7. Post-gadolinium fat suppressed T1-weighted spin echo Transverse and coronal
If gadolinium contrast is used, whether intravenous or intra-articular, pre- and post-con-
trast T1-weighted sequences in each appropriate plane will be needed for optimal evalu-
ation/comparison. Table A25.1.2 Equipment Parameters

Coil type Quadrature extremity coil

MR EXAMINATION OF THE ANKLE AND HINDFOOT BASIC Gradient coil strength 25 mT/m (or whatever the system permits)
PROTOCOL Cardiac gating No
Most reports on MR imaging of the ankle and foot available in the literature are based on
studies performed on high-strength (1.5 T) magnets (Kneeland, 1994; Zlatkin, 1999). The Respiratory gating No
following sequences described herein are based on the author’s experience with a GE 1.5 Peripheral monitoring For safety only
T scanner, but are expected to be applicable to other equipment manufacturers. The entire Respirator/oxygen If required by the patient
protocol takes ∼21 min to perform. If post-gadolinium imaging is obtained, 8 additional Motion cushions Useful
min are required.
A screening form should be filled and signed by each patient or legal guardian prior to
Table A25.1.1 lists the six sequences necessary to perform the Basic Protocol along with scanning in a magnetic resonance system. Generally, standard screening forms (APPENDIX
one optional sequence. Stepwise instructions for performing the imaging protocol are 1) are used for all patients scanned in a magnetic resonance system.
provided. This protocol should be easily acquired on most clinical systems.
The presence of ferromagnetic material in a patient may represent a health hazard to the
Table A25.1.2 lists the hardware necessary to perform the procedure, along with the patient while he or she is inside the magnet and/or deteriorate image quality. The safety of
appropriate parameters. The available gradient strength will depend on the scanner, and certain ferromagnetic materials can be found in Shellock and Kanal (1996).
the echo times given below may be varied accordingly (the smaller gradient strength, the Patients with previous metal exposure to the eyes should have radiographs of the orbits in
longer the echo time for a particular scan). order to exclude metallic foreign bodies that have not been removed prior to placing them
in the magnetic field.
When using mid/low field scanners, adjustments to the imaging protocols need to be made
in order to compensate for the lower signal-to-noise ratio (SNR) related to lower-field A friend or family member, who can sit in the room during the scan and comfort the patient
as needed, may accompany patients into the magnet room. This companion must be
strength and gradient strength. In general, MR examinations obtained in mid/low field
screened as well to ensure the absence of loose metal objects on the body or clothing.
machines require longer acquisition times and a larger field of view. Signal abnormalities
related to increased water content in tissue must be evaluated using STIR sequences since 2. If the procedure is a research protocol, have the patient sign any necessary consent
frequency-selective fat saturation cannot be obtained when using these systems. forms.

NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.
Materials
Extravascular contrast agent (e.g., Omniscan, Magnevist, Prohance)
Normal saline (0.9% NaCl), sterile
Set up patient and equipment
1. Interview the patient to assess for contraindications such as cardiac pacemakers or
other implants containing ferromagnetic materials. Also, determine if the patient will
need sedation medication requiring the use of monitoring equipment.
Contributed by Jenny Bencardino and Zehava S. Rosenberg
Current Protocols in Magnetic Resonance Imaging (2002) A25.1.1-A25.1.12
Copyright © 2002 by John Wiley & Sons, Inc.
3. Have the patient remove jewelry and change into a gown to eliminate any metal that
might be found in clothing.
All personal belongings should be secured properly during the examination.
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating
and image artifacts.
5. Inform the patient about what will occur during the procedure, what he or she will
experience while in the magnet, and how to behave.
a. If earphones or headphones are used to protect the ears from the loud sounds
produced by the gradients, the patient will be asked to wear these, but will be able
to communicate with you at any time during the imaging.
MRI of the Ankle b. The patient will be given a safety squeeze-bulb or similar equipment to request
Ankle and Foot and Hindfoot assistance at any time (demonstrate how this works).
A25.1.1 A25.1.2
Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging
 c. For good results, the patient should not talk, and should avoid or minimize other
movement, during each scan—i.e., as long as the banging sounds continue.
Between scans, talking is allowed in most cases, but should be avoided when
comparative positional studies are being performed; the patient will be informed
when this is the case.
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.

6. Ask the patient to remain perfectly still during imaging acquisition and also explain
to him or her not to move between imaging sequences.
7. Have the patient lie on the table, with the feet toward the bore of the magnet. For
imaging of the ankle and hindfoot, the authors recommend to lie the patient supine
with the ankle in neutral position. Either before or right after the patient lies down,
set up any triggering devices or other monitoring equipment that is to be used.
The ankle in neutral position has an inherent 10° to 20° of plantar flexion and 10° to 30°
of external rotation. This mild degree of plantar flexion is useful for eliminating the magic
angle effect when examining the flexor and extensor tendons of the ankle and hindfoot.
Increasing the thickness of the fat planes between the peroneus brevis and longus tendons,
and displacing the flexor and peroneal tendons away from the bones are other advantages
of imaging the ankle in plantar flexion.
Mechanical stabilization of the ankle is paramount for good image quality. Immobilization
should be achieved by using foam sponges between the ankle and the coil and Velcro straps
on the leg outside the coil.
8. Center the ankle to be examined in the extremity coil using the infrared light pointing
to the ankle joint level.
9. Advance the patient table to isocenter.
Once this step has been performed, so long as the patient does not move on the table, the
table itself can be moved and then placed in the same position as before without jeopard-
izing the positioning of one scan relative to another.
10. If the patient is unable to hold still, provide an appropriate sedative if indicated.

Table A25.1.3 Parameters for Scout Scan

Patient position Supine

Scan type Gradient echo
Imaging plane (orientation) Three planes Figure A25.1.1 Transverse imaging plane. (A) Using the sagittal localizer, transverse T1-weighted spin echo
images are prescribed covering the Achilles tendon and entire calcaneus. The ankle is in neutral position with 20°
Central slice or volume center Magnet isocenter
of plantar flexion. (B) Transverse T1-weighted image through the distal tibial plafond. Note exquisite anatomic detail.
Echo time (TE) 4.2 msec (C) Transverse T2-weighted FSE image at the same level as B.
Repeat time (TR) 12.1 msec
Flip angle (FA) 30°
Fields of view (FOVx, FOVy) 240 mm, 240 mm Sequence 1: Three plane positioning scout
Resolution (Δx, Δy) 1.88 mm, 1.88 mm 11. To verify the patient’s position, run the scout scan to ensure correct location of the
Number of data points collected (Nx, Ny) 128, 128 ankle in 3-D using the imaging parameters provided in Table A25.1.3.
Slice thickness (Δz) 4 mm
This sequence will be used to set up coverage of the area of interest. Most MR machines
Number of slices 9 (3 in each of 3 cardinal planes) can be programmed to acquire the scout images automatically after coil tuning or after the
Slice gap 1 mm patient has been placed in the isocenter.
Number of acquisitions (Nacq) 1
Swap read and phase encoding No Sequence 2: Transverse T1-weighted spin echo
Slice location Not applicable 12. Using the images generated in sequence 1, prescribe transverse images to provide
Saturation pulses Not applicable complete coverage of the ankle and hindfoot (Fig. A25.1.1). Run sequence 2 accord-
MRI of the Ankle
Scan time 25 sec Ankle and Foot and Hindfoot ing to Table A25.1.4.
A25.1.3 A25.1.4
Current Protocols in Magnetic Resonance Imaging Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging
 Table A25.1.4 Parameters for Transverse T1-Weighted Spin Echo Table A25.1.6 Parameters for Coronal T1-Weighted Spin Echo

Patient position Supine Patient position Supine

Scan type Spin echo Scan type Spin echo
Imaging plane (orientation) Transverse Imaging plane (orientation) Coronal
Central slice or volume center Position for whole ankle coverage Central slice or volume center Position for whole ankle coverage
Echo time (TE) 22 msec Echo time (TE) 22 msec
Repeat time (TR) 900 msec Repeat time (TR) 750 msec
Flip angle (FA) 90° Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 140 mm, 100 mm Fields of view (FOVx, FOVy) 140 mm, 100 mm
Resolution (Δx, Δy) 0.27 mm, 0.39 mm Resolution (Δx, Δy) 0.27 mm, 0.45 mm
Number of data points collected (Nx, Ny) 512, 256 Number of data points collected (Nx, Ny) 512, 224
Slice thickness (Δz) 4 mm Slice thickness (Δz) 4 mm
Number of slices 22 Number of slices 18
Slice gap 1 mm Slice gap 1 mm
Number of acquisitions (Nacq) 1 Number of acquisitions (Nacq) 2
Read direction Anterior to posterior Read direction Left to right
Scan time 3 min, 4 sec Scan time 4 min, 17 sec

Table A25.1.5 Parameters for Transverse T2-Weighted FSE This scan is positioned as in sequence 2. The coverage should include the ankle joint and
hindfoot structures as well. This sequence will provide demonstration of alterations in
Patient position Supine water content characteristic of most pathologic conditions in the foot and ankle. Fat
Scan type Fast spin echo suppressed imaging may be utilized to highlight the contrast differences between the low
Imaging plane (orientation) Transverse signal tendinous and ligamentous structures and the abnormal-free water.
Central slice or volume center Position for whole ankle coverage
Perform system shim as recommended by manufacturer as this is a fat-saturation sequence.
Echo time (TE) 90 msec
Echo train length (ETL) 6
Sequence 4: Coronal T1-weighted spin echo
Repeat time (TR) 3500 msec
14. Use the transverse T1-weighted images to position the coronal scans. Set parameters
Flip angle (FA) 90°
for the coronal T1-weighted spin echo as given in Table A25.1.6 and run sequence 4.
Fields of view (FOVx, FOVy) 140 mm, 100 mm
Resolution (Δx, Δy) 0.55 mm, 0.52 mm Some angulation is needed to provide true coronal anatomic imaging. The imaging plane
Number of data points collected (Nx, Ny) 256, 192 should be parallel to the anterior margin of the talar dome (Fig. A25.1.2). The coverage
Slice thickness (Δz) 4 mm should include the entire Achilles tendon posteriorly.
Number of slices 22 This plane is particularly useful for evaluating osteochondral lesions of the talar dome,
Slice gap 1 mm sinus tarsi and subtalar joint abnormalities and injuries of the collateral ligaments.
Number of acquisitions (Nacq) 2
Read direction Anterior to posterior Sequence 5: Coronal fat-suppressed proton density–weighted fast spin echo
Fat suppression Yes 15. Set parameters for the coronal fat-suppressed proton density–weighted FSE sequence
Scan time 2 min, 55 sec as indicated in Table A25.1.7 and run sequence 5.
This scan is just positioned as sequence 4. Perform system shim as recommended by
The sagittal scout scan is the best plane for planning this transverse sequence. The coverage manufacturer, as this is a fat saturation sequence.
provided should then be checked on the coronal scout. This transverse sequence is used to
position subsequent scans accurately. Sequence 6: Sagittal turbo STIR
When Achilles tendon tears are encountered, care should be taken to cover the Achilles 16. Set parameters for the sagittal turbo STIR sequence as indicated in Table A25.1.8 and
tendon in its entirety. Often, retraction of the proximal edge of the torn tendon requires run sequence 6.
additional imaging of the calf region. In such cases, a body coil should be used to ensure
The T1-weighted transverse image is used to position the sagittal scans. The imaging plane
complete coverage of the entire Achilles tendon and its myotendinous junction. In acute
should be aligned with the long axis of the calcaneus (Fig. A25.1.3). The lateral wall of the
tears, imaging in plantar flexion may approximate the torn ends and can provide the
calcaneus can be used as a reference landmark to ensure appropriate angulation.
surgeon with information regarding the length of the remaining tendon gap.
The sagittal plane is particularly useful for evaluating Achilles tendon ruptures. The entire
Sequence 3: Transverse T2-weighted FSE Achilles tendon as well as its myotendinous junction should be covered.
13. Set parameters for the T2-weighted FSE sequence as indicated in Table A25.1.5 and
run sequence 3. MRI of the Ankle
Ankle and Foot and Hindfoot

A25.1.5 A25.1.6
Current Protocols in Magnetic Resonance Imaging Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging
 Table A25.1.7 Parameters for Coronal Fat Suppressed Proton Density Weighted
Fast Spin Echo

Patient position Supine

Scan type Fast spin echo
Imaging plane (orientation) Coronal
Central slice or volume center Midline
Echo time (TE) 20 msec
Echo train length (ETL) 4
Repeat time (TR) 4000 msec
Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 140 mm, 100 mm
Resolution (Δx, Δy) 0.55 mm, 0.52 mm
Number of data points collected (Nx, Ny) 256, 192
Slice thickness (Δz) 4 mm
Number of slices 25
Slice gap 1.5 mm
Number of acquisitions (Nacq) 2
Read direction Left to right
Fat suppression Yes
Scan time 6 min, 32 sec

Table A25.1.8 Parameters for Sagittal Turbo STIR

Patient position Supine

Scan type STIR fast spin echo
Imaging plane (orientation) Sagittal
Central slice or volume center Midline
Echo time (TE) 21 msec
Echo train length (ETL) 6
Repeat time (TR) 4200 msec
Inversion time (TI) 130 msec
Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 150 mm, 150 mm
Resolution (Δx, Δy) 0.59 mm, 0.78 mm
Number of data points collected (Nx, Ny) 256, 192
Slice thickness (Δz) 4 mm
Number of slices 14
Slice gap 1 mm
Number of acquisitions (Nacq) 2
Read direction Cranial to caudal
Scan time 4 min, 16 sec

Sequence 7: Post-gadolinium fat-suppressed T1-weighted spin echo (optional)

When osteomyelitis or neoplastic processes are suspected, additional imaging following
i.v. administration of gadolinium is recommended.
Figure A25.1.2 Coronal imaging plane. (A) Based on a transverse T1-weighted image, the coronal plane is prescribed
parallel to the anterior margin of the talar dome. Care should be taken to cover the entire Achilles posteriorly. (B) Coronal 17. Pull the patient out of the magnet.
T1-weighted image through the posterior aspect of the talus. (C) Coronal fat suppressed proton density weighted FSE
image in a different patient. Note a small osteochondral lesion of the tibia (short arrow). 18. Establish an i.v. line from which the contrast agent can be injected, and attach this
line securely to the patient so that movement into or out of the magnet will not pull
MRI of the Ankle at the patient’s arm.
and Hindfoot

A25.1.7 A25.1.8
Current Protocols in Magnetic Resonance Imaging Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging
 It is preferable to insert the line prior to imaging and to leave the patient in the magnet so
that there is no intervening motion between the scans run before contrast agent injection
and those run after injection.
A B
19. Repeat step 9.
20. If the patient moved, run sequence 1 again to ensure the position of the patient.
21. Leaving the patient in the magnet, inject the contrast agent, and flush the line with
10 ml saline.
A dose of 0.1 mmol/kg of contrast agent is usually given.
22. Run sequence 7 according to Table A25.1.9 with a transverse orientation first and
then coronal.

COMMENTARY
Background information tudinal tears of these structures, early diagnosis
MR imaging based on its unique soft-tissue by kinematic MR imaging may help to prevent
contrast resolution, noninvasiveness, and mul- this complication. High-resolution kinematic
tiplanar capabilities has been applied to the MR images can be obtained using a device
assessment of tendinous, ligamentous, and os- incorporating dual 5-in. circular surface coils.
seous abnormalities of the foot and ankle Transverse short TR gradient echo (GRASS:
(Erickson and Johnson, 1997). In general, T1- gradient refocused acquisition in the steady
weighted images provide high anatomical de- state) images with TR/TE of 6.1 msec/2.5 msec
tail while T2-weighted images are useful to and 5-mm slice thickness have been used.
evaluate alterations in water content charac- The ligaments of the ankle joint are best
teristic of most pathologic conditions. Short tau evaluated on transverse and coronal images.
inversion recovery (STIR) and fat-suppressed Early MR studies advocated imaging with the
Figure A25.1.3 Sagittal imaging plane. (A) Based on a transverse T1-weighted image, the sagittal plane should be imaging may be utilized to highlight the con- foot in both dorsiflexion and plantar flexion to
aligned with the long axis of the calcaneus. The lateral wall of the calcaneus is used as a reference landmark to ensure trast differences between the low signal tendi- optimally visualize each ligament in its entirety
appropriate angulation. (B) Sagittal STIR image. Edematous changes of the pre-Achilles fat pad are noted (asterisk) in nous and ligamentous structures and the abnor- on a single cut. Conversely, recent evidence
this patient with diagnosis of anterior Achilles peritendinitis. mal-free water. These sequences are also very based on correlation of MR imaging and
useful for assessment of bone marrow edema. anatomic findings, suggest that coronal and
MR imaging of the foot and ankle should be transverse slices with the ankle taped in 10° to
Table A25.1.9 Parameters for Post-Gadolinium Fat-Suppressed T1-Weighted obtained in the transverse, sagittal, and coronal 20° of dorsiflexion are sufficient to provide
Spin Echo planes. The transverse and sagittal planes pro- detailed evaluation of the anatomy of the
vide most of the information necessary to assess lateral and medial collateral ligaments
Patient position Supine the tendons at the ankle. The transverse images (Muhle et al., 1999). In the authors’ experi-
Scan type Spin echo are optimal for evaluating tendon morphology, ence, however, routine examinations with the
Imaging plane (orientation) Transverse and coronal longitudinal splits, tendon sheath fluid, and ankle in neutral position are usually sufficient
Central slice or volume center Position for whole-ankle coverage adjacent soft tissues such as overlying reti- for diagnosing most ligamentous abnormali-
Echo time (TE) Minimum nacula. The sagittal images are most useful for ties of the ankle.
Repeat time (TR) 600 msec depicting disease of the Achilles tendon. Routine MR imaging is adequate for visu-
Flip angle (FA) 90° Oblique transverse imaging, perpendicular to alizing most osseous lesions of the hindfoot.
Fields of view (FOVx, FOVy) 140 mm, 100 mm the long axis of the tendons has been recom- However, the use of two 3-in. planar coils
mended but the authors find it unnecessary. placed adjacent to the medial and lateral
Resolution (Δx, Δy) 0.55 mm, 0.52 mm
Kinematic MR imaging in the transverse plane malleoli has been recommended for high reso-
Number of data points collected (Nx, Ny) 256, 192
has been proposed as an ancillary method for lution imaging of the talar dome in patients with
Slice thickness (Δz) 3 mm dynamic visualization of the peroneal tendons suspected osteochondral lesions (Verhaven et
Number of slices 22 during dorsiflexed and plantar-flexed ankle al., 1991). The field of view can be reduced to
Slice gap 1 mm movements (Shellock et al., 1997). This tech- 8 cm using acquisition matrices of 256 by 192
Number of acquisitions (Nacq) 2 nique allows identification of chronic occult to 256 by 256. Coronal fat suppressed T2-
Read direction Anterior to posterior (for transverse); subluxation and dislocation. Since repetitive weighted fast spin echo (Table A25.1.10) or T2*
left to right (for coronal) attritional trauma of the peroneal tendons in and gradient echo sequences (Table A25.1.11) are
Fat suppression Yes out of the fibular groove has been associated best for evaluating the subchondral bone and
Scan time 3 min, 54 sec MRI of the Ankle with subsequent development of partial longi- status of the articular cartilage.
Ankle and Foot and Hindfoot

A25.1.9 A25.1.10
Current Protocols in Magnetic Resonance Imaging Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging
 Table A25.1.10 Coronal T2-Weighted Fast Spin Echo Sequence fat-suppressed T1-weighted images, which ity: Evaluation with MR arthrography, MR im-
demonstrate marrow enhancement in viable aging and stress radiography. Radiology
192:189-194.
Patient position Supine bone and lack of enhancement in necrotic frag-
Scan type Fast spin echo ments. Erickson, S.J. and Johnson, J.E. 1997. MR imaging
of the ankle and foot. Radiol. Clin. N. Am.
Imaging plane (orientation) Coronal Depiction of bone marrow signal abnormal-
35:163-192.
Central slice or volume center Midline ity manifested as low signal intensity on T1-
Kneeland, J.B. 1994. Technical considerations for
45 msec weighted images and T2-weighted hyperinten-
Echo time (TE) magnetic resonance imaging of the ankle and
Echo train length (ETL) 4 sity images in continuity with a skin ulcer or foot. Magn. Reson. Imaging Clin. N. Am. 2:23-
adjacent subcutaneous abscess are indicative of 28.
Repeat time (TR) 4000 msec
osteomyelitis. Distinguishing features with Mesgarzadeh, M., Sapega, A.A., Bonakdarpour, A.,
Flip angle (FA) 90°
neuroarthropathy include bone marrow Revesz, G., Moyer, R.A., Maurer, A.H., and Al-
Fields of view (FOVx, FOVy) 100 mm, 100 mm changes in the absence of adjacent soft tissue burger, P.D. 1987. Osteochondritis dissecans:
Resolution (Δx, Δy) 0.39 mm, 0.52 mm infection. Gadolinium-enhanced MR imaging Analysis of mechanical stability with radiogra-
Number of data points collected (Nx, Ny) 256, 192 phy, scintigraphy, and MR imaging. Radiology
using fat-suppressed T1-weighted images in the
165:775-780.
Slice thickness (Δz) 4 mm sagittal and short axis planes may provide ad-
Number of slices 25 Morrison, W.B., Schweitzer, M.E., Batte, W.G.,
ditional information regarding the extent of soft
Rad ack, D.P., and Russel, K.M. 1998.
Slice gap 1.5 mm tissue abnormality as well as depiction of sinus Osteomyelitis of the foot: Relative importance of
Number of acquisitions (Nacq) 2 tracts and abscesses (Morrison et al., 1998). primary and secondary MR imaging signs. Ra-
Read direction Left to right diology 201:625-632.
Fat suppression Yes Critical Parameters and Muhle, C., Frank, L.R., Rand, T., Yeh, L., Wong,
Troubleshooting E.C., Skaf, A., Dantas, R.W., Haghighi, P.,
Scan time 6 min, 32 sec
Motion artifacts can be reduced by carefully Trudell, D., and Resnick, D. 1999. Collateral
ligament of the ankle: High resolution MR im-
placing and securing the ankle joint within the
aging with a local gradient coil and anatomic
Table A25.1.11 Coronal T2*-Gradient Echo Sequence extremity coil. Wedging sponges should be correlation in cadavers. Radiographics 19:673-
used in order to provide stability. If these meas- 683.
Patient position Supine ures fail, conscious sedation can be adminis- Shellock, F.G. and Kanal, E. 1996. Magnetic Reso-
Scan type 2-D gradient echo tered. If sedation is not possible, shortening the nance: Bioeffects, Safety, and Patient Manage-
Imaging plane (orientation) Coronal protocol can be attempted by reducing the ac- ment, 2nd ed. Lippincott-Raven, Philadelphia.
Central slice or volume center Midline quisition matrix. This will, however, decrease Shellock, F.G., Feske, W., and Frey, C. 1997.
resolution. Peroneal tendons: Use of kinematic MR imaging
Echo time (TE) 15 msec
Phase encoding artifact may arise from the of the ankle to determine subluxation. J. Magn.
Repeat time (TR) 400 msec Reson. Imaging 7:451-454.
posterior neurovascular bundle. On transverse
Flip angle (FA) 15° Vahlensieck, M., Peterfy, C.G., Wischer, T., Som-
images, the flow related artifact could obscure
Fields of view (FOVx, FOVy) 100 mm, 100 mm mer, T., Lang, P., Schlippert, U., Genant, H.K.,
tendinous or ligamentous structures. In such
Resolution (Δx, Δy) 0.45 mm, 0.20 mm and Schild, H.H. 1996. Indirect MR arthrogra-
cases, changing the phase encoding direction phy; optimization and clinical applications. Ra-
Number of data points collected (Nx, Ny) 224, 512 from anterior/posterior to right/left usually diology 200:249-254.
Slice thickness (Δz) 3 mm solves this problem. Verhaven, E.F, Shahabpour, M., Handelberg, F.W.J.,
Number of slices 25 Vaes, P.H., and Opdecam, P.J. 1991. The accu-
Slice gap 3 mm (interleave) Anticipated Results racy of three-dimensional magnetic resonance
Number of acquisitions (Nacq) 2 This protocol is designed to assess ligamen- imaging in the diagnosis of ruptures of the lateral
ligaments of the ankle. Am. J. Sports Med.
Read direction Left to right tous, tendinous, and osseous abnormalities that
19:583-587.
Fat suppression No affect the ankle joint. The use of post gadolin-
Zlatkin, M.B. 1999. Techniques for MR imaging of
Scan time 5 min, 14 sec ium imaging is indicated when osteomyelitis
joints in sports medicine. Magn. Reson. Imaging
and neoplastic processes are suspected. Clin. N. Am. 7:1-21.

MR imaging diagnosis of fragment stability
has relied on T2-weighted images. A low signal
intensity line in the interface between the nor-
mal bone and the osteochondral fragment has
been related to healing and stability as opposed
to a high signal intensity interface, which indi-
cates fluid in between the nonattached part of
the fragment and the donor site (Mesgarzadeh
et al., 1987). However, the inability to discrimi-
nate between fluid and granulation tissue on
T2-weighted images, has led some investigators
to utilize MR arthrography in this setting
(Chandnani, 1994). Direct MR arthrography,
post intraarticular contrast injection (Vahlen-
sieck et al., 1996; Allman et al., 1999), as well
as indirect MR arthrography, post-intravenous
contrast injection, have also been recom-
mended for assessing abnormalities of the ar-
ticular cartilage.
Because osteochondral lesions may be com-
plicated by osteonecrosis, assessment of frag-
ment viability is also of significance in the MR
imaging analysis. Additional evaluation may be
obtained using post-intravenous gadolinium
Literature Cited
Allman, K.H., Schafer, O., Haver, M., Winterer, J.,
Laubenberger, J., Reichelt, A., and Uhl, M. 1999.
Contributed by Jenny Bencardino
Massachusetts General Hospital
Indirect MR arthrography of the unexercised
Boston, Massachusetts
glenohumeral joint in patients with rotator cuff
tears. Invest. Radiol. 34:435-440.
Zehava S. Rosenberg
Chandnani, V.P., Harper, M.T., Ficke, J.R., New York University Medical Center
Gagliardi, J.A., Rolling, L., Christensen, K.P.,
and Hansen, M.F. 1994. Chronic ankle instabil-
New York, New York
MRI of the Ankle
Ankle and Foot and Hindfoot

A25.1.11 A25.1.12
Current Protocols in Magnetic Resonance Imaging Supplement 5 Supplement 5 Current Protocols in Magnetic Resonance Imaging
Abnormalities of the Osseous Structures of UNIT A26.1 Table A26.1.1 Equipment Parameters for MRI of the Hips

the Hip and Peri-Articular Soft Tissues Coil type
Gradient coil strength
MRI is the most accurate imaging technique for the detection and assessment of a large Cardiac gating
number of disorders arising in the bones of the hip or the surrounding soft tissues including Peripheral gating
osteonecrosis, transient osteopenia, fractures, soft tissue injuries, and tumors. All of these Respiratory gating
disorders can be readily assessed without the use of a contrast agent. The evaluation of Respirator
intra-articular structures, in particular the acetabular labrum, however, is probably best Oxygen
accomplished with the use of the direct injection of a Gd-chelate contrast agent into the Motion cushions
hip joint (which is discussed in UNIT A26.2). Use of contrast agents
Torso array coil
25 mT/m (or whatever the system permits)
No
For safety only
No
If required by patient
If required by patient
Useful
No

The protocol given below is utilized on a 1.5-Tesla system for the evaluation of disease
in the region of the hip. Some modifications must be made when it is used on low-field
(0.23 to 0.3 T) systems as will be discussed below.
MR IMAGING OF THE HIPS
Although MR imaging of the hips for any of these disorders can be satisfactorily
performed with either the whole body radiofrequency (RF) coil or a dedicated surface
coil, the torso-array coil is generally used. Following the acquisition of a scout sequence
in the transverse plane, the first two sequences used are STIR (short tau inversion
recovery) and short TE spin echo sequences of both hips obtained in the coronal or oblique
coronal plane. Although a fat-saturated, long TR/TE fast (or turbo) spin echo (FSE)
sequence can be substituted for the STIR sequence on high-field systems, it appears that
the STIR sequence is slightly more sensitive to the presence of abnormalities within the
marrow or the surrounding soft tissues and should be used. Only moderate size acquisition
matrices (e.g., 192 by 256) are employed. In the authors’ experience, the use of higher
resolution matrices (e.g., 384 by 512) does not generally aid in the assessment of most
disorders and prolongs study time significantly. These are followed by short TE spin echo
and fat saturated, long TE fast-spin echo sequences obtained in the transverse plane. For
the detection and assessment of disease in general, these sequences are the most important
and will generally suffice. In those cases in which osteonecrosis (ON) is seen in the
femoral head on the coronal sequences, a fat-saturated intermediate TE sagittal FSE
sequence is acquired to help estimate the size of the focus and to look for signs of cortical
collapse. The authors also evaluate the articular cartilage and acetabular labrum with this
sequence, although its accuracy for these purposes in the absence of intra-articular
contrast agent is unknown. On a high-field system with a torso array or similar surface
coil in place, the sagittal sequence is performed at much higher resolution than the
previously described coronal and transverse sequences. On a low-field system or when a
torso array coil has not been applied, the resolution chosen for the sagittal series is the
same as that used for the prior series.
On high-field systems (>1.0 T), this entire protocol will require ∼30 min to perform.
Table A26.1.1 lists the hardware necessary to perform the procedure, along with appro-
priate parameters. The available gradient strength will depend on the scanner, and the
echo times given below should be varied accordingly (the smaller the gradient strength,
the longer the echo time for a particular scan).
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.
Contributed by J. Bruce Kneeland, Carolyn Kaut Roth, James Garrison, and Anthony Testa
Current Protocols in Magnetic Resonance Imaging (2002) A26.1.1-A26.1.17
Copyright © 2003 by John Wiley & Sons, Inc.
Set up patient and equipment
1. Interview (screen) the patient to ensure that he or she has no contraindications such
as cardiac pacemakers or other implants containing ferromagnetic materials. Also be
BASIC sure to find out if the patient has any health conditions that may require the presence
PROTOCOL of special emergency equipment during the scanning procedure, or necessitate any
other precautions. The patient should be questioned as to the location of the pain, in
particular the side and whether the pain is actually in the hip or, e.g., lower in the
thigh.
Generally, standard screening forms (APPENDIX 1) are used for all patients scanned in a
magnetic resonance system.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(1996) for discussion of what implants may be safely scanned using magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as if they were to undergo MR imaging themselves to ensure the absence of
loose metal objects on the body or clothing.
2. If the procedure is a research protocol, have the patient sign any necessary consent
forms.
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
4. Inform the patient about what will occur during the procedure, what he or she will
experience while in the magnet, and how to behave, including the following:
a. If earphones or headphones are used to protect the ears from the loud sounds
produced by the gradients, the patient will be asked to wear these, but will be able
to communicate with you at any time during the imaging.
b. The patient will be given a safety squeeze-bulb or similar equipment to request
assistance at any time (demonstrate how this works).
c. For good results the patient should not talk, and should avoid or minimize other
movement, during each scan—i.e., as long as the banging sounds continue.
Between scans, talking is allowed in most cases, but should be avoided when
comparative positional studies are being performed; the patient will be informed
The Osseous when this is the case.
Structures of the
Hip and d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
Peri-Articular
Hip Soft Tissues
A26.1.1 A26.1.2
Supplement 8 Supplement 8 Current Protocols in Magnetic Resonance Imaging
 5. Have the patient lie down on the table with his or her feet toward the machine. Either Table A26.1.2 Transverse Scout
before or right after the patient lies down in the supine position, set up any triggering
devices or other monitoring equipment that is to be used. Patient position Supine
Scan type 3-D short TR gradient echo
6. Center the patient in a torso array coil or other form of dedicated radiofrequency coil Imaging plane (orientation) Transverse
at the hip(s) where the key information is desired. The superior end of the torso array Central slice or volume center Halfway between the anterior/superior
coil should be located at the iliac crest and the inferior end should be well below the iliac spine and the symphysis pubis
greater trochanter. Echo time (TE) 1.5 msec (or minimum)
Receiver bandwidth (RBW) ±32 kHz
7. Start to move the patient into the magnet center. Locate the sagittal alignment light
Repeat time (TR) 6 msec (or minimum)
along the mid-sagittal plane of the patient. The transverse alignment light should be
Flip angle (FA) 30°
centered approximately halfway between the anterior/superior iliac spine and the
Fields of view (FOVx, FOVy) 480 mm, 480 mm
symphysis pubis because the femoral head lies in that location. The coronal alignment
Resolution (Δx, Δy) 1.88 mm, 3.75 mm
light should fall approximately midway between the front and back of the patient.
Number of data points collected (Nx, Ny) 256, 128
Place him or her into the center of the magnet.
Slice thickness (Δz) 10 mm
Once this step has been performed, so long as the patient does not move on the table, the Number of slices 15–20
table itself can be moved and then replaced in the same position as before without Slice gap 0 mm
jeopardizing the positioning of one scan relative to another. Number of excitations (NEX) 2
8. Sedation is not performed routinely and should be reserved for severe claustrophobia. Number of acquisitions (Nacq) 1
Swap read and phase encoding No
Sequence 1: Rapid transverse scout Read direction Right–left
9. Obtain a scout series in the transverse plane using the sequence given in Table A26.1.2 Slice locations From above iliac crest to below greater
(Fig. A26.1.1). trochanter
Flow compensation No
Sequence 2: Coronal fast-STIR No phase wrap (NPW) No
10. Obtain a coronal or oblique coronal fast-STIR sequence using the parameters given Chemical saturation No
in Table A26.1.3. The plane of the slices should be parallel to an imaginary line joining Spatial saturation No
corresponding locations on the left and right hips on the transverse images (Figs. Scan time ∼23–31 sec (depends on size of patient)
A26.1.2 and A26.1.3). The field of view (FOV) should be slightly larger than the soft
tissues surrounding the hips.
As is true for all multi-slice 2-D techniques, the value of TR should be adjusted depending
on the size of the patient. The number of slices is determined by the TR, TE, and, in the case
of fast-spin echo sequences, the number of echos in the echo train (i.e., echo train length).
In the case of T2-weighted and STIR sequences, increasing the TR does not alter the contrast
and it should be increased so that only a single acquisition will cover the entire volume
unless it pushes the acquisition time to the point at which patient motion is likely to interfere
with image quality (>10 min).
The authors’ system automatically interleaves two or more sequences (doubling or tripling
the acquisition time) to cover the appropriate distance if the TR chosen does not permit it
to be covered in a single acquisition. For this reason, adjustment of TR while studying the
calculated time of acquisition is performed prior to each sequence. The TI will be ∼150
msec at 1.5 T but should be adjusted as part of the prescan procedure for the sequence.
Note that the appropriate value of TI to null the signal from fat decreases with decreasing
field strength and that most systems will slightly adjust the value during the pre-scan to
minimize the signal from fat. Choose the value recommended by the equipment manufac-
turer or the applications specialist.
As with all sequences, the use of a lower field strength magnet may require changes in the
sequence parameters (increased FOV or slice thickness, decreased acquisition matrix,
greater Nacq or NEX, and others) to compensate for the lower signal-to-noise ratio (SNR).
This is particularly true for the STIR sequence, which has a low SNR compared to FSE
The Osseous
sequences. Structures of the Figure A26.1.1 A representative slice from the transverse scout series obtained using the
Hip and parameters listed in Table A26.1.2
Peri-Articular
Hip Soft Tissues

A26.1.3 A26.1.4
Current Protocols in Magnetic Resonance Imaging Supplement 7 Supplement 7 Current Protocols in Magnetic Resonance Imaging
 Table A26.1.3 Coronal Fast-STIR

Patient position Supine

Scan type 2-D inversion recovery fast spin echo
Imaging plane (orientation) Coronal or oblique coronal
Central slice or volume center Through the femoral heads
Echo time (TE) ∼70 msec
Receiver bandwidth (RBW) ±16 kHz
Echo train length (ETL) 8
Repeat time (TR) ∼3500 msec
Inversion time (TI) ∼150 msec (determined at prescan)
Flip angle (FA) 180° (or default)
Fields of view (FOVx, FOVy) 380 mm, 380 mm (should be chosen to
include soft tissues lateral to
hips—may be larger or smaller)
Resolution (Δx, Δy) 1.48 mm, 1.98 mm (but depends on
FOV)
Number of data points collected (Nx, Ny) 256, 192
Slice thickness (Δz) 4 mm
Number of slices 30–40 (depends on size of patient) Figure A26.1.2 An image from transverse scout series with locations for the images of both the
Slice gap 0.5 mm coronal series (sequences 2 and 3) indicated.
Number of excitations (NEX) 2
Number of acquisitions (Nacq) 2
Read direction Superior–inferior
Slice locations From posterior to the SI (sacroiliac)
joints to anterior to symphysis pubis
Flow compensation Yes
No phase wrap (NPW) Yes
Scan time ∼7 min (will depend on factors
discussed in text)

Sequence 3: Coronal short TR/TE spin echo

11. Obtain a coronal or oblique coronal short TR/TE SE sequence using the parameters
given in Table A26.1.4 (Fig. A26.1.4). The FOV should be the same as that used for
coronal fast-STIR sequence.
The number of slices will be determined by TR and TE. TR can be increased to permit more
slices but, if T1 weighting is to be maintained, should not be increased >800 msec at 1.5 T
(less at lower field strengths). Because of the automatic interleaving of acquisitions, some
adjustment of TR with calculated acquisition times is generally performed.

Sequence 4: Transverse short TE spin echo

12. Use T1-weighted coronal images to prescribe transverse images (Fig. A26.1.5).
Obtain a short TE transverse series of images using the parameters given in Table
A26.1.5 (Fig. A26.1.6). The FOV should be the same as with above coronal se-
quences.
In contra-distinction to the T1-weighted coronal sequence above, no effort is made to Figure A26.1.3 A representative coronal fast-STIR image obtained with parameters listed in Table
provide T1 weighting, so that TR is adjusted to whatever length is required to image the A26.1.3.
region of interest in one acquisition as discussed in the steps above. The phase encoding
and read directions are “swapped” so that the phase ghosts from the anterior wall are
projected in the mediolateral direction to avoid hips. The Osseous
Structures of the
Hip and
Peri-Articular
Hip Soft Tissues

A26.1.5 A26.1.6
Current Protocols in Magnetic Resonance Imaging Supplement 7 Supplement 7 Current Protocols in Magnetic Resonance Imaging
 Table A26.1.4 Coronal T1-Weighted Spin Echo

Patient position Supine

Scan type Spin echo
Imaging plane (orientation) Coronal or oblique coronal
Central slice or volume center Mid-pelvis through lesser trochanter
Echo time (TE) 8 msec (or minimum)
Receiver bandwidth (RBW) ±16 kHz
Repeat time (TR) 600–800 msec (depending on size of
patient)
Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 380 mm, 380 mm (should be chosen to
include soft tissues lateral to
hips—may be larger or smaller)
Resolution (Δx, Δy) 1.48 mm, 1.98 mm (depends on FOV)
Number of data points collected (Nx, Ny) 256, 192
Slice thickness (Δz) 4 mm
Number of slices 30–40 (depends on size of patient)
Slice gap 0.5 mm
Number of excitations (NEX) 1
Number of acquisitions (Nacq) 2
Read direction Superior–inferior
Slice location From posterior to SI joints to anterior
to the symphysis pubis
Figure A26.1.5 A T1-weighted coronal image with locations of the images from both transverse
No phase wrap (NPW) Yes series (sequences 4 and 5) indicated.
Scan time ∼4–6 min (depends on TR, Ny, and Nacq)
Table A26.1.5 Transverse Short TE Spin Echo

Patient position Supine

Scan type 2-D spin echo
Imaging plane (orientation) Transverse
Central slice or volume center Mid-pelvis through lesser trochanter
Echo time (TE) 8 msec
Receiver bandwidth (RBW) ±16 kHz
Repeat time (TR) 600–1000 msec (depends on size of
patient; no attempt to maintain T1
weighting)
Flip angle (FA) 90°
Fields of view (FOVx, FOVy) 380 mm, 380 mm (depends on size of
patient)
Resolution (Δx, Δy) 1.48 mm, 1.98 mm (depends on FOV)
Number of data points collected (Nx, Ny) 256, 192
Slice thickness (Δz) 5 mm
Number of slices 30–40 (depends on size of patient)
Slice gap 1 mm
Number of excitations (NEX) 1
Number of acquisitions (Nacq) 2
Swap read and phase encoding Yes
Read direction Anterior–posterior
Slice locations From above iliac crests to below lesser
The Osseous trochanters
Figure A26.1.4 A representative coronal T1-weighted spin echo image obtained with parameters Structures of the No phase wrap (NPW) Yes
listed in Table A26.1.4. Hip and Spatial saturation Superior and inferior
Peri-Articular
Hip Soft Tissues Scan time ∼4–6 min (depends on TR, Ny, and Nacq)

A26.1.7 A26.1.8
Current Protocols in Magnetic Resonance Imaging Supplement 7 Supplement 7 Current Protocols in Magnetic Resonance Imaging
Figure A26.1.6 A representative short TE spin echo transverse slice obtained using the parame- Figure A26.1.7 A representative fat-saturated T2-weighted fast-spin echo transverse image
ters listed in Table A26.1.5. obtained with the parameters listed in Table A26.1.6.

Table A26.1.6 Transverse T2-Weighted Fat-Saturated Fast-Spin Echo

Patient position Supine

Scan type 2-D fast spin echo
Imaging plane (orientation) Transverse
Central slice or volume center From above iliac crests to below lesser
trochanters
Echo time (TE) 90 msec
Receiver bandwidth (RBW) ±16 kHz
Echo train length (ETL) 8
Repeat time (TR) 6000 msec (depending on factors
discussed in text)
Flip angle (FA) 90° (or default)
Fields of view (FOVx, FOVy) 380 mm, 380 mm (same FOV as above
sequences)
Resolution (Δx, Δy) 1.48 mm, 1.98 mm
Number of data points collected (Nx, Ny) 256, 192
Slice thickness (Δz) 5 mm
Number of slices 30–40 (depends on TR)
Slice gap 1 mm Figure A26.1.8 A short TE transverse slice with location of the images from the sagittal series
Number of excitations (NEX) 1 indicated.
Number of acquisitions (Nacq) 2
Swap read and phase encoding Yes
Read direction Anterior–posterior
Slice location From above iliac crests to below lesser
trochanters
No phase wrap (NPW) Yes
Chemical saturation Yes (fat)
The Osseous
Spatial saturation Yes (superior and inferior) Structures of the
Scan time ∼5 min (depends on TR) Hip and
Peri-Articular
Hip Soft Tissues

A26.1.9 A26.1.10
Current Protocols in Magnetic Resonance Imaging Supplement 7 Supplement 7 Current Protocols in Magnetic Resonance Imaging
 Table A26.1.7 Sagittal Fat-Saturated Intermediate-Weighted FSE Sequence 5: Transverse fat-saturated long TE FSE
13. Run a long TE FSE sequence using the parameters in Table A26.1.6 (Fig. A26.1.7).
Patient position Supine The FOV should be the same as with the coronal sequences.
Scan type Fast spin echo
Imaging plane (orientation) Sagittal Fat saturation is used because of the relatively bright signal seen from fat with T2-weighted
FSE sequences.
Central slice or volume center Through lesser trochanter
Echo time (TE) 40 msec Selection of precise TR is governed by the same considerations concerning multiple
Receiver bandwidth (RBW) ±16 kHz acquisitions discussed above for fast-STIR sequence and will generally be performed at
Echo train length (ETL) 8 the time of scan.
Repeat time (TR) 5000 msec (depends on factors On low-field systems, fat-saturation cannot be used, thus, the authors typically combine
discussed in text) sequences 4 and 5 into one single long TR, double spin echo sequence with both short and
Flip angle (FA) 90° (or default) long TEs.
Fields of view (FOVx, FOVy) 200 mm, 200 mm
Resolution (Δx, Δy) 0.78 mm, 1.04 mm Sequence 6: Sagittal fat-saturated intermediate-weighted FSE (optional)
Number of data points collected (Nx, Ny) 256, 192 14. Use transverse short TE images to prescribe sagittal images (Fig. A26.1.8). If
Slice thickness (Δz) 4 mm osteonecrosis is seen in the femoral head on the coronal or transverse images, run a
Number of slices 20–25 long TR, intermediate TE FSE sequence using the parameters in Table A26.1.7 (Fig.
Slice gap 0.5 mm A26.1.9).
Number of excitations (NEX) 3
The combination of fat saturation and an intermediate TE makes the articular surface more
Number of acquisitions (Nacq) 2 visible for assessment of contour abnormalities. The selection of TR is determined by the
Swap read and phase encoding Yes same considerations used for the selection of TR in regard to the fast-STIR coronal and
Read direction Anterior–posterior T2-weighted fast-spin echo transverse sequences that were discussed above.
No phase wrap (NPW) Yes
For low-field systems, a short TR/TE SE sequence with the same resolution as that used with
Chemical saturation Yes (fat)
the other sequences is acquired (see Table A26.1.5 with FOVx = FOVy = 200 mm).
Spatial saturation No
Scan time ∼5 min (depends on TR)
COMMENTARY
Background Information mester of pregnancy but is more commonly
seen in middle-aged and elderly men. The dis-
Osteonecrosis and transient osteopenia ease is self-limited but can result in a very
Osteonecrosis (ON) of the femoral head painful hip for up to 9 months.
represents the death of bone that occurs in the
absence of trauma. Although its etiology is Fractures
uncertain, there are several well-established Fractures of the hip or pelvis can result from
pre-disposing factors, of which the most com- either single traumatic events or from the ap-
mon is the administration of steroids. plication of chronic repetitive trauma (stress
In the absence of treatment, most cases of fractures). Plain films are always the first im-
ON of the femoral head progress to collapse of aging technique that should be chosen for the
the cortical surface with the subsequent devel- detection of fractures but they may fail to detect
opment of rapidly progressive osteoarthritis. In them. Fractures of the sacrum are most com-
view of the young age of most patients (30 to monly stress fractures and have generally been
50), this poses a major problem because of the described in osteoporotic women or in women
limited lifetime of hip prostheses. who have undergone radiation therapy for
Many orthopedists believe that surgical in- gynecologic tumors. More recently, they have
tervention, using such techniques as core de- been identified in female athletes, particularly
compression when performed before the onset those who are amenorrheic. More recently, sub-
of cortical collapse, can halt progression of the chondral stress fractures of the femoral head
disease. For this reason, there is considerable have been described in elderly women that,
importance attached to the early diagnosis of clinically, may potentially be confused with
ON. MRI is believed to be the best means of ON. Although radionuclide bone scans have
establishing the diagnosis prior to collapse. traditionally been used to detect radiographi-
The Osseous Transient osteopenia, also known as tran- cally occult fractures, most investigators be-
Figure A26.1.9 A representative intermediate weighted fast spin echo image. Structures of the sient bone marrow edema syndrome, is a poorly lieve that MRI is as sensitive, more specific,
Hip and
Peri-Articular understood but self-limited disorder that was and, in older patients, displays earlier positivity
Hip Soft Tissues originally described in women in the third tri- than bone scans.
A26.1.11 A26.1.12
Current Protocols in Magnetic Resonance Imaging Supplement 7 Supplement 7 Current Protocols in Magnetic Resonance Imaging
 The early diagnosis of these radiographi-
cally occult fractures permits the institution of
appropriate therapy at an earlier stage, often
avoiding the more aggressive treatment neces-
sary for more advanced fractures.
Soft tissue abnormalities
Abnormalities of the soft tissues surround-
ing the pelvis represent a diverse group of
disorders. All of these disorders occur else-
where in the musculoskeletal system and in-
clude muscle injury, tendon degeneration, bur-
sitis, and soft tissue tumors. In the authors’
experience, the patient’s major complaint is
pain referring to the hip and the nature of the
disorder is often not known. The significance
of timely diagnosis of soft tissue abnormalities
with MRI is heavily dependent on the nature of
the abnormality.
Critical Parameters and
Troubleshooting
The use of MRI for the detection of osseous
and peri-articular soft tissue abnormalites is
straightforward and accurate studies can be
performed on almost any commercial system.
The FOV is chosen to include the hips and the
adjacent soft tissue that could also give rise to
symptoms in the region of the hip. TR is prob-
ably the most critical parameter for each of the
Figure A26.1.10 A coronal T1-weighted image through the hips of a patient with bilateral ON. In
the patient’s left femoral head, a low signal interface surrounds the focus of ON with the signal
intensity of fat (arrow). In the right femoral head, the focus of ON demonstrates diffuse low signal
(arrowhead).
Current Protocols in Magnetic Resonance Imaging
sequences and its selection involves the factors
that were detailed above for the individual se-
quences. The use of array coils or other surface
coils is not necessary for the MR diagnosis of
any of these disorders although occasionally
the higher resolution attainable with these coils
may be useful to distinguish between two dis-
orders with a similar appearance such as
osteonecrosis and the much less common sub-
chondral insufficiency fracture. The added
value of obtaining additional images at higher
resolution is unproven.
Anticipated Results
Osteonecrosis and transient osteopenia
ON is seen as a well-marginated geographic
focus of variable signal in the weight-bearing
surface of the femoral head with a well-defined
interface. The interface has a low signal on a
T1-weighted sequence (sequence 3; Fig.
A26.1.10) and demonstrates parallel high- and
low-signal lines on long TE sequences (“dou-
ble-line sign”). The region of the femoral head
surrounded by the abnormal curvilinear margin
most commonly demonstrates the same signal Figure A26.1.11 A fast-STIR coronal image through the hips in a patient with transient bone
marrow edema syndrome showing diffuse edema (e) in the left femoral head and neck.
intensity as marrow fat on all sequences. Less
commonly, these foci may demonstrate low
signal on T1-weighted images and high signal
on T2-weighted images or low signal on both
Figure A26.1.12 A fast-STIR coronal image through both ischiopubic rami demonstrates a
high-signal fracture (arrow) with edema in the surrounding bone and soft tissue.
The Osseous
Structures of the
Hip and
Peri-Articular
Hip Soft Tissues
A26.1.13 A26.1.14
Supplement 7 Supplement 7 Current Protocols in Magnetic Resonance Imaging
 short and long TE sequences (Mitchell et al.,
1987; Lang et al., 1988). In some cases, there
is a zone of surrounding bony edema that is seen
as diffuse low signal on T1-weighted images
and high signal on T2-weighted images. The
presence of bony edema has been claimed by
some investigators to be associated with hip
pain.
There have been several publications that
have investigated the value of measuring the
size of the lesion. All have used different tech-
niques with varying degrees of quantitation and
all have agreed that the larger the lesion, the
greater the probability of collapse even with
performance of core decompression surgery
(Beltran et al., 1990; Lafforgue et al., 1993).
There is no agreement on the best technique to
measure size but the percentage (by quartiles)
of the weight-bearing surface involved can be
readily estimated.
Transient osteopenia is seen as abnormal
signal (low on T1-weighted and bright on T2-
weighted and STIR sequences) that extends
throughout the femoral head and neck but with-
out the more discrete focus of abnormal signal
with a well-defined margin that is typical of ON
(Fig. A26.1.11; Vande Berge et al., 1999).
Figure A26.1.13 A coned-down fast-STIR coronal image illustrates the appearance of a tear (T)
of the pectineus muscle. Fractures
Fractures are most commonly seen as linear
foci with low signal on all pulse sequences,
although in some cases, a high signal linear
focus is seen on T2-weighted images (Fig.
A26.1.12). There is generally ill-defined, ab-
normal signal in the surrounding marrow (low
signal on short TE sequences, high signal on
long TE and STIR sequences) that is believed
to represent marrow edema. With the use of the
STIR sequence, the marrow edema may be very
bright and obscure the fracture line (Quinn and
McCarthy, 1993). When the fracture occurs in
the subchondral region of the femoral head, it
can be difficult to distinguish from ON, al-
though the double-line sign is less commonly
seen (Yamamoto et al., 2001).
In addition to identifying the presence of
pelvic fractures, MRI can also identify the pres-
ence of associated soft tissue injuries such as
muscle tears (see below; Bogost et al., 1995).
Although MRI is the best technique for the
detection of subtle fractures, CT (computed
tomography) with reformatting or surface ren-
dering is considered to be the most accurate
The Osseous technique for determining the spatial relation-
Structures of the ship among displaced fragments and for class-
Hip and
Figure A26.1.14 A fast-STIR coronal image demonstrating a partial tear of the gluteal tendon Peri-Articular ifying complex acetabular fractures.
(arrow) with an accompanying trochanteric bursitis (arrowhead). Hip Soft Tissues
A26.1.15 A26.1.16
Current Protocols in Magnetic Resonance Imaging Supplement 7 Supplement 7
Soft tissue abnormalities
In view of the wide range of possible diag-
noses, it is not possible to discuss the MR
appearance of every disorder. Benign and ma-
lignant soft tissue tumors have a wide variety
of appearances that depend on the nature of the
tumor (Sundaram et al., 1988). Muscle injury
is seen as increased signal on T2-weighted or
STIR sequences with varying degrees of dis-
ruption of muscle architecture (DeSmet and
Best, 2000; Fig. A26.1.13). Degeneration and
tears of the gluteal tendons are seen as high
signal on T2-weighted and STIR images within
or completely interrupting the normal low sig-
nal of the tendons (Kingzett-Taylor et al., 1999;
Fig. A26.1.14). Bursitis (iliopsoas, trochan-
teric) is seen as homogeneous high signal on
T2-weighted and STIR sequences in the appro-
priate location. Trochanteric bursitis often ac-
companies gluteal tendon degeneration but
may also occur in its absence (Fig. A26.1.14).
The significance of bursitis must be interpreted
in light of its location and clinical symptoms.
Literature Cited
Beltran, J., Knight, C.T., Zuelzer, W.A., Morgan,
J.P., Shwendeman, L.J., Chandnani, V.P., Mo-
sure, J.C., and Shaffer, P.B. 1990. Core decom-
pression for avascular necrosis of the femoral
head: Correlation between long-term results and
pre-operative MR staging. Radiology 175:533-
536.
Bogost, G.A., Lizerbram, E.K., and Crues, J.V. 3rd.
1995. MR imaging in evaluation of suspected hip
fracture: Frequency of unsuspected bone and
soft-tissue injury. Radiology 197:263-267.
DeSmet, A.A. and Best, T.M. 2000. MR imaging of
the distribution and location of acute hamstring
injuries in athletes. A.J.R. 174:393-399.
Kingzett-Taylor, A., Tirman, P.F., Feller, J.,
McGann, W., Prieto, V., Wischer, T., Cameron,
J.A., Cvitanic, O., and Genant, H.K. 1999. Tend-
inosis and tears of the gluteus medius and
minimus muscles as a cause of hip pain: MR
imaging findings. A.J.R. 173:1123-1126.
Lafforgue, P., Dahan, E., Chagnaud, C., Schiano, A.,
Kasbarian, M., and Acquaviva, P.C. 1993. Early-
stage avascular necrosis of the femoral head: MR
imaging for prognosis in 31 cases with at least 2
years of follow-up. Radiology 187:199-204.
Lang, P., Jergesen, H.E., Moseley, M.E., Block, J.E.,
Chafetz, N.I., and Genant, H.K. 1988. Avascular
necrosis of the femoral head: High-field-strength
MR imaging with histologic correlation. Radiol-
ogy 169:517-524.
Mitchell, D.G., Rao, V.M., Dalinka, M.K., Spritzer,
C.E., Alavi, A., Steinberg, M.E., Fallon, M., and
Kressel, H.Y. 1987. Femoral head avascular ne-
crosis: Correlation of MR imaging, radiographic
staging, radionuclide imaging, and clinical find-
ings. Radiology 162:709-715.
Current Protocols in Magnetic Resonance Imaging
Quinn, S.F. and McCarthy, J.L. 1993. Prospective
evaluation of patients with suspected hip fracture
and indeterminate radiographs: Use of T1-
weighted MR images. Radiology 187:469-471.
Shellock, F.G. 1996. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Sundaram, M., McGuire, M.H., and Herbold, D.R.
1988. Magnetic resonance imaging of soft tissue
masses: An evaluation of 53 histologically
proven tumors. Magn. Reson. Imaging 6:237-
248.
Vande Berge, B.C., Malghem, J.J., Lecouvet, F.E.,
Jamart, J., and Maldague, B.E. 1999. Idiopathic
Current Protocols in Magnetic Resonance Imaging
bone marrow edema lesions of the femoral head:
Predictive value of MR imaging findings. Radi-
ology 212:527-535.
Yamamoto, T., Schneider, R., and Bullough, P.G.
2001. Subchondral insufficiency fracture of the
femoral head: Histopathologic correlation with
MRI. Skeletal Radiol. 30:247-254.
Contributed by J. Bruce Kneeland, Carolyn
Kaut Roth, James Garrison, and Anthony
Testa
University of Pennsylvania
Philadelphia, Pennsylvania
Hip
A26.1.17
Supplement 7
Hip MR Arthrography for Acetabular Labral UNIT A26.2
Tears
Tears of the acetabular labrum are an uncommon cause of hip pain. A snapping sound
may also accompany labral tears, although this sign is by no means specific for that
disorder and may be found with a variety of abnormalities most commonly the snapping
of the iliotibial band over the greater trochanter.
IMAGING OF ACETABULAR TEARS BASIC
PROTOCOL
It is the authors’ opinion that the MR assessment of labral tears should be performed as
an MR arthrogram, employing an array or surface coil and using higher resolution than
that employed for routine MR examinations of the hip MR. The authors also believe that
these studies should not be attempted on systems with static field strength <1 Tesla,
because of the considerable demands placed on these systems to evaluate the labrum, in
particular, the need for high resolution and fat saturation. (Not all investigators agree with
the necessity of using arthrography, however. At least one investigator maintains that
intra-articular contrast is not needed if sufficiently high resolution is employed. A
different group of investigators performed MR arthrograms on a group of patients using
a 0.5-T system for some patients and a 1.0-T system for others. These investigators did
not describe any difference in accuracy between the two systems, although, they did not
directly address this issue.) Under fluoroscopic guidance, a small amount of iodinated
contrast agent is injected to confirm intra-articular location of the needle. A small amount
of dilute Gd-chelate in saline, bupivicaine, and betamethasone is injected into the joint.
Bupivicaine, a long-acting anesthetic, is used to confirm the intra-articular origin of the
pain, as well as to provide short-term relief of symptoms. Betamethasone, a long-acting
steroid, is used to provide somewhat longer-term symptomatic relief. Following the
acquisition of a scout sequence in the coronal plane, fat-saturated, T1-weighted, fast-spin
echo sequences in the transverse, coronal, and sagittal planes are obtained. Fast spin echo
(FSE) sequences are used because fat saturation, when used with short TE FSE sequences,
increases the study time much less than when used with short TE conventional spin echo
sequences (CSE). The authors also set the second echo of the echo train in the FSE
sequence to two times the minimum TE as the effective TE because of reduced blurring
compared with use of the first echo. See Kowalchuk et al. (2000) for further discussion
of the rationale for using the second echo of a fast-spin echo sequence as well as
demonstrating its accuracy for the assessment of knee meniscal tears. To keep this
effective TE <30 msec while using the second echo of the train requires relatively strong
gradients and wide bandwidth. This type of sequence cannot be performed on all systems.
In particular, if the gradients are not sufficiently strong, fat-saturated, T1-weighted CSE
sequences are used and more time is taken to perform the study. Other investigators have
advocated the use of 3-D, T1-weighted, short TR, gradient sequences. The authors have
had no experience with these sequences for this application.
The authors typically use a small FOV (200 to 240 mm) and a 256 by 192 acquisition
matrix. These are followed by a long TE FSE sequence with fat saturation, obtained in
either the coronal or the transverse plane. This last sequence is primarily used to look for
disease in the marrow or surrounding soft tissues. When performed with FSE sequences,
this imaging protocol should require <30 min from start to finish.
Table A26.2.1 lists the hardware necessary to perform the procedure, along with appro-
priate parameters. The available gradient strength will depend on the scanner and
determines whether to use an FSE sequence or a CSE sequence as described above.
Hip
Contributed by J. Bruce Kneeland, Carolyn Kaut Roth, James Garrison, and Anthony Testa A26.2.1
Current Protocols in Magnetic Resonance Imaging (2003) A26.2.1-A26.2.11
Copyright © 2003 by John Wiley & Sons, Inc. Supplement 8
 Table A26.2.1 Equipment Parameters for MR Arthrography of the 4. Perform a hip arthrogram under fluoroscopic guidance. Inject a small amount of
Hips for Acetabular Labral Tears iodinated contrast agent into the hip joint using standard technique to establish the
intra-articular location of the needle. Inject into the hip a solution consisting of 0.2
Coil type Torso array coil (or similar local coil)
ml Gd-chelate, 5 ml saline, 1 ml (6 mg) betamethasone, and 5 ml bupivicaine MPF.
Gradient coil strength 25 mT/m (or whatever the system permits)
Pre-mix this solution in a small saline bottle by withdrawing 15 ml from a 20-ml vial
Cardiac gating No
and adding the other compounds in order to reduce time in the fluoroscopic suite.
Peripheral gating For safety only
Inject the entire 11.2 ml of the solution into the hip followed by an injection of up to
Respiratory gating No
10 ml of saline. This last saline injection may be halted before the full 10 ml is
Respirator If required by patient
administered if either the patient experiences pain or considerable backpressure is
Oxygen If required by patient
felt.
Motion cushions Useful
Use of contrast agents Yes (intra-articular injection performed The patient’s assessment of the change in hip pain following the injection of the betamatha-
with fluoroscopic guidance prior to MR) sone and bupivicaine is used by the referring physician to confirm that intra-articular
structures especially the labrum are the cause of pain, as well as to deliver short-term relief
from symptoms.
NOTE: Be sure that technologists and nurses have immediate access to any emergency
The MPF formulation is used to prevent precipitation with the betamethasone.
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen. 5. Take the patient by wheelchair to the MR system.
Materials 6. Inform the patient about what will occur during the procedure, what he or she will
Betamethasone experience while in the magnet, and how to behave, including the following:
Bupivicaine MPF (methyl paraben–free) a. If earphones or headphones are used to protect the ears from the loud sounds
Gd-chelate produced by the gradients, the patient will be asked to wear these, but will be able
Iodinated contrast agent to communicate with you at any time during the imaging.
Saline
b. The patient will be given a safety squeeze-bulb or similar equipment to request
Set up patient and equipment assistance at any time (demonstrate how this works).
1. Interview (screen) the patient to ensure that he or she has no contraindications such c. For good results, the patient should not talk, and should avoid or minimize other
as cardiac pacemakers or other implants containing ferromagnetic materials. Also be movement during each scan—i.e., as long as the banging sounds continue.
sure to find out if the patient has any health conditions that may require the presence Between scans, talking is allowed in most cases, but should be avoided when
of special emergency equipment during the scanning procedure, or necessitate any comparative positional studies are being performed; the patient will be informed
other precautions. The patient should be questioned as to the location of the pain, in when this is the case.
the particular side. d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
Generally, standard screening forms (APPENDIX 1) are used for all patients scanned in a 7. Have the patient lie down on the table with his or her feet toward the machine. Either
magnetic resonance system.
before or right after the patient lies down in the supine position, set up any triggering
The presence of any ferromagnetic metals may be a health hazard to the patient when he devices or other monitoring equipment that is to be used.
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see 8. Center the patient in a torso array coil or other form of dedicated radiofrequency coil
Shellock (1996) for a discussion of what implants may be safely scanned using magnetic at the hip(s) where the key information is desired. Generally, the top of the torso array
resonance. coil is at the iliac crest and the bottom just below the greater trochanter.
Patients may be accompanied into the magnet room by a friend or family member, who can 9. Center positioning light at about the midpoint between the anterior superior iliac spine
sit in the room during the scan and comfort the patient as needed. This companion must and the symphysis pubis and put him or her into the center of the magnet. This usually
be treated as if he/she was having the MR examination him- or herself to ensure the absence falls to the center of the torso array coil.
of loose metal objects on the body or clothing. Because of the need for a sterile field, the
authors recommend that the companion not be permitted in the fluoroscopic suite. If the Once this step has been performed, so long as the patient does not move on the table, the
companion must be in the fluoroscopic suite to perform the procedure, appropriate table itself can be moved and then replaced in the same position as before without
radiation protection (i.e., a lead apron) must be worn and the observer is kept at a jeopardizing the positioning of one scan relative to another.
reasonable distance from the sterile field.
10. Sedation is not needed for routine cases and should be reserved for severe claustro-
2. If the procedure is a research protocol, have the patient sign any necessary consent phobics.
forms.
Hip MR
Sequence 1: Rapid coronal scout
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
Arthrography for 11. Run a scout series in the coronal plane using the parameters in Table A26.2.2. (Fig.
Acetabular that might be found in clothing.
A26.2.1).
Labral Tears Hip

A26.2.2 A26.2.3
Supplement 8 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 8
 Table A26.2.2 Coronal Scout

Patient position Supine

Scan type 3-D short TR gradient echo
Imaging plane (orientation) Coronal
Central slice or volume center Halfway between the anterior/superior
iliac spine and the symphysis pubis
Echo time (TE) 1.5 msec (or minimum)
Receiver bandwidth (RBW) ±32 kHz
Repeat time (TR) 6 msec (or minimum)
Flip angle (FA) 30°
Fields of view (FOVx, FOVy) 480 mm, 480 mm
Resolution (Δx, Δy) 1.88 mm, 3.75 mm
Number of data points collected (Nx, Ny) 256, 128
Slice thickness (Δz) 10 mm
Number of slices 15–20
Slice gap 0 mm
Number of excitations (NEX) 2
Number of acquisitions (Nacq) 1
Swap read and phase encoding No
Read direction Right–left
Slice locations From above iliac crest to below greater Figure A26.2.2 An image from coronal scout series with locations for the images of both of the
trochanter following transverse series (sequences 2 and 5) indicated.
Flow compensation No
No phase wrap (NPW) No Table A26.2.3 Transverse Fat-Saturated T1-Weighted FSE
Chemical saturation No
Spatial saturation No Patient position Supine
Scan time ∼23–31 sec (depends on size of patient) Scan type Fast spin echo
Imaging plane (orientation) Transverse
Central slice or volume center Centered on the hips
Echo time (TE) ∼20 msec (second echo of the echo
train set at two times the minimum TE
as the effective echo time)
Receiver bandwidth (RBW) ±32 kHz
Echo train length (ETL) 4
Repeat time (TR) 600–800 msec
Flip angle (FA) 90° (or default)
Fields of view (FOVx, FOVy) 200–240 mm, 200–240 mm
Resolution (Δx, Δy) 0.78–0.94 mm, 1.04–1.25 mm
Number of data points collected (Nx, Ny) 256, 192
Slice thickness (Δz) 4 mm
Number of slices 25–30
Slice gap 0.5 mm
Number of excitations (NEX) 3
Number of acquisitions (Nacq) 2 (interleaved series)
Read direction Right–left
Slice location From posterior to the SI (sacroiliac)
joints to anterior to symphysis
pubis—include all areas with Gd
contrast agent
Figure A26.2.1 A representative slice from the coronal series obtained using the parameters
listed in Table A26.2.2. No phase wrap (NPW) Yes
Hip MR Chemical saturation Yes (fat)
Arthrography for Spatial saturation No
Acetabular
Labral Tears Scan time ∼3 min (depends on TR) Hip

A26.2.4 A26.2.5
Supplement 8 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 8
 Table A26.2.4 Coronal Fat-Saturated T1-Weighted FSE

Patient position Supine

Scan type Fast spin echo
Imaging plane (orientation) Coronal
Central slice or volume center Centered on the hip
Echo time (TE) ∼20 msec (second echo of the echo
train set at two times the minimum TE
as the effective echo time)
Receiver bandwidth (RBW) ±32 kHz
Echo train length (ETL) 4
Repeat time (TR) 600–800 msec
Flip angle (FA) 90° (or default)
Fields of view (FOVx, FOVy) 200–240 mm, 200–240 mm
Resolution (Δx, Δy) 0.78–0.94 mm, 1.04–1.25 mm
Number of data points collected (Nx, Ny) 256, 192
Slice thickness (Δz) 4 mm
Number of slices 25–30
Slice gap 0.5 mm
Number of excitations (NEX) 3
Number of acquisitions (Nacq) 2 (interleaved series)
Figure A26.2.3 A representative image from the transverse, fat-saturated, T1-weighted FSE
sequence obtained with parameters listed in Table A26.2.3. Read direction Superior–inferior
Slice location From just anterior to the hip to just
posterior
No phase wrap (NPW) Yes
Chemical saturation Yes (fat)
Scan time ∼3 min (depends on TR)

Figure A26.2.4 An image from the transverse, fat-saturated, T1-weighted FSE sequence with
locations for coronal series (sequence 3) indicated.

Figure A26.2.5 A representative image from the coronal, fat-saturated, T1-weighted FSE se-
quence obtained with parameters listed in Table A26.2.4.
Hip MR
Arthrography for
Acetabular
Labral Tears Hip

A26.2.6 A26.2.7
Supplement 8 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 8
 Figure A26.2.6 An image from the transverse, fat-saturated, T1-weighted FSE sequence with
locations for sagittal series (sequence 4) indicated.

Table A26.2.5 Sagittal Fat-Saturated T1-Weighted FSE

Patient position Supine

Scan type Fast spin echo
Imaging plane (orientation) Sagittal
Central slice or volume center Centered on the hip Figure A26.2.7 A representative image from the sagittal, fat-saturated, T1-weighted FSE se-
Echo time (TE) ∼20 msec (second echo of the echo quence obtained with parameters listed in Table A26.2.5.
train set at two times the minimum TE
as the effective echo time)
Receiver bandwidth (RBW) ±32 kHz Sequence 2: Transverse T1-weighted, fat saturated, FSE
Echo train length (ETL) 4 12. Use scout (coronal) images to prescribe transverse images (Fig. A26.2.2). Run a
Repeat time (TR) 600–800 msec transverse, fat-saturated, T1-weighted FSE sequence using the parameters given in
Flip angle (FA) 90° (or default) Table A26.2.3 (Fig. A26.2.3).
Fields of view (FOVx, FOVy) 200–240 mm, 200–240 mm The rationale for using FSE sequences with the second echo of the echo train set at two
Resolution (Δx, Δy) 0.78–0.94 mm, 1.04–1.25 mm times the minimum TE as the effective TE was explained above. If the gradient strength is
Number of data points collected (Nx, Ny) 256, 192 not sufficient to permit a TE of ≤30 msec, a conventional spin echo sequence with the
Slice thickness (Δz) 4 mm minimal TE achievable on the system should be employed. TR must be sufficiently short to
Number of slices 25–30 ensure T1-weighting.
Slice gap 0.5 mm
Number of excitations (NEX) 3 Sequence 3: Coronal T1-weighted, fat-saturated, FSE
Number of acquisitions (Nacq) 2 (interleaved series) 13. Use previously obtained transverse images to prescribe coronal images (Fig.
Read direction Superior–inferior A26.2.4). Run a coronal, fat-saturated, T1-weighted FSE sequence using the parame-
Slice location From just medial to the acetabulum to ters given in Table A26.2.4 (Fig. A26.2.5).
the lateral margin of the surrounding This sequence is identical to sequence 2 except for the orientation.
muscles
No phase wrap (NPW) Yes Sequence 4: Sagittal T1-weighted, fat-saturated, FSE
Chemical saturation Yes (fat) 14. Use previously obtained transverse images to prescribed sagittal images (Fig.
Scan time ∼3 min (depends on TR) A26.2.6). Run a sagittal, fat-saturated, T1-weighted FSE sequence using the parame-
Hip MR
Arthrography for ters given in Table A26.2.5 (Fig. A26.2.7).
Acetabular
Labral Tears Hip

A26.2.8 A26.2.9
Supplement 8 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 8
 Table A26.2.6 Transverse T2-Weighted Fat-Saturated Fast-Spin Echo

Patient position Supine

Scan type
Imaging plane (orientation)
Central slice or volume center
Echo time (TE)
Receiver bandwidth (RBW)
Echo train length (ETL)
Repeat time (TR)
Flip angle (FA)
Fields of view (FOVx, FOVy)
Resolution (Δx, Δy)
Number of data points collected (Nx, Ny)
Slice thickness (Δz)
Number of slices
Slice gap
Number of excitations (NEX)
Number of acquisitions (Nacq)
Swap read and phase encoding
Read direction
Slice location
No phase wrap (NPW)
Chemical saturation
Spatial saturation
Scan time
Sequence 5: Transverse, fat-saturated, T2-weighted FSE
15. Use scout (coronal) images to prescribe transverse images (Fig. A26.2.2). Run a
transverse, fat-saturated, T2-weighted FSE sequence using the parameters given in
Table A26.2.6. (A representative image can be seen as Fig. A26.1.7.)
COMMENTARY
Background Information
Acetabular labral tears are an uncommon but
significant cause of hip pain that is amenable
to surgical treatment. They are often accompa-
nied by a snapping sound but this is also seen
with other disorders of the hip.
Critical Parameters and
Troubleshooting
The authors believe that the use of a high-
field (i.e., ≥1 T) system and MR arthrography
in conjunction with surface or array coils to
obtain the highest possible resolution are im-
portant factors to demonstrate labral abnor-
Hip MR malities. However, there have been no studies
Arthrography for to prove that any of these factors are critical to
Acetabular
Labral Tears obtain accurate diagnoses.
2-D fast-spin echo
Transverse
Centered on the hip
90 msec
±16 kHz
8
6000 msec
90° (or default)
380 mm, 380 mm
1.48 mm, 1.98 mm
256, 192
5 mm
30–40 (depends on TR)
1 mm
1
2
Yes
Anterior–posterior Figure A26.2.8 A coned down, coronal fat-saturated, T1-weighted FSE image of the right hip
demonstrates the presence of a tear of the superior acetabular labrum (arrowhead).
From above iliac crests to below lesser
trochanters
Yes
increased FOV. The technical factors given ertheless, it does cast some doubt on the clinical
Yes (fat) above work well on the system described in this significance of what have been considered to
Yes (superior and inferior) unit but may have to be altered for different represent labral tears.
∼5 min (depends on TR) systems.
Literature Cited
Anticipated Results Czerny, C., Hofmann, S., Urban, M., Tschauner, C.,
A small number of cases of surgically con- Neuhold, A., Pretterklieber, M., Recht, M.P., and
Kramer, J. 1999. MR arthrography of the adult
firmed acetabular labral tears have been re-
acetabular capsular-labral complex: Correlation
ported in the radiology literature (Czerny et al., with surgery and anatomy. A.J.R. 173:345-349.
1999). The anterosuperior segment of the
Kowalchuk, R., Kneeland, J.B., Dalinka, M.R.,
labrum is the region most commonly involved Siegelman, E.S., and Dockery, W.D. 2000. MRI
by a tear. Multiple different appearances of of the knee: Valve of short echo time fast spin-
tears diagnosed with MR arthrography have echo using high performance gradients versus
been described and include absence, fragmen- conventional spin-echo for the detection of me-
niscal tears. Skeletal Radiol. 29:520-524.
tation, linear defects, clefts, and intrasubstance
The rationale for the use of the second echo high signal (Fig. A26.2.8). It is of interest that Shellock, F.G. 1996. Pocket Guide to MR Proce-
of the echo train for the short-TE, fat-saturated, dures and Metallic Objects. Lippincott-Raven,
similar appearing morphologic abnormalities
Philadelphia.
FSE sequence with this second echo set to the have been described in the labrum on non-
smallest value possible (i.e., two times the mini- arthrographic studies in asymptomatic sub-
mum echo) is discussed above. Note that the jects. They are noted to occur with increased
authors use a 32-kHz bandwidth to help achieve Contributed by J. Bruce Kneeland, Carolyn
frequency in older subjects and most com-
this minimum effective TE. If the TE cannot be Kaut Roth, James Garrison, and Anthony
monly in the anterosuperior segment. Because Testa
kept at ≤30 msec, it is probably better to used these subjects were not studied with arthrogra- University of Pennsylvania
a conventional spin echo sequence. phy, the relationship between these findings Philadelphia, Pennsylvania
The number of excitations will depend on and those seen at arthrography is unclear. Nev-
the particular coil used, the FOV, the band-
width, and the size of the patient. For heavy-set
patients, choose the largest FOV in the given
range (240 mm) or even increase it up to 280
mm. An increased number of excitations (i.e.,
more than the NEX = 3 given above) may be
used in conjunction with, or in place of, the Hip

A26.2.10 A26.2.11
Supplement 8 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 8
Traumatic and Overuse Injuries of the Elbow UNIT A27.1

With the advance of technology, MR images of the elbow provide exquisite soft tissue
detail and can be obtained using a variety of clinical MR systems of varying strengths.
MR is the modality of choice for evaluation of ligament, tendon, and muscle pathology
of the elbow that typically results from traumatic or overuse injuries. In addition, MR is
superior in evaluating bone marrow and chondral surfaces.

IMAGING OF THE ELBOW BASIC

PROTOCOL
Optimal imaging of the elbow is performed using a dedicated phased array elbow or
extremity coil. The high signal-to-noise ratio (SNR) provides excellent resolution of soft
tissue structures for the small surface area. Additionally, a dedicated elbow or extremity
coil allows for more comfortable patient positioning, which will decrease motion artifacts.
Routine imaging of the elbow should include all three imaging planes. The recommended
sequences for a high-field 1.5-T magnet are: coronal proton density (PD)–weighted and
fat-suppressed T2-weighted fast spin echo images (FSE) or short-tau inversion recovery
(STIR); transverse PD-weighted and T2-weighted images (which may be dual echo as
specified in the tables below or may be done as two separate sequences); and sagittal
fat-suppressed T2-weighted images. The proton density weighted images provide optimal
contrast of soft tissue for the best depiction of anatomy. The longer TE sequences
(T2-weighted) are critical because of their sensitivity for pathology such as marrow edema
and chondral surface abnormalities, in addition to ligament, tendon, and muscle pathology
such as sprains, strains, or tears. An anatomy (PD) and a pathology (T2-weighted)
sequence in each of the three planes is ideal for full evaluation of the elbow (Edelman et
al., 1996). A T2-weighted sequence without fat suppression in one of the three planes is
helpful to assist in differentiating true high-signal fluid from severe degeneration that may
also appear high in signal on a fat-suppressed T2-weighted image. When fat saturation is
used, relatively light T2-weighting (e.g., TE of 40 to 60 msec) should be used for better
signal-to-noise ratio and anatomic visualization. These sequences are also directly appli-
cable for a 1.0-T machine. With lower-field scanners that may not have frequency-selec-
tive fat saturation available, an STIR sequence should be substituted, although the
signal-to-noise ratio will be less. Fast or turbo spin-echo (FSE or TSE) sequences may
be used to advantage over conventional spin-echo sequences, since they provide better
resolution in the same amount of time with more repetitions. FSE or TSE imaging allows
a shortened exam time, which may be needed with difficult or restless patients to optimize
the quality of the images. The disadvantage of FSE/TSE imaging is the bright signal from
fat, obscuring pathology that is negated by the companion fat-suppressed or STIR
sequence in the same plane. Each of the three imaging planes may individually be the
best for demonstrating specific anatomic structures; therefore, imaging in all three planes
is recommended in order to view the three-dimensional anatomy and to achieve accuracy
in diagnoses in an area of fine anatomic detail (Holtz et al., 1998).

Additional Considerations
In the postoperative patient, artifacts from the presence of hardware or micrometallic
material are reduced by fast spin-echo sequences, due to the multiple 180° pulses that
distinguish them from conventional spin echo sequences. However, it is difficult to
achieve homogenous fat saturation in the presence of post-surgical hardware or micromet-
allic artifacts. Therefore, an STIR sequence is preferred over fat-suppressed T2-weighted
images to maximize image quality. Lastly, increases in the bandwidth (i.e., read gradient
strength) serve to reduce metal artifacts.
Elbow
Contributed by Jana Crain and Charles Ho A27.1.1
Current Protocols in Magnetic Resonance Imaging (2003) A27.1.1-A27.1.10
Copyright © 2003 by John Wiley & Sons, Inc. Supplement 9
 Intra-articular saline or diluted gadolinium contrast administration for an MRI arthrogram
may be helpful for visualizing intra-articular bodies in the absence of joint fluid. With
saline injection, the aforementioned standard sequences are still used. In the post-diluted
gadolinium contrast setting, the fat-suppressed T1-weighted sequence (see sequence 3) is
added. For a brief discussion of MR arthrogram, please see the introduction to, and step
1 in, the Basic Protocol in UNIT A22.2.
The elbow is best imaged in near-anatomic position with the arm by the patient’s side and
the forearm in supination. This positioning optimizes demonstration of the collateral
ligaments and common extensor and flexor tendons in the coronal plane. Pronation of the
forearm places the radio-ulnar joint at substantial rotation relative to the humeral/ulnar
joint and markedly degrades imaging. Imaging in each plane should be planned from
scouts in both of the other two planes and should be oriented to the elbow, not to the
orthogonal planes of the body or scanner. A standard exam can be performed in approxi-
mately 30 to 45 min using a 1.5-T scanner. Mid- and low-field machines generally require
45 to 60 min.
Table A27.1.1 lists the hardware necessary to perform the procedure, along with appro-
priate parameters. The available gradient strength will depend on the scanner, and the
echo times given in other tables below may be varied accordingly (the smaller the gradient
strength, the longer the echo time for a particular scan).
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.
Set up patient and equipment
1. Interview (screen) the patient to ensure that there are no contraindications for the
MRI exam such as cardiac pacemakers, cerebral aneurysm clips, biostimulaters,
metallic bodies within the orbit, and any technology sensitive to magnetic fields.
Question the patient regarding medical conditions that may require emergency care
or equipment, in addition to conditions that would inhibit the patient from being able
to complete the exam.
Generally, standard screening forms (see APPENDIX 1) are used for all patients scanned in
a magnetic resonance system.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(2001) for discussion of what implants may be safely scanned using magnetic resonance.
Orthopedic hardware is not a contraindication; however, ferromagnetic and even nonfer-
romagnetic material in this hardware will produce artifacts that may degrade the images.
Adjustments to the technique may be made as discussed below.
Table A27.1.1 Equipment Parameters for MRI of the Elbow
Coil type Phased-array extremity coil
Gradient coil strength 25 mT/m (or whatever the system permits)
Cardiac gating No
Peripheral gating No
Respiratory gating No
Oxygen No
Traumatic and Motion cushions Useful
Overuse Injuries
of the Elbow Use of contrast agent No
A27.1.2
Supplement 9 Current Protocols in Magnetic Resonance Imaging
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This accompanying
person must be screened as well to ensure the absence of loose metal objects on the body
or clothing.
2. If the procedure is a research protocol, have the patient sign any necessary consent
forms.
3. Have the patient remove all jewelry and clothing items or accessories that may contain
metal and change into a gown to eliminate the presence of any metal that might be
found in clothing. Remove pants due to the positioning of the arm adjacent to the hip.
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating.
5. Inform the patient about what will occur during the procedure, what he or she will
experience while inside the magnet, and how to behave, including the following:
a. If earplugs or headphones are used to protect the ears from the loud sounds
produced by the gradients, the patient will be asked to wear these, but will be able
to communicate with you at any time during the imaging.
b. Establish a method of communication with the patient that is clear to the patient
using either an intercom or signaling device.
c. In order to obtain good results, the patient should not talk, and should avoid or
minimize other movement, during each scan—i.e., as long as the banging sounds
continue. Between scans, talking is allowed in most cases, but should be avoided
when comparative positional studies are being performed; the patient will be
informed when this is the case.
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
6. Tape a surface marker (a cod liver oil or vitamin E pill) over the site of pain, if
localized, or over the palpable abnormality.
The cod liver oil pill shows up better on all sequences.
7. Have the patient lie supine with the forearm in supination at his or her side. Place the
elbow, in full extension, into the RF coil with the antecubital fossa in the center of
the coil; the patient’s arm (not necessarily the body) should be straight and parallel
to the scanning table; the arm will be parallel to the main field for a cylindrical magnet
and perpendicular to the main field for a vertical-field magnet. Place sandbags on the
arm and lateral to the arm in order to prevent potential motion of the arm.
A small-angle sponge may be placed under the hand/forearm for very slight flexion of the
elbow joint, which may allow for more comfort and therefore reduce potential motion.
8. Advance the scan table until the centering light is at the center of the RF coil. Evaluate
for right/left and anterior/posterior offsets and use these values for localizer setup.
Advance the patient to the isocenter of the magnet.
Once this step has been performed, so as long as the patient does not move on the table,
the table itself can be moved and then returned to the same position as before without
jeopardizing the positioning of one scan relative to another.
9. If the patient is unable to lie still, provide an appropriate sedative.
Sequence 1: Localizer
10. Run the localizer to ensure the correct location of the elbow in three orientations
using the imaging parameters in Table A27.1.2.
Elbow
A27.1.3
Current Protocols in Magnetic Resonance Imaging Supplement 9
 Table A27.1.2 Primary Clinical Imaging Parameters for Sequence 1 (Localizer) Table A27.1.3 Primary Clinical Imaging Parameters for Sequence 2 (Coronal PD)

Patient position Supine Patient position Supine

Scan type 2-D gradient echo Scan type Fast spin echo (FSE)
Imaging plane (orientation) Three planes (if unavailable use Imaging plane (orientation) Coronal oblique
transverse plane) Central slice or volume center Center of joint
Central slice or volume center Center of joint Echo time (TE) 33 msec
Echo time (TE) 5 msec Echo train length (ETL) 5
Repeat time (TR) 15 msec Repeat time (TR) 3000 msec
Flip angle (FA) 40° Flip angle (FA) 180°a
Fields of view (FOVx, FOVy) 200 mm, 200 mm Fields of view (FOVx, FOVy) 100 mm, 100 mm
Resolution (Δx, Δy) 0.78 mm, 1.56 mm Resolution (Δx, Δy) 0.39 mm, 0.39 mm
Number of data points collected (Nx, Ny) 256, 128 Number of data points collected (Nx, Ny) 256, 256b
Slice thickness (Δz) 5 mm Slice thickness (Δz) 3.5 mm
Number of slices 9 total (3 per plane) Number of slices 16–20
Slice gap 1.5 mm Slice gap 0.4 mm
Number of acquisitions (Nacq) 1 Number of acquisitions (Nacq) 1
Scan time 23 sec Scan time 4 min, 53 sec
aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this sequence is 90°.
bAn additional 80% phase oversampling is applied.

A B Table A27.1.4 Primary Clinical Imaging Parameters for Sequence 3 (Coronal

T1-Weighted)

Patient position Supine

Scan type Fast spin echo
Imaging plane (orientation) Coronal oblique
Central slice or volume center Center of joint
Echo time (TE) 17 msec
Echo train length (ETL) 3
Repeat time (TR) 493 msec
Flip angle (FA) 180°a
Fields of view (FOVx, FOVy) 110 mm, 110 mm
Resolution (Δx, Δy) 0.43 mm, 0.54 mm
Number of data points collected (Nx, Ny) 256, 204b
Slice thickness (Δz) 3.5 mm
Number of slices 16
Slice gap 0.4 mm
Number of excitations (NEX) 2c
Number of acquisitions (Nacq) 2
Scan time 3 min, 24 sec
Figure A27.1.1 (A) Transverse scout for coronal images. Coronal images are obtained parallel to a line connecting the
aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this sequence is 90°.
humeral condyles anteriorly. (B) Sagittal scout for coronal images. Coronal images are obtained parallel to a line along the
bAn additional 50% phase oversampling is applied.
long axis of the upper arm and forearm in the sagittal plane.
cThe number of concatenation is set to be 2. This means that only half of the total slices will be excited

during a given repeat time.

Sequence 2: Coronal proton density
11. Use the localizer images to center at the joint line. Prescribe off transverse or sagittal the gadolinium-based contrast agent is injected during the MR arthrogram, as stated in
scout images in order to obtain coronal images (Fig. A27.1.1). the introduction.
Sample images are provided in Figures A27.1.1A and B.
13. Run sequence 3 according to Table A27.1.4.
12. Run sequence 2 according to the parameters in Table A27.1.3.
Sequence 4: Coronal T2-weighted FSE with fat suppression
Sequence 3: Coronal T1-weighted (optional)
Traumatic and 14. Run sequence 4 according to Table A27.1.5.
This sequence may be substituted for sequence 2 if a shorter scan time or a T1-weighted
Overuse Injuries
of the Elbow signal is desired (see Critical Parameters and Troubleshooting). This sequence is used if Elbow

A27.1.4 A27.1.5
Supplement 9 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 10
 Table A27.1.5 Primary Clinical Imaging Parameters for Sequence 4 (Coronal
Fat-Suppressed T2-Weighted) A B
Patient position Supine
Scan type Fast spin echo
Imaging plane (orientation) Coronal oblique
Central slice or volume center Center of joint
Echo time (TE) 44 msec
Echo train length (ETL) 5
Repeat time (TR) 3000 msec
Flip angle (FA) 180°a
Fields of view (FOVx, FOVy) 110 mm, 110 mm
Resolution (Δx, Δy) 0.43 mm, 0.43 mm
Number of data points collected (Nx, Ny) 256, 256b
Slice thickness (Δz) 3.5 mm
Number of slices 16
Slice gap 0.4 mm
Number of acquisitions (Nacq) 1
Fat suppression Yes
Scan time 4 min, 53 sec Figure A27.1.2 (A) Coronal scout for setting up transverse images. Transverse images should be obtained tangentially to
aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this sequence is 90°. the humeral condyle surface. Transverse images should extend from the distal humeral metaphysis (above the epicondyles)
bAn additional 80% phase oversampling is applied. to the proximal radial metaphysis, notably to include the biceps attachment to the radial tuberosity. (B) Sagittal scout for
planning transverse images, which should be obtained orthogonally to the long axis of the elbow in the sagittal plane.

Table A27.1.6 Primary Clinical Imaging Parameters for Sequence 5 (Coronal

STIR) Table A27.1.7 Primary Clinical Imaging Parameters for Sequence 6 (Transverse
PD- and T2-Weighted)
Patient position Supine
Patient position Supine
Scan type Inversion recovery FSE
Scan type Dual echo PD- and T2-weighted FSE
Imaging plane (orientation) Coronal oblique
Imaging plane (orientation) Transverse
Central slice or volume center Center of joint
Central slice or volume center Center of joint
Echo time (TE) 30 msec
Echo time (TE) 15 msec and 92 msec
Echo train length (ETL) 7
Echo train length (ETL) 5
Repeat time (TR) 4100 msec
Repeat time (TR) 3000 msec
Inversion time (TI) 140 msec
Flip angle (FA) 150°a
Flip angle (FA) 180°
Fields of view (FOVx, FOVy) 110 mm, 110 mm
Fields of view (FOVx, FOVy) 110 mm, 110 mm
Resolution (Δx, Δy) 0.43 mm, 0.43 mm
Resolution (Δx, Δy) 0.43 mm, 0.54 mm
Number of data points collected (Nx, Ny) 256, 256b
Number of data points collected (Nx, Ny) 256, 205a
Slice thickness (Δz) 3.5 mm
Slice thickness (Δz) 3.5 mm
Number of slices 20
Number of slices 16
Slice gap 0.4 mm
Slice gap 0.4 mm
Number of acquisitions (Nacq) 1
Number of acquisitions (Nacq) 1
Scan time ∼5 min
Scan time 3 min, 45 sec
aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this sequence is 90°.
aAn additional 80% phase oversampling is applied.
bAn additional 80% phase oversampling is applied.

Sequence 5: Coronal STIR (optional) Sequence 6: Transverse dual echo or proton density– and T2-weighted
This sequence may be substituted for sequence 4 as an alternative fat-suppression 16. Prescribe off coronal or sagittal scout images in order to obtain transverse images
sequence, e.g., for mid/low-field scanners for which frequency-selective fat-saturation fat (see Fig. A27.1.2).
suppression is not available, or for high-field scanners where field homogeneity is suspect
and fat saturation is nonuniform, as discussed below. 17. Run Sequence 6 according to Table A27.1.7.
Sequence 7: Sagittal proton density
Traumatic and 15. If this sequence is chosen to be substituted for sequence 4, run it according to the
18. Prescribe off transverse or coronal scout images in order to obtain sagittal images
Overuse Injuries parameters in Table A27.1.6.
of the Elbow (see Fig. A27.1.3). Elbow

A27.1.6 A27.1.7
Supplement 10 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 10
 Table A27.1.9 Primary Clinical Imaging Parameters for Sequence 8 (Sagittal
T2-Weighted Fat-Suppressed)
A B
Patient position Supine
Scan type Fast spin echo
Imaging plane (orientation) Sagittal
Central slice or volume center Center of joint
Echo time (TE) 44 msec
Echo train length (ETL) 5
Repeat time (TR) 3000 msec
Flip angle (FA) 180°a
Fields of view (FOVx, FOVy) 110 mm, 110 mm
Resolution (Δx, Δy) 0.43 mm, 0.43 mm
Number of data points collected (Nx, Ny) 256, 256b
Slice thickness (Δz) 3.5 mm
Number of slices 20
Slice gap 0.4 mm
Number of acquisitions (Nacq) 1
Fat suppression Yes
Figure A27.1.3 (A) Transverse scout for setting up sagittal images. Sagittal images are obtained perpendicularly to the Scan time 4 min, 53 sec
coronal plane. The first and last images should include the epicondyle tips. (B) Coronal scout for planning sagittal images. aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this sequence is 90°.
Sagittal images should be parallel to the long axis of the elbow in the coronal plane of the scout image. bAn additional 80% phase oversampling is applied.

Table A27.1.8 Primary Clinical Imaging Parameters for Sequence 7 (Sagittal PD)
COMMENTARY
Patient position Supine
Background Information Routine elbow imaging can be helpful in di-
Scan type Fast spin echo
The most common injuries of the elbow occur agnosing the presence and etiology of neuropa-
Imaging plane (orientation) Sagittal with repetitive microtrauma or acute trauma, usu- thies. Ulnar neuritis may be seen in the clinical
Central slice or volume center Center of joint ally during varus or valgus stress. The classic and setting of medial epicondylitis. Inflammation of
Echo time (TE) 33 msec most common clinical presentation is lateral epi- the ulnar, radial, or median nerves commonly
Echo train length (ETL) 5 condylitis, commonly known as tennis elbow. occurs due to entrapment or frictional trauma
Repeat time (TR) 3000 msec Due to overload in the varus position (backhand from adjacent anomalous musculature, bursitis,
Flip angle (FA) 180°a tennis position), there is degeneration of the com- bony spurring, ganglion cysts, tumors, or fracture
Fields of view (FOVx, FOVy) 110 mm, 110 mm mon extensor tendon. Medial epicondylitis, also fragments. Inflammation of the nerve results in
Resolution (Δx, Δy) 0.43 mm, 0.43 mm known as golfer’s and pitcher’s elbow, is secon- intrasubstance edema which is easily recognized
Number of data points collected (Nx, Ny) 256, 256b dary to overload of the flexor-pronator group on the fat-suppressed T2-weighted or STIR se-
Slice thickness (Δz) 3.5 mm occurring with repetitive valgus maneuvers. MRI quences. Sequelae from neuritis may include
Number of slices 20 findings in these conditions may include tendi- edema and/or atrophy of the musculature due to
nosis, partial or complete tendon tears, underlying denervation.
Slice gap 0.3 mm
collateral ligament degeneration or tearing, and Contusions, fractures, stress-related edema,
Number of acquisitions (Nacq) 1
bone marrow edema. MRI is useful in identifying and osteochondral lesions may result from repeti-
Scan time 4 min, 53 sec the presence and extent of tendon tears in addition tive microtrauma or acute macrotrauma to the
aThe system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this sequence is 90°.
to underlying pathology that will serve as a guide elbow. Osteochondral lesions may result from
bAn additional 80% phase oversampling is applied.
to clinical management. Fluid-filled tears are re- chronic impaction, typically lateral impaction of
vealed nicely on longer TE sequences (T2- the capitellum. The T2-weighted sequences are
19. Run sequence 7 according to Table A27.1.8. weighted), with or without fat suppression. particularly useful in finding chondral surface
Less common tendon injuries involving the abnormalities, as well as reactive subchondral
Sequence 8: Sagittal T2-weighted with fat suppression biceps and triceps are well evaluated on MRI. edema. Common fractures that can be identified
20. Repeat step 18. Sagittal and transverse images are particularly and further evaluated on MRI are radial head
21. Run Sequence 8 according to Table A27.1.9. useful in evaluation of these tendons. The ma- fractures in adults and supracondylar or physeal
jority of these injuries are tendon rupture. Im- fractures in children. Anterior coronoid process
The fields of view (FOV) should be adjusted to the size of the patient. The FOV values in
aging is valuable to determine the location of fractures may be radiographically occult. Loose
Table A27.1.9 are suggested starting points applicable to the average-size patient. With
the tear and the distance of retraction. Some- bodies may result from frictional trauma or acute
smaller patients, the FOV should be decreased. With larger patients, the FOV can be
Traumatic and times the plane of imaging must be expanded fractures and can be identified on MR images
Overuse Injuries increased if necessary.
to include the proximally retracted torn tendon. (Sonin et al., 1996; Steinbach et al., 1997).
of the Elbow Elbow

A27.1.8 A27.1.9
Supplement 10 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 10
 Critical Parameters and
Troubleshooting
Intra-articular bodies
Intra-articular bodies may be difficult to visu-
alize in the absence of a joint effusion. In this
setting, intra-articular saline or dilute gadolinium
administration may be helpful in localizing bod-
ies and in evaluating the chondral surfaces that
serve as donating sites. Post-contrast imaging is
performed as noted above.
Presence of orthopedic hardware or
post-surgical/post-traumatic metallic artifact
An STIR sequence is preferred for better ho-
mogeneity. It is difficult to obtain homogeneous
fat-saturation sequences. Increasing the band-
width also serves to reduce metal artifacts, but a
notable increase may be needed, e.g., from the
usual 13 kHz to ∼64 kHz.
Presence of a cast
MRI can be performed in the presence of an
arm cast, which may preclude the use of a flex
coil. A T1-weighted sequence (Sequence 3) may
be substituted for the standard non–fat suppressed
PD sequence to evaluate cortical margins better
and to increase the conspicuity of a fracture line.
Imaging of the elbow in flexion because of a cast
or inability to extend can still be done. When
obtaining coronal images of the upper arm, addi-
tional anterior extent of imaging will provide
transverse images of the forearm. Conversely,
inferior extension of transverse images through
the upper arm will provide coronal images
through the forearm. Center slice selection may
be patient-specific but should still be chosen near
the joint line.
Patient positioning
It is critical to position the arm as closely to
the isocenter as possible to ensure that the elbow
is within the shimming region (DSV, or diameter
sphere volume) of the magnet. Have the patient
shift as far as possible to the contralateral side.
The contralateral arm may be flexed over the head
to assist in attaining this position, if it is comfort-
able for the patient. If the patient is unable to shift
due to his or her own size, or due to the locked
position of the elbow coil, be sure that the patient’s
arm, but not necessarily the body, is straight and
aligned parallel to the table.
Traumatic and
Overuse Injuries
of the Elbow
A27.1.10
Supplement 10
In the absence of a dedicated elbow coil or
if the patient’s arm is unable to fit in the
surface coil
A knee coil may be used as an alternative to a
surface coil. The positioning of the patient
changes as a result. The patient lies semiprone in
the “superman position” with the arm being im-
aged in the extended position overhead. This is
generally a more difficult position for the patient
to remain in for the duration of the exam and may
result in motion-degraded images. Also, in this
position the forearm invariably is pronated, with
resulting suboptimal anatomic demonstration in
the coronal plane images.
Anticipated Results
The goals of imaging the elbow, like any other
joint, include determining the source of the pa-
tient’s pain, instability, or other clinical problem,
as well as defining the degree of injury as well as
the structures involved, which will assist the cli-
nician in determining the appropriate pathway of
management or treatment. With the development
and improvement of dedicated surface coils and
new techniques, MR imaging continues to be the
most definitive imaging modality in evaluating
soft tissue and marrow abnormalities of the elbow
that result from acute and chronic injury.
Acknowledgement
Special thanks to Phillip Becklund, Dr. Rajesh
Sethi, Margie Garbazino, and Susan Bybee for
their time and contributions.
Literature Cited
Edelman, R.R., Hesselink, J.R., and Zlatkin, M.B.
1996. Clinical Magnetic Resonance Imaging.
W.B. Saunders, Philadelphia.
Holtz, P., Erickson, S.J., and Homquist, K. 1998.
MR imaging of the elbow: Technical considera-
tions. Semin. Musculoskelet. Radiol. 2:121-132.
Shellock, F.G. 2001. MR Procedures: Health Effects
and Safety. CRC Press, Florida.
Sonin, A.H., Tutton, S.M., Fitzgerald, S.W., and
Peduto, A.J. 1996. MR imaging of the adult
elbow. Radiographics 16:1323-1336.
Steinbach, L.S., Fritz, R.C., Tirman, P.F., and Uff-
man, M. 1997. Magnetic resonance imaging of
the elbow. Eur. J. Radiol. 25:223-241.
Contributed by Jana Crain and Charles Ho
National Orthopedic Imaging Associates
San Francisco, California
Current Protocols in Magnetic Resonance Imaging
Masses of the Elbow and Forearm UNIT A27.2
MRI is valuable for mass localization and evaluation because of its multiplanar capa-
bilities and soft-tissue detail. While in many instances the imaging findings may not be
specific, MRI can be very sensitive. The technique used plays a vital role in the pathway
to the differential diagnosis and staging of a mass lesion by defining the precise anatomic
involvement, relationships, and intrinsic characteristics.
IMAGING OF THE ELBOW BASIC
PROTOCOL
As in routine imaging of the elbow, a dedicated extremity coil is preferred. The se-
quences should include T1 - and T2 -weighted sequences for tissue characterization. Fat-
suppression images should also be included to increase sensitivity as well as for possible
fat tissue characterization. Images should be obtained in at least two and preferably all
three planes for anatomic demonstration and involvement. The transverse plane should al-
ways be included. Additional long axis planes would include sagittal for antero-posterior
masses and coronal for medio-lateral masses; if possible, both are generally helpful.
T1 -weighted images are vital in tumor evaluation because of the anatomic soft tissue and
cortical detail and for demonstration of fat or hemorrhage components. It is this anatomic
and tissue definition that may aid the diagnosis or in determining the diagnostic category,
in addition to providing a map for possible surgical management. A fat-saturated T2 -
weighted or STIR (short-tau inversion recovery) sequence is also recommended. STIR
sequences should be substituted for the fat-saturated sequences to achieve more homo-
geneous fat suppression when field uniformity may be poor. With lower-field scanners,
where fat saturation sequences may not be available, STIR sequences generally still
provide good fat suppression. The fat-suppressed longer TE (T2 -weighted) and STIR
sequences are highly sensitive to marrow edema. A T2 -weighted sequence without fat
suppression in one of the three planes is helpful to assist in demonstrating true high-signal
fluid in a cystic or necrotic lesion (Kneeland, 1997; Steinbach et al., 1997a).
Table A27.2.1 lists the hardware necessary to perform the procedure, along with ap-
propriate parameters. The available gradient strength will depend on the scanner, and
the echo times given in other tables below may be varied accordingly (the smaller the
gradient strength, the longer the echo time for a particular scan).
In summary, our standard protocol consists of coronal and sagittal T1 -weighted and
fat-saturated T2 -weighted (or STIR) images, transverse proton density- and T2 -weighted
images, as detailed below. Total exam time is ∼25 min.
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.
Table A27.2.1 Equipment Parameters for MRI of the Elbow
Coil type Phased array extremity coil
Gradient coil strength 25 mT/m (or whatever the system permits)
Cardiac gating No
Peripheral gating No
Respiratory gating No
Oxygen No
Motion cushions Useful
Use of contrast agent No (but used in Alternate Protocol)
Elbow
Contributed by Jana Crain and Charles Ho A27.2.1
Current Protocols in Magnetic Resonance Imaging (2007) A27.2.1-A27.2.10
C 2007 by John Wiley & Sons, Inc.
Copyright  Supplement 14
 Set up patient and equipment Table A27.2.2 Primary Clinical Imaging Parameters for Sequence 1 (Localizer)
1. Interview (screen) the patient to ensure that there are no contraindications for the Patient position Supine
MRI exam such as cardiac pacemakers, cerebral aneurysm clips, biostimulators, Scan type 2-D gradient echo
and metallic bodies within the orbit, or other electronic devices sensitive to magnetic
Imaging plane (orientation) Three planes (if unavailable use
fields. Question the patient regarding medical conditions that may require emergency transverse plane)
care or equipment, in addition to conditions that would inhibit the patient from being
Central slice or volume center Center of mass
able to complete the exam.
Echo time (TE ) 5 msec
Generally, standard screening forms (see APPENDIX 1) are used for all patients scanned
Repeat time (TR ) 15 msec
in a magnetic resonance system.
Flip angle (FA) 40◦
The presence of any ferromagnetic metals may be a health hazard to the patient when
Fields of view (FOVx , FOVy ) 200 mm, 200 mm (varies with the
he or she is inside the magnet, and will also affect the imaging. If in doubt as to the
mass)
exact composition of the items, it is best to exclude patients with any metal implants; see
Shellock (2001) for discussion of what implants may be safely scanned using magnetic Resolution (x, y) 0.78 mm, 1.56 mm
resonance. Orthopedic hardware is not a contraindication. However, ferromagnetic and Number of data points collected (Nx , Ny ) 256, 128
even nonferromagnetic material in this hardware will produce artifacts that may degrade Slice thickness (z) 5 mm
the exam. Adjustments to the technique may be made as discussed below.
Number of slices 9 (3 per plane)
Patients may be accompanied into the magnet room by a friend or family member, who Slice gap 1.5 mm
can sit in the room during the scan and comfort the patient as needed. This companion
Number of acquisitions (Nacq ) 1
must be screened as well to ensure the absence of loose metal objects on the body or
clothing. Scan time 23 sec

2. If the procedure is a research protocol, have the patient sign any necessary consent
forms.
3. Have the patient remove all jewelry and clothing items or accessories that may
contain metal and change into a gown to eliminate the presence of any metal that
might be found in clothing. Remove pants due to the positioning of the arm adjacent
to the hip.
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating
and image artifacts.
5. Inform the patient about what will occur during the procedure, what he or she will
experience while in the magnet, and how to behave, including the following:

a. If earplugs or headphones are used to protect the ears from the loud sounds
produced by the gradients, the patient will be asked to wear these, but will be able
to communicate with you at any time during the imaging.
b. Establish a method of communication with the patient using either an intercom or
signaling device.
c. In order to obtain good results, the patient should not talk, and should avoid or
minimize other movement, during each scan—i.e., as long as the banging sounds
continue. Between scans, talking is allowed in most cases, but should be avoided
Figure A27.2.1 (A) Transverse scout for coronal images. Oblique coronal images are obtained parallel to a line connecting
when comparative positional studies are being performed; the patient will be the humeral condyles anteriorly. (B) Sagittal scout for coronal images. Oblique coronal images are obtained parallel to a
informed when this is the case. line along the long axis of the upper arm and forearm in the sagittal plane.
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
and perpendicular to the main field for a vertical-field magnet. Place sandbags on
6. Tape a surface marker (a cod liver oil or vitamin E pill) over the symptomatic site or the arm and lateral to the arm in order to prevent potential motion of the arm.
palpable abnormality. Do not compress a superficial mass.
A small-angle sponge may be placed under the hand/forearm for very slight flexion of the
The cod liver oil pill shows up better on STIR sequences. elbow joint, which may allow for more comfort and therefore reduce potential motion.
7. Have the patient lie supine with the forearm in supination at his or her side. Place the 8. Advance the scan table until the centering light is at the center of the RF coil. Evaluate
elbow in full extension into the RF coil with the antecubital fossa in the center of the for right/left and anterior/posterior offsets and use these values for localizer setup.
Masses of the coil; the patient’s arm (not necessarily the body) should be straight and parallel to
Elbow and Advance the patient to the isocenter of the magnet.
Forearm the scanning table; the arm will be parallel to the main field for a cylindrical magnet Elbow

A27.2.2 A27.2.3
Supplement 14 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 14
 Table A27.2.3 Primary Clinical Imaging Parameters for Sequence 2 (Coronal
T1 -Weighted)
Patient position Supine
Scan type Fast spin echo (FSE)
Imaging plane (orientation) Coronal oblique
Central slice or volume center Center of mass
Echo time (TE ) 17 msec
Echo train length (ETL) 3
Repeat time (TR ) 493 msec
Flip angle (FA) 180◦a
Fields of view (FOVx , FOVy ) 110 mm, 110 mm
Resolution (x, y) 0.43 mm, 0.54 mm
Number of data points collected (Nx , Ny ) 256, 204b
Slice thickness (z) 3.5 mm
Number of slices 16
Slice gap 0.4 mm
Number of excitations (NEX) 2c
Number of acquisitions (Nacq ) 2
Scan time 3 min, 24 sec
a The system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this sequence
Figure A27.2.2 (A) Coronal scout for setting up transverse images. Transverse images should be obtained tangentially
is 90◦ . to the humeral condyle surface. Images may need to extend beyond routine parameters to include the entire mass lesion.
b An additional 50% phase oversampling is applied.
c The number of concatenation is set to be 2. This means that only half of the total slices will be excited
(B) Sagittal scout for planning transverse images, which should be obtained orthogonally to the long axis of the elbow in
the sagittal plane.
during a given repeat time.

Table A27.2.4 Primary Clinical Imaging Parameters for Sequence 3 (Coronal STIR)
Table A27.2.5 Primary Clinical Imaging Parameters Sequence 4 (Transverse PD-
Patient position Supine and T2 -Weighted)
Scan type Inversion recovery, fast spin echo
Patient Position Supine
Imaging plane (orientation) Coronal oblique
Scan type Fast spin echo, dual echo (PD- and
Central slice or volume center Center of mass T2 -weighted)
Echo time (TE ) 30 msec Imaging plane (orientation) Transverse
Echo train length (ETL) 7 Central slice or volume center Center of mass
Repeat time (TR ) 4100 msec Echo time (TE ) 15 msec and 92 msec
Inversion time (TI ) 140 msec Echo train length (ETL) 5
Flip angle (FA) 180◦ Repeat time (TR ) 3000 msec
Fields of view (FOVx , FOVy ) 110 mm, 110 mm Flip angle (FA) 150◦a
Resolution (x, y) 0.43 mm, 0.54 mm Fields of view (FOVx , FOVy ) 110 mm, 110 mm
Number of data points collected (Nx , Ny ) 256, 205a Resolution (x, y) 0.43 mm, 0.43 mm
Slice thickness (z) 3.5 mm Number of data points collected (Nx , Ny ) 256, 256b
Number of slices 16 Slice thickness (z) 3 mm
Slice gap 0.4 mm Number of slices 16
Number of acquisitions (Nacq ) 1 Slice gap 0.3 mm
Scan time 3 min, 45 sec Number of acquisitions (Nacq ) 1
a An additional 80% phase oversampling is applied.
Scan time ∼5 min
a The system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this sequence
Once this step has been performed, so long as the patient does not move on the table, is 90◦ .
the table itself can be moved and then returned to the same position as before without b An additional 80% phase oversampling is applied.

jeopardizing the positioning of one scan relative to another.

9. If the patient is unable to lie still, provide an appropriate sedative.

Masses of the
Sequence 1: Localizer
Elbow and 10. Run the localizer (scout) to ensure the correct location of the elbow in three
Forearm Elbow
orientations using the imaging parameters in Table A27.2.2.
A27.2.4 A27.2.5
Supplement 14 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 14
 Table A27.2.7 Primary Clinical Imaging Parameters for Sequence 6 (Fat-
Suppressed T2 -Weighted)
Patient position Supine
Scan type Fast spin echo
Imaging plane (orientation) Sagittal
Central slice or volume center Center of mass
Echo time (TE ) 44 msec
Echo train length (ETL) 5
Repeat time (TR ) 3000 msec
Flip angle 180◦a
Fields of view (FOVx , FOVy ) 110 mm, 110 mm
Resolution (x, y) 0.43 mm, 0.43 mm
Number of data points collected (Nx , Ny ) 256, 256b
Slice thickness (z) 3.5 mm
Number of slices 20
Slice gap 0.4 mm
Number of acquisitions (Nacq ) 1
Fat suppression Yes
Scan time 4 min, 53 sec
a The system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this sequence
Figure A27.2.3 (A) Transverse scout for setting up sagittal images. Sagittal images are obtained perpendicular to coronal
plane. The first and last images should include the epicondyle tips. (B) Coronal scout for planning sagittal images. Sagittal is 90◦ .
b An additional 80% phase oversampling is applied.
images should be parallel to the long axis of the elbow in the coronal plane of the scout image.

Table A27.2.6 Primary Clinical Imaging Parameters for Sequence 5 (Sagittal Sequence 2: Coronal T1 -weighted
T1 -Weighted) 11. Use the localizer images to center at the joint line. Prescribe off transverse or sagittal
Patient position Supine scout images, in order to obtain coronal images (see Fig. A27.2.1). Run sequence 2
Scan type Fast spin echo according to the parameters in Table A27.2.3.
Imaging plane (orientation) Sagittal Refer to the sample images in Figures A27.2.1A and B.
Central slice or volume center Center of mass
Sequence 3: Coronal STIR
Echo time (TE ) 17 msec
Echo train length (ETL) 3 12. Run sequence 3 according to the parameters in Table A27.2.4.
Repeat time (TR ) 493 msec Sequence 4: Transverse proton density weighted and T2 -weighted
Flip angle (FA) 180◦a 13. Prescribe off coronal or sagittal scout images in order to obtain transverse images
Fields of view (FOVx , FOVy ) 110 mm, 110 mm (see Fig. A27.2.2).
Resolution (x, y) 0.43 mm, 0.54 mm
Number of data points collected (Nx , Ny ) 256, 204b 14. Run sequence 4 according to the parameters in Table A27.2.5.
Slice thickness (z) 3.5 mm Sequence 5: Sagittal T1 -weighted
Number of slices 16
15. Prescribe off transverse or coronal scout images in order to obtain sagittal images
Slice gap 0.4 mm (see Fig. A27.2.3).
Number of excitations (NEX) 2c
Number of acquisitions (Nacq ) 2 16. Run sequence 5 according to the parameters in Table A27.2.6.
Scan time 3 min, 24 sec
Sequence 6: Sagittal fat-suppressed T2 -weighted
a The system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this sequence
17. Run sequence 6 according to the parameters in Table A27.2.7.
is 90◦ .
b An additional 100% phase oversampling is applied.
Also, an STIR sequence may be substituted for the sagittal fat saturated T2 -weighted
c The number of concatenation is set to be 2. This means that only half of the total slices will be excited
sequence. An STIR sequence may be an alternative fat-suppression sequence, e.g., for
during a given repeat time. mid/low-field scanners for which frequency-selective fat-saturation or fat suppression
is not available, or for high-field scanners where field homogeneity is suspect and fat
saturation is nonuniform.

Masses of the
Elbow and
Forearm Elbow

A27.2.6 A27.2.7
Supplement 14 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 14
ALTERNATE ADDITIONAL FAT-SUPPRESSED SEQUENCES AND INTRAVENOUS Table A27.2.8 Primary Clinical Imaging Parameters for Sequence 7 (Transverse
PROTOCOL POST-CONTRAST SEQUENCES Fat-Suppressed T1 -Weighted)

Post-contrast images are optional and in many cases do not contribute to the diagno- Patient position Supine
sis. However, intravenous gadolinium contrast agent may help to differentiate cystic or Scan type Fast spin echo
necrotic lesions from solid masses by defining a nodular or peripheral enhancing compo- Imaging plane (orientation) Transverse
nent in a cystic lesion versus more diffuse enhancement in a solid mass. In some cases, Central slice or volume center Center of joint
conversely, initial imaging without contrast agents may reveal the diagnosis and exclude Echo time (TE ) 19 msec
the need for post-contrast imaging. Contrast agents may also be helpful in defining ex- Echo train length (ETL) 3
tent of a mass, but at times may be misleading for mass margins, as microscopic tumor
Repeat time (TR ) 699 msec
involvement may extent beyond the macroscopic enhancing region of a mass. In general,
Flip angle (FA) 150◦a
the authors do not use contrast agents for evaluation of a mass lesion.
Fields of view (FOVx , FOVy ) 110 mm, 110 mm
Standard, single-dose gadolinium contrast agent may be administered as an intravenous Resolution (x, y) 0.43 mm, 0.43 mm
injection. Post-contrast T1 -weighted images (with fat suppression if available) can then Number of data points collected (Nx , Ny ) 256, 256b
be performed in transverse and coronal and/or sagittal planes. Pre-contrast T1 -weighted Slice thickness (z) 3 mm
images should also be performed for comparison. Number of slices 16
If a gadolinium i.v. contrast agent is to be used, the standard protocol is as follows: first, Slice gap 0.3 mm
pre-contrast coronal or sagittal T1 -weighted images (depending on orientation of mass), Number of excitations (NEX) 2c
and transverse proton-density- and T2 -weighted images, then, post-contrast coronal, Number of acquisitions (Nacq ) 1
sagittal, and transverse T1 -weighted images with fat suppression. The total scan time Fat suppression Yes
with contrast sequences is 30 to 45 min, including injection time. Scan time 4 min, 2 sec
a The system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this sequence
Materials
is 90◦ .
b An additional 100% phase oversampling is applied.
Normal saline (0.9% NaCl), sterile
c The number of concatenation is set to be 2. This means that only half of the total slices will be excited
Gadolinium-based MR contrast agent (e.g., Magnevist, Omniscan, or Prohance)
during a given repeat time.
21- to 23-G butterfly needle
Table A27.2.9 Primary Clinical Imaging Parameters for Sequence 8 (Coronal and/or
Set up patient and equipment Sagittal Fat-Suppressed T1 -Weighted)
1. Set up patient as in the Basic Protocol, steps 1 to 6.
Patient position Supine
2. Access a vein in the forearm not being imaged and, using a 21- to 23-G butterfly Scan type Fast spin echo
needle, establish an intravenous line through which the contrast agent can be injected. Imaging plane (orientation) Coronal or sagittal
It is preferable to insert the line prior to imaging and to leave the patient in the magnet, so Central slice or volume center Center of joint
that there is no intervening motion between the scans run before contrast agent injection Echo time (TE ) 19 msec
and those run after injection. Echo train length (ETL) 3
3. Finish setting up the patient by performing steps 7 to 9 of the Basic Protocol. Repeat time (TR ) 699 msec
Flip angle 180◦a
4. Run sequences 1, 2 (or 5), and 4 in the Basic Protocol.
Fields of view (FOVx , FOVy ) 110 mm, 110 mm
5. Without removing the patient from the magnet, inject the contrast agent and flush Resolution (x, y) 0.43 mm, 0.43 mm
the intravenous line with 10 ml saline solution. Number of data points collected (Nx , Ny ) 256, 256b
A standard single dose of 0.1 mmol/kg of contrast agent is usually given. Slice thickness (z) 3.5 mm
Number of slices 16
Sequence 7: Transverse fat-suppressed T1 -weighted
Slice gap 0.4 mm
6. Run sequence 7 according to the imaging parameters in Table A27.2.8.
Number of excitations (NEX) 2c
Sequence 8: Coronal fat-suppressed T1 -weighted Number of acquisitions (Nacq ) 1
7. Run sequence 8 according to the imaging parameters in Table A27.2.9. Fat suppression Yes
Sequence 9: Sagittal fat-suppressed T1 -weighted Scan time 4 min, 2 sec
a The system displays the flip angle of the refocusing pulse. The flip angle of the first pulse of this sequence
8. Run sequence 9 according to the imaging parameters in Table A27.2.9.
is 90◦ .
b An additional 100% phase oversampling is applied.
Masses of the
Elbow and c The number of concatenation is set to be 2. This means that only half of the total slices will be excited
Forearm during a given repeat time. Elbow

A27.2.8 A27.2.9
Supplement 14 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 14
 COMMENTARY
Background Information
Musculoskeletal tumors of the elbow are
unusual. By category, there are benign and ma-
lignant soft tissue and osseous masses. More
commonly, benign masses occur as a result of
acute trauma, chronic overuse, or arthropathy,
and may simulate a tumor or more aggressive
process. These include a retracted mass of torn
tendon tissue, bursal collections, synovial or
ganglion cysts, loose bodies, and scar tissue.
Post-traumatic myositis ossificans presents as
an area of soft-tissue swelling and/or hema-
toma that eventually undergoes a characteris-
tic ossification pattern and may mimic a tumor.
Synovial disorders (such as pigmented villon-
odular synovitis and synovial chondromato-
sis), gout, hemophilia, and neuropathic joints
frequently present with masses about the joint.
The most commonly occurring benign soft tis-
sue tumors include lipomas, vascular tumors
such as hemangiomas, and benign neural tu-
mors, all of which often have definitive imag-
ing characteristics and do not require post-
contrast imaging. Malignant soft tissue tumors
of the elbow are generally sarcomas and in-
clude liposarcoma, malignant fibrous histio-
cytoma, and synovial sarcoma. These tumors
have nondefinitive imaging characteristics, but
MR imaging may be vital in the evaluation of
the extent of soft tissue involvement for tumor
staging, as well as post-operative evaluation.
Benign and malignant osseous masses are
generally better evaluated on plain film and
computed tomography for specificity. How-
ever, MR imaging plays an important role in
evaluation of soft tissue structures involved
and for extent of bone/marrow involvement.
The majority of bone tumors (50%) occur in
the ulna, with 20% occurring in the radius,
20% in the humerus, and 10% in the soft tis-
sues. Overall, benign osseous tumors are more
common. Benign tumors that can occur on the
elbow, although unusual, include osteoblas-
tomas, osteochondromas, chondroblastomas,
and enchondromas. Additionally, fibrous dys-
plasia, giant cell tumors, aneurismal bone
cysts, and unicameral bone cysts may oc-
cur. Subtle MR imaging findings of periosti-
tis, nidus, and marrow edema may diagnose
an unsuspected osteoid osteoma. The most
common malignant osseous lesion about the
elbow is metastasis. Plasmacytomas and mul-
tiple myeloma are the most common primary
bone tumors, but much more commonly oc-
Masses of the
Elbow and cur in the axial and proximal appendicular
Forearm skeleton. Other primary malignant bone tu-
mors include osteosarcoma, chondrosarcoma,
A27.2.10
Supplement 14
Ewing’s sarcoma, and primary malignant lym-
phoma. MR imaging is most useful in delin-
eating and staging these tumors (Ma, 1999;
Steinbach et al., 1997a,b).
Critical Parameters and
Troubleshooting
Field of view
Masses may extend beyond the normal
fields of view (FOV) which may require the
FOV to be increased to include all of the abnor-
mality. Alternatively, additional imaging of the
proximal humerus or distal forearm may pro-
vide the optimal evaluation without the sacri-
fice of resolution. In these instances, the elbow
joint line should be included at the edge of
the additional sagittal and/or coronal images
to show the extent and location of the mass
relative to the joint line.
Contrast
The need for post-contrast imaging is vari-
able, as discussed in the Alternate Protocol.
Anticipated Results
The role of MRI in evaluating masses oc-
curring on the elbow is to define the lesion by
its location, intrinsic characteristics, and ex-
tent including structures involved. This infor-
mation will lead the radiologist and the clini-
cian to the diagnosis, differential diagnosis, or
diagnostic category. It will also provide pre-
operative assessment or staging of the elbow
when needed.
Acknowledgement
Special thanks to Phillip Becklund, Dr.
Rajesh Sethi, and Susan Bybee for their time
and contributions.
Literature Cited
Kneeland, J.B. 1997. MR Imaging of the elbow:
Technical considerations. Magn. Reson. Imag-
ing Clin. N. Am. 5:439-442.
Ma, L.D. 1999. Magnetic resonance imaging of
musculoskeletal tumors: Skeletal and soft tissue
masses. Curr. Probl. Diagn. Radiol. 28:29-62.
Shellock, F.G. 2001. MR Procedures: Health Ef-
fects and Safety. CRC Press, Boca Raton, Fla.
Steinbach, L.S., Fritz, R.C., Tirman, P.F., and
Uffman, M. 1997a. Magnetic resonance imag-
ing of the elbow. Eur. J. Radiol. 25:223-241.
Steinbach, L.S., Anderson, S., and Panicek, D.
1997b. MR imaging of musculoskeletal tumors
in the elbow region. MRI Clin. N. Am. 5:619-
653.
Contributed by Jana Crain and Charles Ho
National Orthopedic Imaging Associates
San Francisco, California
Current Protocols in Magnetic Resonance Imaging
Contrast-Enhanced Renal MRA UNIT A28.1
Magnetic resonance angiography (MRA) scans can be run at a range of field strengths BASIC
with varying degrees of success. The success of the technique in imaging arteries is based PROTOCOL
on several factors and will be discussed in Critical Parameters and Troubleshooting.
Although the pulse sequence parameters are defined with respect to 1.5 T scanners, the
actual parameters differ according to available gradient strength. Therefore, TR, TE, and
bandwidths are adjusted for maximum vascular signal. The advantage in using higher
gradients is to obtain images with higher temporal resolution, a useful tool in time-re-
solved imaging studies. The entire protocol consists of four different pulse sequences.
Sequence 2 is used to survey kidneys for additional morphologic information that is not
obtained from MRA sequences. The total scan time should be less than or equal to about
30 min.
Imaging hardware necessary for performing renal MRA is listed in Table A28.1.1.
Available gradient strength will depend on the scanner, and accordingly, the echo times,
repeat times, and bandwidths may be varied.
NOTE: Be sure that technologists and nurses have immediate access to any emergency
equipment that may be relevant to a given study, or that may be needed for a particular
patient, such as crash carts or oxygen.
Materials
Normal saline (0.9% NaCl), sterile
Extravascular contrast agent (e.g., Magnevist, Omniscan, or Prohance), volume
estimated using patient weight (usual amount is based on a single dose, 0.1
mmol/kg)
Disposable contrast-agent syringes and i.v. line
Power injector
Set up patient and equipment
1. Interview and screen the patient to ensure that he or she has no contraindications to
an MR examination. Contraindications include a cardiac pacemaker or defibrillator,
intracerebral aneurysm clip, or ferromagnetic materials in or near vital structures,
including the eyes. Also be sure to find out if the patient has any health condition that
may require the presence of emergency equipment such as cardiac monitor or oxygen.
Since a contrast agent will be injected, obtain an allergy history.
Table A28.1.1 Equipment and Parameters for Renal Artery MRA
Coil type Phased-array or torso-array coil with at
least 4 channels
Gradient coil strength 25 mT/m for Vision (40 mT/m for Sonata)
Cardiac gating No
Peripheral gating No
Oxygen Yes, if patient finds it difficult to hold
his/her breath for >15 sec, then
administer 2 liters via a nasal cannula
Power injector Yes
Contrast agents Yes (e.g., Magnevist, Omniscan, or
Magnetic
Prohance) Resonance
Angiography
Contributed by Anil N. Shetty and Kostaki G. Bis A28.1.1
Current Protocols in Magnetic Resonance Imaging (2003) A28.1.1-A28.1.18
Copyright © 2003 by John Wiley & Sons, Inc. Supplement 11
 Generally, standard screening forms (see APPENDIX 1) are used for all patients scanned in
a magnetic resonance system.
The presence of any ferromagnetic metals may be a health hazard to the patient when he
or she is inside the magnet, and will also affect the imaging. If in doubt as to the exact
composition of the items, it is best to exclude patients with any metal implants; see Shellock
(2001) for a discussion of what implants may be safely scanned using magnetic resonance.
Patients may be accompanied into the magnet room by a friend or family member, who can
sit in the room during the scan and comfort the patient as needed. This companion must
be screened as well to ensure the absence of loose metal objects on the body or clothing,
as well as other items as described above.
2. If the procedure is a research protocol, have the patient sign any consent form as
required by the hospital research administration.
3. Have the patient remove all jewelry and change into a gown to eliminate any metal
that might be found in clothing.
4. Have the patient wash off any mascara and other makeup to avoid local tissue heating.
5. Inform the patient about what will occur during the procedure, what he or she will
experience while in the magnet, and how to behave, including the following:
a. If earplugs and headphones are used to protect the ears from the loud sound
produced by the gradients, the patient will be asked to wear these devices, but will
be able to communicate with you at any time during the imaging.
b. The patient may be given a safety squeeze-bulb or similar equipment to request
assistance at any time; demonstrate how this works.
c. For good results, the patient should not talk, and should avoid or minimize other
movement during each scan—i.e., as long as the banging sounds continue.
Between scans, talking is allowed in most cases, but should be avoided when
comparative positional studies are being performed; the patient will be informed
when this is the case.
d. Nevertheless, the patient may call out at any time if he or she feels it necessary.
6. Educate the patient on how to hold his or her breath at the end of deep inspiration.
Practice the breath-hold method with the patient. Advise the patient of the importance
of not moving during the data acquisition periods and not taking deep breaths during
the non-breath-hold acquisitions.
Patients are instructed to take a breath in, blow it out, take a breath in again, relax,and
hold it. This allows images to be obtained in mid-respiratory cycle (∼15 to 20 sec will be
required for each breath-hold for each 3-D acquisition). Some of the newer sequences, such
as true FISP (true fast imaging with steady-state free precession), may require shorter
breath-holds. Breath holding is critical to obtaining motion-free images. Most patients find
it easier to hold their breath following a maximum inspiration cycle. At the authors’
institution, patients are asked to hold their breath following a deep inspiration.
7. Have the patient mount onto the table. Either before or right after the patient lies
down, set up any triggering devices or other monitoring equipment that is to be used.
8. Establish an intravenous (i.v.) line from which the contrast agent can be injected, and
attach the i.v. line securely to the patient so that movement into or out of the magnet
will not pull at the patient’s arm.
Contrast-
Enhanced
Renal MRA
A28.1.2
Supplement 11 Current Protocols in Magnetic Resonance Imaging
It is preferable to insert the i.v. line prior to imaging and to leave the patient in the magnet,
so that there is no intervening motion between the scans run before and after contrast agent
injection. Also, it is preferable to keep-veins-open (KVO) so that there are no air bubbles
in the i.v. line.
9. Estimate the amount of contrast agent to be used based on patient weight. Set up both
contrast-agent and saline syringes on the injector.
A general rule is to use 1 ml contrast agent for every 10 lb. of patient weight when injecting
as a single dose (0.1 mmol/kg), followed by 15 ml saline to flush the contrast agent.
10. Once the patient is on the table, place the anterior phased-array or torso-array coil.
Center the patient’s kidneys in the phased-array coil and fix it with the Velcro or
buckle straps provided by the manufacturer. Make sure that the phased-array body
coil (and spine coil if necessary) are plugged in.
For some systems, it will be important that the patient lie within the center of the spine
coil,which may be the posterior elements for the phased-array coil system.
11. If needed, place a pillow or other support under the knees to make the patient more
comfortable.
12. Use the laser light available on most systems to center the phased-array coil or
torso-array coil at the “0" location and move the patient into the magnet.
Once this step has been performed, so long as the patient does not move on the table, the
table itself can be moved and then returned to the same position as before without
jeopardizing the positioning of one scan relative to another.
13. Assess the need for supplementary oxygen to improve the breath-holding capacity.
If the patient is not able to hold his or her breath, administer oxygen (2 liters) through
a nasal cannula.
For severely ill patients who cannot hold their breath, cannulas are necessary for basic
breath-hold studies. If the patient still cannot hold his/her breath during the scanning,
imaging can be performed during free breathing in which scan time is considerably reduced
using low acquisition matrix size (number of data points collected) as described in
Anticipated Results (breath-holding). This will improve temporal resolution at the expense
of spatial resolution.
Sequence 1: Rapid three-plane positioning localizers
14. Ask the patient to take a breath in, blow it out, take a breath in again, relax, and
hold it.
15. To validate the correct position for imaging, run the localizer (or pilot) scan of the
system according to Table A28.1.2.
This is done using fast-gradient-recalled imaging sequences with three orientations,
prescribed with multiple slices in each orientation. Refer to the manufacturer’s specifica-
tions for standard default settings and swap as necessary. It is preferable to obtain
localizers in a breath-hold mode. The subsequent placement of an imaging slab will be
more accurate with breath-hold localizers.
Sequence 2: Transverse true-FISP imaging
16. Bring up the sequence for a rapid transverse scan using the parameters defined in
Table A28.1.3.
At the authors’ institution, a true FISP sequence is used to survey both kidneys in a
transverse orientation. This allows for visualizing any disease processes, such as cysts in
the kidneys. Magnetic
Resonance
Angiography
A28.1.3
Current Protocols in Magnetic Resonance Imaging Supplement 11
 Table A28.1.2 Clinical Imaging Parameters for Sequence 1 (Rapid Localizers) Table A28.1.4 Imaging Parameters for Sequence 3 (Bolus Arrival Time
Estimation)
Patient position Supine
Scan type 2-D gradient echo Patient position Supine
Imaging plane (orientation) Transverse, sagittal, coronal Scan type 2-D gradient echo
Central slice or volume center Xyphoid Imaging plane (orientation) Transverse
Echo time (TE) 6.5 msec (1.56 msec)a Central slice or volume center Renal hilum
Receiver bandwidth (RBW) 390 Hz/pixel (781 Hz/pixel)a Echo time (TE) 2.4 msec (1.66 msec)a
Repeat time (TR) 15 msec (3.12 msec)a Repeat time (TR) 5.8 msec (500 msec)a
Delay time (TD) 0 msec Inversion time (TI) 300 msec (250 msec)a
Flip angle (FA) 30° (55°)a Delay time (TD) 0 msec
Fields of view (FOVx, FOVy) 450 mm, 450 mm (400 mm, 400 mm)a Flip angle (FA) 12° (20°)a
Resolution (Δx, Δy) 1.76 mm , 3.52 mm (1.56 mm, 1.68 Fields of view (FOVx, FOVy) 380 mm, 380 mm (400 mm, 400 mm)a
mm)a Resolution (Δx, Δy) 1.48 mm, 2.97 mm (1.56 mm, 3.13
Number of data points collected (Nx, Ny) 256, 128 (256, 238)a mm)a
Display matrix (Dx, Dy) 256, 256 Number of data points collected (Nx, Ny) 256, 128
Slice thickness (Δz) 10 mm Display matrix (Dx, Dy) 256, 256
Number of slices 5 Slice thickness (Δz) 10 mm
Slice gap Variable Number of slices 1
Number of acquisitions (Nacq) 1 Slice gap Not applicable
Swap read and phase encoding No Number of acquisitions (Nacq) 1 (2)a
Saturation pulses No Number of repetitions 40
Scan time 10 sec (12 sec)a Swap read and phase encoding No
aInformation in parentheses refers to parameters for a Sonata scanner. Scan time for the Sonata is based Saturation pulses No
on 15 slices. Scan time ∼40 sec (40 sec)a
aInformation in parentheses refers to parameters for a Sonata scanner. The T on the Sonata is based on
R
echoplanar sequence that includes 128 lines in 0.5 sec (1 slice). The number of acquisitions is doubled to
Table A28.1.3 Imaging Parameters for Sequence 2 (True-FISP Survey of Kidneys) make 1 sec per slice.

Patient position Supine

Scan type 3-D gradient echo 17. Use the pilot (scout) images run earlier to locate both kidneys and set up imaging in
Imaging plane (orientation) Transverse a transverse orientation.
Central slice or volume center Xyphoid
18. Let the patient know you are ready to image during breath-hold state. Repeat step 14
Echo time (TE) 2.3 msec (1.56 msec)a
and begin the scan during suspended respiration following inhalation.
Receiver bandwidth (RBW) 650 Hz/pixel (781 Hz/pixel)a
Repeat time (TR) 4.8 msec (3.12 msec)a Some of the newer sequences, such as true FISP, may require shorter breath-holds. In cases
Delay time (TD) 0 msec where multiple breath-holds are needed, assess the need for supplementary oxygen to
Flip angle (FA) 70° (55°)a improve the breath-holding capacity, and, if necessary, administer 2 liters of oxygen via a
Fields of view (FOVx, FOVy) 350 mm, 350 mm (400 mm, 400 mm)a nasal cannula. Advise the patient of the importance of not moving during the acquisition.
Resolution (Δx, Δy) 1.37 mm, 1.37 mm (1.56 mm, 1.68
mm)a Sequence 3: Estimation of bolus arrival time (BAT)
Number of data points collected (Nx, Ny) 256, 256 (256, 238)a 19. Using orthogonal localizers (coronal and transverse), set up the imaging sequence
Display matrix (Dx, Dy) 256, 256 (256, 256)a with parameters as described in Table A28.1.4 for estimating the bolus arrival time.
Slice thickness (Δz) 6 mm (6 mm)a At the author’s institution, a rapid imaging pulse sequence is used with a nonselective
Number of slices 15 (15)a inversion pulse to null blood.
Slice gap 1.2 mm (1.2 mm)a
20. Position a single 10-mm slice in a transverse orientation near the renal hila.
Number of acquisitions (Nacq) 1
Swap read and phase encoding No 21. Set up the power injector so that 2 ml of contrast agent will be injected followed by
Saturation pulses No 15 ml of saline. Set up the injection rate at 2.5 ml/sec for both contrast agent and
Scan time 18 sec (11 sec)a saline injections.
aInformation in parentheses refers to parameters for a Sonata scanner.
The imaging allows for one image per second. By setting the number of repetitions in scans
to ∼40, upon visual inspection of each of these images, one can see the signal changing in
Contrast- Magnetic
Enhanced the abdominal aorta as the contrast agent transits through that section. Resonance
Renal MRA Angiography

A28.1.4 A28.1.5
Supplement 11 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 11
 Table A28.1.5 Sequence 4 for Acquiring Pre-Contrast MASK Images Table A28.1.6 Imaging Parameters for Sequence 5 (Contrast-Enhanced Imaging)

Patient position Supine Patient position Supine

Scan type Spoiled 3-D-gradient echo Scan type Spoiled 3-D-gradient echo
Imaging plane (orientation) Coronal Imaging plane (orientation) Coronal
Central slice or volume center Xyphoid Central slice or volume center Xyphoid
Echo time (TE) 1.8 msec (1.22 msec)a Echo time (TE) 1.8 msec (1.22 msec)a
Receiver bandwidth (RBW) 390 Hz/pixel (390 Hz/pixel)a Receiver bandwidth (RBW) 390 Hz/pixel (390 Hz/pixel)a
Repeat time (TR) 4.6 msec (3.37 msec)a Repeat time (TR) 4.6 msec (3.37 msec)a
Delay time (TD) 0 msec Delay time (TD) 0 msec
Flip angle (FA) 25° (25°)a Flip angle (FA) 25° (25°)a
Fields of view (FOVx, FOVy) 420 mm, 420 mm (390 mm, 341 mm)a Fields of view (FOVx, FOVy) 420 mm, 420 mm (390 mm, 341 mm)a
Resolution (Δx, Δy) 0.82 mm, 1.62 mm (0.80 mm, 1.40 Resolution (Δx, Δy) 0.82 mm, 1.62 mm (0.80 mm, 1.40
mm)a mm)a
Number of data points collected (Nx, Ny) 512, 260 (512, 246)a Number of data points collected (Nx, Ny) 512, 260 (512, 246)a
Display matrix (Dx, Dy) 512, 512 (512, 512)a Display matrix (Dx, Dy) 512, 512 (512, 512)a
Slice thickness (Δz) 2.5–3.00 mm; 1.25–1.5 mm after Slice thickness (Δz) 2.5–3.00 mm; 1.25–1.5 mm after
interpolation (3.1 mm interpolated to 2 interpolation (3.1 mm interpolated to 2
mm)a mm)a
Number of slices 20 (40 after interpolation); 26 (40 after Number of slices 20 (40 after interpolation); 26 (40 after
fractional interpolation) fractional interpolation)
Slice gap 0 Slice gap 0
Number of acquisitions (Nacq) 1 (or 2) Number of acquisitions (Nacq) 1
Number of repetitions 1 Number of repetitions 2 (no pause between measurements)
Swap read and phase encoding No Swap read and phase encoding No
ZIP 2 Yes ZIP 2 Yes
Saturation pulses No Saturation pulses No
Scan time 23 sec (16 sec)ab Scan time 23 sec (16 sec)a
aInformation in parentheses refers to parameters for a Sonata scanner. aInformation in parentheses refers to parameters for a Sonata scanner.
bThe scan times are doubled if N
acq = 2.

Sequence 5: Contrast-enhanced MRA

22. Instruct the patient to take in a deep breath and exhale, take in another deep breath,
relax, and hold it. Simultaneously start the injection and initiate the measurement
while the patient is holding his/her breath.
This procedure is preferably performed using a breath-hold mode. Even with proper
hyperventilation, patients may not be able to hold their breath beyond 25 sec. In such cases,
advise the patient to commence slow and shallow breathing when he/she cannot hold
his/her breath any longer.
On newer scanners (e.g., Siemens’ Sonata or Symphony or GE’s Signa Advantage), one
can reconstruct images in real time as they are acquired. This helps to visually confirm the
arrival of the contrast agent. At this point, abort the scan and stop the patient from holding
his/her breath.
Sequence 4: Pre-contrast MASK imaging
23. Using localizer or scout images, place a 3-D imaging slab in a coronal orientation
with the parameters described in Table A28.1.5 for acquiring MASK imaging for
subtraction.
24. Have the patient hyperventilate so that he/she is ready to hold his/her breath. Repeat
step 14. Start the scan according to the parameters listed in Table A28.1.5.
Immediately following the MASK image reconstruction, the patient is ready to receive a
contrast agent for post-gadolinium (Gd) imaging using the same pulse sequence parame-
ters used in MASK imaging.
25. Set up the pulse sequence with the number of acquisitions (Nacq) = 1 and number of
repetitions (or measurements) = 2.
Typically, two repetitions are performed to provide both an arterial and a delayed phase,
which includes both arteries and veins. The first repetition (or first measurement) provides
the arterial phase and the second repetition (or second measurement) provides the venous
phase.
26. Set up the power injector with the remaining contrast agent (estimated based on
patient weight in step 9) followed by 15 ml of saline at an injection rate of 2.5 ml/sec
for both.
27. Estimate measurement start time (delay time) based on the method described in
Critical Parameters and Troubleshooting (see Effects of Bolus Timing; Equation
A28.1.3); or in UNIT A10.1.

In some institutions, a MASK is obtained with 2 acquisitions (or NEX = 2) to provide better On a power injector, one can set this delay time as a countdown time to remind the operator
subtraction of background tissue signal. Remember that the scan time is doubled in that case. when to initiate the pulse sequence.
Contrast- Magnetic
Enhanced On a Sonata scanner using partial Fourier 6/8 along phase encoding direction, the scan Resonance
Renal MRA time is further reduced to 16 sec. Angiography

A28.1.6 A28.1.7
Supplement 11 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 11
 28. Inform the patient to be ready to hold his/her breath. Inject the contrast agent. During
the delay time, between the start of the injection and the start of the measurement, T 1 (msec)
hyperventilate the patient, repeat step 14, and run sequence 5 according to the
parameters described in Table A28.1.6. 50

The power injector automatically flushes saline after the injection of the contrast agent. 0.12
Two repetitions (measurements) are performed back-to-back without any delay between
measurements. Therefore, the total scan time is doubled. Make sure that the patient is
sufficiently hyperventilated to hold his/her breath for the entire duration of the measure- 0.15
ment.
750
The first measurement provides images of arteries only and the second measurement 0.10

Signal (a.u.)
provides images with arteries and veins.
1200
Post-processing
0.05
29. Prior to subjecting the data to a vessel rendering routine, perform the subtraction of
the pre-contrast MASK data (from sequence 4) from post-contrast imaging data (from
sequence 5).
5 10 15 20
If multiple measurements are performed, subtract MASK data from each of these sets.
TR (msec)
30. Input these subtracted data to a vascular rendering program such as the maximum
intensity projection (MIP), which is available on the scanner.
Figure A28.1.1 A simulation of the variation in signal intensity as a function of TR. At a relatively
Alternatively, volume rendering is used to render vessels using surface rendering loga- short TR, spins experience a relatively large number of RF pulses while within the slice. This results
rithms with some translucency. With tortuous vascular anatomy, this method has an in a slow recovery of longitudinal magnetization causing saturation in their signal. Therefore, as TR
advantage over MIP in accurately reflecting vascular pathology. is reduced, the relative contrast between unenhanced blood (T1 = 1200) and tissue (T1 = 750) is
considerably reduced. Experimental TR for rapid imaging is generally held for <5 msec. With a TR
<5 msec, there is no noticeable contrast between blood and tissue (long T1). However, shortening
COMMENTARY T1 (∼50 msec) dramatically improves contrast between blood and tissue. Abbreviation: a.u., arbitrary
Background Information ily due to T1 reduction and is achieved by units.
The prevalence of renal artery stenosis is introducing a paramagnetic contrast agent. The
∼45% in patients with peripheral vascular dis- end result is the reduction in signal loss from result, spins in the blood experience many RF dependent. This may be helpful in imaging that
ease (Iglesias et al., 2000). Renal artery stenosis intra-voxel phase dispersion and from in-plane pulses [n = d/(vTR), where n = no. of RF pulses will acquire a larger FOV (coronal or sagittal
has long been recognized as a cause of hyper- saturation. The principal advantage is that it and d = slice thickness] while inside the same orientation) with a minimum number of parti-
tension and progressive renal insufficiency. allows for placing the imaging slab in a coronal slice (also see UNIT B7.3). This leads to saturation tions.
Such a high prevalence demands techniques orientation covering a larger field-of-view and of spin signal with a smaller contribution from
that will provide sensitivity and specificity bet- thinner partitions. In addition, the RF-spoiled inflow enhancement. The mathematical simu- Pulse sequence parameter optimization
ter than 90%. Over the years, the technical 3-D-gradient recalled echo (GRE) sequence, lation of signal with an ultrashort TR (Fig. The paramagnetic contrast agent is a gad-
developments in the field of MR angiography together with T1 shortening effects from the A28.1.1) shows that when TR is reduced con- olinium chelate, an extracellular agent, and its
have tremendously improved the ability to contrast agent, leads to a further reduction in siderably, the contrast between blood and the concentration in blood rapidly decreases fol-
visualize vessels with sensitivity and specific- the surrounding stationary tissue signal, thus surrounding tissue is very poor prior to injec- lowing injection during longer scan times. Ap-
ity well over 93% (Holland et al., 1996; Hany providing an MR luminogram. tion (long T1) and dramatically improves fol- proximately 50% of the injected contrast agent
et al., 1998; Schoenberg et al., 1998; Lee et al., lowing gadolinium injection (short T1). is washed out from the blood pool during the
1999). Over the past 10 years, MRA has been Critical Parameters and However, blood/tissue contrast can be dra- first pass, with ∼80% of the contrast agent
suggested as a cost-effective alternative to digi- Troubleshooting matically increased using a paramagnetic con- washing out of the blood pool within the first
tal subtraction angiography (DSA) and Dop- trast agent. The increase in signal is due to the 5 min. Since, the concentration of gadolinium
pler sonography. Why use a contrast agent? fact that with contrast agent injection, there is is maximum during the first pass; it is best to
Among available MRA techniques, con- The majority of pulse sequences used in a dramatic reduction in T1 of blood spins, caus- image arteries during this phase.
trast-enhanced magnetic resonance angiogra- dynamic 3-D imaging require a short TR to ing a rapid recovery of its magnetization. This The optimization of the pulse sequence is,
phy (CE-MRA) is more robust and reproduc- complete a measurement in a breath-held pe- rapid recovery of magnetization between RF therefore, dependent on several key parame-
ible compared to traditional non-contrast-en- riod. There are several consequences of reduc- pulses is crucial in increasing the net longitu- ters: (1) the effective relaxation times (T1 and
hanced time-of-flight and phase-contrast ing the TR and TE that can work against the dinal magnetization available during the sub- T2) of blood spins during the first passage of
techniques. By virtue of its insensitivity to flow inflow enhancement of the signal. When the sequent application of RF pulses, which results gadolinium agent during circulation; (2) the
artifacts and slab orientation, CE-MRA is an repeat time, TR, is shortened, the product of in an overall increase in signal (Fig. A28.1.2). circulation time and injection rate to determine
ideal method to precisely depict the arterial flow velocity (v) and TR, which is the distance Since it is the reduction in T1 and not the the exact time location of the peak bolus signal;
Contrast- Magnetic
Enhanced
morphology in the assessment of renal artery traveled within a slice, becomes much time-of-flight effect that is the key to signal and (3) the correct flip angle that provides the Resonance
Renal MRA disease. The vascular signal increase is primar- smaller than the thickness of the slice. As a improvement, imaging is truly orientation-in- maximum signal (based on fixed short TR). Angiography

A28.1.8 A28.1.9
Supplement 11 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 11
 T 1 (msec)
TR = 3.0msec
10
0.3
0.2
Signal (a.u.)
40
0.1
80
100
1200
10 20 30 40 50 60
Flip angle (°)
Figure A28.1.2 A simulation of the variation in overall signal intensity as a function of RF flip angle
at different longitudinal relaxation times, T1. The optimum flip angle for the peak signal shifts towards
high angles as T1 is shortened as shown by the dashed line. The results from this simulation show
that a higher flip angle is suitable for a shorter T1 (from a faster injection) and a longer TR, whereas
a shorter flip angle is suitable at a longer T1 (from slow injection rates) and a short TR. The dashed
line passing through the peaks shows the shift in optimal flip angle as T1 is varied. This simulation
is based on a fixed TR = 3.0 msec. Abbreviation: a.u., arbitrary units.
Although the above conditions vary from pa- gadolinium chelate and is described by the
tient to patient, a qualitative understanding of following equation (Johansson and Ahlstrom,
their effects can be understood by using the 1998; Shetty et al., 2000):
following simple model (Earls et al., 1996; post-Gd pre-Gd
R1,2 = R1,2 + r 1,2 × Cfp
Hany et al., 1997).
T1 and T2 at the region of interest (ROI) Equation A28.1.1
during the first-pass where Rpre−Gd
1,2 is the relaxation rates (inverse of
When the measurement time onset is set to T1 and T2) prior to injection and Rpost−Gd
1,2 is the
coincide with the first-pass contrast bolus arri- resulting relaxation rates (inverse of T1 and T2)
val time at the ROI, the overall signal is based following injection, and r1,2 are the longitudinal
on the effective relaxation times (T1, T2) of (r1) and transverse (r2) relaxivities of a param-
protons in the blood at the ROI during data agnetic agent. Cfp is the first-pass concentration
acquisition. Since the recovery of magnetiza- of gadolinium in the bloodstream. The relaxiv-
tion is based on the relative T1, the shorter the ity of commercial gadolinium chelates are con-
T1, the stronger the signal. The relative change stant over a wide range of magnetic field
in signal (contrast) is based on the differences strength and is between ∼4 liter/sec/mM (r1)
in T1 of the blood and tissue following contrast and ∼6 liter/sec/mM (r2) (see also UNIT B6.4). The
agent injection. The effectiveness of gadolin- first-pass arterial Gd concentration (Cfp) is an
ium in reducing T1 and T2 of blood is based on important factor that determines the overall T1
its relaxivity (r1,2), a relationship between the and T2 at the time of data acquisition and is
relaxation rate and the concentration of the related to the circulation status of the patient
contrast agent. In addition, the circulatory con- (expressed in cardiac output, φ), concentration
dition of the patient also influences the change of the contrast agent (C), and the injection rate
in relaxation times. The correlation between T1, (Trate). Because of the inverse relationship be-
Contrast-
Enhanced T2, post-injection and the T1, T2, prior to injec- tween Cfp and φ, it is important to ensure that
Renal MRA tion is based on the first-pass concentration of when a patient is inside the magnet, he/she is
A28.1.10
Supplement 11 Current Protocols in Magnetic Resonance Imaging
0.4
0.3
0.2
T1, T2 (sec)
T1
0.1 T2
1 2
Injection rate (ml/sec)
Figure A28.1.3 The relative changes in T1 and T2 as a function of injection rate. First-pass values
of T1 and T2 do not change significantly when the injection rate is increased beyond 2.5 ml/sec. A
reasonable injection rate would be between 1.5 and 2.5 ml/sec.
more relaxed so that cardiac output remains low
(Brandfonbrener et al., 1955). In most patients,
this can easily be achieved with simple instruc-
tions and education, along with music.
The corresponding expression for T1 and T2
during the peak arterial flow at the region of
interest is given by the following equation
(Shetty et al., 2000):
φ × Txpre-Gd
Txpost-Gd =
φ + Txpre-Gd × rx × Cfp × Trate
Equation A28.1.2
where, x = 1,2 refers to T1 and T2. These tran-
sient values of T1 and T2 are based on other
variables that are intrinsic and extrinsic to
blood. Figure A28.1.3 shows the relative
changes in T1 and T2 during first pass at differ-
ent injection rates.
Amount of contrast agent
The literature cites various investigators us-
ing a range of contrast agent strategies. One
approach is to administer a fixed amount (40
ml) regardless of patient weight. While this
may help in simplifying the injection protocol,
it may not be economical for patients who do
not require that volume of contrast agent. Many
have advocated imaging with doses of 0.1 or
0.2 mmol/kg. A lower-dose regimen works best
with perfect bolus timing, a rapid injection rate,
Current Protocols in Magnetic Resonance Imaging
3 4
and faster scan times. Higher doses may help
increase signal with proper injection rates to
coincide with the scan duration. At the authors’
institution, a single-dose (0.1 mmol/kg) con-
trast agent protocol is calculated based on the
weight of the patient. A general rule is to use 1
ml for each 10 lb. of patient weight.
Optimum injection rate
It is clear from Fig. A28.1.3 that at an injec-
tion rate beyond ∼2.5 ml/sec, there are no sig-
nificant changes in T1 and T2 values. It is also
possible that at an injection rate higher than 3
ml/sec, signal loss from T2* effect will start to
dominate and overwhelm the gain from T1
enhancement effects. A higher injection rate
necessitates the use of shorter imaging time
(with lower doses) to fit the injection duration.
The other consequence of shortening the scan
time by reducing TR is the increase in band-
width (causing higher noise). A reasonable rate
of injection is ∼2.5 ml/sec with a single dose.
For a consistent injection rate throughout the
scan period, an MR-compatible power injector
is recommended.
Optimum TE
Scanners equipped with high-performance
gradients allow for a much shorter TE. The
advantage of having a short TE is that the sus-
ceptibility-induced dephasing (T2*) artifacts
Magnetic
are greatly reduced. This may be useful in Resonance
Angiography
A28.1.11
Supplement 11
 imaging the lungs, where the air-tissue inter-
face may cause magnetic susceptibility artifact.
With a renal MRA, it is helpful in minimizing
metallic susceptibility artifact originating from
surgical clips or stents. However, shortening the
TE may have an inadvertent effect on widening
the bandwidth, causing a lower signal-to-noise
ratio. Also, a choice of shorter TE may have an
effect on fat-water signal cancellation, causing
a signal loss artifact at tissue boundaries. An
appropriate choice of TE is the minimum
achievable by the scanner at a reasonable re-
ceiver bandwidth of ∼350 to 450 Hz/pixel for
an MRA sequence (sequences 4 and 5).
Optimum TR
The reduction in overall scan time can be
achieved by reducing the TR. Most scanners
allow for a TR as short as 1.5 msec. However,
the reduction in TR is a direct effect of widening
the receiver bandwidth. The reduction of TR is
also achieved by using narrow RF pulses. Most
newer scanners provide shorter RF pulse
lengths for 3-D MRA. However, this has an
effect of aliasing along the slice-select direc-
tion, rendering many end partitions useless.
Selection of TR is made based on the required
scan time for 512 sampling points to fit within
the injection duration. Frequently, the scan time
with a very short TR can be matched with a
shorter injection duration and an appropriate
rate of injection. Although a short TR provides
minimum signal, and, consequently, a lower
signal-to-noise ratio (SNR), the effect of rap-
idly injecting the contrast agent creates a coun-
terbalancing increase in the signal-to-noise ra-
tio (SNR).
Optimum flip angle
The optimum flip angle depends upon the
TR/T1 ratio, which depends on the injection rate.
While any reasonable choice of flip angle may
provide a good signal, it may not be optimal
given the other parameters. At the authors’
institution, a flip angle of 25° is used with a TR
∼4 to 5 msec, whereas a flip angle of 15° is used
with a shorter TR (Fig. A28.1.2).
Bolus arrival (transit) time
The peak enhancement is visualized with the
arrival of a bolus during the first pass. The peak
signal is found when the bolus arrival coincides
with the time corresponding to the central k-
space of the data acquisition. Thus, the knowl-
edge of the bolus arrival time at the intended
Contrast-
Enhanced target is an important first step in the optimiza-
Renal MRA tion of the protocol. This is particularly chal-
A28.1.12
Supplement 11
lenging in contrast-enhanced MRA, as the bo-
lus arrival time varies from patient to patient 120
depending on circulation time and cardiac out- 16 ml bolus injection
put. Other timing strategies include automated 2 ml test bolus injection
bolus detection (Ho and Foo, 1998) and MR
100
fluoroscopy (Willman et al., 1997). The
authors’ approach was to use a small amount
(∼2 ml) of contrast agent for estimating bolus
80
arrival time and use the remaining amount for
the full contrast agent injection as described in
steps 21 and 26.
60
Effects of bolus timing
Signal (a.u.)
The signal component of the data is con-
tained in the central portion of k-space (Shetty 40
et al., 1998). Also, the signal peaks during
maximum arterial concentration, which is
reached during the first-pass phase of arterial 20
circulation. In an ideal situation, the contrast
agent must be delivered in such a way that the
arterial phase occurs during the central portion 0
of k-space. Based on the transit time of the 5 10 15 20 25 30 35
bolus, one can set up imaging parameters in the Time (sec)
sequence delayed to acquire the data during the
peak signal. The scan time and the type of scan
order in lines will determine the delay time Figure A28.1.4 The bolus arrival time differs with different amounts of injected contrast agent as
from the injection. Due to the symmetric tem- shown in this graph. The test bolus (2 ml) plot is shown in square boxes and the injection of an
poral location of k-space, the scan delay time amount (16 ml) is shown in diamond boxes. Both are injected at 1.5 ml/sec followed by a 15-ml
was set up for a reverse centric sequence as: saline flush at the same rate. The shift in peak position increases with injection duration and is
approximated in Equation A28.1.3. Abbreviation: a.u., arbitrary units.
Tdelay = TBAT − Tacq / 2 + (injection duration ) / 2
( )
Equation A28.1.3
where Tdelay is the measured sequence delay A B
time from the start of the injection, TBAT is the Δ Δ
experimentally determined bolus arrival time,
k=0 k=0
and Tacq is the data acquisition time for a single
measurement. Notice that the term (injection k k
duration)/2 was added to Tdelay in order to ac-
count for the delayed arterial peak with full-
dose injection when compared with test bolus
t/Δ t /Δ
dose (Fig. A28.1.4). The possible options are 0 1/4 1/2 3/4 1 0 1/4 1/2 3/4 1
shown in Figure A28.1.5, in which the acquisi-
tion of the central 2⁄3 of the k-space is shown
with respect to the bolus profile. In all, the
signal is based on the relative position of the
central k-space with respect to the bolus arrival
Figure A28.1.5 (A) With a reverse centric ordered pulse sequence where the center of k-space lies in
(Ito et al., 1997). If the central k-space is the middle of the acquisition window, the injection rate is adjusted so that its duration spreads across the
spanned too early or too late with respect to the acquisition window. (B) With a tight bolus, injection rate is increased and its duration is shorter than the
bolus peak, the signal will be poor for arteries. acquisition window but spreads at least 2/3 of the range. This is an ideal situation. Δ, duration of data
The best situation is when the contrast agent is acquisition; t/Δ, shows fraction of k-space covered.
present throughout the data acquisition, which
is usually the case when slow infusion bolus is
used (Fig. A28.1.5A). In this case, high signal
is maintained for all k-space data. However, a
Magnetic
long injection time is necessary to span the Resonance
entire k-space, requiring more contrast agent. Angiography
A28.1.13
Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 11
 Figure A28.1.6 Normal bilateral renal arteries. The bolus arrival timing was estimated by placing
a transverse slice at the mid field-of-view or center of the coronal image. The injection rate was 2.5
ml/sec and the amount of Gd contrast agent was based on a single dose strength (0.1 mmol/kg)
and patient weight. The bolus injection of contrast agent was followed by a 15-ml saline injection at
the same rate. Figure A28.1.7 An image of a bilateral renal artery stenosis. The left renal artery (arrow) appears
occluded; however, a slight hypoperfusion in the left kidney possibly indicates a critical stenosis and
On the other hand, an ideal situation is when a of all image slices to pick the highest (maxi- not an occlusion. This was confirmed by viewing source images. In addition, iliac artery stenosis at
tight bolus injection is used with a bolus spread mum) intensity and project it on a plane per- the origin (arrowhead) is visualized along with an aneurysm of the aorto-iliac junction.
that enhances 2/3 of the k-space including the pendicular to the ray. By orienting the ray at an
central k-space frequencies (Fig. A28.1.5B). angle, another projection image is formed. The vides images similar to those obtained in a achieved at the expense of an increase in scan
projection images are obtained at various view digital subtraction angiography (DSA; Watan- time. Another way to reduce the fat signal is to
Anticipated Results angles from 0° to 360°. These projection im- abe et al., 1998). Most CE-MRA methods use use a TE that corresponds to a fat-water out-of-
ages can be viewed in a cine loop to provide a a 3-D-FLASH (fast low angle shot)-type pulse phase image to minimize the fat signal. The
Post-processing of data 3-D rotational perspective. Sometimes, choos- sequence with RF spoiling. The purpose of RF downside of this technique is the presence of
The maximum intensity projection (MIP) ing a target volume is helpful in minimizing the spoiling is to remove any residual transverse prominent signal loss artifact at the fat-tissue
processing of subtracted data for rendering ves- processing time and removing the presence of magnetization following the data acquisition, boundaries. Using MASK images to subtract
sels can be performed on a main console, on a other high-intensity structures that are not ves- which helps reduce background tissue signal. the surrounding background tissue signal ap-
satellite console, or on a stand-alone worksta- sels. A major drawback in MIP is the stepladder This type of spoiling further improves T1 con- pears to be appropriate given the short TR.
tion. Most scanners and workstations do pro- effect when a view angle is away from the trast against reduced background signal. Al- Another reason for using the subtraction tech-
vide standard software that provides MIP proc- normal projection to the slices. This becomes though the background suppression is not fully nique is to remove signal enhancement from
essing (Figs. A28.1.6 and A28.1.7). In addition, progressively worse as the view angle is in- achievable with RF spoiling alone, additional collecting systems that were present due to a
some workstations provide the capability to creased (Anderson et al., 1990). background suppression is achieved by using small amount of contrast agent injected during
perform volume rendering of the data. MIP the subtraction method at the expense of a loss the test bolus procedure.
images are produced with a ray-tracing algo- Subtraction process in SNR. Many institutions use an additional fat
rithm. Images are stacked together and a ray is Use of pre-Gd MASK images to remove saturation pulse to suppress the fat signal, Breath-holding
Contrast- Magnetic
Enhanced cast through a single pixel in the same location surrounding unenhanced soft tissue signal pro- which has an isointense appearance with blood In some cases where patients find it very Resonance
Renal MRA signal following the injection. However, this is difficult to hold their breath despite administer- Angiography

A28.1.14 A28.1.15
Supplement 11 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 11
 Table A28.1.7 Sequence for Acquiring Phase Images with ECG Triggering for
Flow Quantificationa

Patient position Supine

Scan type 2-D-phase contrast cine gradient-echo
Imaging plane (orientation) Perpendicular to vessel in question
Central slice or volume center Xyphoid
Echo time (TE) 5 msec (3.2 msec)b
Receiver bandwidth (RBW) 195 Hz/pixel (391 Hz/pixel)b
Number of lines per segment 3 (3–7)b
Repeat time (TR) 24 msec (28 msec)b (temporal
resolution)
Delay time (TD) 0 msec
Flip angle (FA) 30° (30°)b
Fields of view (FOVx, FOVy) 180 mm, 135 mm (180 mm, 135 mm)b
Resolution (Δx, Δy) 0.70 mm, 0.75 mm (0.70 mm, 0.75
mm)b
Number of data points collected (Nx, Ny) 256, 180 (256, 180)b
Display matrix (Dx, Dy) 256, 256 (256,256)b
Slice thickness (Δz) 6 mm (5 mm)b
Number of slices 1
Slice gap Not applicable
Number of acquisitions (Nacq) 2 (3)b
Swap read and phase encoding No
Number of cardiac phasesc (90% R-to-R interval)/TR (e.g., if
R-to-R interval = 800 msec, number of
cardiac phases = 720/24 = 30)
Vascular options Velocity encoding (Venc) = 150 cm/sec
ECG gating Yes
Scan timed ∼3 min, 31 sec (∼2 min, 14 sec to 5
min)b
aIf flow quantification is suggested, the patient will require ECG leads placed on his/her chest. Follow the
Figure A28.1.8 An image from a patient with a left renal artery stent. The presence of a stent standard procedure of placing leads on chest to obtain good ECG signal.
(arrow) was visualized from the loss of signal. CE-MRA is a useful technique in visualizing the bInformation in parentheses refers to parameters for a Sonata scanner.
presence of a stent, but not very useful in evaluating stenosis within the stent. This is accomplished cBased on the number of lines per segment. On a Vision scanner, the number of segments is fixed at 3,
by further imaging using a triggered phase-contrast technique. The flow is estimated at the level of whereas it is variable on a newer Sonata scanner. By increasing the number of lines per segment per R-to-R
the stent and is compared with flow in the opposite kidney at the same level. interval, the temporal resolution is increased, but the number of cardiac phases is decreased.
dThis time is dependent on the R-to-R interval of the patient.

ing oxygen and hyperventilation, one might
consider reducing the scan time to ∼3 to 4 sec
by reducing either the number of slices or the
acquisition matrix size. The latter change will
severely compromise the spatial resolution.
The advantage of using short-time acquisition
techniques is that time-resolved imaging can be
generated. In this case, the estimating of bolus
arrival time may not be crucial. With the num-
ber of repetitions increased to 7 to 8 and the
patient freely breathing, at least one, if not
several, of the repetitions (measurements) will
acquire data during peak arterial concentration.
Presence of stent
Contrast- The presence of any metal tends to dephase
Enhanced the spin signal significantly, causing a wide
Renal MRA
halo of dark signal around the metal location
(Fig. A28.1.8). MRA is still a robust technique
for evaluating in-stent stenosis. Although, there
is signal loss around the region where the stent
is placed, one can use phase-contrast imaging
to assess the flow beyond a stent. Using an
electro-cardiograph (ECG)-triggered phase-
contrast imaging sequence (see Table A28.1.7),
multiple images (different phases of the same
imaging slice) are obtained during the entire
cardiac cycle. The selection of a proper velocity
encoding (Venc), at 150 cm/sec, is important to
eliminate any aliasing. This is compared to flow
at the same level in the renal artery from a
normal kidney. A prolonged rise time (defined
as time from onset of systole to the first systolic
peak) beyond 70 msec is considered to be indi-
rect evidence for stenosis within the stent based
on the Doppler literature for the assessment of
renal artery stenosis (Desberg et al., 1990). At
the authors’ institution, flow assessment is per-
formed in patients with renal artery stents for
excluding in-stent stenoses.
Literature Cited
Anderson, C.M., Saloner, D., Tsuruda, J.S., Shapeero,
L.G., and Lee, R.E. 1990. Artifacts in maximum
intensity projection display of MR angiograms.
A.J.R. Am. J. Roentgenol. 154:623-629.
Brandfonbrener, M., Landowne, M., and Shock,
N.W. 1955. Changes in cardiac output with age.
Circulation 12:557-566.
Desberg, A.L., Paushter, D.M., Lammert, G.K.,
Hale, J.C., Troy, R.B., Novick, A.C., Nally, J.V.,
Jr., and Weltevreden, A.M. 1990. Renal artery
stenosis: Evaluation with color Doppler flow
imaging. Radiology 177:749-753.
Earls, J.P., Rofsky, N.M., Decorato, D.R., Krinsky,
G.A., and Weinreb, J.C. 1996. Breath-hold sin-
gle-dose gadolinium-enhanced three-dimen-
sional MR angiography: Usefulness of a timing
examination and MR power injector. Radiology
201:705-710.
Hany, T.F., McKinnon, G.C., Leung, D.A., Pfam-
matter, T., and Debatin, J.F. 1997. Optimization
of contrast timing for breath-hold three-dimen-
sional MR angiography. J. Magn. Reson. Imag-
ing 7:551-556.
Hany, T.F., Lueng, D.A., Pfammatter, T., and Deba-
tin, J.F. 1998. Contrast-enhanced magnetic reso-
Magnetic
nance angiography of the renal arteries. Invest.
Radiol. 9:653-659. Resonance
Angiography

A28.1.16 A28.1.17
Supplement 11 Current Protocols in Magnetic Resonance Imaging Current Protocols in Magnetic Resonance Imaging Supplement 11
 Ho, V.B. and Foo, T.K. 1998. Optimization of gad-
olinium-enhanced magnetic resonance angiog-
raphy using an automated bolus-detection algo-
rithm (MR smartprep). Invest. Radiol. 33:515-
523.
Holland, G.A., Dougherty, L., Carpenter, J.P., Gold-
en, M.A., Gilfeather, M., Slossman, F., Schnall,
M.D., and Axel, L. 1996. Breath-hold ultrafast
three-dimensional gadolinium-enhanced MR
angiography of the aorta and the renal and other
visceral abdominal arteries. A.J.R. Am. J.
Roentgenol 166:971-981.
Iglesias, J.I., Hamburger, R.J., Feldman, L., and
Kaufman, J.S. 2000. The natural history of inci-
dental renal artery stenosis in patients with aor-
toiliac vascular disease. Am. J. Med. 109:642-
647.
Ito, K., Kato, J., Okada, S., and Kumazaki, T. 1997.
K-space filter effect in three-dimensional con-
trast MR angiography. Acta Radiol. 38:172-175.
Johansson, L.O.M. and Ahlstrom, H.K. 1998. Cor-
relation between dose rate and T1 in blood at
Gd-enhanced MR angiography. Acta Radiol.
39:579-582.
Lee, V.S., Rofsky, N.M., Krinsky, G.A., Stemerman,
D.H., and Weinreb, J.C. 1999. Single dose
breath-hold gadolinium-enhanced three-dimen-
sional MR angiography of the renal arteries.
Radiology 211:69-78.
Prince, M.R. 1994. Gadolinium enhanced MR aor-
tography. Radiology 191:155-164.
Schoenberg, S.O., Prince, M.R., Knopp, M.V., and
Allenberg, J.R. 1998. Renal MR angiography.
Magn. Reson. Imaging Clin. N. Am. 2:351-370.
Shellock, F.G. 2001. Pocket Guide to MR Proce-
dures and Metallic Objects. Lippincott-Raven,
Philadelphia.
Shetty, A.N., Bis, K.G., Vrachliotis, T.G., Kirsch,
M., Shirkhoda, A., and Ellwood, R.E. 1998. Con-
trast-enhanced 3D-MRA with centric ordering
k-space: A preliminary experience in imaging the
Contrast-
Enhanced
Renal MRA
A28.1.18
Supplement 11
abdominal aorta, renal and peripheral vascula-
ture. J. Magn. Reson. Imaging 8:603-615.
Shetty, A.N., Bis, K.G., Kirsch, M., Weintraub, J.,
and Laub, G. 2000. Contrast-enhanced breath-
hold three-dimensional magnetic resonance
angiography in the evaluation of renal arteries:
Optimization of techniques and pitfalls. J. Magn.
Reson. Imaging 12:912-923.
Watanabe, Y., Dohke, M., Okumura, A., Amoh, Y.,
Ishimori, T., Oda, K., and Dodo, Y. 1998. Dy-
namic subtraction MR angiography: First-pass
imaging of the main arteries of the lower body.
A.J.R. Am. J. Roentgenol. 170:357-360.
Willman, A.H., Riederer, S.J., King, B.F., Debbins,
J.P., Rossman, P.J., and Ehman, R.L. 1997.
Fluoroscopically triggered contrast-enhanced
three-dimensional MR angiography with ellipti-
cal centric view order: Application to the renal
arteries. Radiology 205:137-146.
Key References
Prince, M.R., Grist, T.M., and Debatin JK. 2002. 3D
Contrast MR Angiography. 3rd revised ed.
Springer-Verlag, New York.
Excellent presentation of technical aspects and
clinical applications of contrast-enhanced MRA
methods.
Shellock, F.G. and Kanal, E. 2001. Magnetic Reso-
nance Procedures: Health Effects and Safety.
CRC Press, Boca Raton, Fla.
Covers important patient management issues re-
lated to MR procedures and lists metallic implants
that have been tested for MR compatibility and other
MR safety issues.
Contributed by Anil N. Shetty and
Kostaki G. Bis
William Beaumont Hospital
Royal Oak, Michigan
Current Protocols in Magnetic Resonance Imaging
Magnetic Resonance Imaging of the UNIT A29.1
Gastrointestinal Tract
The early focus in body MR imaging was on imaging solid visceral organs, such as the BASIC
liver and kidney. MRI of the extrahepatic abdomen, including the gastrointestinal (GI) PROTOCOL
tract, posed significant challenges related to motion artifacts and long exam times.
However, new advances in MR imaging hardware and software have overcome all of these
challenges, so that MR imaging of the GI tract has evolved to become a powerful and
robust imaging technique. The improved efficiency of new MR scanners and software
allows one to image the abdomen and pelvis rapidly using breath-hold sequences. The
excellent contrast conspicuity of MR imaging when combined with gadolinium chelates
and intraluminal contrast material facilitates depiction of mural diseases of the stomach,
small intestine, and colon (Fig. A29.1.1; Kettritz et al., 1995a,b; Low and Francis, 1997;
Lee et al., 1998; Regan et al., 1998; Low et al., 1999, 2002; Prassopoulos et al., 2001;
Schunk, 2002; Manglinte et al., 2003). This unit will review clinical applications and MR
protocols and techniques for MR imaging of the gastrointestinal tract, including ischemic,
inflammatory, and infectious intestinal diseases.
Intraluminal Contrast Agents
Adequate distension of the small bowel and colon will improve depiction of subtle
diseases of the bowel wall (Low and Francis, 1997; Low et al., 1999, 2002; Prassopoulos
et al., 2001; Schunk, 2002). Collapsed bowel will hide abnormalities or may be mistaken
for diseased bowel. There are many different intraluminal contrast agents available. The
authors’ experience has shown the benefits of a biphasic intraluminal agent that is bright
on T2-weighted images and dark on T1-weighted images. A dark or low-signal intraluminal
agent on the gadolinium-enhanced SPGR (spoiled gradient-echo) T1-weighted images is
very useful, as it helps to accentuate enhancement of the adjacent diseased bowel wall. A
A B
Figure A29.1.1 (A) Transverse SS-FSE image through the pelvis shows distension of the bowel
with water-soluble contrast material. (B) Transverse fat-suppressed, gadolinium-enhanced SPGR
image through the pelvis shows moderate thickening and marked enhancement of the terminal
ileum in the right lower quadrant (arrow). In retrospect, this same loop of ileum shows mural
thickening on the SS-FSE image (A). The conspicuity of enhancement with gadolinium on (B) makes
it much easier to appreciate the abnormality. Good bowel distension and homogeneous fat
suppression are important features of this protocol.
Gastrointestinal
Tract
Contributed by Russell N. Low A29.1.1
Current Protocols in Magnetic Resonance Imaging (2004) A29.1.1-A29.1.13
Copyright © 2004 by John Wiley & Sons, Inc. Supplement 12
 high-signal intraluminal agent on T1-weighted images could obscure subtle mural en- Table A29.1.1 Equipment Specifications Needed to Perform MRI of