Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Myocardial
Perfusion Imaging
(Revised Edition)
A Technologist’s Guide
Contributors
Abreu, Carla (London) Koziorowski, Jacek (Linköping)
Acampa, Wanda (Naples) Lezaic, Luka (Ljubljana)
Assante, Roberta (Naples) Mann, April (South Hadley)
Ballinger, James (London) Medolago, Giuseppe (Bergamo)
Fragoso Costa, Pedro (Oldenburg) Pereira, Edgar (Almada)
Figueredo, Sergio (Lisbon) Santos, Andrea (Alverca do Ribatejo)
Geão, Ana (Lisbon) Vara, Anil (Brighton)
Ghilardi, Adriana (Bergamo) Zampella, Emilia (Naples)
Holbrook, Scott (Gray)
Contents
EANM
Foreword 4
Introduction 5
Borut Vidovič
Chapter 1
State of the Art in Myocardial Imaging 6
Wanda Acampa, Emilia Zampella and Roberta Assante
Chapter 2
Clinical Indications 16
Luka Lezaic
Chapter 3
Patient Preparation and Stress Protocols 23
Giuseppe Medolago and Adriana Ghilardi
Chapter 4
Multidisciplinary Approach and Advanced Practice 35
Anil Vara
Chapter 5
Advances in Radiopharmaceuticals for Myocardial Perfusion Imaging 42
James R. Ballinger and Jacek Koziorowski
Chapter 6
SPECT and SPECT/CT Protocols and New Imaging Equipment 54
Chapter 7
PET/CT Protocols and Imaging Equipment (*) 62
April Mann and Scott Holbrook
Chapter 8
Image Processing and Software 77
Sérgio Figueiredo and Pedro Fragoso Costa
Chapter 9
Artefacts and Pitfalls in Myocardial Imaging (SPECT, SPECT/CT and PET/CT) 109
Ana Geão and Carla Abreu
Imprint126
3
Foreword
The EANM Technologist Committee was dural workflow and need to cooperate with
created more than 20 years ago. From the different professionals including nurses, phy-
outset it worked not only to improve the pro- sicians, cardiologists and radiopharmacists.
fessional expertise of nuclear medicine tech-
nologists (NMTs) in Europe but also to assist The current book is specifically aimed at
in raising the quality of Nuclear Medicine radiographers and technologists working
clinical practice. Over the past two decades, or intending to work in a Nuclear Medicine
it has developed continuously to become an department performing MPI; it will also have
important group within the EANM. value for other healthcare professionals
working or willing to work in this challenging
With the above-mentioned goals of the environment. As you will see, some chapters
Committee in mind, in early 2004 the idea from the previous edition of the guide have
of providing a series of books on imaging for been revised and updated by the authors
technologists was conceived. By September while new chapters and new contributors
2004, thanks to the hard work of the EANM have been added to extend and further im-
Technologist Sub-committee on Education, prove the quality of the book.
it was possible to publish the first Technolo-
gist’s Guide, dedicated to myocardial perfu- I would like to thank all those who have
sion scintigraphy. The success of this first contributed to the realisation of this project,
guide led the Technologist Committee to whether as authors or reviewers, without
propose a new book every year, starting a whom the book would not have been pos-
series of guides that is still ongoing. sible. I also wish to thank SNMMITS (Society
of Nuclear Medicine and Molecular Imaging
The technological advancements in Nuclear Med- Technologist Section) and EANM Cardiovas-
icine and Molecular Imaging of the past decade cular Committee for their help and high-
have necessitated an upgrading of this pioneer- quality contributions.
ing series, and the EANM Technologist Commit-
tee accordingly decided to expand it to encom- Special thanks are due to Helen Ryder, Vanessa
pass recent developments in scanner technology, Veloso Jerónimo and Borut Vidovič for their in-
radiopharmaceuticals and stress tests. credible enthusiasm and competence in deal-
ing with the editorial duties and organisation-
Myocardial perfusion imaging (MPI) has a al work. Finally, I remain extremely grateful to
leading role in cardiological diagnostic pro- the EANM Board, the EANM Executive Office,
tocols due to its great efficiency in detecting the Technologist Committee and all the other
ventricular perfusion defects with an almost EANM committees involved in the project.
non-invasive procedure. MPI is, however, a
challenging methodology for technologists, Giorgio Testanera
who may cover various roles in the proce- Chair, EANM Technologist Committee
4
Introduction
Borut Vidovič
EANM
Cardiovascular diseases are the leading cause and the imaging equipment, with image
of morbidity and mortality in the developed processing and software. The final chapter
world, and their frequency is also increasing elucidates the causes and effects of potential
in less developed countries [1]. Reliable and artefacts and pitfalls in myocardial perfusion
rapid diagnosis is important to reduce the imaging.
number of deaths and allow introduction of
appropriate therapy at a very early stage of I would like to thank all the authors who have
the disease. taken the time to write the chapters and all of
my fellow editors who have helped to create
Given that the last EANM Technologist’s this book. I hope all professionals who work
Guide entitled Myocardial Perfusion Imaging in the area of Nuclear Medicine Cardiology or
was released way back in 2004, it is certainly are interested in this topic will enjoy reading
time for this new book. During the interven- and using the book.
ing period, Nuclear Medicine Cardiology has
made great progress, with the development Reference
of new radiopharmaceuticals for myocardial 1. http://epp.eurostat.ec.europa.eu/cache/ITY_OFFPUB/
perfusion imaging and the introduction of KS-30-08-357/EN/KS-30-08-357-EN.PDF, accessed July
2014
new imaging equipment with new post-
processing programs.
5
Chapter 1
State of the Art in Myocardial Imaging
Wanda Acampa, Emilia Zampella and Roberta Assante
6
Chapter 1 State of the Art in Myocardial Imaging
EANM
the potential to decrease patient radiation plifies the light signal into an electrical signal.
exposure. Low-dose protocols have been This requires a large number of photons to
shown to yield high-quality images while be incident at the detectors. Novel semicon-
limiting radiation exposure to as little as 1–2 ductor CZT SPECT uses detectors that direct-
mSv [2]. The last three decades have also wit- ly convert photon energy into an electrical
nessed changes in radionuclide use (less du- signal, eliminating the photomultiplier tube
al-isotope imaging) and imaging sequences and resulting in better energy resolution and
(more stress/rest imaging with the option of a much more compact detector. This new
stress only) and improvements in processing geometric design and the new detector ma-
software that incorporate iterative recon- terial, combined with novel reconstruction
struction techniques. algorithms, have led to improved diagnostic
performance of CZT SPECT, with reduction of
New-generation cardiac scanners: acquisition times and radiation dose (4.2 mSv
cadmium-zinc-telluride detectors vs 11.8 mSv) [3] without any significant sac-
In new-generation dedicated cardiac ultra- rifice in accuracy. Low-dose protocols have
fast-acquisition scanners, eight detectors been shown to yield high-quality images
surrounding the patient simultaneously im- and several studies have not only revealed a
age the heart. These new designs vary in good concordance with results provided by
respect of the number and type of scanning a conventional Anger camera but also dem-
or stationary detectors and whether NaI, CsI onstrated that the physical characteristics of
or cadmium-zinc-telluride (CZT) solid-state CZT cameras, such as their enhanced count
detectors are used. They all have in common sensitivity, have the potential to improve on
the potential for a five- to tenfold increase in and therefore modify the results from Anger
count sensitivity at no loss of, or even a gain SPECT.
in, resolution, resulting in the potential for ac-
quisition of a scan within 2 min or less if the SPECT/CT cameras
patient is injected with a standard dose. Hybrid systems combining SPECT and com-
puted tomography (CT) acquisition are
The introduction of modern dedicated gam- available. SPECT/CT devices can perform
ma cameras allows some limitations of Anger CT acquisitions for attenuation correction
SPECT, such as the lengthy acquisition time of SPECT data and to estimate the coronary
and radiation exposure, to be overcome. calcium score. For clinical applications, these
Traditional Anger SPECT uses collimators to devices permit 3D fusion display of SPECT
direct photons towards a scintillation crystal, and CT images, thereby allowing accurate
and a photomultiplier tube converts and am- assignment of a stenotic coronary vessel to
7
the corresponding stress-induced myocar- The prognostic power of MPS has been
dial perfusion defect [4]. SPECT/CT increases extensively evaluated, and several studies
the diagnostic accuracy of SPECT imaging have demonstrated that it adds incremen-
in patients with suspected and known CAD tal prognostic value to the information ob-
and myocardial perfusion defects at SPECT tained from patient clinical variables, stress
imaging. Moreover, in patients with normal test and angiographic findings. SPECT has
perfusion at SPECT and suspected three- a high negative predictive value and pa-
vessel disease, SPECT/CT can provide useful tients with normal MPS have an excellent
information on risk stratification [4]. prognosis, with an annual cardiac event rate
<1% [7]. In these patients, in the absence of
Diagnostic and prognostic value of MPS new symptoms and clinical conditions such
In clinical practice, MPS is used to establish as diabetes and known CAD, MPS may not
a diagnosis of CAD and to provide accurate need to be repeated for 3–5 years. Moreover,
risk stratification. Diagnostic accuracy of the in patients with abnormal MPS, the severity
method is influenced by several variables, and extent of perfusion may provide further
such as pre-test likelihood of disease. Accu- prognostic information, considering that the
racy in the identification of CAD has been annual event rate increases in accordance
demonstrated to be highest in patients with with both the severity and the extent of fixed
intermediate pre-test CAD likelihood. Similar- or reversible perfusion defects [8]. One of the
ly, when performed for the purpose of prog- benefits of the wealth of evidence on prog-
nostic evaluation in patients with suspected nostic evaluation by means of MPS is that the
or known CAD, MPS offers the greatest ben- data can be easily integrated into risk-based
efit in those at intermediate risk [5]. On the patient management algorithms. A review
basis of MPS, a large majority of intermedi- of evidence suggests that high-risk findings
ate-risk patients may be shifted to lower-risk are obtained on MPS in patients whose ex-
cohorts (when the test result is negative) or pected rate of major adverse cardiac events
higher-risk cohorts (in the setting of a mod- is 3%–5% or more and patients with moder-
erately to severely abnormal perfusion scan). ately to severely abnormal perfusion abnor-
Published reports note that, on average, 53% malities or a summed stress score greater
of patients have normal or low-risk myocar- than 8. Patients with a high-risk post-stress
dial perfusion imaging results. Accordingly, it LVEF lower than 45% are at an elevated risk of
is expected that approximately half of inter- major adverse cardiac events.
mediate-risk patients will be at low risk after
testing, with an expected annual mortality of Risk stratification is widely used in the prog-
approximately 0.6% [6]. nostic assessment of patients with a variety
of clinical disorders, on the unquestioned
8
Chapter 1 State of the Art in Myocardial Imaging
EANM
assumption that the intensity of treatment those with reduced post-stress LVEF and se-
should be proportionate to the risk of an ad- vere ischaemia [14].
verse event. A recent review evaluated the
comparative ability of stress MPS risk mark- The estimation of prognosis with an accu-
ers using various iterative and risk classifica- rate and reproducible modality, such as MPS,
tion approaches [9]. Other studies [10,11] allows a more precise linkage with efforts
have compared analytical approaches to to reduce patient risk and therapeutic risk.
assess the added value of MPS variables in In fact, a nuclear-based focus for decision-
estimating CAD outcomes in asymptomatic making concentrates on the physiological
diabetic patients. At multivariable analysis, significance of the disease state and its rela-
post-stress LVEF and stress MPS ischaemia tionship to event risk. In the long-term prog-
were independent predictors of CAD death nostic evaluation of patients with suspected
or myocardial infarction (MI) [11]. Moreover, or known CAD, serial stress MPS makes an
the addition of MPS results to a prediction important contribution to clinical decision-
model based on traditional risk factors and making. This is important in such patients
ECG stress test data significantly improved when comparing a repeat study with a pre-
the classification of risk. In particular, the vious one: the interpreting physician should
addition of MPS data to pre-test CAD like- pay close attention to whether a new perfu-
lihood was important in estimating risk of sion defect has developed or whether the
CAD death or MI and effective in reclassify- extent of pre-existing ischaemia has sub-
ing a sizeable proportion (55%) of patients, stantially increased or decreased. Published
yielding a significant net improvement com- reports have clearly shown that quantitative
pared with traditional approaches to prog- assessment of perfusion abnormality on MPS
nostication [11]. More recent data showed has a minimal intrinsic variability [15]. Soft-
that also different functional parameters ware and scan reproducibility is extremely
obtained by MPS, such as TID, help to iden- high because of the high degrees of automa-
tify patients with higher mortality [12,13]. tion of the algorithms. Different studies have
In particular, among patients with diabetes, provided quantitative criteria for ascertain-
when only subjects with post-stress LVEF ing by means of MPS whether a meaningful
>45% and no ischaemia were considered, change has occurred in an individual patient
TID was a strong predictor of cardiac events and these criteria should help guide clinical
at long-term follow-up, with an event rate management decisions [15,16]. In the COUR-
of 0.21% per year in those without TID and AGE nuclear sub-study, a change of ≥5% was
4.9% per year in those with TID [13]. Interest- used as the criterion for a significant serial
ingly, it emerged that the beneficial effect change in ischaemic total perfusion defect in
of revascularisation on hazard ratio is more an individual patient [17].
evident in patients with TID, particularly in
9
Cardiac PET/CT reduction in the time required to complete
PET/CT systems were initially dedicated to rest and stress studies compared with MPS
oncology imaging. The increasing evidence studies. Dynamic acquisitions are performed
of their clinical usefulness in cardiology is to obtain perfusion data while static perfu-
contributing in advancing the clinical role sion images, as well as ECG-gated images,
of PET/CT in cardiovascular imaging. Sev- are acquired for the evaluation of LV func-
eral technical advantages account for the tion and wall motion. Modern PET systems
improved image quality and diagnostic abil- are able to perform “list mode” acquisition,
ity of PET compared with SPECT, including where spatial data for every coincident pair
the high spatial resolution of reconstructed of photons are recorded sequentially with
images (heart–background ratio), the high very high temporal resolution along with
sensitivity in the identification of small con- the associated electrocardiographic phase,
centrations of radiotracers, and above all providing maximal flexibility for the offline
the high temporal resolution, which allows reconstruction of summed, gated or dynam-
dynamic sequences to be obtained in order ic images. Through list-mode acquisitions,
to describe tracer kinetics and perform ab- multiple images are reconstructed from a
solute measurements of myocardial blood single acquisition. With this approach, image
flow (MBF). This last-mentioned feature also acquisition starts with the bolus injection of
results in improved detection of multivessel the radiopharmaceutical and continues for
CAD. Moreover, the appropriate attenua- 6–7 min after 82Rb injection and 10–20 min
tion correction obtained with simultaneous after 13N-ammonia injection. Stress testing is
acquisition PET/CT technology decreases most commonly performed with pharmaco-
false-positives, increasing specificity, and logical agents.
provides a potential opportunity to delin-
eate the anatomical extent and physiologi- Diagnostic and prognostic value of PET
cal severity of coronary atherosclerosis in The average weighted sensitivity for detec-
a single setting. At the moment, cardiac tion of at least one coronary artery with
PET/CT using perfusion tracers represents >50% stenosis is 90% (range 83–100%),
the gold standard for quantifying myocar- while the average specificity is 89% (range
dial absolute perfusion (ml/min/g), at both 73–100%). The corresponding average posi-
rest and stress acquisition, and coronary flow tive predictive value (PPV) and negative pre-
reserve, defined as the ratio between MBF dictive value (NPV) are 94% (range 80–100%)
at peak stress and MBF at rest. Several PET and 73% (range 36–100%), respectively, and
myocardial perfusion tracers are available for the overall diagnostic accuracy is 90% (range
clinical use, such as rubidium-82 (82Rb) and 84–98%) [18]. Myocardial perfusion PET has
nitrogen-13 (13N) ammonia. The short physi- higher diagnostic accuracy than SPECT [19].
cal half-life of PET perfusion tracers allows a This is because coronary vasodilator reserve
10
Chapter 1 State of the Art in Myocardial Imaging
EANM
is globally reduced in patients with diffuse a novel finding in this study was that LVEF
CAD, limiting detection of multivessel coro- reserve is a significant independent predic-
nary artery stenosis. Diagnosis of multivessel tor of these events [22]. In particular, LVEF re-
CAD can be helped by PET measurements of serve is inversely and independently related
MBF (ml/min/g) and coronary flow reserve. to the risk of events.
It has been demonstrated that estimates of
coronary vasodilator reserve by PET are in- Comprehensive overviews of the role of
versely and non-linearly related to the sever- [18F]-2fluoro-2-deoxy-D-glucose (FDG) PET in
ity of stenosis [20]. Segments with reversible the assessment of hibernating myocardium
ischaemia and coronary stenosis had re- have recently been published (e.g. [23]). In
duced hyperaemic MBF. Even segments that the setting of persistent low-flow ischaemia,
had significant coronary stenosis, but were in which perfusion defects may show little
negative for ischaemia, had reduced hyper- evidence of reversibility, the ischaemic car-
aemic MBF compared with non-stenotic diac myocytes will utilise glycolysis and FDG
segments. Thus, MBF quantification by 82Rb uptake will increase, producing the classic
PET may provide additional diagnostic infor- mismatch pattern between blood flow, de-
mation in patients with CAD [21]. However, termined by perfusion PET, and metabolism.
it is important to note that myocardial perfu- FDG PET assesses cellular metabolism, and
sion quantitation with PET methods cannot thus also cellular integrity. As recently point-
help in differentiating decreased perfusion ed out, techniques that detect presence of
due to epicardial coronary artery stenosis cellular integrity are likely to be more sensi-
from that due to microvascular dysfunc- tive but less specific than techniques that
tion. Hybrid PET/coronary CT angiography detect contractile reserve, like dobutamine
may be particularly helpful in this regard. It CMR, because a critical myocardial mass
has been demonstrated that 82Rb PET MPI needs to be viable for a functional response
provides independent and incremental to occur [24].
information for the prediction of cardiac
events and all-cause death. Risk-adjusted Other myocardial imaging modalities
survival analysis showed that both statisti- Echocardiography
cal (increased chi-square and ROC analysis) Echocardiography represents the most fre-
and clinical enhanced risk stratification was quently performed cardiac imaging investi-
achieved on the basis of the ischaemic bur- gation and is an invaluable tool for accurate
den and scar on PET MPI. The percentage of evaluation of cardiac structure and function.
ischaemic or scarred myocardium on 82Rb Numerous modalities are available for assess-
PET MPI added incremental value to the ment of cardiac morphology and function:
clinical data and rest LVEF in predicting car- M-mode has a high temporal resolution and
diac event and all-cause death. Importantly, permits accurate depiction of rapidly moving
11
structures; 2D transthoracic echocardiogra- Compared with MPS, SE was found to dis-
phy allows identification of structural abnor- play significantly superior specificity for
malities; transoesophageal echocardiogra- detection of CAD, while nuclear perfusion
phy permits more detailed views of cardiac had better sensitivity [27]. The ischaemic
structures; and Doppler echocardiography is cascade can explain this finding, whereby
able to identify the direction, velocity, ampli- the perfusion abnormalities precede wall
tude and timing of blood flow through the motion abnormalities.
heart.
Coronary CT angiography
Stress echocardiography Coronary CT angiography has evolved into
Stress echocardiography (SE) has proved a a robust and non-invasive tool for the as-
useful diagnostic and prognostic tool in pa- sessment of CAD, being able to identify cor-
tients with known or suspected CAD. Several onary anatomy, presence of obstructive and
advances have helped to define this role, non-obstructive CAD and plaque charac-
including the introduction of synchronised teristics. With the latest multislice CT scan-
display of rest and stress images side by ners (64- and post-64-slice CT), coronary
side using digital acquisition, the improve- CT angiography has been reported to have
ment of endocardial detection by tissue har- high diagnostic and prognostic value for
monic imaging, the use of contrast agent to prediction of major cardiac events. It pro-
achieve LV enhancement and detection of vides non-invasive visualisation of the coro-
myocardial perfusion, 3D imaging and myo- nary artery with a high sensitivity (>95%)
cardial tissue quantitative techniques (tissue and good specificity (82%) for detection of
Doppler, myocardial strain and strain rate coronary artery stenosis [28]. Despite this,
imaging). Stress ischaemia can be induced the presence of anatomically significant
by exercise, pharmacological agents (such coronary stenosis does not always equate
as dobutamine and vasodilator agents, e.g. to functional significance. An alternative ap-
dipyridamole or adenosine) or transoesoph- proach to assessment of the functional sig-
ageal atrial pacing [25]. Evaluation of SE im- nificance of a coronary stenosis is CT-based
aging is performed using qualitative assess- fractional flow reserve (CT-FFR) quantifica-
ment based on comparison of rest and stress tion. This method employs computational
images for global and regional dysfunction. fluid dynamics to calculate pressure gradi-
Four response patterns are described: “nor- ents across coronary lesions on standard CT
mal”, “inducible ischaemia”, “fixed scar (ne- coronary angiograms. Use of CT-FFR data
crosis)” and “stunned” or “biphasic response”, improves diagnostic accuracy and discrimi-
as may be seen in dobutamine SE viability nation of functionally significant stenosis
studies [26]. over CT angiography alone [29]. Other
12
Chapter 1 State of the Art in Myocardial Imaging
EANM
limitations of coronary CT angiography in- analysis most commonly employs dobuta-
clude inferior temporal resolution, motion- mine stress. Advantages of CMR are good
related artefacts and high false positive re- specificity and sensitivity in detecting perfu-
sults due to severe calcification, presence of sion and wall motion abnormalities, concur-
coronary stents and coronary artery bypass rent assessment of cardiac structures and
grafts, and arrhythmia. function and lack of radiation [30]. However,
CMR has several limitations, including cost,
Calcium score test duration, variable centre expertise, pa-
The amount of coronary artery calcium, tient claustrophobia, paramagnetic implants
based on CT and traditionally expressed as and gadolinium-associated nephrogenic sys-
the calcium score, is a marker of coronary temic fibrosis. For the assessment of coronary
atherosclerosis, correlated with the coronary anatomy, magnetic resonance angiography
atherosclerotic plaque burden. The coronary is used, but its operating characteristics are
artery calcium score is associated with the inferior to those of other available modalities.
risk of future cardiovascular events, and the
detection of coronary artery calcification New advances: PET/MR
with cardiac CT has been used for risk strati- The first available whole-body system for
fication. Although coronary artery calcium PET/MR hybrid imaging was introduced in
scoring reflects the coronary atherosclerotic 2010, based on two separate MR and PET im-
plaque burden, the absence of coronary ar- agers in one room. This system was followed
tery calcification does not exclude obstruc- by a fully integrated hybrid PET/MR system
tive CAD with non-calcified plaque. (Biograph mMR, Siemens AG, Healthcare Sec-
tor, Erlangen, Germany) that enables simulta-
Cardiac magnetic resonance neous PET/MR data acquisition. Integrated
Magnetic resonance imaging has rapidly PET/MR appears a promising diagnostic tool
developed into a versatile tool for investi- for evaluation of cardiovascular disease be-
gating CAD, being able to evaluate cardiac cause it combines the ability of MR to pro-
structure and ventricular function and to duce high-resolution images and provide in-
detect myocardial perfusion defects or in- formation about anatomy, function, flow and
farction scar. CMR stress protocols are very perfusion with the strength of PET in quanti-
useful for evaluation of myocardial perfu- fying physiological and metabolic processes.
sion and wall motion. For the assessment of
myocardial perfusion, first-pass myocardial
enhancement with gadolinium is performed
using predominantly stress induced by a va-
sodilator, while stress-induced wall motion
13
References Chapter 1
References
1. Shaw LJ, Hage FG, Berman DS, Hachamovitch R, Iskan- 10. Acampa W, Petretta M, Evangelista L, Daniele S, Xhoxhi
drian A. Prognosis in the era of comparative effective- E, De Rimini ML, et al. Myocardial perfusion imaging and
ness research: where is nuclear cardiology now and risk classification for coronary heart disease in diabetic
where should it be? J Nucl Cardiol. 2012;19:1026–43. patients. The IDIS study: a prospective, multicentre trial.
Erratum in: J Nucl Cardiol. 2012;19:1092–3. Eur J Nucl Med Mol Imaging. 2012;39:387–95.
2. Nakazato R, Berman DS, Hayes SW, Fish M, Padgett R, Xu 11. Acampa W, Petretta M, Daniele S, Del Prete G, Assante
Y, et al. Myocardial perfusion imaging with a solid-state R, Zampella E, Cuocolo A. Incremental prognostic value
camera: simulation of a very low dose imaging protocol. of stress myocardial perfusion imaging in asymptomatic
J Nucl Med. 2013;54:373–9. diabetic patients. Atherosclerosis. 2013;227:307–12.
3. Garcia EV, Faber TL, Esteves FP. Cardiac dedicated ultra- 12. Petretta M, Acampa W, Daniele S, Petretta MP, Nappi
fast SPECT cameras: new designs and clinical implica- C, Assante R, et al. Transient ischemic dilation in SPECT
tions. J Nucl Med. 2011;52:210–7. myocardial perfusion imaging for prediction of severe
coronary artery disease in diabetic patients. J Nucl Car-
4. Schindler TH, Magosaki N, Jeserich M, Oser U, Krause diol. 2013;20:45–52.
T, Fischer R, et al. Fusion imaging: combined visualiza-
tion of 3D reconstructed coronary artery tree and 3D 13. Petretta M, Acampa W, Daniele S, Petretta MP, Plaitano
myocardial scintigraphic image in coronary artery dis- M, Cuocolo A. Transient ischemic dilation in patients
ease. Int J Card Imaging. 1999;15:357–68. with diabetes mellitus: prognostic value and effect on
clinical outcome after coronary revascularization. Circ
5. Shaw LJ, Iskandrian AE. Prognostic value of gated myo- Cardiovasc Imaging. 2013;6:908–15.
cardial perfusion SPECT. J Nucl Cardiol. 2004;11:171–85.
14. Acampa W, Cantoni V, Green R, Cuocolo R, Petretta MP,
6. Shaw LJ, Hendel R, Borges-Neto S, Lauer MS, Alazraki N, Orlandi M, Petretta M. Prognostic value of stress myocar-
Burnette J, et al.; Myoview Multicenter Registry. Prog- dial perfusion imaging in asymptomatic diabetic patients.
nostic value of normal exercise and adenosine (99m)Tc- Current Cardiovascular Imaging Reports. 2014;7:9268.
tetrofosmin SPECT imaging: results from the multicenter
registry of 4,728 patients. J Nucl Med. 2003;44:134–9. 15. Xu Y, Hayes S, Ali I, Ruddy TD, Wells RG, Berman DS, et al.
Automatic and visual reproducibility of perfusion and
7. Hachamovitch R, Hayes S, Friedman JD, Cohen I, Shaw function measures for myocardial perfusion SPECT. J
LJ, Germano G, Berman DS. Determinants of risk and its Nucl Cardiol. 2010;17 :1050–7.
temporal variation in patients with normal stress myo-
cardial perfusion scans: what is the warranty period 16. Mahmarian JJ, Moyé LA, Verani MS, Bloom MF, Pratt CM.
of a normal scan? J Am Coll Cardiol. 2003;41:1329–40. High reproducibility of myocardial perfusion defects
in patients undergoing serial exercise thallium-201 to-
8. Sharir T, Germano G, Kang X, Lewin HC, Miranda R, Co- mography. Am J Cardiol. 1995;75:1116–9.
hen I, et al. Prediction of myocardial infarction versus
cardiac death by gated myocardial perfusion SPECT: 17. Shaw LJ, Berman DS, Maron DJ, Mancini GB, Hayes SW,
risk stratification by the amount of stress-induced isch- Hartigan PM, et al.; COURAGE Investigators. Optimal
emia and the poststress ejection fraction. J Nucl Med. medical therapy with or without percutaneous coronary
2001;42:831–7. intervention to reduce ischemic burden: results from
the Clinical Outcomes Utilizing Revascularization and
9. Petretta M, Cuocolo A. Prediction models for risk clas- Aggressive Drug Evaluation (COURAGE) trial nuclear
sification in cardiovascular disease. Eur J Nucl Med Mol substudy. Circulation. 2008;117:1283–91.
Imaging. 2012;39:1959–69.
18. Di Carli MF, Murthy VL. Cardiac PET/CT for the evalua-
tion of known or suspected coronary artery disease.
Radiographics. 2011;31:1239–54.
14
References Chapter 1
EANM
19. Bateman TM, Heller GV, McGhie AI, Friedman JD, Case 28. Paech DC, Weston AR. A systematic review of the clinical
JA, Bryngelson JR, et al. Diagnostic accuracy of rest/ effectiveness of 64-slice or higher computed tomogra-
stress ECG-gated Rb-82 myocardial perfusion PET: com- phy angiography as an alternative to invasive coronary
parison with ECGgated Tc-99m sestamibi SPECT. J Nucl angiography in the investigation of suspected coronary
Cardiol. 2006;13:24–33. artery disease. BMC Cardiovasc Disord. 2011;11:32.
20. Anagnostopoulos C, Almonacid A, El Fakhri G, Curillova 29. Min JK, Leipsic J, Pencina MJ, Berman DS, Koo BK, van
Z, Sitek A, Roughton M, et al. Quantitative relationship Mieghem C, et al. Diagnostic accuracy of fractional
between coronary vasodilator reserve assessed by 82Rb flow reserve from anatomic CT angiography. JAMA.
PET imaging and coronary artery stenosis severity. Eur J 2012;308:1237–45.
Nucl Med Mol Imaging. 2008;35:1593–1601.
30. Nandalur KR, Dwamena BA, Choudhri AF, Nandalur MR,
21. Yoshinaga K, Katoh C, Manabe O, Klein R, Naya M, Sakaki- Carlos RC. Diagnostic performance of stress cardiac
bara M, et al. Incremental diagnostic value of regional magnetic resonance imaging in the detection of coro-
myocardial blood flow quantification over relative per- nary artery disease: a meta-analysis. J Am Coll Cardiol.
fusion imaging with generator-produced rubidium-82 2007;50:1343–53.
PET. Circ J. 2011;75:2628–34.
26. Pellikka PA, Nagueh SF, Elhendy AA, Kuehl CA, Sawada
SG; American Society of Echocardiography. American
Society of Echocardiography recommendations for per-
formance, interpretation, and application of stress echo-
cardiography. J Am Soc Echocardiogr. 2007;20:1021–41.
15
Chapter 2
Clinical Indications
Luka Lezaic
16
Chapter 2 Clinical Indications
EANM
least expressed with thallium-201 (201Tl), more able to detect diffuse reduction in coronary
so with technetium-99m (99mTc)-sestamibi flow reserve.
and most with 99mTc-tetrofosmin. At exercise
and particularly with pharmacological stress Differences in kinetics of clinically used trac-
(which increases regional blood flow approxi- ers do not result in clinically detectable dif-
mately 5 times above rest, while exercise-in- ferences. Overall, the substantial body of
duced stress increases it to 2–3 times above evidence accumulated over four decades
rest), mild-to-moderate narrowing of coro- shows that for detection of significant isch-
nary vessels may go undetected. Fortunately, aemic heart disease, MPS yields a sensitivity
significant (>70%) coronary stenoses result in of approximately 85% with a slightly lower
a sufficient difference in myocardial uptake specificity of approximately 70% [5–11]. Im-
to be detected at imaging. In addition to de- provement in specificity can be achieved by
tecting myocardial ischaemia, MPS is able to the use of ECG-gated imaging and/or attenu-
assess its location (and therefore predict the ation correction. ECG-gated imaging allows
major epicardial vessel involved), extent (the assessment of LV volumes, ejection fraction
amount of myocardium involved) and sever- and regional contractility (wall motion and
ity (level of ischaemia). MPS can be quantified: thickening). Regional contractility analysis
the left ventricle is divided into 17 segments is used for differentiation of attenuation ar-
using a universally accepted model for cardiac tefacts from ischaemic changes through re-
imaging [3] and segmental perfusion deficit is tained regional contraction, improving speci-
scored on a 0 (no defect) to 4 (no perfusion) ficity (approaching 85%) [12]. Furthermore,
scale by visual analysis or, preferably, by com- extensive balanced ischaemia can be de-
parison to a normal database. Quantification tected through relative reduction of ejection
adds to confidence, reproducibility and ob- fraction at stress and/or increase in volume
jective assessment of the degree of perfusion of the left ventricle (transient ischaemic dila-
abnormality [4]. However, if all major epicardi- tion [13]). Attenuation correction also serves
al vessels are equally affected (similar degree to differentiate attenuation artefacts from
of stenosis), no relative difference in coronary ischaemic changes and has been shown to
flow reserve exists in different parts of the improve specificity and overall accuracy [14].
myocardium and therefore no ischaemia is Both techniques require constant quality
detected, although coronary flow reserve is control to provide reliable information.
reduced (“balanced” ischaemia). In this case,
absolute quantification of myocardial blood Risk assessment/prognosis in ischaemic
flow using positron-emitting flow tracers heart disease
[rubidium-82 (82Rb), nitrogen-13 ammonia Arguably the most important role of MPS is
(13NH3) or oxygen-15 labelled water (H215O)] risk assessment – identification of patients
and positron emission tomography (PET) is at risk for major adverse cardiac events
17
(myocardial infarction, sudden cardiac An effective risk assessment and manage-
death). Pooled data from numerous studies ment strategy is essential in patients with
suggest that patients with normal or near- known CAD treated by revascularisation
normal MPS are at a very low risk for a major [by either percutaneous coronary interven-
adverse cardiac event (defined at <1% per tion (PCI) or coronary artery bypass grafting
year). Furthermore, the extent (amount of (CABG)]. Improvements in revascularisation
myocardium affected) and severity (level techniques (such as drug-eluting stents and
of ischaemia) of reversible perfusion abnor- arterial grafts) have reduced the number of
malities can further stratify patients in terms post-procedural major cardiac events, but
of risk for a cardiac event. However, these a significant number of patients remain at
findings cannot be generalised to all pa- high risk. After PCI, repeated occurrence of
tients referred for MPS. While patients with- symptoms may be related to restenosis at
out known prior CAD or known myocardial a treated site or to disease progression at a
infarction and normal MPS have been re- different site; typically, symptoms within 6
ported to be at low annual risk (0.6%) for an months of the procedure occur due to re-
adverse cardiac event [15], the predictive stenosis and at 9 months and beyond due
value of MPS may be lower in specific pa- to progression of disease elsewhere. No-
tient subsets (those with known CAD [16], tably, only about half of the patients with
the elderly [17], women [18] and patients restenosis exhibit symptoms. MPS is able to
with cardiac risk factors, particularly diabe- detect ischaemia due to restenosis, residual
tes [19] and chronic kidney disease [20]). stenoses and disease progression (repre-
Nevertheless, in most clinical contexts in senting high risk of a major adverse cardiac
which patients are referred to MPS, addi- event) and should thus be performed in
tional risk stratification can be provided by symptomatic patients. As ischaemia is fre-
the results of the examination. An addition- quently silent in patients after revascularisa-
al advantage of MPS is the ability to guide tion, MPS may have a role in asymptomatic
therapeutic strategy according to quan- patients after PCI, taking clinical variables
titative assessment of ischaemic burden into consideration [23]. MPS should not be
[21]. Observational data have shown that performed sooner than 3 (ideally 6) months
patients with a large amount of ischaemic after PCI to allow recovery of normal en-
myocardium (exceeding 10% of total myo- dothelial function and avoid false positive
cardium) benefit from revascularisation, ischaemic findings. In patients after CABG,
while in patients with a small amount of similar risk stratification is achieved by late
ischaemic myocardium appropriate medi- (≥5 years post procedure) MPS, again with
cal therapy alone may suffice for improve- consideration of clinical variables (par-
ment in survival [22]. ticularly diabetes, where early MPS may be
18
Chapter 2 Clinical Indications
EANM
considered) [24]. Repeated revascularisa- myocardium is present, revascularisation is
tion in these patients significantly reduces likely to lead to improvement in symptoms,
cardiovascular risk [25]. myocardial function and overall survival [28].
In patients considered for a major operative The role of MPS in viability assessment is
procedure, MPS aids in assessing risk of peri- well established. It relies on the physiological
operative cardiac mortality. It is indicated in principle that tracer uptake in myocardium
patients with moderate-to-high clinical risk is related to preservation of the integrity of
and reduced exercise tolerance. In the pres- cell membranes and metabolism with main-
ence of reduced left ventricular ejection frac- tenance of the internal cellular milieu (trans-
tion (LVEF), ischaemic changes signify a high membrane pumps for ion analogues in the
perioperative risk [26]. case of 201Tl and transmembrane gradient in
mitochondria in the case of 99mTc-labelled
Viability assessment tracers) [29]. Patient preparation and imag-
Severe ischaemia leads to transient regional ing protocols are adapted to the selected
contractile dysfunction (“stunning”), which myocardial perfusion tracer. With 201Tl, a rest-
persists beyond the ischaemic episode. Re- redistribution protocol with delayed imaging
peated episodes of stunning and progres- after 4 h is usually employed. If 99mTc-labelled
sion of CAD lead to changes in regional myo- tracers are used, patient preparation with
cardial perfusion, metabolism and eventually nitrates is performed; induction of vasodila-
structure. Resulting myocardial adaptation tion and increased collateral flow results in
through regional reduction in contractil- an increase in tracer uptake and augments
ity (hypokinesis or akinesis) and metabolism detection of viable myocardium. Segmental
shift (from free fatty acids to glucose as the tracer uptake is quantified to identify areas
main metabolic substrate) is termed “hiber- in which the amount of viable myocardium
nation”. Importantly, hibernating myocardi- present is too small to result in a clinically
um can regain function upon restoration of observable benefit. Relative uptake of >50%
regional blood flow [27]. of maximum myocardial uptake at rest and
≥10% increase in uptake at redistribution
Myocardial functional impairment (degree of with 201Tl and relative uptake >50–60% of
reduction in LVEF) is an important predictor maximum myocardial uptake on rest imag-
of survival in patients with ischaemic cardio- ing with 99mTc-labelled tracers are considered
myopathy. If coronary anatomy is amenable to signify viable myocardium.
to a revascularisation procedure, assessment
of myocardial viability should be performed. ECG-gated SPECT is an important integral
If a significant amount of viable, hibernating part of viability assessment. By definition,
19
hibernating myocardium is regionally dys- in conjunction with myocardial perfusion as-
functional (hypokinetic or akinetic). Similarly sessment with SPECT or, preferentially, PET
to its role in the detection of ischaemic heart tracers.
disease, ECG-gated SPECT assists in differ-
entiation between regionally dysfunctional Conclusion
myocardium and attenuation artefacts. With Myocardial perfusion scintigraphy is a well-
the addition of low-dose dobutamine, the established tool for assessment of various
contractile reserve of dysfunctional seg- clinical manifestations of ischaemic heart dis-
ments can be assessed for complementary ease. Information provided by the technique
determination of viable myocardium [30]. In is highly dependent on appropriate indica-
doubtful cases and advanced heart failure tion, patient selection and implementation
due to ischaemic heart disease (ischaemic of the examination; ultimately, these factors
cardiomyopathy), the gold standard for vi- determine its clinical usefulness.
ability assessment remains fluorine-18 (18F)
fluorodeoxyglucose (FDG) PET, which iden- Acknowledgement. The author would like to
tifies viable myocardium by the presence of thank Barbara Guzic Salobir, MD, PhD, for use-
preserved glucose metabolism, performed ful comments during finalisation of the text.
20
References Chapter 2
EANM
References
1. Hesse B, Tägil K, Cuocolo A, Anagnostopoulos C, Bardiés 8. Kim C, Kwok YS, Heagerty P, Redberg R. Pharmacologic
M, Bax J, et al.; EANM/ESC Group. EANM/ESC procedural stress testing for coronary disease diagnosis: A meta-
guidelines for myocardial perfusion imaging in nuclear analysis. Am Heart J. 2001;42:934–44.
cardiology. Eur J Nucl Med Mol Imaging. 2005;32:855–
97. 9. Heijenbrok-Kal MH, Fleischmann KE, Hunink MG. Stress
echocardiography, stress single-photon-emission com-
2. Task Force Members, Montalescot G, Sechtem U, Achen- puted tomography and electron beam computed to-
bach S, Andreotti F, Arden C, Budaj A, et al. 2013 ESC mography for the assessment of coronary artery disease:
guidelines on the management of stable coronary a meta-analysis of diagnostic performance. Am Heart J.
artery disease: the Task Force on the management of 2007;154:415–23.
stable coronary artery disease of the European Society
of Cardiology. Eur Heart J. 2013;34:2949–3003. 10. O’Keefe JH Jr, Barnhart CS, Bateman TM. Comparison of
stress echocardiography and stress myocardial perfu-
3. Cerqueira MD, Weissman NJ, Dilsizian V, Jacobs AK, Kaul sion scintigraphy for diagnosing coronary artery disease
S, Laskey WK, et al.; American Heart Association Writing and assessing its severity. Am J Cardiol. 1995;75:25D–
Group on Myocardial Segmentation and Registration for 34D.
Cardiac Imaging. Standardized myocardial segmenta-
tion and nomenclature for tomographic imaging of the 11. Fleischmann KE, Hunink MG, Kuntz KM, Douglas PS.
heart. A statement for healthcare professionals from the Exercise echocardiography or exercise SPECT imaging?
Cardiac Imaging Committee of the Council on Clinical A meta-analysis of diagnostic test performance. JAMA.
Cardiology of the American Heart Association. Circula- 1998;280:913–20.
tion. 2002;105:539–42. 12. Smanio PE, Watson DD, Segalla DL, Vinson EL, Smith WH,
4. Berman DS, Kang X, Gransar H, Gerlach J, Friedman JD, Beller GA. Value of gating of technetium-99m sestamibi
Hayes SW, et al. Quantitative assessment of myocardial single-photon emission computed tomographic imag-
perfusion abnormality on SPECT myocardial perfusion ing. J Am Coll Cardiol. 1997;30:1687–92.
imaging is more reproducible than expert visual analy- 13. McLaughlin MG, Danias PG. Transient ischemic dila-
sis. J Nucl Cardiol. 2009;16:45–53. tion: a powerful diagnostic and prognostic finding of
5. Health Quality Ontario. Single photon emission com- stress myocardial perfusion imaging. J Nucl Cardiol.
puted tomography for the diagnosis of coronary artery 2002;9:663–7.
disease: an evidence-based analysis. Ont Health Technol 14. Hendel RC. Attenuation correction: eternal dilemma
Assess Ser. 2010;10:1–64. or real improvement? Q J Nucl Med Mol Imaging.
6. Mowatt G, Vale L, Brazzelli M, Hernandez R, Murray A, 2005;49:30–42.
Scott N, et al. Systematic review of the effectiveness 15. Klocke FJ, Baird MG, Lorell BH, Bateman TM, Messer
and cost-effectiveness, and economic evaluation, of JV, Berman DS, et al.; American College of Cardiology;
myocardial perfusion scintigraphy for the diagnosis American Heart Association Task Force on Practice
and management of angina and myocardial infarction. Guidelines; American Society for Nuclear Cardiology.
Health Technol Assess. 2004;8(30):iii-iv, 1–207. ACC/AHA/ASNC guidelines for the clinical use of cardiac
7. Imran MB, Pálinkás A, Picano E. Head-to-head com- radionuclide imaging--executive summary: a report of
parison of dipyridamole echocardiography and stress the American College of Cardiology/American Heart
perfusion scintigraphy for the detection of coronary Association Task Force on Practice Guidelines (ACC/
artery disease: a meta-analysis. Comparison between AHA/ASNC Committee to Revise the 1995 Guidelines
stress echo and scintigraphy. Int J Cardiovasc Imaging. for the Clinical Use of Cardiac Radionuclide Imaging).
2003;19:23–8. Circulation. 2003;108:1404–18.
21
16. Parisi AF, Hartigan PM, Folland ED. Evaluation of exercise 24. Acampa W, Petretta M, Evangelista L, Nappi G, Luongo L,
thallium scintigraphy versus exercise electrocardiogra- Petretta MP, et al. Stress cardiac single-photon emission
phy in predicting survival outcomes and morbid cardiac computed tomographic imaging late after coronary
events in patients with single- and double-vessel dis- artery bypass surgery for risk stratification and estima-
ease. Findings from the Angioplasty Compared to Medi- tion of time to cardiac events. J Thorac Cardiovasc Surg.
cine (ACME) Study. J Am Coll Cardiol. 1997;30:1256–63. 2008;136:46–51.
17. Hachamovitch R, Kang X, Amanullah AM, Abidov 25. Rizzello V, Poldermans D, Schinkel AF, Biagini E, Boersma
A, Hayes SW, Friedman JD, et al. Prognostic implica- E, Elhendy A, et al. Outcome after redo coronary artery
tions of myocardial perfusion single-photon emission bypass grafting in patients with ischaemic cardiomy-
computed tomography in the elderly. Circulation. opathy and viable myocardium. Heart. 2007;93:221–5.
2009;120:2197–206.
26. Weinstein H, Steingart R. Myocardial perfusion im-
18. Cerci MS, Cerci JJ, Cerci RJ, Pereira Neto CC, Trindade aging for preoperative risk stratification. J Nucl Med.
E, Delbeke D, et al. Myocardial perfusion imaging is a 2011;52:750–60.
strong predictor of death in women. JACC Cardiovasc
Imaging. 2011;4:880–8. 27. Rahimtoola SH. The hibernating myocardium. Am Heart
J. 1989;117:211–21.
19. Bourque JM, Patel CA, Ali MM, Perez M, Watson DD,
Beller GA. Prevalence and predictors of ischemia and 28. Schinkel AF, Bax JJ, Poldermans D, Elhendy A, Ferrari R,
outcomes in outpatients with diabetes mellitus referred Rahimtoola SH. Hibernating myocardium: diagnosis and
for single-photon emission computed tomography patient outcomes. Curr Probl Cardiol. 2007;32:375–410.
myocardial perfusion imaging. Circ Cardiovasc Imag- 29. Schinkel AF, Poldermans D, Elhendy A, Bax JJ. Assess-
ing. 2013;6:466–77. ment of myocardial viability in patients with heart fail-
20. Al-Mallah MH, Hachamovitch R, Dorbala S, Di Carli MF. ure. J Nucl Med. 2007;48:1135–46.
Incremental prognostic value of myocardial perfusion 30. Leoncini M, Sciagrà R, Bellandi F, Maioli M, Sestini S, Mar-
imaging in patients referred to stress single-photon cucci G, et al. Low-dose dobutamine nitrate-enhanced
emission computed tomography with renal dysfunc- technetium 99m sestamibi gated SPECT versus low-
tion. Circ Cardiovasc Imaging. 2009;2:429–36. dose dobutamine echocardiography for detecting
21. Hachamovitch R, Di Carli MF. Methods and limitations reversible dysfunction in ischemic cardiomyopathy. J
of assessing new noninvasive tests: Part II: Outcomes- Nucl Cardiol. 2002;9:402–6.
based validation and reliability assessment of noninva-
sive testing. Circulation. 2008;117:2793–801.
22
Chapter 3
Patient Preparation and Stress Protocols
Giuseppe Medolago and Adriana Ghilardi
EANM
Introduction
Non-invasive nuclear cardiology techniques The rationale is that an increase in oxygen
for diagnosis and risk stratification of coro- demand (exercise) or in coronary blood flow
nary artery disease (CAD) employ both exer- (pharmacological stress) causes regional
cise and pharmacological stressors to induce hypoperfusion or dysfunction in ischaemic
flow heterogeneity or functional abnormali- myocardial segments supplied by stenotic
ties attributable to myocardial ischaemia. coronary vessels (Fig. 1).
BICYCLE
PHYSIOLOGICAL
TREADMILL
STRESS MODALITY
DOBUTAMINE
PHARMACOLOGICAL
DIPYRIDAMOLE
ADENOSINE
23
related to heart workload [calculated by mul- registered at rest, at the end of each stage
tiplying the heart rate (HR) by the systolic and also during recovery. ECG monitoring is
blood pressure (BP) at the peak of exercise]. mandatory throughout the test.
In patients without CAD, exercise test causes
vasodilation and increases coronary blood Usually drugs that may interfere with the ex-
flow to 2–2.5 times above baseline levels. ercise test should be interrupted some days
beforehand, though they may be maintained
Bicycle ergometry to evaluate efficacy of therapy.
Bicycle protocols involve incremental work-
loads calibrated in watts (W) or kilopond (kp) 250 W
metres/minute [1 W = 6 kp]. Electronically 200 W x
braked bicycles are preferred because they x
150 W
provide a constant workload regardless of x
changes in the pedalling rate (60-80 rpm). 100 W
x
During the test the resistance to the pedal- 50 W
24
Chapter 3 Patient Preparation and Stress Protocols
EANM
Safety and risks. Bicycle ergometry is an ex- ceutical is injected at the peak of exercise.
tremely safe procedure. The risk is deter- The patient is required to keep on walking at
mined by the clinical characteristics of the a minor ramp angle for some more minutes
patient. In non-selected patient populations (recovery) to obtain almost resting values of
the mortality is negligible and the morbidity HR and BP.
less than 0.05%; the risk of complications is
greater in patients with known CAD (infarc- The patient is prepared with a standard 10-
tion, multivessel disease) [4]. Indications, to 12-lead ECG setup. HR, BP and ECG are
contraindications and limitations to exercise registered at rest, at the end of each stage
testing are shown in Tables 1 and 2. Exercise and also during recovery. ECG monitoring is
testing should always be undertaken un- mandatory throughout the test.
der the supervision of a physician properly
trained to perform such a test and able to Usually drugs that may interfere with the ex-
deal with any emergency. ercise test should be interrupted some days
beforehand, though they may be maintained
Treadmill to evaluate the efficacy of therapy.
The treadmill protocol should be consistent
with the patient’s physical capacity. Several Standard Bruce Protocol
standardised treadmill exercise protocols
STAGE DURATION TOTAL SPEED GRADE
exist (motor driven with variable speed and (min) TIME (mile/h) (%)
gradient steepness). The Bruce multistage
1 3 1.7 10
maximal treadmill protocol is the most wide-
2 3 6 2.5 12
ly used [4].
3 3 9 3.4 14
25
Disadvantages of both exercise tests. BP mea- In patients without CAD, dipyridamole in-
surements are often difficult to obtain and fusion causes vasodilation and increases
ECG tracings may be subject to more motion coronary blood flow 3–5 times above base-
artefacts. line levels. By contrast, in patients with sig-
nificant CAD, the resistance vessels distal to
Pharmacological testing the stenosis are already dilated, potentially
Pharmacological stress is increasingly em- even maximally, to maintain normal resting
ployed as an alternative to exercise testing. flow. Infusion of dipyridamole in the adjacent
Most patients referred to the nuclear cardi- myocardium, supplied by normal vessels,
ology laboratory are unable to perform a di- causes a substantial increase in blood flow,
agnostic exercise test owing to orthopaedic, producing heterogeneity in the myocardial
neurological, systemic or vascular disease. In blood flow (coronary steal).
these patients, the presence of CAD can be
evaluated on the basis of pharmacological Dipyridamole is infused over a 4-min period
vasodilation or chronotropic effect. at a dose of 0.56 mg/kg diluted in normal
saline solution. Maximal dilatory effect is
Three main types of pharmacological stress- achieved at 3–4 min after completion of the
or are distinguished: infusion: a slight increase in heart rate and
decrease in systolic blood pressure may oc-
1. Dipyridamole
cur. Radiopharmaceutical is injected at the
2. Adenosine/regadenoson 7th–8th minute of the infusion (Fig. 4). In
3. Dobutamine some laboratories, dipyridamole infusion
is combined with handgrip exercise to re-
Dipyridamole infusion protocol duce background activity of the radiotracer
More clinical experience has been gained in the abdomen. After i.v. administration of
with dipyridamole (a synthetic indirect vaso- the radiopharmaceutical, if necessary, the
dilator) than with the other pharmacological dipyridamole antidote aminophylline can
tests [5]. When intravenously infused, dipyri- be administered intravenously in order to re-
damole blocks cellular uptake (in vascular verse quickly any undesirable dipyridamole-
endothelium and red blood cell membranes) associated side effects and/or stress blood
of the natural vasodilator adenosine (syn- flow discrepancy.
thesised and released by endothelial cells),
which regulates coronary blood flow to meet
myocardial metabolic demands (by reacting
with specific receptors which stimulate vas-
cular smooth muscle cell relaxation).
26
Chapter 3 Patient Preparation and Stress Protocols
EANM
clinical characteristics of the patient referred
DIPYRIDAMOLE PROTOCOL
for the procedure (Table 3). Dipyridamole in-
Total dose: 0.56 mg / kg fusion should be undertaken under the su-
Inject Tracer
pervision of a physician properly trained to
perform such a test and able to deal with any
emergency.
0 2 4 6 8
BP x x x x x x x x x Absolute contraindications. The absolute con-
HR x x x x x x x x x
Cont ECG x...............x................x................x................x............x..........x.............x...........x
traindications to dipyridamole testing are:
1. Bronchospasm
Figure 4: Dipyridamole infusion protocol
(from [3]) 2. Drug intolerance
The patient is prepared with a standard 10- Limitations. Like any other drug, in some pa-
to 12-lead ECG setup in the supine or semi- tients (non-responders) dipyridamole may
orthostatic (stress table) position. HR, BP and display only slight or moderate pharmaco-
ECG are registered at rest and every minute logical efficacy, thus reducing the accuracy
throughout the test and also during recov- of the stress testing.
ery. ECG monitoring is mandatory through-
out the test. Adenosine infusion protocol
Adenosine, unlike dipyridamole, is a natural
Usually cardiovascular drugs (calcium antag- vasodilator, synthesised (from ATP) in the vas-
onists, nitrates) should be interrupted some cular endothelium and metabolised through
days before the pharmacological test, as active cellular uptake and enzymatic deg-
should caffeine and theophylline (adenosine radation in myocardial cells and vascular
receptor blockers). smooth cells very quickly (the half-time of ex-
ogenously infused adenosine is about 10 s).
The dipyridamole protocol is particularly
well suited for patients with left bundle In the heart, endogenous and exogenous
branch block, as it has a false positive rate of adenosine has an important role in the nat-
only 2–5% in such patients, compared with ural regulation of the coronary flow (vasodi-
30–40% for exercise testing [6]. lation) and cardiac demand (lowering BP) by
stimulating A2 purine receptors directly. It
Safety and risks. Even if side effects with di- inhibits noradrenaline release from sympa-
pyridamole are often more severe and more thetic nerve endings, reduces AV node con-
difficult to control, dipyridamole infusion is a duction velocity and has negative inotropic
safe procedure. The risk is determined by the and chronotropic effects by stimulating
27
A1 purine receptors in the sinoatrial and the In some laboratories, adenosine infusion is
atrioventricular node. combined with low-level exercise to reduce
background activity of the tracer in the ab-
The regional and systemic vascular effects dominal viscera. At the end of adenosine
of adenosine occur early (within 20–30 s) infusion and after intravenous administra-
and quickly disappear after discontinuation tion of the radiopharmaceutical, the antidote
of the infusion (T1/2 in plasma is about 15 s). aminophylline may be administered intrave-
Maximal effect has been observed invasively nously in order to reverse quickly any unde-
after 60 s and continues as long as the drug sirable adenosine-associated side effects.
is infused. These metabolic characteristics ex-
plain the lower rate of most side effects with The patient is prepared with a standard 10-
adenosine in comparison with dipyridamole to 12-lead ECG setup in the supine or semi-
(Table 3). orthostatic (stress table) position. HR, BP and
ECG are registered at rest and every minute
Adenosine is infused over a 4- to 6-min pe- during the whole test and also during recov-
riod at a t the dose of 140 mg/kg per minute. ery. ECG monitoring is mandatory through-
A shorter-duration adenosine infusion, last- out the test.
ing 4 min, has been found to be equally ef-
fective for the detection of CAD compared to Discontinuation of therapy is similar to that
the 6-min infusion. Injection is at 3 min for a with dipyridamole.
6-min protocol and at 2 min for a 4-min pro-
tocol (Fig. 5). Adenosine testing is the protocol of choice
in patients with significant arrhythmias and
with a history of psychiatric illness. Further-
Adenosine Protocol
more, adenosine testing is safe for stress
140 microg / min (63.6 microg / P / min) during 4 min
(2 min 2 model)
testing early after acute myocardial reaction
[8, 9].
Inj Tracer
Safety and risks, absolute contraindications
and side effects. These are all similar to those
for dipyridamole testing, even if broncho-
0 2 4 5 min spasm is less frequent.
BP x x x x x x
HR x x x x x x
Cont ECG x..............x.............. x.................x.............x..............x
28
Chapter 3 Patient Preparation and Stress Protocols
EANM
Regadenoson protocol Advantages of regadenoson. Regadenoson
Recently regadenoson, a selective A2A ad- has the following advantages:
enosine receptor agonist, has become avail-
• Selective A2A adenosine receptor agonist
able. Regadenoson is less likely to cause the
side effects associated with the other known • Simplified dosing
receptors of adenosine. It is administered -- 10-s injection
as a single standard dose of 400 µg in 5 ml -- Single unit dosing
over a full 10 s – with no dose adjustment -- No weight determination or dose
for weight – into a peripheral vein using a preparation
22-gauge or larger needle. This should be fol- -- No need for an infusion pump
lowed by a 5-ml saline flush, also over 10 s. -- No need for an extension line or
Radiopharmaceutical is injected 10–20 s after intravenous fluids
the flush. In total it thus takes less than 1 min -- Reduces supplies and waste
to complete the full administration protocol • Improved tolerability over adenosine
(Fig. 6).
• Alternative to dipyridamole stress
for patients susceptible to a
Unlike adenosine or dipyridamole, regade-
bronchonconstrictive response
noson is not contraindicated in patients with
asthma or COPD. Dobutamine infusion protocol
The dobutamine stress protocol is a demand/
supply type protocol simulating a physical
Regadenoson Protocol
stress test. Dobutamine infusion produces
adenosine 140 mcg/kg/min iv infusion systolic wall motion abnormalities because
6 min it increases myocardial oxygen demand by
➡
Dipyridamole 0.56 mcg/kg/min iv infusion increasing heart rate, blood pressure and
4 min within 3-5 min myocardial contractility. This results in a mis-
➡
regadenoson 400 mcg/5ml iv injection match between oxygen supply and demand
1 min in the presence of a functionally significant
➡
➡
29
Indications: Dobutamine quickly clears from the blood
Doputamine
· Obstructive Lung Disease (T1/2 about 2 min). It should be emphasised
mg / kg / min · Xanthine Derivative Intake
· Caffeine Intake that it is relatively common (15–20% of pa-
tients) to observe a fall in blood pressure at
40 + ...... ATROPINE ......
higher doses of dobutamine, both in pa-
30
End Points:
tients with and in those without CAD, due
20 · Peak dose to a mechanoreceptor reflex initiated in the
· Peak HR
· Symptoms: left ventricle. This reaction does not carry the
10
Angina, Dyspnoea same significance as a fall in blood pressure
5 · Severe ST Depres. during exercise testing. If symptoms occur,
BASELINE · BP - SYST > 220
· BP - DIA > 120 simple leg elevation will help; occasionally, in
0 2 4 6 8 10 min · BP > 30
· Arrhythmia the presence of severe symptoms, a low dose
BP & ECG
of beta-blocker antidote is needed.
Figure 7: Dobutamine infusion protocol
All side effects (Table 4) or severe symptoms
Dobutamine is infused at incremental doses are usually easily controlled with low intrave-
of 5, 10, 20, 30 and 40 mg/kg per minute at nous doses of antidote beta-blockers.
3-min intervals, until symptoms or target HR
is reached. If target HR cannot be reached Ongoing treatment with beta-blockers is of-
by dobutamine infusion alone (most often ten a problem when using the dobutamine
owing to ongoing beta-blocker medica- protocol, because it may be very difficult (or
tion), adjunctive low intravenous doses of impossible) to reach the target HR even after
atropine (0.25–0.50 mg/push) are necessary addition of atropine. In this situation, a phar-
(Fig. 7). Radiotracer is injected when PMHR is macological vasodilation protocol (dipyri-
reached. damole or adenosine) should be used.
The patient is prepared with a standard 10- Contraindications to dobutamine testing. Ab-
to 12-lead ECG setup in the supine or semi- solute or relative contraindications are:
orthostatic (stress table) position. HR, BP and
ECG are registered at rest and every minute 1. Presence of severe arrhythmias
throughout the test and also during recov- 2. Presence of psychiatric disorders
ery. ECG monitoring is mandatory through-
out the test. Table 5 summarises the properties of all the
stressors (adenosine, dipyridamole and do-
butamine).
30
Chapter 3 Patient Preparation and Stress Protocols
EANM
Pharmacological stress for PET/CT RB
82
RB
82
20 mCi
N-ammonia
13
ting better diagnosis and risk stratification of 10–30 mCi 10–30 mCi
6 min 6 min Total time: ~30 min
patients with CAD. Use of the PET radiophar- CTAC† Rest PET CTAC† Stress PET
maceuticals rubidium-82 (82Rb), nitrogen-13 Pharm stress
31
References Chapter 3
32
References Chapter 3
EANM
Table 2: Limitations to exercise testing Table 2: Limitations to exercise testing
AV block 0 10
Non-cardiac
Headache 12.2 35
Dizziness 11.8 ?
Nausea 4.6 ?
Flushing 3.4 29
Pain (non-
2.6 ?
specific)
Dyspnoea 2.6 15
Paraesthesia 1.3 ?
Fatigue 1.2 ?
Dyspepsia 1.0 ?
33
Table 4: Reported side effects (% of patients) of intravenous dobutamine infusion (from [3])
Cardiac Non-cardiac
34
Chapter 4
Multidisciplinary Approach and Advanced Practice
Anil Vara
EANM
Introduction
A multidisciplinary team (MDT) with re- cardiologists, technologists, nurses, health-
sponsibility for formulation of patient care care assistants and cardiac surgeons, who
pathways is composed of members from dif- jointly provide an integrated care approach
ferent healthcare professions who have spe- appropriate for that patient. Within nuclear
cialised skills and expertise. A patient being medicine, too, such an approach is adopted,
examined for cardiology-related issues will with the MDT team consisting of the profes-
be seen by a group of general practitioners, sional groups shown in Fig. 1.
Technologist
Technical professional in
nuclear medicine.
Responsible for radioisotope
administration, processing,
scanning and advice Clinician
Physicist
Medical professional in
Scientific professional in nuclear medicine.
nuclear medicine. Responsible for justification
Responsible for radiation of scan, prescribing, medical
protection and expertise on advice, reporting and stress
physics leads
PATIENT
Radiopharmacy Nurses
Chemistry/pharmacy Nursing professional in
professional in nuclear nuclear medicine.
medicine. Responsible for patient
Responsible for preparation of care, drug administration,
radiopharmaceuticals and Administration clinical observations and
advice Administrative/ emergency support.
managerial professional
in nuclear medicine.
Responsible for back office
functions, operational
management,
administrative staff
and process
Figure 1: The professional groups involved in the MDT team and their roles
35
Nature of the multidisciplinary approach vanced practice” is the term used to describe
in myocardial perfusion imaging the situation where technologists develop
Myocardial perfusion imaging (MPI) is a com- a set of competencies that allows them to
plex examination that is often very daunting perform tasks previously within the bound-
for the patient, given the discomfort during aries of other professionals. Advanced prac-
exercise stress testing and the requirement tice is impacting on MPI as it is becoming
to keep still for long periods and to lie flat more common for technologists to adopt
for the scan. The MDT approach to MPI reas- enhanced roles after learning the appropri-
sures the patient that they are being cared ate set of nuclear cardiology competencies
for by professionals with the requisite com- (as set out, for example, for the “Nuclear
petencies and with clear responsibility for Medicine Advanced Associate” in the SNM’s
individual roles during the MPI exam. Fig- Nuclear Medicine Advanced Associate Curricu-
ure 1 outlines the key professions and briefly lum Guide, 1st Edition [1]; see Appendix).
describes how they contribute to the overall
MPI exam. The introduction of advanced practice is
not unique to technologists, as many other
The main advantages and disadvantages professionals such as radiographers and
of the MDT approach in MPI are outlined in nurses have also increased their contribu-
Table 1. Overall, the advantages do outweigh tion to patient care through the adoption of
the disadvantages, and it is important to re- advanced practice. In the United Kingdom,
member that the MDT approach is consid- the introduction of the NICE 73 appraisal in
ered more effective given the higher levels the management and diagnosis of angina
of quality and efficiency that it can provide. resulted in large numbers of patients being
In addition, there are many emerging chal- referred for MPI. The guidelines outlined an
lenges to the delivery of healthcare in Eu- expectation that service provision to 4000
rope, and without an MDT approach there is patients per million population would be
a high risk that MPI exams will be overtaken achieved. This level of provision, coupled
by other diagnostic approaches and become with the aim of performance of MPI within 6
obsolete. weeks following referral, would have placed
tremendous pressure on hospital services in
What is advanced practice? the absence of modification of practices. The
Recently, boundaries between the profes- challenge led nuclear medicine services to
sions represented in MDTs have started to adopt new ways of working through service
break down, and it is becoming more com- configuration and skill mix within existing
mon for professionals responsible for health- professional groups. Through the adoption
care delivery to have overlapping respon- of an enhanced technologist role in stress
sibilities (e.g. in the United Kingdom). “Ad- testing, the nuclear cardiology service was
36
Chapter 4 Multidisciplinary Approach and Advanced Practice
EANM
able to effectively increase the capacity for lished a competency framework that pro-
stress sessions and allow clinicians to focus vides a useful guide for training.
their resources on reporting, clinical devel-
opments/research, and development of the The fundamental competency framework
clinical service. set out by the BNMS [3] should ensure the
technologist can take responsibility for di-
The nursing profession had already estab- verse specific aspects of a stress test:
lished an advanced practice role through
the acquisition of skills in cardiac physiol- • Communication – with patients, the MPI
ogy, and nurses could easily gain further team, clinical referrers and other profes-
competencies in advising on and managing sionals
clinical issues arising from pharmacological
• Clinical assessment – taking a patient
and exercise-induced stress. However, such
history, examination, justification and au-
nursing support has often been limited to
thorisation, obtaining patient consent,
large hospitals where nurses rotate through
examination of the patient’s clinical notes
all of the imaging modalities. In contrast,
and drug charts
technologists remain solely responsible to
the nuclear cardiology department, which • Physical preparation – venous cannulation,
facilitates skill development and acquisition ECG and blood pressure interpretation,
of the competencies required for advanced stress equipment and pharmaceuticals
practice roles in the performance of stress
• Selection of the appropriate stress test –
studies. It remains the case, however, that
correct method, individual variations,
these advanced practice roles are adopted
plans for adverse events
only to complement the MDT involved in
MPI in order to enhance patient care, reduce • Conducting a stress test – safety, observa-
costs and provide more efficient service [2]. tions and interpretation, handling radionu-
clides, knowledge and safe administration
Advanced practice competencies for of stress drugs, use of equipment, patient
technologists in MPI needs and documentation
Technologists with an advanced practice
role within nuclear cardiology would nor- The appendix to this chapter provides a de-
mally have acquired competencies to an tailed list of the competencies listed in the
enhanced level in nuclear medicine [1]. The Nuclear Medicine Advanced Associate Cur-
core competencies required for such a role riculum Guide, 1st edition [1], which represent
have been published by the British Nuclear good guidelines for delivery of a suitable
Medicine Society (BNMS) and the Society of technologist training programme. In the
Nuclear Medicine, both of which have pub- United Kingdom, some academic centres
37
have developed courses that deliver the core In the United Kingdom, technologists are not
competencies for advanced practice within yet state registered and therefore advanced
nuclear cardiology for a wide group of pro- practice can often pose difficulties in respect
fessionals, including technologists, radiog- of assurance to the employing organisation
raphers and nurses. These courses comprise and the patient, compared with other pro-
both theoretical and work-based practical fessional groups such as radiographers and
components, with case studies, assessments nurses. This should not, however, deter the
and written assignments. technologist from developing a career in
advanced practice. In the United Kingdom,
Possibility of building a career ladder for certain professional groups will provide in-
the technologist practising MPI demnity as well as an infrastructure to en-
Technologists partaking in advanced prac- sure maintenance of continuing professional
tice within nuclear cardiology are profes- development. The technologist’s role in MPI
sionally accountable to their nuclear medi- might be further advanced to include au-
cine consultant, who must ensure that the thorisation of request forms, prescription of
technologist’s practice is routinely audited to drugs as part of MPI and possibly even re-
reassure patients that they are receiving an porting of MPI scans. Such a level of practice
appropriate quality of care. It must be em- for technologists could be accomplished by
phasised that such technologists are oper- ensuring that state registration is available for
ating as part of an MDT. They should always nuclear medicine technologists and by em-
have adequate medical support available bedding the technologist’s role more deeply
and must work in accordance with the writ- within the MDT delivering the MPI service.
ten protocols and seek medical advice when
difficult issues arise.
38
Appendix Chapter 4
EANM
Appendix
16. Establish intravenous access
Detailed nuclear cardiology competencies
for a Nuclear Medicine Advanced Associate 17. Identify and administer the appropriate medica-
tions for commonly occurring cardiac arrhythmias
as set out in the SNM’s Nuclear Medicine Ad- under the direction of the supervising physician
vanced Associate Curriculum Guide 1st Edition. 18. Perform cardiac compression or defibrillate patient
if required
1. Successfully complete Advanced Cardiac Life Sup- 19. Facilitate the ordering of laboratory tests or other
port credentialing tests as needed for a cardiac arrest event under the
2. Assess normal electrocardiogram to determine pa- direction of the supervising physician
tient safety for stress testing 20. Facilitate admission of the patient to the hospital
3. Assess abnormal electrocardiographic conduction if necessary
in preparation for stress testing 21. Provide indicated intervention for non-cardiac
4. Develop procedural policies and standards for pre- emergency events
cardiac arrest emergencies that might occur within 22. Manage crash cart for compliance
the department as directed by institutional policy
and practice standards 23. Follow the appropriate guidelines in implement-
ing regulation for managing the department’s
5. Identify the signs and symptoms of symptomatic crash cart
bradycardia and symptomatic tachycardia
24. Inventory crash cart components according to in-
6. Follow a step-by-step course of action for the pa- stitutional policy
tient who develops asymptomatic bradycardia or
tachycardia while in office (before, during or after 25. Properly dispose of expired drugs
the stress test) 26. Replace expired drugs
7. Follow a step-by-step course of action for the pa- 27. Perform quality assurance testing on defibrillator
tient who develops signs and symptoms of brady- and document results
cardia or tachycardia while in office (before, during
or after the stress test) 28. Take comprehensive patient history and evaluate
for patient pathology
8. Identify the proper medications and dosages for
stable cardiac rhythms 29. Interview the patient and document on depart-
ment form a complete past and current cardiac
9. List contraindications and precautions of common history
cardiac medications
30. Review non-cardiac history for prevalence to study
10. Follow a step-by-step approach to handling an ST requested
elevated myocardial infarction
31. Perform physical assessment
11. Follow a step-by-step approach to handling a
stroke situation 32. Evaluate patient laboratory biochemical markers
relevant to cardiac pathology
12. Follow a step-by-step approach to handling other
patient incidents 33. Review most recent laboratory test results relevant
to cardiovascular diseases
13. Identify and delegate personnel to perform various
tasks in preparation for cardiac emergencies 34. Order relevant blood tests if necessary (including
pregnancy testing)
14. Incorporate the appropriate guidelines into de-
partmental policies and procedures 35. Evaluate patient medications for contraindications
to stress testing
15. Develop procedural policies and standards for
cardiac arrest emergencies that occur within the 36. Understand contraindications to each type of
department as directed by institutional policy and stress test and evaluate for each
practice standards and provide indicated interven-
tion for a cardiac emergency event
39
37. Review patient medications for contraindications 54. Administer pharmacological agents
to exercise stress testing
55. Monitor patient response to pharmacological
38. Obtain patient informed consent as required for agents and treat the patient appropriately in the
nuclear cardiology procedures according to state event of an adverse effect
law and hospital policy
56. Analyse results of the stress test and imaging por-
39. Understand the ethical and legal underpinnings of tion of the examination and prepare a preliminary
informed consent description of findings for the supervising physi-
cian
40. Determine the capability of the patient to give in-
formed consent 57. Create a preliminary description of findings detail-
ing the results of the stress portion of the test
41. Explain the procedure to the patient, including all
components of a valid informed consent 58. Examine rotating raw data from both stress and
resting image acquisitions and evaluate image
42. Obtain the patient’s or guardian’s signature quality
43. Conduct treadmill testing per all protocol options 59. Review data for incidental finding outside of the
under the direction of the supervising physician heart
44. Prepare the patient for exercise protocol 60. Compare and contrast stress versus resting pro-
45. Determine type of exercise stress test cessed images for perfusion defects
46. Monitor electrocardiographic tracings and blood 61. Determine whether the heart-to-lung ratio and
pressure for specific pathology and cardiac events transient ischaemic dilation are abnormal
during stress testing 62. Evaluate the wall motion of stress and resting im-
47. Use the appropriate termination protocols ages for ejection fraction and kinetic abnormalities
48. Prescribe and administer interventional drugs for 63. Review and evaluate bull’s eye polar maps and
pharmacologic stress under the direction of the summed stress scores
supervising physician 64. Create a preliminary description of findings detail-
49. Explain the indications and contraindications for ing the results of the imaging portion of the test
each pharmacological stress agent 65. Facilitate or recommend patient-specific cardiac-
50. Identify the physiological action of each pharma- related procedures based on nuclear cardiology
cological agent as it relates to stress testing examination results (outcome management) ac-
cording to the supervising physician
51. Calculate total dose, volume and dose rate for
each of the most common pharmacological stress 66. Order or facilitate scheduling of complementary
agents diagnostic procedures as indicated
52. Set up drug administration pump
53. Prepare pharmacological agents for administration
utilising sterile technique
40
References Chapter 4
EANM
References
1. Advanced Practice Task Force of the SNMTS. Nuclear 3. British Nuclear Medicine Society. Clinical competence in
Medicine Advanced Associate curriculum guide, 1st myocardial perfusion scintigraphic stress testing. 2005.
ed. SNM, 2008.
Advantages Disadvantages
Working as a cohesive team improves the scope Teams can become fragmented owing to
for innovation and developments within the competition among the various professionals,
different professions, thereby enhancing the resulting in loss of cohesiveness and impairment
quality of MPI of ability to pursue an integrated approach
41
Chapter 5
Advances in Radiopharmaceuticals for Myocardial Perfusion Imaging
James R. Ballinger and Jacek Koziorowski
42
Chapter 5 Advances in Radiopharmaceuticals for Myocardial Perfusion Imaging
EANM
1970s as a potassium analogue. Its 3-day half- history. Deutsch et al. continued work with
life made it convenient to supply, but limited the phosphines and developed furifosmin,
the administered activity because of concerns which Mallinckrodt brought to market, but
about radiation dosimetry. The initial distri- its properties were definitely inferior to those
bution of 201Tl reflects regional MPI; however, of sestamibi. Radiochemists at the University
over time there is redistribution from storage of Cardiff developed another monovalent
depots in other tissues, resulting in the filling cation, tetrofosmin (Myoview), with Amer-
in of transient defects. This property allows sham International and it became a rival to
stress and rest images to be obtained at differ- sestamibi. Ironically, it turned out that these
ent times after a single injection. tracers entered the heart by passive diffusion
rather than active transport, though their re-
At the time, 201Tl was relatively expensive, so tention was due to the positive charge caus-
there was interest in finding a technetium- ing them to localise in the mitochondria.
99m (99mTc)-labelled tracer which would be
suitable for MPI. In the late 1970s a group The group of Nunn et al. at Squibb devel-
of inorganic chemists began to explore the oped the boronic acid adduct of technetium
chemistry of technetium and to develop oxime (BATO) compound teboroxime (Car-
compounds in which the 99mTc atom was diotec), a neutral complex which showed
hidden inside the molecule and functional good myocardial extraction but rapid wash-
groups on the exterior of the molecule deter- out, meaning that SPECT imaging had to
mined its biological properties. This was un- be performed soon after injection. It was li-
like the metal chelates which had been used censed in the United States at the same time
previously. For targeting the Na+/K+ ATPase, it as sestamibi but dropped out of the market
was believed that a monovalent cation was within a few years. Ironically, its properties are
required. The first promising compound was well suited to the solid state cardiac cameras
bis(1,2-dimethylphosphino)ethane (DMPE), now in use. Pasqualini et al. at CIS Bio Interna-
developed by Deutsch et al. in Cincinatti. Im- tional developed 99mTc(N)-N-ethoxy-N-ethyl
ages of a dog heart were published in Science dithiocarbamate (NOEt), a neutral complex
but the compound failed in humans due to which showed good extraction and reten-
binding to a protein in blood which did not tion by the myocardium. Although it was
allow activity to clear from the circulation. very promising, it was not brought to market
Davison and Jones in Boston developed because sestamibi and tetrofosmin were too
the isonitrile complexes, of which sesta- well established by that time.
mibi (Cardiolite) was brought to market by
DuPont Pharma, becoming the most com- PET imaging of myocardial perfusion can
mercially successful radiopharmaceutical in be performed using 13N-ammonia and
43
15
O-water, as described below, though both tion of 201Tl reflects perfusion, due to its high
have limitations including the requirement extraction, and thus ischaemic and infarcted
for a nearby cyclotron due to their short regions are evident as defects (low activity).
half-lives. As an analogue of 201Tl, generator- However, 201Tl is not retained in its initial dis-
produced rubidium-82 (82Rb) was marketed tribution and as the patient rests it diffuses
by Squibb (later Bracco) and became widely into regions which had been ischaemic due
used in the United States. Its use is gradu- to increased demand. Repeat imaging 4 h
ally increasing in Europe. More recently, Lan- after injection provides an image in which
theus Medical Imaging has developed an only infarcted areas will be seen as defects.
18
F-labelled mitochondrial targeting agent, However, the imaging properties of 201Tl are
flurpiridaz, which may come to market in the not ideal, with its gamma photons in low
near future. abundance and the more abundant X-rays
being subject to attenuation and scatter. Fur-
Mechanisms of accumulation thermore, its long half-life and poor dosim-
The tracers used for MPI accumulate in the etry limit the amount of activity which can
myocardium by different mechanisms. They be administered and hence the image qual-
will be categorised with respect to their ity. It has largely been replaced by the 99mTc
mechanisms of localisation rather than their agents, though it is useful during times of
types of emission (i.e. SPECT vs PET). The three molybdenum-99 shortage.
main mechanisms are active transport, pas-
sive diffusion and mitochondrial targeting. 82
Rb is a generator-produced positron emit-
ter which is taken up by the same mecha-
Active transport nism as 201Tl [6,7]. However, its half-life is too
The Na+/K+ ATPase pump regulates the bal- short for redistribution images to be per-
ance of these two cations in the heart. formed so separate injections are required
for stress and rest images, but the entire pro-
Both 82Rb and 201Tl are biological analogues cedure, including attenuation correction by
of potassium and are thus extracted by the CT, takes less than 30 min. The generator is
same mechanism as potassium from the relatively expensive and must be replaced
blood pool and transported into the myo- monthly, so a high throughput of patients is
cyte (Fig. 1). 201Tl has a slower clearance time required to make the test cost effective [7].
from the myocardium than potassium; this However, the quantitative nature of the re-
facilitates the imaging of the myocardium sults obtained more than justify the cost of
by gamma cameras. In a standard MPI pro- the test by producing a more accurate diag-
cedure, 201Tl is injected when the patient is nosis with savings on alternative treatments
undergoing peak stress, either exercise or [7]. 82Rb is eluted from the generator with sa-
pharmacological [4,5]. The initial distribu- line and infused directly into the patient.
44
Chapter 5 Advances in Radiopharmaceuticals for Myocardial Perfusion Imaging
EANM
Rb rubidium chloride
82
45
Passive diffusion glutamine [7,10] (Fig. 3). After clearance of
Oxygen-15 labelled water (15O-water) is activity from the lungs and blood pool, imag-
freely diffusible and efficiently extracted by ing commences 3–5 min post injection and
the heart, but rapid washout occurs, which continues up to 20 min. Following a delay of
necessitates dynamic imaging and sub- 30–60 min for decay of the radionuclide, a
traction of blood pool activity [8] (Fig. 2). stress protocol is performed.
Each set of rest or stress images requires 5
min of dynamic acquisition, with a period
of about 10 min between injections. 15O-
water is generally administered as a bolus,
although 15O-carbon dioxide can be inhaled
and converted to 15O-water by carbonic an-
hydrase in vivo. Following bolus injection,
the first-pass information can be used to
delineate the blood pool in order to pro-
duce subtraction images of the myocardi-
um. Alternatively, 15O-carbon monoxide can
be given by inhalation to generate a blood
pool image. Neither approach is ideal. The
first-pass information from a slow bolus will Figure 2: Mechanism of myocardial accumula-
overestimate blood pool activity because of tion of 15O-water
early extraction into the myocardium, while
patient motion between the perfusion and
blood pool images can result in artefacts
[9]. Overall, image quality is poorer with
15
O-water than with the other PET agents
discussed.
46
Chapter 5 Advances in Radiopharmaceuticals for Myocardial Perfusion Imaging
EANM
15
O-water Advantages: Convenient for small number of
Production: Cyclotron patient studies; absolute quantification can
be performed
Physical half-life: 2 min
Disadvantages: Requires nearby cyclotron;
Maximum positron energy: 1.7 MeV
rest-stress protocol requires 1 h between
First-pass extraction: 100% tests to allow for decay; intense liver activity
Half-life of retention in myocardium: Washes can mask inferior wall of heart
out freely
Routes of excretion: Irrelevant with 2-min Mitochondrial targeting
half-life The two most commercially successful and
widely used 99mTc-labelled agents for MPI
Typical administered activity: 2200 MBq
both target the mitochondria, though this
Effective dose: 2.6 mSv was not intentional (Fig. 4). They were de-
Advantages: Does not require high patient veloped as lipophilic cations, intended to
throughput; absolute quantification can be be substrates for Na+/K+ ATPase as a trans-
performed; high extraction, not affected by port pump. However, in vitro studies early in
flow; due to short half-life, studies can be their development demonstrated that their
repeated after 15 min with physiological uptake was not inhibited by pretreatment
intervention with ouabain, which blocks Na+/K+ ATPase. It
Disadvantages: Requires on-site cyclotron; was later determined that both compounds
blood pool activity must be subtracted; poor enter cells by passive diffusion due to their
counting statistics due to short half-life lipophilicity. Once in the cell they home to
the mitochondria, where they are trapped
N-ammonia
13 by their positive charge. Both sestamibi and
Production: Cyclotron tetrofosmin consist of a 99mTc core hidden
within a lipophilic structure with methoxy
Physical half-life: 10 min
or ethoxy groups on the exterior to fine tune
Maximum positron energy: 1.2 MeV the biodistribution of the complex. Follow-
First-pass extraction: 80% ing intravenous injection at rest or during
Half-life of retention in myocardium: 4 h peak exercise or pharmacological stress,
both are extracted quickly into the heart;
Route of excretion: Hepatobiliary
however, imaging cannot begin until activity
Typical administered activity: 740 MBq has cleared from the lungs and clearance has
Effective dose: 1.7 mSv begun from the liver [1,5,11,12]. This occurs
47
somewhat more rapidly with tetrofosmin, al-
lowing imaging after rest or pharmacological
stress injection within 30–45 min whereas a
45- to 60-min delay is required with sesta-
mibi. The time pattern is similar for exercise
stress, with a 10-min delay for tetrofosmin
and a 20-min delay for sestamibi before im-
aging begins [13]. Both undergo extensive
hepatobiliary excretion while tetrofosmin
also has a degree of urinary elimination. Both Figure 5: Mechanism of myocardial accumula-
are reported to be excreted unchanged. tion of 18F-flurpiridaz
48
Chapter 5 Advances in Radiopharmaceuticals for Myocardial Perfusion Imaging
EANM
Disadvantages: Preparation requires heating; Physical half-life: 109 min
hepatobiliary excretion delays imaging and Maximum positron energy: 1.6 MeV
can compromise cardiac resolution
First-pass extraction: 90%
99m
Tc-tetrofosmin Half-life of retention in myocardium: 6 h
Production: Prepared from kit using Effective dose: 6 mSv (rest + stress)
99m
Tc-pertechnetate eluted from generator Advantages: Available as unit dose;
and incubation at room temperature high extraction, no flow limitation
Physical half-life: 6 h Disadvantages: Expensive for high
Principal gamma emissions (abundance): throughput
140 keV (89%)
Production and quality assurance aspects
First-pass extraction: 55%
Licensed products
Half-life of retention in myocardium: 4.5 h 201
Tl is a fully licensed product which arrives
Routes of excretion: Hepatobiliary and renal ready to use. The end user only needs to
Typical administered activity: 800 MBq check the documentation and verify the ra-
dioactivity measurement. 18F-flurpiridaz will
Effective dose: 6 mSv
eventually be in the same class.
Advantages: High-quality images;
room temperature preparation is more 99m
Tc products prepared on site
convenient; faster background clearance Sestamibi and tetrofosmin are prepared
than sestamibi, allowing earlier imaging on site or at a central radiopharmacy from
Disadvantages: Slightly lower first-pass licensed kits and a licensed 99Mo/99mTc gen-
extraction than sestamibi erator under national regulations. Appropri-
ate facilities are required and radiochemical
F-flurpiridaz (aka flupiridaz)
18
purity testing may be performed.
Chemical name: 2-tert-butyl-4-chloro-
5-[(4-(2-[18F]fluoroanylethoxymethyl)phenyl)
methoxy]pyridazin-3-one, BMS747158-02,
RP1012, Lantheus Medical Imaging)
Production: Cyclotron
49
PET products prepared on site MPI with PET offers a number of real or po-
15
O-water and 13N-ammonia are prepared in tential advantages. There is higher spatial
a local cyclotron. Because of the short half- and temporal resolution of images, with few-
lives, quality testing is performed on a test er attenuation defects. Diagnostic accuracy
production immediately prior to the clinical is reported to be higher and thus better for
production. 15O-water is particularly prob- risk stratification[15]. In most cases the pro-
lematic as it can be produced adjacent to the cedure is shorter (30–90 min). PET/CT allows
patient and infused directly. the potential for complementary information
such as calcium scoring [14]. However, on the
Generator-produced PET products downside there are higher capital costs and,
82
Rb is obtained from a generator and in- at present, more limited availability [14,15].
fused directly into the patient. Problems have
occurred with breakthrough of the parent The properties of the discussed SPECT and
radionuclide 82Sr and radiocontaminant 85Sr. PET agents are compared in Tables 1 and 2.
As a result it is now imperative that radionu-
clidic purity testing is performed on a test Current research in MPI tracers
elution prior to clinical use each day. There is great interest in developing MPI trac-
ers for PET using the generator-produced
Summary and future directions radionuclides copper-62 (62Cu) (half-life 10
MPI with SPECT is an extensively validated min) and gallium-68 (68Ga) (half-life 68 min).
test which is useful for cost-effective risk 62
Cu-pyruvaldehyde-bis(N4-methylthiosemi-
stratification and patient management[14]. carbazone (62Cu-PTSM) produced good myo-
It is widely available in the outpatient set- cardial images in patients but accumulation
ting using relatively inexpensive technol- was non-linear with relation to blood flow
ogy. Protocols are standardised and there are due to binding to albumin in the circulation
excellent procedural and clinical utilization [16]. An optimal 68Ga-labelled agent remains
guidelines published by professional medi- to be identified. A series of 68Ga-labelled sali-
cal societies internationally [14]. However, cylaldimine ligands failed to show suitable
SPECT suffers the limitations that the radio- behaviour in a pig model [17]. However, the
tracers are not optimal and are subject to at- immense potential benefits of generator-
tenuation artefacts,the test is time inefficient produced PET radionuclides will ensure that
(stress-rest requires 4 h) and attenuation cor- this research continues.
rection is not robust and, as a result, may un-
derestimate the extent of CAD [7].
50
Chapter 5 Advances in Radiopharmaceuticals for Myocardial Perfusion Imaging
EANM
Alternatives to flurpiridaz are under develop- or exclusion of CAD [21]. CMR cannot be
ment, including BFPET (4-18F-fluorophenyl) performed in most patients with pacemak-
triphenylphosphonium ion, FluoroPharma), ers and implantable cardiac devices. Unlike
a lipophilic cationic mitochondrial targeting PET/CT, these methods cannot provide abso-
agent currently in phase 2 clinical trials. Other lute quantification (mL/g per minute), which
18
F-labelled cations are also under develop- is necessary to calculate the myocardial per-
ment [18]. fusion reserve (MPR), and have lower sensitiv-
ity for the detection of balanced ischaemia,
In contrast, there has been relatively little multi-vessel disease and other pathological
recent interest in novel SPECT agents as the conditions that cannot be easily visualised
market is felt to be mature, with sestamibi by relative perfusion [7]. MDCT/DSCT has so
and tetrofosmin being well established. Re- far proven to be clinically feasible only for the
cently an iodine-123 labelled analogue of ro- detection of CAD (stenosis), and has shown
tenone has shown promising characteristics limited success with MPI [21]. There is also
for mitochondrial targeting [19]. concern about the radiation dose associated
with the CT techniques.
Competing modalities
Competing modalities include cardiac mag- Prospects
netic resonance (CMR), echocardioradiog- MPI with SPECT is a well-established proce-
raphy (ECR) and multi-detector computed dure used widely around the world. MPI with
tomography (MDCT) or dual-source CT PET/CT offers some potential advantages but
(DSCT) [14,20]. CMR and ECR provide good at a cost. Competing modalities are being
diagnostic accuracy but variable intra- and developed, but it will be a number of years
inter-observer agreement for the presence before these replace MPI.
51
References Chapter 5
References
1. Notghi A, Low CS. Myocardial perfusion scintigraphy: 12. Kelly JD, Forster AM, Higley B, Archer CM, Booker FS,
past, present and future, Br J Radiol. 2011;84:S229-36. Canning LR, et al. Technetium-99m-tetrofosmin as a new
radiopharmaceutical for myocardial perfusion imaging.J
2. Reyes E, Wiener S, Underwood SR. Myocardial perfusion Nucl Med.1993;34:222-7.
scintigraphy in Europe 2007: a survey of the European
Council of Nuclear Cardiology. Eur J Nucl Med Mol Imag- 13. BaggishAL, Boucher CA. Radiopharmaceutical
ing.2012;39:160-4. agents for myocardial perfusion imaging. Circula-
tion.2008;118:1668-74.
3. Yu M, Nekolla SG, SchwaigerM, Robinson SP. The next
generation of cardiac positron emission tomography 14. Flotats A, Knuuti J, Gutberlet M, Marcassa C, Bengel FM,
imaging agents: discovery of flurpiridaz F-18 for detec- Kaufmann PA, et al., Cardiovascular Committee of the
tion of coronary disease.SeminNucl Med.2011;41:305-13. EANM, the ESCR and the ECNC. Hybrid cardiac imaging:
SPECT/CT and PET/CT. A joint position statement by the
4. Ritchie JL, Albro PC, Caldwell JH, TrobaughGB, Ham- European Association of Nuclear Medicine (EANM), the
ilton GW. Thallium-201 myocardial imaging: a com- European Society of Cardiac Radiology (ESCR) and the
parison of the redistribution and rest images.J Nucl European Council of Nuclear Cardiology (ECNC).Eur J
Med.1979;20:477-83. Nucl Med Mol Imaging.2011;38:201-12.
5. Hesse B, Tägil K, Cuocolo A, Anagnostopoulos C, 15. McArdle BA, Dowsley TF, deKemp RA, Wells GA, Bean-
Bardiés M, Bax J, et al. and EANM/ESC Group. EANM/ lands RS.Does rubidium-82 PET have superior accuracy
ESC procedural guidelines for myocardial perfusion to SPECT perfusion imaging for the diagnosis of ob-
imaging in nuclear cardiology.Eur J Nucl Med Mol Im- structive coronary disease? A systematic review and
aging.2005;32:855-97. meta-analysis.J Am CollCardiol.2012;60:1828-37.
6. Selwyn AP, Allan RM, L’Abbate A, Horlock P, Camici P, Clark 16. Herrero P, Hartman JJ, Green MA, Anderson CJ, Welch
J, et al. Relation between regional myocardial uptake MJ, Markham J, Bergmann SR. Regional myocardial
of rubidium-82 and perfusion: absolute reduction of perfusion assessed with generator-produced copper-
cation uptake in ischemia. Am J Cardiol.1982;50:112-21. 62-PTSM and PET.J Nucl Med.1996;37:1294-300.
7. Machac J. Cardiac positron emission tomography imag- 17. Tarkia M, Saraste A, Saanijoki T, Oikonen V, Vähäsilta T,
ing.SeminNucl Med.2005;35:17-36. Strandberg M, et al. Evaluation of 68Ga-labeled tracers
8. Bergmann SR, Herrero P, Markham J, WeinheimerCJ, for PET imaging of myocardial perfusion in pigs.Nucl
Walsh MN.Noninvasive quantitation of myocardial Med Biol.2012;39:715-23.
blood flow in human subjects with oxygen-15-labeled 18. Kim DY, Kim HJ, Yu KH, Min JJ. Synthesis of [18F]-la-
water and positron emission tomography.J Am Col- beled (2-(2-fluoroethoxy)ethyl)tris(4-methoxyphenyl)
lCardiol.1989;14:639-52. phosphoniumcation as a potential agent for positron
9. Bergmann SR. Clinical applications of myocardial perfu- emission tomography myocardial imaging.Nucl Med
sion assessments made with oxygen-15 water and posi- Biol.2012;39:1093-8.
tron emission tomography. Cardiology.1997;88:71-9. 19. Wei L, Bensimon C, Lockwood J, Yan X, Fernando P,
10. Schelbert HR, Phelps ME, Hoffman EJ, Huang SC, SelinCE, Wells RG, et al. Synthesis and characterization of 123I-
Kuhl DE. Regional myocardial perfusion assessed with CMICE-013: a potential SPECT myocardial perfusion
N-13 labeled ammonia and positron emission comput- imaging agent.Bioorg Med Chem.2013;21:2903-11.
erized axial tomography. Am J Cardiol.1979;43:209-18. 20. Greenwood JP, Maredia N, Younger JF, Brown JM, Nixon
11. Wackers FJ, Berman DS, Maddahi J, Watson DD, Beller GA, J, Everett CC, et al. Cardiovascular magnetic resonance
Strauss HW, et al. Technetium-99m hexakis 2-methoxyiso- and single-photon emission computed tomography
butyl isonitrile: human biodistribution, dosimetry, safety, for diagnosis of coronary heart disease (CE-MARC): a
and preliminary comparison to thallium-201 for myo- prospective trial.Lancet.2012;379:453-60.
cardial perfusion imaging.J Nucl Med.1989;30:301-11.
52
References Chapter 5
EANM
21. Pakkal M, Raj V, McCann GP. Non-invasive imaging in
coronary artery disease including anatomical and func-
tional evaluation of ischaemia and viability assessment.
Br J Radiol.2011;84:S280-95.
Tables
Availability Long half-life makes supply Kit preparation with boiling Kit preparation at room
convenient temperature
Image Poor emissions, low count Optimal for gamma Optimal for gamma
quality density cameras cameras
Imaging Stress imaging must begin Stress and rest on same or Stress and rest on same or
flexibility immediately different days different days
Imaging Stress and rest data from Separate injections at stress Separate injections at stress
protocol single injection and rest and rest
Image High positron Good image quality High positron Good image quality
quality energy, poor energy
counting statistics
Imaging Good for small Good for small Unlimited number Available as unit
flexibility numbers of patients numbers of patients of patients dose; no special
infrastructure re-
quired
53
Chapter 6
SPECT and SPECT/CT Protocols and New Imaging Equipment
Andrea Santos and Edgar Lemos Pereira
54
Chapter 6 SPECT and SPECT/CT Protocols and New Imaging Equipment
EANM
Image optimisation tion [5, 6]. The properties of detectors used in
In order to maximise the quality of the im- SPECT are summarised in Table 2.
ages acquired, with the lowest absorbed
dose possible, awareness of some important Administered activity and count statistics
factors is essential. The activity administered in each part of the
exam (rest/stress, one-day/two-day proto-
Gamma camera detectors col) affects the image quality, if a standard
Nowadays, inorganic crystals are used in nu- time per projection is established. If the time
clear medicine to detect gamma rays. Ideally, per projection can be adapted, depending
the detector should have high detection effi- on the left ventricle count rate, the adminis-
ciency, good energy resolution, good intrinsic tered activity may be less important than if a
spatial resolution and low dead time. The most rigid time per view protocol is established. In
commonly used crystal in SPECT systems is the latter case, the administered activity has
NaI(Tl). More recently, a CdZnTe (CZT) solid to be better adapted to the patient’s weight
state camera has been developed with better or body mass index (BMI). In larger patients,
energy resolution than NaI(Tl) cameras. This higher activity must be used and in patients
newer technology, due to its better physics with lower BMI, lower activity must be ad-
characteristics (better sensibility and efficien- ministered in order to attain similar total
cy), is considered the best choice for MPI as it counts on every SPECT study acquired in the
enables both faster SPECT acquisition (reduc- same department. In the one-day protocol,
ing motion and maximising patient through- the second injection should be given with an
put) and lower radiopharmaceutical adminis- activity that triples the count rate [7–10].
tered activity. Moreover, recently developed
gamma cameras use multiple-detector gan- According to EANM’s guidelines, the recom-
tries that can be used in dedicated systems, mended administered activities for both
like cardiac SPECT dedicated gamma cameras, 99m
Tc agents and 201Tl are as follows:
because the equipment’s architecture is opti-
mised to acquire only one type of study. One 99m
Tc-sestamibi or -tetrofosmin [1]:
example of the specific design on such equip- -- Two-day protocol: 600–900 MBq/study
ment relies on the detector and/or patient
-- One-day protocol: 400–500 MBq for the first
positioning system. The so-called D-SPECT
injection and, as mentioned above, three
systems consist of several collimated CZT
times greater for the second injection.
modules. A fast scout scan is first performed
to locate the heart and the detector modules 201
Tl:
can confine their sampling to the heart. The
main advantages of these systems are their in- -- Stress redistribution: 74–111 MBq
crease sensitivity and improved spatial resolu- -- Re-injection: 37 MBq
55
Dedicated cameras smaller and can acquire images with a small-
Dedicated cameras are already being used er pixel matrix size. NaI(Tl) crystals can be
for cardiac SPECT. Image acquisition is fo- used, but the solid state detectors are more
cused on the heart of the patient with highly commonly used in these dedicated cameras,
efficient acquisition. These detectors are such as CZT [7].
Figure 2 A – C: A Conventional gamma camera with dual-detector system (from Garcia et al. [8]).
B Dedicated NaI(Tl) or CZT multiple detector gamma camera (from Madsen [16]). C Multiple-
pinhole collimator (from Eagle Heart Imaging, LLC)
Several new types of gamma camera, with dif- commonly used in MPI. In addition, iterative
fering characteristics, are available from various reconstruction algorithms are available and
vendors. There are dedicated gamma cameras, can improve image quality, one example be-
with or without CT, smaller equipment, in some ing wide beam reconstruction. This subject is
cases with three detection heads, cameras with further explored in another chapter.
multiple detectors and even dedicated multi-
pinhole collimators (Fig. 2). Some equipment Attenuation correction
allows acquisition with the patient seated, Patient body attenuation is one of the factors
potentially reducing patient movement and, that can degrade image quality, especially in
therefore, artefacts. With these dedicated sys- patients with higher BMI. For example, soft
tems, the acquisition can be faster (2–4 min vs tissue can absorb or scatter radiation. In order
the conventional 15–20 min), enhancing the to minimise the effect of attenuation, it is fun-
patient’s tolerance and consequently allowing damental to perform attenuation correction.
better image quality [8, 9]. This compensation can be done by apply-
ing a mathematical algorithm that provides
Software information about known and developed
Improved software can also enhance the attenuation maps. Numerous attenuation
image quality. Filtered back-projection is correction algorithms have been proposed;
56
Chapter 6 SPECT and SPECT/CT Protocols and New Imaging Equipment
EANM
the most commonly used algorithms are the is allowed, without reduction of the study
Chang algorithm (from 1978) and the maxi- quality. One can define 8–32 images per car-
mum-likelihood expectation-maximisation diac cycle (usually 8–16), depending on the
(MLEM) reconstruction algorithm [10]. desired information. The larger the number
of images per cycle, the longer the exam will
There is also a more patient-specific method take. It is important that the heart rate re-
to correct or compensate attenuation: esti- mains stable and consistent during the entire
mation of patient-specific attenuation maps. procedure; an arrhythmic/dysrhythmic heart
This method is more accurate as it develops rate can significantly decrease the accuracy
a patient-personal attenuation map. This type of the study [1, 4, 5].
of attenuation correction can be applied by
co-registration of the maps of another modal- Cardiac SPECT/CT
ity (such as computed tomography), or ob- SPECT/CT combines SPECT MPI with a CT
tained by the estimation of attenuation maps scan in the same gantry. It allows acquisition
with the emission data. Use of the attenua- of a SPECT scan and/or a CT scan, registration
tion map generated by the CT image is one and fusion of the two images, and hybrid
of the most accurate methods of attenuation imaging, which adds specificity. Hybrid im-
correction but it must be very carefully used ages can also be achieved by registering and
owing to both the increased radiation burden fusing scans from different equipment. Ade-
and the need for perfect emission and trans- quacy of the registration process is crucial, to
mission alignment, systematically controlled. avoid reconstruction artefacts. Some authors
In the absence of the latter, motion artefacts recommend the acquisition of a different CT
between the images can cause a mismatch scan at stress and rest for the purpose of at-
between the attenuation correction map tenuation correction [11–14].
from CT and the emission image [10].
Attenuation correction can be achieved
Gating with CT attenuation maps from linear at-
Gated studies should be acquired whenever tenuation coefficients. Table 3 summarises
possible despite the slight increase in acqui- the general characteristics of CT for non-
sition time. This technique allows evaluation diagnostic and diagnostic purposes. When
of functional parameters such as ejection used for a non-diagnostic purpose, CT is
fraction, left ventricle volumes and regional performed only for attenuation correction
wall motion and improves the diagnostic as an alternative to the mathematical meth-
accuracy of perfusion imaging. For gating ods previously discussed (it offers a more
purposes, only the systolic signal is needed accurate method to correct attenuation,
(usually the R wave), as it is not a diagnostic but at the cost of an increase in radiation
ECG. Usually, a 20% window of acceptance burden). If CT is performed for a diagnostic
57
purpose, it maintains the ability to correct Calcium scoring CT and CTCA are comple-
attenuation but can also be used to assess mentary exams and the information provid-
calcium scores or coronary angiography. ed by each one should be considered in con-
junction with the results of MPI in order to
Diagnostic calcium scoring allows the detec- maximise the information available regard-
tion of calcified plaques within the coronary ing possible heart disease in each patient.
vessels and can thus help in the early detec-
tion of coronary artery disease. Since the Radiation dose
presence of calcified plaques does not mean Nowadays, the reduction of radiation doses
that the myocardium is already in ischaemia, related to myocardial scintigraphy is an im-
the information provided by this method is portant issue. New reconstruction methods
complementary to that acquired by MPI. and modern detectors and/or collimators
optimised for cardiac SPECT may provide a
CT coronary angiography (CTCA) pro- solution. Exclusion of the rest study when
vides detailed anatomical information on stress is normal can also appreciably reduce
coronary vessels, which, when combined the radiation dose. CT for the purpose of at-
with the functional information offered by tenuation correction (non-diagnostic) does
SPECT, increases diagnostic accuracy. CTCA not increase the radiation dose significantly.
is performed in several steps. Prior to CTCA, In any case, physicians and technologists
technologists should instruct the patient on must always pay due attention to adjusting
breath-hold technique. First, a scout scan is CT protocols to the patient’s BMI, respecting
performed for anatomical localisation of the the ALARA principle.
heart, to define the start and end location
of the CTCA scan. The scan should extend The estimated radiation dose due to CTCA
from 2.5–5 cm above the aortic root to be- ranges from 8 to 18 mSv. However, modern
low the bottom of the heart, in accordance cardiac CT protocols (step-and-shoot ECG
with the scout image. It is important that triggering, ECG-controlled current modula-
scan length is optimised for the purpose of tion that reduces the tube current by 80%
dose optimisation. The aim is to minimise the during systole and body mass-adapted tube
distance scanned from above to below the voltage that reduces the tube voltage to
heart, while avoiding excessive restriction of 100 kV in patients under 90 kg) allow reduc-
the scan volume. CT coronary angiography tion of the radiation dose by 60–80%. Some
is a contrast-enhanced scan: the scan starts studies have reported a radiation dose of
about 10 s after intravenous contrast admin- 5.4 mSv at stress-only hybrid 99mTc-tetrofos-
istration as a single CT slice, with acquisition min SPECT/CT [11].
of one image every 2 s until a decrease in
opacification of the aorta is observed [13, 14].
58
References Chapter 6
EANM
References
1. Hesse B, Tägil K, Cuocolo A, Anagnostopoulos C, Bar- 9. Sharir T, Slomka PJ, Berman DS. Solid-state SPECT tech-
dies M, Bax J, et al. EANM/ESC procedural guidelines for nology: fast and furious. J Nucl Cardiol. 2010;17:890–6.
myocardial perfusion imaging in nuclear cardiology. Eur
J NUcl Med Mol Imaging. 2005;32:855–97. 10. King M, Glick S, Pretorius P. Attenuation, scatter and
spatial resolution compensation in SPECT. In: Wernick
2. Strauss HW, Miller DD, Wittry MD, Cerqueira MD, Garcia M, Aarsvold J. Emssion tomograph: The fundamen-
EV, Iskandrian AS, et al. Procedure guideline for myo- tals of PET and SPECT. USA: Elsevier Academic Press;
cardial perfusion imaging 3.3. J Nucl Med Technol. 2004:473–85.
2008;36:155–61.
11. Flotats A, Knuuti J, Gutberlet M, Marcassa C, Bengel FM,
3. Lecoultre R, Jorge JP. Preparation and use of imaging Kaufmann PA, et al. Hybrid cardiac imaging: SPECT/CT
equipment. In: Myocardial perfusion imaging – a tech- and PET/CT. A joint position statement by the European
nologist’s guide. EANM. 2004:29–35. Association of Nuclear Medicine (EANM), the European
Society of Cardiac Radiology (ESCR) and the European
4. Holly TA, Abbott BG, Al-Mallah M, Calnon DA, Cohen MC, Council of Nuclear Cardiology (ECNC). Eur J Nucl Med
DiFilippo FP, et al. ASNC imaging guidelines for nuclear Mol Imaging 2011;38:201–12.
cardiology procedures. J Nucl Cardiol. 2010.
12. Delbeke D, Coleman RE, Guiberteau MJ, Brown ML, Royal
5. Pedroso de Lima J. Física em Medicina Nuclear: Temas HD, Siegel BA, et al. Procedure guidelines for SPECT/CT
e aplicações. Coimbra: Coimbra University Press; imaging 1.0. SNM. 2006.
2008:159–86.
13. Dorbala S, Di Carli MF, Delbeke D, Abbara S, DePuey
6. Arlt R, Ivanov V, Parnham K. Advantages and use of EG, Dilsizian V, et al. SNMMI/ASNC/SCCT guideline
CdZnTe detectors in safeguard measurements. Ac- for cardiac SPECT/CT and PET/CT 1.0. J Nucl Med.
cessed in January, 2014. Available from http://www. 2013;54:1485–507.
evproducts.com/pdf/Applications/Advantages%20
and%20use%20of%20CdZnTe.pdf 14. Mariani G, Flotats A, Israel O, Kim EE, Kuwert T. Clinical
applications of SPECT/CT: New hybrid nuclear medicine
7. Frans J, Wackers T. Cardiac single-photon emission imaging system. IAEA-TECDOC-1597. 2008.
computed tomography myocardial perfusion imag-
ing: finally up to speed. J Am Coll Cardiol. 55:1975–8. 15. Patton JA, Turkington TG. SNM SPECT/CT physical prin-
ciples and attenuation correction. J Nucl Med Technol.
8. Garcia EV, Faber TL, Esteves FP. Cardiac dedicated ultra- 2008;36:1–10.
fast SPECT cameras: new designs and clinical implica-
tions. J Nucl Med. 2011;52:210–7. 16. Madsen MT. Recent advances in SPECT imaging. J Nucl
Med. 2007;48:661–73.
59
Table 1: Gamma camera and imaging parameters for SPECT
Nr. projections
Pixel matrix
Collimator
Projection
Rotation
Protocol
window
Isotope
Energy
Energy
Angles
Zoom
time
180º 1st acquisition:
3–6º; 25 s/frame
1 day LEHR (360º if 64
45º RAO 2nd acquisition:
99m
Tc or 140 keV 15–20% 3-head or 20 s/frame
to
LEGP gamma 128 45º LPO
64×64 1.0
2 day 25 s/frame
camera)
or (higher if
70 keV 20–30%
128×128 necessary)
180º
LEGP
32
(static
or
201
Tl 1 day or 20–25 s/frame
167 keV 20%
images if
LEHR
64
necessary)
Table 2: Properties of detectors used in SPECT (adapted from Arlt et al. [6])
60
References Chapter 6
EANM
Table 3: Acquisition modalities: CT scanners for hybrid imaging [11–13, 15]
Non-diagnostic
Scan type Diagnostic
(low resolution)
Depending on equipment,
Tube currenta Depending on equipment and purpose, 20–500 mA
usually from 1 to 2.5 mA
a
Dose optimisation, tube current and voltage can be lowered for patients with lower body mass index. However,
manufacturer and guideline recommendations should be consulted to guarantee that the scan stills accomplishes its
purpose [13]
61
Chapter 7
PET/CT Protocols and Imaging Equipment
April Mann and Scott Holbrook
62
Chapter 7 PET/CT Protocols and Imaging Equipment
EANM
Rubidium-82 decays by positron emission via the Na+/K+-ATPase transporter similar to
and associated gamma emission. Once pro- 201
Tl. 82Rb has been recorded to have an ex-
duced, it has a half-life (T1/2 of 75 s and, since traction rate of 50–60% at normal flow rates.
it is generator produced, an on-site cyclotron However, its extraction is documented to be
is not required [1–4]. 82Rb decays into kryp- less than that of 13N-ammonia and decreases
ton-82 (82Kr) by emitting a neutron and a with increased blood flow to as low as 20–
positron. The annihilation photons released 30%. It has also been noted that hypoxia,
following the positron emission have a mean myocardial cell integrity, severe acidosis and
energy of 511 keV and are used for image myocardial ischaemia may also decrease 82Rb
acquisition. The very short half-life allows the extraction rates [1–4].
generator to be eluted every 10 min, mak-
ing 82Rb suitable for repeated and sequential Quality control for rubidium-82
perfusion studies [1–4]. The 82Rb generator is stored in an automatic
infusion system that produces accurate and
The short half-life requires rapid image ac- reproducible constant-activity elution pro-
quisition to occur shortly after tracer admin- files (Fig. 1) [2]. In order to utilise the 82Rb
istration reducing the total study duration to generator, the manufacturer has training and
approximately 30–45 min depending on the education requirements that must be com-
type of PET or PET/CT system and protocol pleted by both physicians and technologists
used for imaging [1, 3]. This time is significantly prior to initiating a cardiac PET programme.
reduced compared with the standard one-day In addition, there are mandatory quality con-
rest/stress technetium-99m (99mTc) labelled trol procedure requirements for radiochemi-
radiotracer study, which takes approximately cal purity and eluate volume that must be
3–4 h depending on the protocol and system performed on the generator and document-
used for image acquisition. Further, the short ed daily before patient use.
half-life, results in a limited radiation exposure
for patients (5.5 mSv for 2220 MBq compared Approximately 75 min is required in order to
with a standard one-day rest/stress 99mTc- complete the 82Rb generator quality control
labelled radiotracer study (10.6–12.0 mSv for using the following steps [2]:
370/1110 MBq) or thallium-201 (201Tl) stress-
rest study (18.8 mSv for 111 MBq) [2]. 1. Elute the generator with 50 mL normal
saline
Rubidium-82 has been well studied as a per-
2. Measure the eluate for the activity of
fusion tracer since the 1950s. As a blood flow 82
Rb (MBq)
tracer, it is rapidly extracted from the blood
and taken up by the myocardium. Since it is 3. Allow sample to sit for an hour and
a potassium analogue, 82Rb uptake occurs measure activity again (Bq)
63
4. Decay the ratio of 85Sr/82Sr from the cali- Use of a generator must STOP at an
bration lot information to current time (R) Expiration Limit:
5. 82Sr = (dose calibrator reading)/[1+(R)(F)] • 17 L for the generator’s cumulative eluate
volume, or
a.
Where F is a constant 0.478 to eliminate
higher counts coming from 85Sr • 42 days post generator calibration date, or
6. THEN, Bq of Sr/MBq of Rb= (N)
82 82
• An eluate 82Sr level of 370 Bq/kBq 82Rb, or
7. The 82Sr/82Rb ratio (N) is multiplied by R to • An eluate 85Sr level of 3700 Bq/kBq 82Rb
get amount of Bq 85Sr/ MBq 82Rb.
64
Chapter 7 PET/CT Protocols and Imaging Equipment
EANM
withholding cardiac medication may not be onstrates an example of an imaging protocol
necessary. performed with a PET/CT system and dipyri-
damole used as the stress agent.
When performing myocardial perfusion imag-
ing with 82Rb, it is recommended to perform
the rest images first in order to reduce any PET/CT Protocol: Hartford Hospital
residual stress effects (ischaemic stunning or RB-82 RB-82
coronary steal. It is also necessary to consider 1850 –2220 MBq 1850 – 2220 MBq
Dipyridamole
that 80% of all useful counts are acquired in 0.56 mg/kg
the first 3 min, 95% in the first 5 min and 97% scout CT-trans
gated
rest
gated
stress
in the first 6 min when imaging with 82Rb.
120 sec 120 sec
Therefore, it is not necessary to acquire im-
ages for longer than 6 min. After the infusion Approx 1 min Approx 7 min Approx 6 min Approx 7 min
65
Stress testing with cardiac PET Stress protocols for cardiac PET imaging are
The best test to evaluate haemodynamic similar and generic to all perfusion tracers.
changes during stress is an exercise test. It The differences for 82Rb and 13N-ammonia are
provides independent prognostic value and related to the half-life and clearance of the
allows for assessment of symptoms and ex- specific tracers [1]. Given the short half-life
ercise capacity. Furthermore, combining ex- (75 s) of 82Rb, physical exercise is not current-
ercise data with perfusion data results in the ly possible. Therefore, only pharmacological
best risk stratification of patients. Standard stress is performed with the 82Rb imaging
clinical indications for stress testing include protocol.
but are not limited to:
Standard stress protocols and patient prepa-
• Evaluation of patients with chest pain or ration should be in compliance with ACC/
other findings suggestive of CAD AHA guidelines for stress testing. Before per-
forming any stress test, it is necessary to en-
• Determination of prognosis and severity
sure patients do not have an absolute or rela-
of disease
tive contraindication to the procedure. When
• Evaluation of effects from medical and performing stress tests in conjunction with
surgical therapy perfusion imaging it is necessary to have at
least two qualified individuals present during
• Screening for latent CAD (only approx. 30%
the procedure: one person to monitor the
of patients with ischaemia have chest pain)
patient and the other individual to perform
• Evaluation of arrhythmias, functional the injection of the radiotracer.
capacity or congenital heart disease
Various types of protocol are used to perform
More specifically, common clinical indica- physical exercise tests, including Bruce, Mod-
tions for stress myocardial perfusion imaging ified Bruce, Naughton and Chung. For PET
include but are not limited to: perfusion imaging, the standard protocol fol-
lowed in most laboratories is the Bruce (Fig.
• Diagnosis of suspected CAD 3). During the procedure, it is necessary to
• Risk stratification of known CAD measure and document the patient’s heart
rate and blood pressure at 2 min into every
• Assessment of medical or surgical stage. The radiotracer should be injected
therapies used to treat known CAD when the patient has achieved at least 85%
• Pre-operative assessment in patients with of the maximum predicted heart rate and is
cardiac symptoms no longer able to continue. As discussed ear-
lier, due to the short half-life of 82Rb, treadmill
• Myocardial viability assessment stress testing in patients is not possible.
66
Chapter 7 PET/CT Protocols and Imaging Equipment
EANM
Speed Grade Time Cum. Adenosine is another pharmaceutical used
Stage for stress testing (Fig. 5). It is a potent coro-
(MPH) (%) (min) Time
nary vasodilator that acts by directly activat-
1 1.7 10 3 3
ing adenosine receptors on cell membrane
2 2.5 12 3 6 surfaces. During infusion it causes blood flow
rates 4–6 times resting blood flow. Similar to
3 3.4 14 3 9
dipyridamole, its use is well validated in the
4 4.2 16 3 12 literature.
5 5.0 18 3 15
6 5.5 20 3 18
7 6.0 22 3 21
* Inject trace 1 min. prior to termination
67
Cardiac PET perfusion imaging with
13
N-ammonia
Considered to be the gold standard for myo-
cardial perfusion imaging due to very high
image resolution and quantitative charac-
teristics, cardiac PET perfusion imaging with
13
N-ammonia has grown in utilisation despite
the relatively short half-life of 13N, which ne-
cessitates the availability of a nearby cyclo-
tron and PET drug manufacturing facilities
Figure 6: Regadenoson infusion protocol (Fig. 7).
Figure 7: A 204-kg male patient with chest tightness and shortness of breath upon exertion.
Despite a very high body mass index, 13N-ammonia cardiac perfusion images provide data of
excellent technical quality
68
Chapter 7 PET/CT Protocols and Imaging Equipment
EANM
Preparation for patients undergoing a 13N- positioning device and the xiphoid process
ammonia cardiac PET perfusion studies is as an anatomical landmark. With a dedicat-
similar to that for other myocardial perfusion ed PET system, a small point source of 13N-
imaging exams. Patients should not have ammonia may also be used to verify correct
anything to eat or drink, except water, for a patient positioning.
minimum of 6 h prior to the scheduled ex-
amination time. Foods containing caffeine The 13N-ammonia should be administered in
should be withheld for at least 12 h prior to less than 30 s, and a delay after injection of
the procedure. Medications such as theoph- 1.5–3 min should occur prior to initiation of
ylline and aminophylline should be withheld imaging. However, if myocardial blood flow
for at least 48 h and beta-blockers for 24 h. is measured, imaging should begin imme-
Patients should be dressed in comfortable diately after injection. A sufficient difference
non-attenuating clothing without metal zip- between stress and rest radiopharmaceuti-
pers or buttons if possible [1]. cal counts may be achieved by decay and/
or administered activities to ensure the data
Stress protocols are the same as previously from the second image set overcome data
mentioned (Fig. 8). The radioactive half-life from the first. Images may be gated to re-
of 13N does potentially allow for treadmill flect myocardial wall motion and should
exercise protocols unlike with 82Rb. 13N-am- be acquired in either static or dynamic list
monia doses are 370–740 MBq per injection. mode. Emission images should be acquired
Once injected, 13N-ammonia accumulates for 10–15 min while the patient is instructed
in myocardium, as with other radiotrac- to remain very still. Transmission imaging for
ers, in proportion to blood flow, extraction attenuation correction is conducted with CT
and retention. 13N-ammonia is transported or 68Ge rod sources and may be completed
intracellularly through active transport via before emission images are acquired. Trans-
the sodium-potassium pump and becomes mission scans obtained with 68Ge rod sources
trapped as 13N-glutamine. Patients should be should take into account the age and energy
positioned supine with arms raised and out of the rod sources. Transmission scan time
of the PET field of view in order to avoid ar- should be increased when the energy of the
tefacts. Electrocardiogram electrodes and a sources is low or the rod sources are older.
blood pressure cuff should be placed on the Data should be processed using filtered
patient’s arm, and care should be taken to back-projection or iterative reconstruction
ensure that intravenous lines will not inter- and examined for artefacts associated with
fere with the study acquisition. Also, in order transmission-emission misregistration or pa-
to ensure that the myocardium is centred tient motion [1].
within the field of view, a scout scan may be
performed utilising CT or the camera’s laser
69
Cardiac metabolic imaging with
N-13 ammonia fluorine-18 fluorodeoxyglucose
Unlike other tissues under normal physiologi-
inject
inject
inject
CT scan Ex CT scan
long chain fatty acid metabolism compared
with 30% from glucose metabolism. The
percentage contribution of substrate me-
0 5 10 15 20 25 60 65 70 75 80 85 90 min
tabolism within the myocardium can shift
depending on various circumstances such
Figure 8: Example of exercise (top) and dipyri-
as the presence of endogenous insulin due
damole 13NH3 PET protocol (bottom)
to exercise or carbohydrate digestion, the
Increased availability and utilisation of 13N-am- administration of insulin in diabetics, and the
monia has led to variations of the aforemen- presence of anaerobic metabolism, as with
tioned protocol in routine clinical practice. ischaemia. Understanding this metabolic
Previously, batches of 13N-ammonia at end- shift is of great importance when engaging in
of-synthesis were commonly produced with a myocardial metabolic imaging. When prop-
total activity of 5550 MBq. New in-target pro- erly utilised, fluorine-18 fluorodeoxyglucose
duction methods now routinely produce total (18F-FDG) cardiac imaging is a valuable tool
batch activities of greater than 37,000–55,500 in distinguishing viable myocardium that has
MBq. Furthermore, cyclotrons equipped with undergone an acute ischaemic event, may
more than one 13N target may be capable of be hypokinetic and is metabolising glucose
producing several batches per hour. Sufficient rather than fatty acids from myocardium that
batches sizes of 13N-ammonia allow for higher has infarcted, is no longer viable and is not
dosing of patients with extremely large body metabolising glucose [1]. In order to evalu-
mass index. For example, patients who weigh ate viability myocardial tissue, it is first neces-
less than 159 kg may receive 370–740 MBq of sary to normalise the metabolic environment
13
N-ammonia compared with an injection of for all of the myocardium to ensure glucose
1110–1665 MBq of 13N-ammonia for patients metabolism will occur within viable tissue.
who weigh greater than 159 kg. Also of note, This may be accomplished through oral ad-
many PET and PET/CT scanners have recom- ministration of glucose after a fasting period
mended maximum patient table weights of resulting in an endogenous release of insulin
204.5 kg, and patients with a very large body or, in the case of diabetics, through the ad-
mass index often benefit from cardiac PET ministration of insulin.
perfusion imaging over SPECT.
70
Chapter 7 PET/CT Protocols and Imaging Equipment
EANM
Fluorine-18 fluorodeoxyglucose, an ana- measurement is greater than 250 mg/dl, no
logue for dietary glucose, is transported further oral glucose loading is necessary.
across cell membranes via glucose transport- However, it may be necessary to administer
er proteins such as GLUT-1. Once in the intra- insulin to achieve the desired effect [5]. Once
cellular space, glucose is acted upon by the the desired blood serum glucose/insulin re-
enzyme hexokinase and converted to two sponse has been achieved, 185–555 MBq 18F-
molecules of pyruvate (glycolysis) that then FDG should be administered.
proceed to the Krebs cycle and subsequent
steps associated with oxidative metabolism. Image acquisition typically begins 45–60
Alternatively, 18F-FDG becomes converted min post injection. Certain patient popula-
to 18F-FDG-6-phosphatase, which is not a tions such as diabetics may present glucose
suitable substrate for oxidative metabolism management challenges because insulin
and cannot pass to subsequent steps associ- resistance may result in the need to delay
ated with glucose metabolism. Additionally, imaging up to 2 or 3 h. Patients should be
18
F-FDG-6-phosphatase becomes trapped instructed to wear comfortable, non-atten-
within the intracellular space and serves as uating clothing and be positioned supine
an excellent marker for relative glucose me- with arms above the head and outside of the
tabolism. camera field of view. Scout scans may be per-
formed with 18F-FDG data or CT data in the
On the day of the scheduled procedure, pa- case of PET/CT. Images may be acquired for
tients should arrive with nothing by mouth 10–30 min in static or dynamic list mode. At-
for at least 6 h prior to the scheduled exami- tenuation correction may be performed im-
nation time. An initial blood serum glucose mediately before or after the emission scan,
measurement should be taken prior to ad- prior to patient movement, using either CT
ministration of glucose. If initial blood serum or 68Ge rod sources. Data are reconstructed
glucose values are less than 250 mg/dl, then with filtered back-projection or iterative re-
the protocol may proceed with 25–100 mg construction [1].
of oral glucose. Subsequent blood serum
glucose measurements should be taken until Cardiac PET images obtained with 18F-FDG
a blood serum glucose level of 100–140 mg/ are often compared with resting perfusion
dl is obtained. The increase in blood serum images in order to identify tissue that was
glucose associated with oral administration not readily detectible based on potassium
of glucose and its subsequent decrease indi- transport or other mechanisms relied upon
cate the release of endogenous insulin by the by cardiac perfusion radiopharmaceuticals
pancreas, creating a suitable environment for but still participates in active glucose metab-
administration and imaging of 18F-FDG. If the olism and is therefore viable. This mismatch
patient’s initial fasting blood serum glucose represents viable myocardium that is either
71
“stunned” or hibernating and would ben- Future directions
efit from revascularisation. Myocardium that The unquestionable benefits of cardiac PET
does not demonstrate glucose metabolism have led to efforts to increase the availabil-
as represented by 18F-FDG accumulation fol- ity of PET radiopharmaceuticals. The require-
lowing a lack of perfusion tracer concentra- ment for a nearby cyclotron for 13N-ammonia
tion represents infarcted tissue (Fig. 9). and the supply costs associated with 82Rb
generators have placed an emphasis on car-
diac PET radiotracers labelled with 18F. The
ideal radiotracer for cardiac PET imaging
would exhibit a high extraction proportional
to regional cardiac flow even at very high
flow rates, have a sufficient half-life to allow
for optimal logistics and work-flow, entail
reasonable radiation exposure for staff and
patients and provide the ability to conduct
quantification and excellent image quality in
Figure 9: Corresponding series of 201Tl rest- all patient populations. The cardiac PET per-
redistribution SPECT and FDG PET short-axis fusion agent [18F]BMS747158 or Flurpiridaz,
slices. The early 201Tl slices (top) show a defect still in clinical trials, has thus far demonstrat-
in the lateral wall and redistribution on late ed these tracer characteristics and has been
201
Tl images (middle). FDG slices (bottom) well tolerated by subjects and provides high-
show increased uptake in the lateral wall resolution cardiac PET perfusion images (Fig.
compared with early 201Tl images. Lateral wall 10). A rotenone analogue, this agent has af-
was akinetic on resting echocardiography and finity for the mitochondrial complex found
improved in function with revascularisation in abundance within myocardial tissue.
18
F-Flurpiridaz has also exhibited excellent
myocardial extraction fraction and flow rate
characteristics [5].
72
Chapter 7 PET/CT Protocols and Imaging Equipment
EANM
Figure 10: Cardiac imaging with [18F]BMS747158. Top row, standard reconstruction. Bottom row,
high-definition cardiac perfusion PET
Figure 11: Representative cardiac PET images of LMI1195 (first and third columns) and [18F]
BMS747158 (second and fourth columns) in rabbits with and without regional (phenol directly
painted on the heart, left four images) and systemic (intravenous injection of 6-OHDA to destroy
cardiac neurons, right four images) cardiac sympathetic denervation. The denervated area in
the heart was clearly detected by imaging with LMI1195 but not [18F]BMS747158 in the same
rabbit. With normal perfusion images in the denervated rabbit, decreased LMI1195 heart uptake
resulted from reduced neuronal function, not changes in myocardial perfusion
73
Future opportunities in cardiac imaging will
offer unique information to clinicians that will
aid in patient stratification for therapy and in-
tervention. The novel biomarker LMI1195 (Fig.
11) has demonstrated the ability to differenti-
ate normal myocardial tissue from tissue that
exhibits neuronal dysfunction and remodel-
ling. Neuronal dysfunction following an isch- A. B.
aemic event has been show to contribute to s
0 25 50 75 100 % Intensity
arrhythmias [5]. Identification of this cardiac
remodelling may assist with patient stratifica- Figure 12 A, B: PET/CT images of free fatty acid
tion for implantable cardioverter defibrillator analogue 18F-fluoro-cyclopropyl hexadecanoic
(ICD). Additional future directions in cardiac acid (18F-CPHA). A Normal myocardium with
PET imaging include detection of ischaemic homogeneous distribution of 18F-CPHA associ-
injury by imaging the metabolic shift from ated with normal free fatty acid metabolism.
18
F-labelled fatty acids to glucose. Areas of B Abnormal myocardium with areas of de-
injured myocardium preferentially metabo- creased fatty acid metabolism from metabolic
lise glucose rather than free fatty acids during shift associated with anaerobic metabolism
and after periods of ischaemia. Initial image and injury
sets associated with flow may be compared
to delayed images associated with metabo-
lism (Fig. 12). Myocardial segments associated
with normal flow but decreased fatty acid
metabolism are potentially indicative of a re-
cent ischaemic event [6].
74
Chapter 7 PET/CT Protocols and Imaging Equipment
EANM
Conclusion considering installation of a PET or PET/CT
Successful implementation of a cardiac PET unit. Patient scheduling may be complicat-
imaging programme requires consideration ed when performing cardiac perfusion and
of many issues. The availability of radiotracer, cardiac viability procedures at a facility also
whether as a result of a local cyclotron with providing oncology and neurology services
sufficient capabilities and staffing or the ac- due to differences in patient preparations,
quisition of 82Rb generator, may be the first imaging start times post dose administration
important factor to consider. Availability of and the requirements for different staff to
13
N-ammonia may provide the opportunity be present. The development of new 18F-la-
to conduct exercise cardiac PET procedures belled cardiac imaging agents may improve
but necessitates planning for adequate image quality, quantification and access to
treadmill space in close proximity to the PET cardiac PET imaging. However, these agents
or PET/CT unit. Stress testing and image ac- may present additional logistical challenges
quisition protocols will be determined by such as radiation exposure to staff, that must
the radiotracer utilised. The half-life of the be addressed. While the establishment of a
administered radiopharmaceutical will deter- cardiac PET service presents unique chal-
mine dosage as well as radiation exposures lenges, they have been successfully ad-
to staff and patients. Adequate shielding, size dressed at many facilities and the benefit to
and weight-bearing characteristics of the patients makes this endeavor worthwhile.
imaging room must be accounted for when
75
References Chapter 7
References
1. Dilsizian V, Bacharach SL, Beanlands RS, et al. PET
myocardial perfusion and metabolism clinical imag-
ing. ASNC Imaging Guidelines for Nuclear Cardiology
Procedures. American Society of Nuclear Cardiology J
Nucl Cardiol. 2009:16.
76
Chapter 8
Image Processing and Software
Sérgio Figueiredo and Pedro Fragoso Costa
EANM
Introduction modynamically significant coronary artery
Myocardial perfusion imaging (MPI) has be- disease (CAD) [1,2].
come a relatively common diagnostic test
in the clinical field of nuclear medicine [1]. Procurement of test results with optimal
In particular, gated single-photon emission quality depends on technical issues relating
computed tomography (G-SPECT) has be- to image acquisition, processing and quan-
come a part of mainstream practice, mostly tification. Despite “soft” inter-lab differences,
being used in the non-invasive assessment a dynamic workflow process similar to that
of patients with known or suspected hae- shown in Fig. 1 is generally followed.
›› Attenuation Correction
DATA ACQUISITION ›› Scatter Correction
›› Resolution Recovery
›› Filtering
IMAGE PROCESSING ›› Reconstruction algorithms
›› Slice orientation
IMAGE QUANTIFICATION
PERFUSION FUNCTION
›› Diagnosis
REPORT ›› Prognosis
Figure 1: Generic cardiac imaging workflow. LVEF, left ventricular ejection fraction; ESV, end-
systolic volume; EDV, end-dystolic volume; RWM, regional wall motion; RWT, regional wall
thickening
77
After the data acquisition, with or without ately aligned in a format that allows ready
attenuation and scatter correction, the pla- comparison of corresponding tomograms
nar projection images should be inspected [6].
immediately by the technologist in order to
identify technical problems that might re- Analytical and iterative reconstruction
quire repetition of the acquisition. This step is Filtered back-projection is an analytical recon-
referred to as “image evaluation” and parame- struction method that has been widely used
ters to consider may include: patient motion, in clinical cardiac G-SPECT because of its sim-
inaccurate ECG gating, hot gastro-intestinal plicity, speed and the fact of being relatively
activity encroaching on the left ventricular simple to compute. However, it has gradually
(LV) inferior wall, injection site, objects sur- declined in importance compared with itera-
rounding the heart and issues related to de- tive methods, mainly because of its inability
tection and mechanical physics (Fig. 1). Only to model basic physical processes in emis-
after proper evaluation of these parameters sion tomography, e.g. attenuation and scat-
and correct validation of the acquisition is ter [7, 8].
it possible to proceed to the next steps of
the flowchart: image reconstruction, image Figure 2A illustrates the practical steps in-
quantification and report. volved in reconstruction by FBP. In a 2D ac-
quisition, each row of projections represents
Image reconstruction the sum of all counts along a straight line
Image reconstruction has traditionally through the depth of the object, and all pro-
been performed using filtered back-projec- jections are organised as a function of the an-
tion (FBP), but reconstruction algorithms gle in a sinogram. The back-projection tech-
such as iterative methods can be applied nique redistributes the number of counts
in MPI [3]. Similarly, filtering techniques uniformly at each particular point along the
help improve image clarity by, potentially, line from which they were originally detect-
removing noise and blur after back-projec- ed, for all pixels and all angles. Star artefacts
tion of the raw data [4]. Recently, improve- and blurring of the image are the main con-
ments in hardware and software have been sequences of this process. In order to elimi-
introduced in an attempt to meet the chal- nate these problems, the projections are fil-
lenges of modern healthcare MPI studies tered with a window filter, in the frequency
through the implementation of resolution domain, before being back-projected onto
recovery (RR) and noise suppression (NS) the image matrix, mathematically using the
algorithms [5]. After reconstruction, re- fast Fourier transform (FFT). The back-pro-
orientation of stress and rest data can be jection takes place in the spatial domain in
displayed and images should be appropri- order to obtain the final image [4, 8].
78
Chapter 8 Image Processing and Software
EANM
A B
SINOGRAM FINAL
PROJECTION
* after the nth iteration
FFT 1D
WINDOW FILTER
INITIAL ESTIMATED COMPARE MEASURED
GUESS PROJECTION PROJECTION
FFT -1
UPDATE
THE
BACK PROJECTION PROJECTIONS
TO CREATE ERROR
A NEW PROJECTION
FINAL IMAGE ESTIMATE
Iterative reconstruction is based on the alge- between the generated and actual projec-
braic reconstruction technique and allows tions, and the process is repeated (hence the
the incorporation of more accurate image term “iterative”) until the differences between
models rather than the Radon model as- the calculated and measured data are small-
sumed in the FBP algorithm (Fig. 2B). These er than a specified preselected value [3, 6].
include attenuation and scatter corrections
as well as collimator response and more re- Diverse iterative algorithms have been de-
alistic statistical noise models [8, 9]. Fuelled veloped, such as the maximum likelihood
by increases in computer speed and perfor- expectation maximisation (MLEM), assum-
mance, these methods are gaining wider ac- ing Poisson noise is present in the projec-
ceptance in the field of nuclear cardiology. tion data but with long computational times.
Currently, the most widely used iterative
These algorithms, as illustrated in Fig. 2B, start technique is based on the ordered subsets
with a rudimentary initial first guess (usually expectation maximisation (OSEM) approach,
by FBP) of the activity distribution, generate which is an accelerated version of MLEM; its
projections from the guess and compare main advantage is the decreased reconstruc-
these projections with the acquired ones. tion time, which is consistent with routine
The guess is refined based on the differences clinical use [8, 10].
79
New trends in iterative image reconstruction
POINT SOURCE
Although it is a well-established medical
imaging modality, MPI suffers from some
fundamental limitations, including long im-
POINT SOURCE
age acquisition time, low image resolution
(depending on scanner/camera photon sen-
sitivity, primarily controlled by the type of
collimator and imaging detection geometry)
and patient radiation dose [11]. Typically, to
balance the binomial resolution–sensitivity
(higher resolution images require lowering
of image sensitivity), the scan time is about
15–20 min for each stress and rest acquisition,
resulting in frequent artefacts due to patient COLLIMATOR
DETECTOR
motion [2, 11]. Additionally, the traditional
dual-head cameras with standard parallel
hole collimators are over 50 years old and SHARP IMAGE BLURRED IMAGE
80
Chapter 8 Image Processing and Software
EANM
For each combination of systems, equip- in order to achieve diagnostically valuable im-
ment-related information for different col- ages. Table 1 lists the main manufacturers and
limators is stored in an additional look-up the reconstruction algorithms applied.
table, which is part of the reconstruction
package and is measured previously to en- Table 1: Most widely available commercial
sure correct modulation of the a priori physi- reconstruction software
cal aspects, while the acquisition parameters
are retrieved from raw projection data [2, 12]. Reconstruction
Software Manufacturer
The modelled physical information (instru- algorithm
mentation and imaging parameters) is then Astonish 3D-OSEM Philips
integrated in the estimated projections of Evolution for OSEM-RR
the overall iterative method, eliminating the GE
Cardiac (MAP)
main degrading effects of the line spread Conjugate
function (LSF) and thereby resulting in im- IQ·SPECT gradient Siemens
proved SNR and enhanced spatial resolution (CG)
[2, 11]. Regarding these new trends in image WBR OSEM-RR UltraSPECT
reconstruction, the initial clinical results that
MAP, Maximum a posteriori
are being reported demonstrate potential for
equivalent diagnostic performance by MPI
scans obtained in half-time sampling mode AstonishTM
[11, 13, 14]. Philips (Milpitas, CA) has developed a fast
SPECT reconstruction algorithm, Astonish,
Commercial software approach that includes corrections for the major fac-
With the recent development of nuclear cardi- tors degrading SPECT image quality [11]. This
ology and its increasing application in the clin- software package is based on a three-dimen-
ical field, the need for processing and acquisi- sional ordered subsets expectation maximi-
tion dedicated software has increased. Apart sation (3D-OSEM) method. Astonish uses a
from the main scanner manufacturers, who distance compensation method to model
develop such software suited to their prod- the varying resolution at different distances
ucts and dedicated cameras, there are also from the detector; such corrections allow the
stand-alone software applications with “solu- visualisation of sharper details. Additionally,
tions” for image reconstruction using con- it corrects for Compton scattering and for
ventional hardware scanners. Most modern photon attenuation by means of the ESSE
software uses iterative reconstruction meth- method described by Kadrmas et al. [15],
ods; each developer uses a specific approach which improves lesion contrast and provides
to produce high-resolution images, reduced a more accurate representation of lesions at
noise acquisitions or shorter acquisition times varying depths [16].
81
Noise suppression – one problem that all it- Noise suppression is achieved with a maxi-
erative methods must address is amplifica- mum a posteriori (MAP) algorithm. This algo-
tion of statistical noise – is attained through rithm is a modification of the EM. It operates
a matched filtering. This process consists of in order to maximise the similarity between
filtering before reconstruction and using an estimated and measured projections and to
identical filter to forward project the informa- ensure that reconstructed images are not too
tion. This avoids the dissipation of features that noisy [5].
could have been “smoothed away” and the in-
creased iteration-associated noise [11, 16]. IQ·SPECTTM
This software package works in synergy with
EvolutionTM for Cardiac a dedicated collimator (SMARTZOOM), the
GE Healthcare (Waukesha, WI) has also design of which enables a cardio-centric
developed a modification of the OSEM al- orbit at a fixed radius of 28 cm. While most
gorithm that incorporates RR named “Evo- software designs are based in the EM opti-
lution for Cardiac” [11]. Focussed on CDR misation method, IQ·SPECT uses a precondi-
modeling, the Evolution approach is able to tioned conjugate gradient (CG) [17]. This al-
achieve resolution recovery by integration gorithm is based on the minimisation of the
of the collimator and detector response in Poisson noise function. In the CG method the
an iterative reconstruction algorithm [13]. aim is not to maximise the likelihood that the
The CDR compensation technique is ac- objective function is an answer to the system
complished by convolving the projected but to minimise the objective function; for
photon ray with the corresponding LSF dur- this purpose, steps conjugate to one another
ing iterative projection and back-projection. are taken in the direction in which the objec-
The following compensation parameters are tive function decreases more quickly [7].
taken into account: collimator hole length
and septa thickness, intrinsic resolution, Similarly to EM, an increase in iterations with
crystal thickness and collimator–detector the CG results in noise accumulation on the
gap. Acquisition parameters, including the produced images, and to diminish the extent
distance from the centre of rotation to the of this the CG is preconditioned with a nearly
collimator for every acquired projection, ob- complete description of the imaging system,
tained from the raw data, also contribute for from emission to collimator acceptance: scat-
modulation [11, 12]. ter and attenuation correction are included, a
vector map of the collimator angles in included
82
Chapter 8 Image Processing and Software
EANM
in the matrix, PSF is modulated based on the Importance of reconstruction filters
radius of rotation and a post filter is applied to Once a Fourier spectrum has been gener-
the final reconstructed images [17, 18]. ated for an image, it can be filtered so that
certain spatial frequencies can be modified,
WBRTM enhanced or suppressed. The filtered spec-
UltraSPECT Limited (Haifa, Israel) has devel- trum can then be inverse transformed, using a
oped a stand-alone workstation (Xpress.car- mathematical operator (inverse Fourier trans-
diac) that utilises the patented Wide Beam form), to generate a filtered image with, for
Reconstruction (WBR) algorithm [11]. The WBR example, sharpened or smoothed features [8].
technique is an OSEM-based algorithm that
models the physics and geometry of the ac- These operations are commonly applied in
quisition for resolution recovery [19] in a similar order to obtain good quality images, and
fashion to the previously described methods. specifically to maximise the SNR, which
Using a segmentation algorithm, the distance describes the relative strength of the sig-
from the detector to the body can be extrapo- nal component compared with noise. The
lated and integrated in the CDR model. Noise SNR is much higher at low spatial frequen-
suppression is achieved using a likelihood cies (broad features that are constant over
function with a combination of the Poisson many pixels, such as large uniform objects)
(high frequency recovery) and Gaussian (high and decreases at higher spatial frequen-
frequency suppression) distributions, accom- cies (features that change over a few pix-
plished by Fourier analysis of a projection to els, such as edges) [1, 3]. Window filters are
determine the approximate SNR, thus avoid- such tools; those most frequently used in
ing post-reconstruction filtering (the standard nuclear cardiology are passive filters, of
approach to overcome noise) [11, 19, 20]. which low-pass and high-pass are repre-
sentative examples [4].
WBR utilises a stand-alone running hardware
platform and can reconstruct data acquired In order to achieve an optimal compromise
from most scanners with standard collimator between the extent of noise reduction and
design [20]. suppression of fine detail, the typical star ar-
tefact from FBP can be reduced by applying
83
a Ramp filter [3]. This mathematical window is
Bw( f ) = 1
( )
a high-pass filter that does not permit low fre- 2n
f
quencies that cause blurring to appear in the 1+
image and enhances the geometric contours fc
of the object. In frequency domain, its math-
ematical function is given by the equation [8]: where f is the spatial frequency domain, f
c
the critical (or cut-off ) frequency and n the
HR(kx ,ky ) = k = (kx2 + ky2) 1/2 order of the filter.
where kx ,ky are the spatial frequencies. The cut-off frequency (or roll-off frequency)
defines the frequency above which the noise
The main disadvantage of the Ramp filter is is eliminated. The Nyquist (Nq) frequency –
that it amplifies the high-frequency statisti- the highest frequency that can be displayed
cal noise present in the measured counts, in an image – is apparently the highest cut-
and as a consequence it is usually combined off frequency for a filter. Its importance is re-
with a low-pass filter [1]. A low-pass filter is lated to the fact that the critical frequency is
the common method to reduce or remove expressed in cycles per pixel (or cm) or as a
statistical noise in an FBP cardiac image be- fraction of the Nq frequency [4, 7]. Typically
cause it allows low spatial frequencies to be the cut-off frequency varies from 0.2 to 1.0
unaltered and attenuates the high frequen- times the Nq frequency, and because of the
cies, where noise predominates. Hanning variability in notations, any measurement
and Butterworth are the most popular expressed in frequency terms must always
low-pass filters used in nuclear cardiology. be accompanied by knowledge of the pixel
Because of the flexibility and ease of de- size [4, 8]. Therefore, a high fcwill improve the
sign, Butterworth filters are the more usual spatial resolution and, consequently, much
choice in nuclear cardiology routine pro- detail can be seen, but the image will remain
cedures [8]. They are characterised by two noisy. Furthermore, a low fc will increase
parameters: the “cut-off frequency” and the smoothing but will degrade image contrast
“order” (or the “power”) and are described, in in the final reconstruction [3, 6] (Fig. 4).
the spatial domain, by the equation:
84
Chapter 8 Image Processing and Software
EANM
Max The order controls the slope of the filter func-
A
tion and characterises the steepness of the
roll-off. A high n will result in a sharp fall, while
a low n is responsible for a smooth fall [8].
Min
Nuclear cardiology images, because of their
BUTTERWORTH FILTER CRITICAL FREQUENCY 0.2 relatively low count statistics and dependence
1.0
on the ramp filter (FBP), tend to have great
amounts of statistical noise. Additionally, bal-
anced selection of parameters n and fc may
0.0
0.0 1.0
CYCLES / cm
2.0
optimise the SNR and improve image quality.
Traditionally, the manufacturer’s recommen-
Max
B dations, e.g. fc =0.5 cycles·cm-1 (n=5 or n=10)
or 0.75 cycles·cm-1, can be chosen [4].
85
to retain a comparable count density/im- rest images, and not in order to maintain a
age appearance in both sets of stress and consistency of appearance (Table 2) [6, 21].
Pre-filter
Butterworth Hanning
Iterative
a
Some individuals prefer to express the critical frequency as a 0–1 numeric range, with 1 being the highest attainable
frequency, or 100% of the Nyquist frequency. Others point out that the Nyquist frequency is, by definition, equivalent to
0.5 cycles per pixel and adopt a 0–0.5 range instead. The same critical frequency can be reported as 0.3 or 0.6 on two
different camera systems. In any event, whether a critical frequency is expressed in cycles per pixel or as a fraction of the
Nyquist frequency, any measurement expressed in frequency terms must, clearly, always be accompanied by knowledge
of the pixel size [4].
b
Usually no pre-filtering is needed, but post-filtering techniques can be applied.
86
Chapter 8 Image Processing and Software
EANM
reorientation are applied, making use of the to be at least as capable as trained operators in
software widely available in clinical practice. achieving improved reproducibility [1, 21]. An
In addition to the “zoom reconstruction” tools, example of correct automatic reorientation is
these automated methods have been shown shown in Fig. 5.
Medicine Laboratory, Lisbon, Portugal
Courtesy of Atomedical, S.A., Nuclear
Figure 5: Correct reorientation and adjusted alignment in a WBR gated-stress half-time acquisition.
87
Stress and rest images should be appropri- the SA plan. The VLA tomograms should be
ately aligned and sequentially adjacent to displayed with septal slices on the left and
each other, and presented in a format that lateral slices on the right. Similarly, the HLA
allows ready comparison of the parallel cor- tomograms should be displayed with inferior
responding tomograms. Conventional stress- slices on the left and anterior slices on the
rest display should show the apical slices to right. Figure 7 shows an example of this tra-
the left and the base at the right, considering ditional display.
Courtesy of Atomedical, S.A., Nuclear Medicine Laboratory, Lisbon, Portugal
Figure 7: Conventional stress-rest display in a WBRTM half-time gated acquisition. Top: Stress-rest
SA slices; middle: stress-rest VLA slices; bottom: stress-rest HLA slices.
Different colour maps can be used depend- ommended [6, 21]. Additionally, the same
ing on the operator’s expertise, but use of a colour map should always be used for rest
linear or a grey colour scale is classically rec- images and stress slices.
88
Chapter 8 Image Processing and Software
EANM
Each study should be displayed with the top Image quantification
of the colour scale at the maximum counts/ Image quantification is an extremely valuable
pixel within the myocardium for each set of tool in MPI and usually involves the extraction
images. Each series (VLA, HLA and SA) may of quantitative parameters derivable from a
be normalised to the brightest pixel in the rest-stress gated perfusion SPECT protocol.
entire image set, providing the most intui- After correct implementation of the left ven-
tively easy way to evaluate the perfusion de- tricle segmentation algorithm, it is possible to
fects. However, this approach is very sensitive obtain the perfusion parameters, i.e. defect
to focal hot spots, so if the pixel with maxi- extent, severity values with polar map display,
mum counts lies outside the myocardium, reversibility stress-rest, percent hypoperfused
care should be taken and manual adjust- myocardium, percent ischaemic myocardium,
ment or masking of extracardiac activity may categorical summed scores (rest, stress and
be required. difference), and even total perfusion deficit.
Likewise, the left ventricle (LV) global func-
Sometimes, the stress-rest display can in- tional parameters, i.e. left ventricle ejection
clude 3D-rendered data sets of regional fraction (LVEF), end-diastolic volume (EDV)
myocardial perfusion. These may help less and end-systolic volume (ESV), and the LV
experienced readers to identify coronary dis- regional parameters, i.e. regional wall motion
tributions associated with perfusion defects and thickening, in conjunction with phase
but should be used only as an adjunct, not a analysis and 3D gated surface display, may
replacement. contribute significantly to the final report.
Dimensionality 3D 3D 3D 3D
Maximal pixel;
Method Gaussian fit Partial volume Gradient
partial volume
89
In order to obtain the LV perfusion and func- Most recently, as described in the section
tional parameters several software packages “Commercial software approach” (above), ven-
are commercially available, among which dors have introduced novel noise reduction
three have become the most popular [1, and/or resolution recovery protocols based on
11]: QPSTM/QGSTM, developed at Cedars-Sinai iterative reconstruction methodology, leading
Medical Centre (Los Angeles, California), ECT- to a significant improvement in image quality.
bTM, developed at Emory University (Atlanta, The above quantification algorithms are being
Georgia), and Corridor 4DMTM, developed at optimised to take account of the new recon-
the University of Michigan (Invia, Ann Arbor, struction methods and, furthermore, the new
Michigan). A fourth method also used com- current fast digital systems.
mercially is the Wackers-Liu CQ software
(GSCQTM), developed at Yale University (New Left ventricle segmentation algorithm
Haven, Connecticut) [1, 11] Table 3 provides a When using a software package, segmenta-
brief summary of these algorithms. tion is a crucial step (Fig. 1). Segmentation re-
fers to the separation of a region or a structure
Further factors have been very important of interest from the remainder of an image. In
in the growth and popularity of these com- nuclear cardiology, this kind of algorithm de-
mercially available software packages: (a) tects the myocardial surface of the LV out of a
they are automated, (b) they integrate dif- myocardial perfusion image [1, 24].
ferent image display techniques, permitting
simultaneous assessment of LV perfusion In general, the segmentation process is de-
and function in one package, and (c) they pendent on the “correctness” of the reori-
are well validated, with excellent reproduc- entation of the SA slices and the geometric
ibility of the results [11, 22]. However, it is boundaries of the LV, based on various com-
important to understand that the imple- putational techniques. The algorithm flow is
mentation of these programs varies from roughly composed of a cascade of consecu-
vendor to vendor and even between ver- tive computational steps, including initial
sions of the same program. It is neverthe- centre of mass calculation and final valve
less well accepted that in general the au- plan fit [1, 25, 26].
tomation in these programs is quite robust
and that they yield similar results [11, 23]. Regardless of the computational tech-
Different image display results can be found niques applied, there are several challenges
in the sections “Polar maps”, “Perfusion im- that the algorithm needs to overcome. In
aging quantification” and “Left ventricular particular, its robustness depends on three
function imaging quantification” (below) , as key criteria: hot gastro-intestinal activity,
examples of these methods. valve plane detection and under-perfused
90
Chapter 8 Image Processing and Software
EANM
tissue. Hot gastro-intestinal activity can attaches to the myocardium itself, the geo-
sometimes be extremely close to myocar- metric contours of the LV are unaffected; in
dial activity and even bind to myocardial contrast, in the example in Fig. 8B, the hot
uptake [26]. Figure 8A shows an example gastro-intestinal activity influences the seg-
in which, although the intestinal activity mentation process.
A B
Courtesy of Atomedical, S.A., Nuclear Medicine Laboratory, Lisbon, Portugal
Figure 8 A, B: Correct LV contour detection (A) and influence of hot gastro-intestinal activity on
the segmentation process (B).
Valve plane detection is responsible for de- of endocardial volume calculation, avoiding
termining the plane, which should enclose atrial uptake at the diastolic-gated intervals
the chamber inner surface for the purpose [26].
91
Regarding under-perfused tissue, in hypo- normal surrounding tissue [26]. The correct
perfused regions of the tomograms the al- calculation of these hypoperfused areas may
gorithm should be able to interpolate the be hardly discernible. The example in Fig. 9
correct myocardial edges according to the illustrates this drawback.
A B
Medicine Laboratory, Lisbon, Portugal
Courtesy of Atomedical, S.A., Nuclear
Figure 9 A, B: Under-perfused tissue criterion. A Correct estimation of the myocardial edges and
corresponding polar maps. B Overestimation of the myocardial delimitation and subsequent
incorrect lesion extension on polar map representation.
Concerning the examples showed above, packages will allow manual override of the
the automatic detection of contours should automatic segmentation (Fig. 10). However,
be careful verified visually for both gated and care should be taken because quantitative
ungated data because errors in LV segmen- results will be associated with operator ac-
tation may lead to significantly erroneous tion.
myocardial quantification [1]. To avoid these
errors, most commercially available software
92
Chapter 8 Image Processing and Software
EANM
Courtesy of Atomedical, S.A., Nuclear Medicine Laboratory, Lisbon, Portugal
Figure 10: Example of manual override segmentation for adjustment and constraint of LV mask
in QPS software. Same patient as in Fig. 8.
93
SA is displayed as a new ring surrounding Accordingly, based on a colour-coded meth-
the previous one. The outermost ring of od, different parameters may be “selected” on
the polar map comprises the basal slices of distinct maps – for example, the percentage
the LV [11, 21]. The polar maps can display of perfusion on rest and stress, the number
the standardised classical nomenclature of standard deviation (SDs) below normal, in-
of 17 segments of the American Society of dicating the severity of any abnormality, and
Nuclear Cardiology/American Heart Associa- the blacked-out areas or extent of the defect
tion (AHA)/American College of Cardiology region, thus creating a white-out reversibil-
or can be segmented by “vessels” [the left ity polar map. Additionally, these polar maps
anterior descending coronary artery (LAD), can be an important tool in determining the
right coronary artery (RCA) and left circum- LV functional parameters, such as regional
flex coronary artery (LCX)] or “walls” (anterior, wall motion, regional wall thickening, %ED
inferior, lateral, septal and apical) (Fig. 11) [27]. perfusion, %ES perfusion and phase analysis.
The 20-segment model may be applied too, More recently, some 3D displays have been
but the 17-segment model is usually recom- adopted in clinical practice.
mended [3, 6].
2 6
13
8 12
14 17 16
9 11
15
3 5
10
LAD
Figure 11: Standard 17-segment polar map
LEFT ANTERIOR DESCENDING CORONARY ARTERY
model with assigment of segments to the vas-
RIGHT CORONARY ARTERY RCA cular territories of the left anterior descending
coronary artery (LAD), right coronary artery
LEFT CIRCUMFLEX CORONARY ARTERY LCX
(RCA) and left circumflex coronary artery (LCX)
94
Chapter 8 Image Processing and Software
EANM
Perfusion imaging quantification
Table 4: Quantitative perfusion measures of myocardial perfusion. Adapted from [11, 21]
95
The SSS represents the perfusion defect seen portant parameter in terms of prognosis [21].
at stress (it equals the sum of the stress scores These semiquantitative scores have been
of all segments). The SRS is considered equal shown to provide important prognostic in-
to the magnitude of a fixed defect (it equals formation. The maximal abnormal score is 68.
the sum of the rest scores of all segments)
and hence represents – in most cases – the For the calculation of the summed scores,
size and severity of a myocardial infarction the basic quantitative software approaches
(although in some cases it may also repre- (shown in Table 3) are similar: regional ra-
sent hibernating myocardium with viability). diotracer uptake is quantitatively compared
Finally, the SDS is the difference between the with normal databases [1, 11] and can be ex-
SSS and the SRS and expresses the magni- pressed by:
tude of ischaemia (reversibility), the most im-
These normal databases are usually gener- Each of the programs is based on slightly
ated from patients with a low likelihood of different models that are used to generate
CAD, i.e. usually less than 5% probability [11]. the quantitative profiles, where the main
Consequently the automatic assignment difference is largely in the display data [11].
score results should be carefully interpreted Examples of application of three commer-
by experienced and authorised professionals. cial programs for quantification in the same
patient are shown in Figs. 12–14. However,
It is also possible to combine defect severity interpretation of summed scores should be
and extent through calculation of the total consistent with the visual analysis of the im-
perfusion deficit, using QPS software. This ages, since there are several possible pitfalls
parameter reflects the extent and severity of in the score calculation [21, 29].
the overall perfusion defect and can be an
important tool for MPI quantification, having
a strong correlation with a variety of widely
used SSS displays [26].
96
Chapter 8 Image Processing and Software
EANM
Courtesy of Atomedical, S.A., Nuclear Medicine Laboratory, Lisbon, Portugal
Figure 12: Perfusion quantification results obtained with QPS processing, suggesting anterolateral
ischaemia in a female patient. The ischaemia can be observed by analysis of the stress-rest
slices, the polar map defect extension at stress and rest, and the whiteout reversibility map.
The summed scores are displayed on the bottom of the image. Data reconstructed with WBR
technology.
97
Courtesy of Atomedical, S.A., Nuclear Medicine Laboratory, Lisbon, Portugal
Figure 13: Perfusion quantification extension results obtained from processing with Corridor
4DM, in the same patient as in Fig. 12. Note the similarities of these results with those of Fig. 12.
98
Chapter 8 Image Processing and Software
EANM
Figure 14: Perfusion quantification output
Courtesy of Atomedical, S.A., Nuclear Medicine Laboratory, Lisbon
Importantly, quantitative interpretation tive polar map display with the score system
should always be viewed as complementary serves to confirm the visual qualitative im-
to visual analysis of the images. Quantita- pression [3, 6].
99
Left ventricular function imaging quantification
Table 5: Global and regional parameters of systolic and diastolic left ventricular function
ESV MFR/3
SWM
SWT
Phase
Regional WM
WT
Phase
LVEF, left ventricular ejection fraction; EDV, end-diatolic volume; ESV, end-systolic volume; SWM, systolic wall motion; SWT,
systolic wall thickening; WM, wall motion; WT, wall thickening; PFR, peak filling rate; TPFR, time to peak filling rate; PER, peak
emptying rate; TPER, time to peak emptying rate; MFR/3, filling rate during first third of diastole
The gated technique is a widely used clinical Commonly a “canonical” shape of the time-
tool that may improve the prognostic power volume curve for an 8- or 16-frame acquisi-
of MPI through measurements of functional tion can also be obtained as a valuable tool
parameters [1, 11]. Gated perfusion images for clinical interpretation. Quantitative phase
can be quantitatively analysed to assess a analysis can also be performed either in a
remarkable number of parameters of cardiac global manner (synchrony of contraction of
function: global, regional, systolic and dia- the LV as a whole) or regionally, as the differ-
stolic (Table 5) [11]. ence between the onset of contraction in dif-
ferent myocardial walls (see section “Ejection
fraction and phase analysis”).
100
Chapter 8 Image Processing and Software
EANM
Different software approaches for the analy- When performing G-SPECT some technical
sis of ventricular function have been devel- considerations can be challenging as a con-
oped. The four most widely accepted com- sequence of the low spatial resolution of the
mercial packages (Table 3), combining sever- images; this is particularly true with respect
al algorithms, have been extensively studied to wall thickening and wall motion, rather
and have been validated for the parameters than global function [3]. It is indeed estimat-
they quantitate [1, 11, 30]. Figures 15 and 16 ed that the overwhelming majority of insti-
show typical examples of these commercial tutions performing G-SPECT imaging today
approaches. also employ quantification to some degree,
as an enhancer tool for the reproducibility of
the measured parameters [1, 11].
Courtesy of Atomedical, S.A., Nuclear Medicine Laboratory, Lisbon
Figure 15: Representative display for Cedars-Sinai Medical Center’s QGS software.
101
Courtesy of Atomedical, S.A., Nuclear Medicine Laboratory, Lisbon
Figure 16 A, B: A Representative display for the University of Michigan’s Corridor 4DM software.
B Representative display for Emory University’s Cardiac Toolbox software.
Ejection fraction and phase analysis interval in the heart, after which the EDV and
LVEF is an important cardiac performance the ESV are identified as the largest and the
parameter and is typically measured using a smallest LV cavity volume, respectively [11],
volume-based rather than a count-based ap- from which the ejection fraction is calculated:
proach. Specifically, the location of the LV en-
docardium is estimated in the 2D or 3D space, %LVEF = (EDV – ESV)EDV x 100
and LV cavity volume is calculated as the ter-
ritory bound by the endocardium and its The different commercially available algo-
valve plane. The process is repeated for every rithms output the LV function in distinct
102
Chapter 8 Image Processing and Software
EANM
displays, as represented in Figs. 15 and 16. Each term in the equation is called a har-
However, all of them have been validated on monic and for each harmonic, A represents
studies showing the agreement between G- its amplitude and P represents its phase.
SPECT and reference standard measurements For example, A0 is called the zero harmonic,
of quantitative LVEF, as described in detail by A1cos(2πft+P1) is called the first harmonic,
Germano et al. [1] and Zaret et al. [11]. A2cos(4πft+P2) is called the second harmonic,
and so on.
Phase analysis, initially described with nu-
clear cardiology methods in the late 1970s, The phase analysis technique measures the
has recently re-emerged for the purposes of first harmonic phase of regional LV count
quantitative measurement of LV mechanical changes throughout the cardiac cycle. This
dyssynchrony from G-SPECT images, assess- phase information is related to the time in-
ment of the likelihood of benefit from car- tervals until different regions in the 3D LV
diac resynchronisation therapy and evalua- myocardial wall start to contract, i.e. onset
tion of internal cardiac defibrillators [11, 31]. of mechanical contraction (OMC). It provides
The phase analysis technique is based on information on the homogeneity of these
the mathematical technique of harmonic time intervals for the entire LV [11, 32, 33].
function decomposition developed by Jean-
Baptiste Fourier (1768-1830), where a peri- The basics of phase analysis are generally
odic function F of t, with frequency f, can be similar in most quantitative software pack-
expressed as: ages, but practical implementations may
vary significantly [11]. Figure 17 shows an ex-
Σ
8
A B
Lisbon, Portugal
Figure 17 A, B: Phase histogram displaying normal pattern (A) and abnormal histogram (B) in a
patient with left bundle branch block.
103
Attenuation correction Transmission-based AC can be attained
The attenuation phenomenon is a result of through a line source system, consisting of
the interaction between the photons emit- a collimated rod source of gadolinium-153
ted by the radiopharmaceutical and the tis- (153Gd), cobalt-57 (57Co), or cerium-139
sue or other materials as the photons pass (139Ce), which can be fixed or rotating
through the body. The degree of attenuation around the patient. The transmission acqui-
increases with the atomic number of the sition provides an attenuation map for the
atom, depends on the tissue density and de- specific radionuclide gamma ray, which is
creases as the inverse of the photon energy. then derived for the imaging photon. This
This interaction is predominant for heavy at- configuration presented drawbacks such as
oms like lead and for low energetic gamma high cost for maintenance of long-lived ra-
rays, X-rays or bremsstrahlung [34]. dionuclide sources and unnecessary radia-
tion exposure for staff and patients and did
Attenuation is particularly challenging for not produce diagnostic information other
MPI because of the variability of the tis- than AC [36]. Recent advances on imaging
sue density in the thoracic region (the at- technology have led to hybrid imaging, in
tenuation values of lungs, myocardial tissue, which a SPECT or PET scanner is coupled to
breasts and diaphragm vary significantly). a CT scanner. CT provides a regional map
This normally produces attenuation-related of attenuation coefficients, specific to each
artefacts, which vary in size and severity de- patient. This works in a similar fashion to the
pending on the attenuator density and loca- above-mentioned technique, with supe-
tion. A typical breast artefact will be present rior spatial resolution and the possibility of
in the lateral wall of the heart, given the over- anatomical correlation. Since the patient’s
lay of the breast tissue and the anterolateral breathing or the constant myocardial mo-
wall of the heart [35]. tion can create discrepant co-registration
between SPECT and CT images, it is man-
To achieve attenuation correction (AC), a de- datory to perform alignment quality control
tailed knowledge of the patient’s anatomy after image acquisition [34].
is fundamental. This can be obtained either
through direct measurements of the patient There has been much controversy on the
using an external radiation source or by com- subject of AC in MPI, mainly owing to the
puting an estimate using body contour de- wide variety of available AC methods, which
lineation based on the acquired data set. makes it difficult to achieve a consensus [37].
104
Chapter 8 Image Processing and Software
EANM
AC images must always be displayed com- An example of AC with the RR algorithms is
plementary to non-corrected images (Fig. shown in Fig. 19.
18), and the decision on which AC method
to use must take into account the patient’s It would be easy to recommend that attenu-
attributes. ation and scatter correction should be ap-
plied in all cases; however, things are not that
No one approach has yet been identified as simple: without a demonstrable change in
clearly definitive, but the interest in hybrid the information that can be sustainably ex-
imaging and the recent RR/NS algorithms tracted from the data, conclusions must be
will likely result in greater reliance on AC. carefully achieved [11].
Courtesy of Klinikum Passau –
Gemeinschaftspraxis
nuklearmedizinische
Figure 18. AC and non-AC images obtained with a SPECT/CT full-time acquisition.
Courtesy of Atomedical, S.A., Nuclear Medicine Laboratory,
A B
Lisbon, Portugal
105
General image reconstruction and processing flow chart (tutorial)
ACQUISITION
YES OK ? NO
STRESS-REST MANUAL
DISPLAY SEGMENTATION
IMAGE QUANTIFICATION
FINAL REPORT
Acknowledgements. The authors have indicat- Fernando Godinho, PhD, Physicist Director of
ed no conflict of interest. While the lead (first) Atomedical, S.A., Lisbon, for the images that
author works with GE Healthcare equipment, they provided, for their comments and for
it is not considered that this has influenced proofreading the text. We would also like to
the writing of the chapter. The authors would thank Priv. Doz. Dr. Wolfgang Römer for kind-
like to thank Guilhermina Cantinho, MD, Clin- ly providing images. Carlos Viana is thanked
ical Director of Atomedical, S.A., Lisbon, and for providing help on illustrations.
106
References Chapter 8
EANM
References
1. Germano G, Berman DS, eds. Clinical gated cardiac 14. Pena H, Cantinho G, Shwartz SC, Figueiredo S, Marona D,
SPECT. Armonk, NY: Blackwell-Futura Publishing; 2006. Monteiro J, et al. Wide beam reconstruction technology:
does it respect myocardial perfusion SPECT functional
2. Garcia EV, Faber TL, Esteves FP. Cardiac dedicated ultra- parameters? J Nucl Cardiol. 2007;14:S108.
fast SPECT cameras: new designs and clinical implica-
tions. J Nucl Med. 2011;52:210-7. 15. Kadrmas DJ, Frey EC, Karimi SS, Tsui BM. Fast implemen-
tations of reconstruction-based scatter compensation
3. Holly TA, Abbott BG, Al-Mallah M, Calnon DA, Cohen MC, in fully 3D SPECT image reconstruction. Phys Med Biol.
DiFilippo FP, et al. Single photon-emission computed 1998;43:857-73.
tomography. J Nucl Cardiol. 2010;17:941-73.
16. Yen J, Song X, Durbin M, Zhao M, Shao L, Garrard J.
4. Germano G. Technical aspects of myocardial SPECT SPECT image quality improvement with Astonish soft-
imaging. J Nucl Med. 2001;42:1499-507. ware. White Paper - Philips Healthcare. 2010.
5. DePuey EG. Advances in SPECT camera software and 17. Hawman P, Ghosh P. IQ·SPECT: A technical and clinical
hardware: Currently available and new on the horizon. overview. White Paper - Siemens AG. 2012.
J Nucl Cardiol. 2012;19:551-81.
18. Vija H. Introduction to xSPECT Technology: Evolving
6. Hesse B, Tägil K, Cuocolo A, Anagnostopoulos C, Bardiés multi-modal SPECT to become context-based and
M, Bax J, et al. EANM/ESC procedural guidelines for myo- quantitiative. White Paper - Siemens Medical Solutions
cardial perfusion imaging in nuclear cardiology. Eur J USA, Inc Molecular Imaging. 2013.
Nucl Med Mol Imaging. 2005;32:855-97.
19. Borges-Neto S, Pagnanelli RA, Shaw LK, Honeycutt E,
7. Bruyant PP. Analytic and iterative reconstruction algo- Shwartz SC, Adams GL, et al. Clinical results of a novel
rithms in SPECT. J Nucl Med. 2002;43:1343-58. wide beam reconstruction method for shortening scan
8. Lyra M, Ploussi A. Filtering in SPECT image reconstruc- time of Tc-99m cardiac SPECT perfusion studies. J Nucl
tion. Int J Biol Imaging. 2011;2011:1-14. Cardiol. 2007;14:555-65.
9. Vandenberghe S, D‘Asseler Y, Van de Walle R, Kauppinen 20. UltraSPECT. Wide beam reconstruction: Breaking the
T, Koole M, Bouwens L, et al. Iterative reconstruction limitation of the line spread function. White Paper - Ul-
algorithms in nuclear medicine. Comput Med Imaging traSPECT. 2007.
Graph. 2001;25:105-11. 21. IAEA. Nuclear cardiology: Guidance and recommen-
10. Zeng GL. Image reconstruction – a tutorial. Comput dations for implementation in developing countries.
Med Imaging Graph. 2001;25:97-103. Vienna: IAEA; 2012.
11. Zaret BL, Beller G. Clinical nuclear cardiology : state of 22. Garcia EV, Faber TL, Cooke CD, Folks RD, Chen J, San-
the art and future directions, 4th ed. St. Louis, London: tana C. The increasing role of quantification in clinical
Elsevier Mosby; 2010. nuclear cardiology: the Emory approach. J Nucl Cardiol.
2007;14:420-32.
12. GE, Healthcare. Evolution for Cardiac - White Paper. 2007.
23. Sharir T, Germano G, Waechter PB, Kavanagh PB, Areeda
13. DePuey EG, Gadiraju R, Clark J, Thompson L, Anstett F, JS, Gerlach J, et al. A new algorithm for the quantita-
Shwartz SC. Ordered subset expectation maximization tion of myocardial perfusion SPECT. II: validation and
and wide beam reconstruction “half-time” gated myo- diagnostic yield. J Nucl Med. 2000;41:720-7.
cardial perfusion SPECT functional imaging: a compari-
son to “full-time” filtered backprojection. J Nucl Cardiol. 24. Germano G, Kiat H, Kavanagh PB, Moriel M, Mazzanti M,
2008;15:547-63. Su HT, et al. Automatic quantification of ejection fraction
from gated myocardial perfusion SPECT. J Nucl Med.
1995;36:2138-47.
107
25. QPS User Manual. Cedars-Sinai Medical Center. 2007. 35. Bateman TM, Cullom SJ. Attenuation correction single-
photon emission computed tomography myocardial
26. GE, Healthcare. Myovation PROTOCOL for XelerisTM perfusion imaging. Semin Nucl Med. 2005;35:37-51.
2 Functional Imaging P&R Systems - Operator Guide.
Technical Publications - Revision 1. Milwaukee, Wiscon- 36. Cherry SR, Sorenson JA, Phelps ME. Physics in nuclear
sin 2006. medicine, 4th ed. Philadelphia: Saunders; 2012.
27. Cerqueira MD, Weissman NJ, Dilsisian V, Jacobs AK, Kaul 37. Hendel RC, Corbett JR, Cullom SJ, DePuey EG, Garcia
S, Laskey WK, et al. Standardised myocardial segmenta- EV, Bateman TM. The value and practice of attenuation
tion and nomenclature for tomographic imaging of the correction for myocardial perfusion SPECT imaging: a
heart. A statement for healthcare professionals from the joint position statement from the American Society of
Cardiac Imaging Committee of the Council on Clinical Nuclear Cardiology and the Society of Nuclear Medicine.
Cardiology of the American Heart Association. J Nucl J Nucl Cardiol. 2002;9:135-43.
Cardiol. 2002;9:240-5.
108
Chapter 9
Artefacts and Pitfalls in Myocardial Imaging (SPECT, SPECT/CT and PET/CT)
Ana Geão and Carla Abreu
EANM
Myocardial perfusion imaging (MPI) is a or artefacts on the reconstructed images.
complex process, subject to a variety of arte- These artefacts typically take the form of
facts and pitfalls related to image processing, alternating concentric hot and cold rings
the patient, and technical factors. The causes and consequently give rise to defects such
and effects of these potential artefacts and as streaks in the reconstructed MIP images
pitfalls need to be understood. Furthermore, (Fig. 1) [1]. Non-uniformities in multiple de-
it is important not only to implement means tector systems do not usually produce com-
of preventing and/or limiting these factors plete rings.
and, when necessary, actions to correct
them, but also to incorporate their influence Satisfactory SPECT demands stringent qual-
into the interpretation of the study. Tech- ity assurance procedures. Field uniformity
nologists play a key role in combating these tolerances of 1% may be required instead of
artefacts and pitfalls, which may arise at any the 5% tolerance for planar imaging [2, 3].
stage in the MPI process and can lead to mis-
diagnosis.
109
the distance of the detectors to this axis of trol of an attenuation correction device can
rotation should not be variable. This requires lead to a MIP artefact such as a truncation ar-
that both the camera head and the yoke ro- tefact due to a positioning error, a cross-talk
tation be carefully set to place the plane of artefact due to the off-peak or window set-
the camera crystal parallel to the axis of rota- up of the transmission source or a low-count
tion. The computer image matrix must also density artefact due to decreased activity of
be correctly aligned with the axis of rotation. the transmission source [1].
Any misalignment in the acquired images,
whatever the cause, will mean blurring and The use of SPECT/CT has improved the at-
loss of resolution in the reconstructed im- tenuation map quality that can be obtained
ages. with sealed transmission sources. However,
the pixel values must be scaled to match
Centre of rotation (COR) is a calibration per- attenuation coefficients appropriate for the
formed regularly to ensure that the frame radionuclide being imaged and the CT and
of reference used by the computer in re- SPECT images must be precisely aligned.
constructing images is aligned with the This alignment should always be verified be-
mechanical axis of rotation of the SPECT. If cause any patient movement or even deep
an offset error exists, spatial resolution and breathing can result in a spatial mismatch
image contrast will be reduced and there and consequently introduce artefacts into
will be increased blurring of the image with the attenuation correction map. Realign-
resultant significant image artefacts, espe- ment can be done to correct some mis-
cially in the apex of MPI images [1]. Such matches. It is also important to keep in mind
an error can also induce, in transverse im- that every artefact in the CT will reverberate
ages, characteristic “doughnut” (360° orbit) into artefacts in the emission scan, such as
or “tuning fork” (180° orbit) artefacts. COR metal objects, which will cause over-correc-
errors are easy to detect with radioactive tion in the SPECT image.
point sources. However, they may be very
difficult to detect with a clinical distribution Processing-related artefacts
of activity [4, 5]. A substantial numbers of artefacts are due
to the processing phase of MPI. A proper
Attenuation correction is an important tool to orientation of the heart is vital to ensure that
distinguish artefacts due to attenuation, but the entire myocardium is included and the
attenuation correction itself can cause arte- axis angles are correct, otherwise significant
facts. First of all, the emission scan and the errors in the perfusion imaging will occur
transmission scan that are used to correct (Fig. 2). In addition, unusual patient charac-
photon attenuation must be perfectly regis- teristics such as dextrocardia may cause pit-
tered with each other. Improper quality con- falls [1, 3].
110
Chapter 9 Artefacts and Pitfalls in Myocardial Imaging (SPECT, SPECT/CT and PET/CT)
EANM
A major source of error in SPECT recon- Incorrect matching of gender, tracer and pro-
struction is data filtering. Filters that are too tocol can lead to quantification errors since
smooth or over-filtering due to the method the quantitative software may misplace the
of summation of gated files in gated SPECT contour incorrectly for perfusion as well as
may result in false negative studies for cor- functional analysis.
onary artery disease and may have a par-
ticularly adverse impact on the detection of Despite the best efforts of the technologist,
small or non-transmural defects. On the oth- artefacts may still arise in response to pro-
er hand, filters that are too coarse may result cessing-related issues. It is therefore crucial
in false positive studies [6]. for the physician to review the raw SPECT cine
display before interpreting the MPI images.
Figure 2: A: Processing images demonstrate incorrect axis alignment in the horizontal long-
axis (HLA) plane on the stress study. Proper alignment is present on the rest study. B: Incorrect
alignment results in an artefactual reversible defect in the lateral wall on perfusion images
(arrowheads). C, D: Processing images (C) and perfusion images (D) from the same study, now
with proper axis selection. Artefactual defect is no longer present. VLA, vertical long axis; SA,
short axis. (From Burrell and MacDonald [3])
111
Patient-related artefacts tion. Typically, two positions can be used, su-
Good patient preparation is mandatory to pine (most commonly used) and prone, the
ensure an optimal study and minimise the latter having been shown to reduce patient
existence of artefacts and pitfalls. A maximal motion and inferior wall attenuation when
safe level of cardiac stressing (physical or compared with supine imaging. Other posi-
pharmacological) optimises the sensitivity tions may be considered according to the
of MPI for ischaemia. In order to achieve that patient’s needs [1].
level of cardiac stressing, the patient should
avoid some substances, such as chocolate When using 180° orbits for SPECT, the pa-
and caffeine; in addition, when medically tient’s left arm must be positioned outside
required, certain cardiac medications should the field of view. This is usually accomplished
be withheld [3]. by placing the left or even both arms above
the head using an arm support device to
Before imaging, radio-opaque material or maximise comfort. When using 360° orbits,
another potential attenuator in the acquired both arms need to be positioned away from
region on the thorax must be removed. How- the patient’s side. Likewise, it is well known
ever, it is not only external material project- that the collimator surface should be as close
ing over the heart during the SPECT acqui- to the patient’s body as possible, to increase
sition that results in focal attenuation: the the image resolution. These requirements
patient’s body itself is responsible for one mean that the technician must be very aware
of the most frequent and discrete source of of the danger that the patient’s position will
artefacts in MPI – attenuation. There are gen- interfere with the gantry radius. It is impor-
erally three sources of soft tissue attenuation tant that patients be positioned similarly for
artefacts (usually fixed defects in both stress both the rest and the stress study, with a con-
and rest images): breast (anteroseptal, anteri- sistent radius of rotation [2].
or and/or anterolateral wall), diaphragm and
‘cold’ bowel (inferior wall) and obese patients Inspection of cine display projection im-
with lateral fat pads (lateral wall). One way ages or even the sinogram and linogram is
of distinguishing attenuation artefacts from one of the most useful means available to
significant perfusion abnormalities, if attenu- physicians to detect patient motion (Fig. 3).
ation correction is not available, is to perform Most processing computers provide motion
a gated study (Table 1) [1, 6]. correction software; nevertheless, motion
correction software does not always correct
Proper patient positioning entails ensur- appropriately and must be used judiciously.
ing patient comfort during the scan, which Prevention of motion is the most effective
will minimise or even avoid many artefacts way to avoid artefacts. Another type of mo-
caused by attenuation as well as patient mo- tion is cardiac or upward creep, which can
112
Chapter 9 Artefacts and Pitfalls in Myocardial Imaging (SPECT, SPECT/CT and PET/CT)
EANM
result from the patient moving or from exag- is more pronounced when post-exercise
gerated diaphragmatic excursion related to images are acquired too soon after exercise
heavy or erratic breathing. This phenomenon [1, 5].
Figure 3: A Two frames from MPI raw data demonstrate offset of the left ventricle (LV) between
frames, indicative of patient motion. B The resultant perfusion images demonstrate a defect in
the apical septum and slight relative offset of lateral and septal aspects of the LV (arrows), along
with a ‘‘tail’’ of activity extending from the LV (arrowheads) as a result of the patient motion.
(From Burrell and MacDonald [3])
Other patient-related pitfalls and artefacts an increase in lung and right ventricular up-
are due to physiological or extracardiac take of tracer as well as an increase in right
causes. Left bundle branch block can cause ventricular volume, causing a D-shaped left
septal perfusion defects, and left ventricular ventricle. Pericardial effusion creates a halo
hypertrophy can lead to underestimation of effect around the myocardium on images
left ventricular ejection fraction due to an and can cause inaccurate calculation of left
excessive blurring effect in the end-systolic ventricular ejection fraction. Another cause
phase; furthermore, markedly increased of pitfalls is non-cardiac incidental findings:
uptake in thickened myocardium can lead for example, the MIP tracer can be concen-
to the erroneous appearance of decreased trated in malignant tumours in the thorax
perfusion to the remaining area of the myo- (lung, breast tissue or lymph nodes), reduc-
cardium. A mitral valve stenosis can lead to ing the image quality [1, 4].
113
The validity of information gained from ECG-
gated SPECT depends on a regular heart rate
and appropriate synchronisation with the
ECG, and in this context proper skin prepa-
ration and placement of electrodes are re-
quired to reduce the noise added to the ECG.
The occurrence of an irregular heart rate re-
sults in missed data during image acquisition
and may lead to errors in calculation of ejec-
tion fraction and display of cardiac function.
These cause improper placement of counts Figure 4 A, B: Comparison of a normal myo-
into the individual temporal bins over the cardial perfusion SPECT sinogram (A) and the
cardiac cycle and inconsistent image recon- striped appearance characteristic of count
structions for each temporal frame. Errors in losses resulting from inadequate gating of
the ECG-gated images can be propagated the SPECT data (B). (From Wheat and Currie [6])
into the perfusion images when the tempo-
ral frames are summed to generate the per- Prominent activity is frequently present in
fusion images. subdiaphragmatic organs adjacent to the
heart. Activity is present in the liver and bow-
Theoretically, a gated SPECT study acquired el as a result of hepatobiliary excretion of
with 8 frames per cardiac cycle is more likely technetium-99m (99mTc)-labelled agents and
to be corrupted by arrhythmias than a study can be present in the stomach due to reflux
acquired with 16 frames per cycle. ECG gat- into the gastric lumen from the duodenum
ing error may be suspected from the inspec- or because of uptake of free 99mTc-pertech-
tion of movie display of planar projection netate by the gastric mucosa. Typically this
images or reconstructed slices but more eas- activity interferes with evaluation of the ad-
ily from a sinogram (Fig. 4). To minimise the jacent inferior wall; only rarely is the lateral
potential for artefacts, ECG-gated studies can wall affected, in the setting of a hiatal hernia
be acquired with an acceptance window de- (Fig. 5) [3].
fining allowable variation between R-R peaks
of the cardiac cycle [7, 8].
114
Chapter 9 Artefacts and Pitfalls in Myocardial Imaging (SPECT, SPECT/CT and PET/CT)
EANM
Activity in subdiaphragmatic organs can in- adjacent myocardium and it is not possible
terfere with evaluation of perfusion in two in any given case to know what the effect
general ways. First, it can result in apparent of this activity has been. Furthermore, both
increased activity in the adjacent inferior wall influences may coexist; as they can exist on
as a consequence of scatter and volume av- either the rest or the stress images, they may
eraging. This can mask a true defect in the result in an artefactual fixed or reversible
inferior wall or may lead to normalisation perfusion defect and the best solution is to
problems throughout the remainder of the avoid adjacent subdiaphragmatic activity al-
myocardium. On the other hand, this adja- together. However, this is not always an easy
cent ‘‘hot’’ activity can result in apparent de- task and measures to reduce the amount
creased activity in the adjacent myocardium of gastrointestinal activity are not consis-
as a result of the reconstruction algorithm tently effective. A low level of exercise may
used in filtered back-projection [7]. decrease splanchnic uptake and drinking a
large amount of fluid or even consumption
Subdiaphragmatic activity can result in ei- of a small fat meal may help to move radioac-
ther increased or decreased activity in the tivity through the gastrointestinal tract [3, 7].
Figure 5: A Anterior raw data frame from a 99mTc-sestamibi study demonstrates activity in
various subdiaphragmatic organs that can interfere with evaluation of perfusion of the inferior
wall. B Left anterior oblique raw data frame from a patient with hiatal hernia and prominent
gastric uptake (arrowheads), which can interfere with evaluation of the lateral wall. (From Burrell
and MacDonald [3])
115
Technical-related artefacts Artefacts in PET/CT
A very important technical issue that can Cardiac PET/CT has gained an important
compromise the quality of MPI is the ad- role in the evaluation of patients with
ministered patient dose in each stress and either known or suspected cardiac disease.
rest phase. Dose recommendations for MPI Adequate quality control of the imaging
are provided in guidelines and should be data is crucial to optimise clinical results. This
adjusted according to patient weight to section specifically outlines the most typical
ensure the appropriate counting statistics appearances of artefacts encountered in
and consequently a good quality study. The cardiac PET/CT, and describes some potential
importance of the correct dose increases in solutions to avoid and correct them. As with
the second phase of a same-day study, since SPECT and SPECT/CT, the most commonly
the second scan will be dominated by activ- seen artefacts are due to the patient,
ity from the first injection if insufficient dose processing and technical factors.
is provided. Nevertheless, there are many
reasons why a low-activity study may occur. Patient motion- and patient-related
Infiltration in the injection site and spilled artefacts
dose in the patient’s clothes not only provide Metallic implants, including the metallic
low counting statistics but may also lead to components of pacemaker and implant-
an erroneous diagnosis of malignancy if “hot able cardioverter defibrillator leads near the
spots” are in the field of view. The insertion of myocardium, may give rise to artefacts on
an intravenous line reduces the possibility of CT transmission images and pose an even
an infiltrated dose, although it is imperative more significant problem than in oncology.
when administering the radiopharmaceuti- Such artefacts on PET images usually result
cal to beware of potential contamination as in either over- or underestimation of the
well as to perform flushing with saline after tracer uptake, so caution is required during
the injection to reduce its retention in the in- the CT-based attenuation correction (CT-AC)
travenous line [1, 3]. procedures [9]. DiFillipo and Brunken [10]
found that the presence of the aforemen-
The various pitfalls technical and equip- tioned metal objects in the field of view of
ment-, processing- and patient-related arte- the heart can introduce artefacts in over 50%
facts in SPECT MPI are summarized in Table 2. of patients.
116
Chapter 9 Artefacts and Pitfalls in Myocardial Imaging (SPECT, SPECT/CT and PET/CT)
EANM
Patient preparation is, as in SPECT MPI, a key Image reconstruction and software
component for successful cardiac PET/CT. Several quantitative measures of myocardial
Glucose metabolism in the heart is influ- perfusion can be derived automatically by
enced by the availability of substrate, myo- software. The quantification of perfusion is
cardial workload and adequacy of myocar- relative in nature and therefore image counts
dial perfusion. Significant myocardial FDG in each study need to be normalised to a
uptake may impede detection of a lesion common level before comparison is made.
that is located either behind or closely at- The local samples of hypoperfusion can be
tached to the heart border. Blood glucose aggregated into regional (per vascular ter-
level and fasting duration before the FDG ritory) or global (per ventricle) measures in
PET scan are two well-known factors that a polar map. The visual assessment of the
influence the myocardial uptake [11]. FDG is amount of ischaemia may be challenging
decreased in normal myocardium, owing to for the observer because there can be dif-
the relatively low glucose and insulin levels ferences between rest and stress [14]. Lima
and high free fatty acid levels, as well as in et al. found that in patients with triple-vessel
diabetics, owing to the low glucose levels. disease, combined perfusion/function analy-
On the other hand, when a patient has a sis resulted in a significantly greater number
shorter period of fasting, the tracer uptake of abnormal segments per patient compared
will be high [12]. with perfusion analysis alone [15]. The fact
that PET can be acquired dynamically allows,
Patient positioning is also important in avoid- through the use of kinetic models, absolute
ing artefacts. Patients are usually imaged measures of myocardial blood flow or myocar-
supine with their arms raised above their dial metabolism. These techniques rely on a
head to prevent streak artefacts from beam dynamic acquisition of the tracer and a mod-
hardening. Also, imaging of large patients el for tracer transport from the blood into the
with the arms down may result in a tight fit cell. Use of a two-compartment model offers
within the gantry and lead to truncation ar- a way of modelling the transport from the
tefacts that appear as linear streaks through blood pool into the myocardium. Ideally, the
the reconstructed images. This issue is usu- first dynamic frame is free of any counts from
ally resolved by using full field of view recon- the radiotracer, so beginning the acquisition
struction [13]. late can lead to an underestimation of the
blood pool and thereby an overestimation of
the myocardial blood flow [3].
117
As in SPECT MPI, excessive subdiaphragmatic ble artefact that results from the difference in
activity from high counts in the liver or bowel scanning times between PET and CT images.
may result in decreased counts in the adja-
cent inferior left ventricular wall when filtered Respiratory, gross physical motion of the patient
back-projection is used for image reconstruc- and cardiac motion are the major causes of
tion. This kind of artefact is not common with misalignment leading to inconsistencies and
PET as iterative methods are generally used hence artefacts when CT-AC is used. By far the
for image reconstruction [13]. most common technique for misalignment
correction is a rigid shift of the emission da-
Attenuation correction taset over the attenuation image to achieve
Accurate attenuation correction is critical for optimal superimposition and co-registration
cardiac PET images. The transmission images of the cardiac free wall and re-reconstruction
and attenuation map should be checked for of the images using the shifted emission data
quality. [3]. The cardiac motion due to heart contrac-
tion and respiration can cause artefacts that
Artefacts may also be introduced when are difficult to compensate. Whilst diagnostic
contrast-enhanced CT data are used for PET CT is acquired during breath-holding and cor-
attenuation correction due to overestima- rected for cardiac motion by prospective ECG
tion of absorption values for PET attenuation triggering, the PET data have to be acquired
correction. Nevertheless, these artefacts can during free breathing and are normally ac-
be minimised by use of post-processing al- quired without gating [13, 16].
gorithms [16]. Misalignment of transmission
and emission images can produce an erro- A potential solution to the effects of possible
neous attenuation map that projects lung respiratory motion when using CT-AC for PET
attenuation parameters onto the heart wall, studies is the use of cine CT, in which the idea
thereby causing underestimation of the at- is to acquire multiple phases of the breathing
tenuation and creating artefactual areas of cycle and average all of the phases together.
hypoperfusion that can be misinterpreted as Another solution to correct the motion ef-
myocardial ischaemia or infarction and con- fects is gating both the CT and the PET. Four-
sequently produce false-positive results. The dimensional (4D) PET and CT imaging (re-
extent and direction of this misalignment spiratory gating) is technically feasible and is
will determine whether artefacts will be ap- realised during free normal breathing. Each
parent in the attenuation-corrected images. gate contains a small portion of the respira-
Over-correction of the inferior wall is a possi- tory cycle and correspondingly reduces the
118
Chapter 9 Artefacts and Pitfalls in Myocardial Imaging (SPECT, SPECT/CT and PET/CT)
EANM
motion present in that gate. It is noteworthy and respiratory cycles. Thus, in combination
that this method assumes that 4D PET and with 4D CT that maps cardiac and respira-
4D CT match well despite the sequential na- tory motion, each gated PET image could be
ture of PET and CT acquisitions, thereby pro- properly attenuation corrected using the ap-
viding an artefact-free image. propriate set of time-dependent CT images
(Fig. 6). However, the signal-to-noise ratio of
In the absence of gating, the PET acquisition, the resulting image is poor since most of the
as a consequence of the lengthy scan dura- acquired PET information and all other respi-
tion, unavoidably averages over many cardiac ratory PET gates are discarded [13, 17, 18].
Courtesy of PET Imaging Centre, King’s
College London
A B C
Figure 6 A–C: Attenuation correction using helical CT (A), average cine-CT (B) and gated PET and
cine CT (C). Note the difference in the anterior wall between the first image and the last two.
119
References Chapter 9
References
1. Alessi AM, Farrell MB, Grabher BJ, Hyun MC, Johnson SG, 11. Ding HJ, Shiau YC, Wang JJ, Ho ST, Kao A. The influences
Squires P. Nuclear cardiology technology study guide. of blood glucose and duration of fasting on myocardial
Reston, VA: Society of Nuclear Medicine; 2010:37–46; glucose uptake [18F]fluoro-2-deoxy-D-glucose. Nucl
99–108; 137– 143. Med Commun. 2002;23:961–5.
2. Christian PE, Waterstram-Rich KM. Nuclear medicine and 12. Lobert P, Brown RK, Dvorak RA, Corbett JR, Kazerooni
PET/CT technology and techniques. 6th ed. Missouri: EA, Wong KK. Spectrum of physiological and pathologi-
Mosby Elsevier; 2007:85–98; 267–313; 487–495. cal cardiac and pericardial uptake of FDG in oncology
PET-CT. Clin Radiol. 2013;68:e59–e71.
3. Burrell S, MacDonald A. Artifacts and pitfalls in myocardi-
al perfusion imaging; J Nucl Med Technol. 2006;34:193– 13. Di Carli MF, Lipton MJ. Cardiac PET and PET/CT imaging.
211. New York: Springer; 2007.
4. Dilsizian V, Narula J. Atlas of nuclear cardiology. 4 ed. 14. Slomka P, Xu Y, Berman D, Germano G. Quantitative
New York: Springer; 2013:1–54; analysis of perfusion studies: Strengths and pitfalls. J
Nucl Cardiol. 2013;19:338–46.
5. DePuey EG. Imaging Guidelines for Nuclear Cardiol-
ogy Procedures. A Report of The American Society 15. Lima RS, Watson DD, Goode AR, Siadaty MS, Ragosta
of Nuclear Cardiology Quality Assurance Committee. M, Beller GA, Samady H. Incremental value of combined
American Society of Nuclear Cardiology; 2006. perfusion and perfusion over perfusion alone by gated
SPECT myocardial perfusion imaging for detection of
6. Wheat J, Currie G. Recognising and dealing with artifact severe three-vessel coronary artery disease. J Am Coll
in myocardial perfusion SPECT. The Internet Journal of Cardiol. 2003;42:64–70.
Cardiovascular Research. 2006; 4(1). Available from:
http://ispub.com/IJCVR/4/1/10112# 16. Schafers KP, Stegger L. Combined imaging of molecular
function and morphology with PET/CT and SPECT/CT:
7. Wackers FJ, Bruni W, Zaret BL. Nuclear cardiology: The image fusion and motion correction. Basic Res Cardiol.
basics. How to set up and maintain a laboratory. New 2008;103:191–9.
Jersey: Humana Press Inc.; 2004:143–225.
17. Daou D. Respiratory motion handling is mandatory to
8. Heller GV, Mann A, Hendel RC. Nuclear cardiology. Tech- accomplish the high-resolution PET destiny. Eur J Nucl
nical applications. New York: McGraw–Hill; 2009:187– Med Mol Imaging. 2008;35:1961–70.
210.
18. Nehmeh SA, Erdi YE. Respiratory motion in positron
9. Case JA, Baterman TM. Taking the perfect nuclear image: emission tomography/computed tomography: A re-
Quality control, acquisition, and processing techniques view. Semin Nucl Med. 2008;38:167–76.
for cardiac SPECT, PET, and hybrid imaging. J Nucl Car-
diol. 2013;20:891–907.
120
EANM
Table 1: Biological and physiological artefacts: their causes, scan appearances and techniques
for eliminating or minimising them (from Wheat and Currie [6])
Breast attenuation Fixed anterior, anteroseptal or Examine gated cine for wall
anterolateral defects motion and wall thickening;
perform 360° reconstruction
Fat chest (not breast) – obese Fixed lateral defect Examine gated cine for wall
patient motion and wall thickening;
perform 360° reconstruction
Splenic flexure (cold) – post Inferolateral defect Perform a prone SPECT and/
barium, ascites or examine gated cine for wall
motion and wall thickening
Left bundle branch block Reversible septal or anteroseptal Dipyridamole stress; examine
defect sparing apex and anterior gated cine for wall motion and
wall wall thickening
Upward creep Reversible inferior and basal Delay scanning until 15 min
inferolateral defects and post exercise and repeat any
possibly reversible anterior study with upward creep
defects
121
Cause of artefacts Scan appearance Solution/comments
Liver activity Inferior or inferolateral defect Examine gated cine for wall
(worse on rest studies and motion and wall thickening;
pharmacological stress studies) delay scanning time post
injection; 360° reconstruction
Apical thinning Fixed apical effect Examine gated cine for wall
motion and wall thickening
122
References Chapter 9
EANM
Table 2: Summary of the various pitfalls and artefacts in SPECT MPI
123
Notes
124
Notes
EANM
125
Imprint
ISBN: 978-3-902785-09-1
Publisher:
European Association of Nuclear Medicine
c/o vereint: Association & Conference Management Ltd.
Hollandstrasse 14, 1020 Vienna, Austria
Phone: +43-(0)1-212 80 30 | Fax: +43-(0)1-212 80 30-9
Email: info@eanm.org | URL: www.eanm.org
Editors:
Ryder, Helen (Dublin)
Testanera, Giorgio (Rozzano, Milan)
Veloso Jerónimo, Vanessa (Almada)
Vidovič, Borut (Munich)
Project Management:
Katharina Leissing
Content:
No responsibility is taken for the correctness of this information.
Information as per date of printing September 2014. Lizenzn
eh
Österrei mer
ch
Umweltz isches
eichen
Green M
Association & Conference Management: & Green
eetings
Events
vereint: Association & Conference Management Ltd.
Hollandstrasse 14, 1020 Vienna, Austria vereint Ltd. is licensee of the
Phone: +43-(0)1-533 35 42 | Fax: +43-(0)1-533 35 42-19 Austrian Eco-label for
„Green Meetings and Green Events“
Email: office@vereint.com | URL: www.vereint.com
kreat
iv
Layout & Design:
kreativ · Mag. Evelyne Sacher-Toporek
Linzer Strasse 358a/1/7, 1140 Vienna, Austria
Phone: +43-(0)1-416 52 27 | Fax: +43-(0)1-416 85 26 G R a f i k & d e s i G n
Email: office@kreativ-sacher.at | URL: www.kreativ-sacher.at Mag. Evelyne Sacher-Toporek
Printing:
Colordruck GmbH
Kalkofenweg 6, 5400 Hallein, Austria
Phone: +43 (0)6245 90 111 26 | Fax: +43 (0)6245 90 111 22
Email: info@colordruck.at | URL: www.colordruck.at
126
EANM