Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Volume 5
Volume 5
Liver Cancer
Edited by
M.A. Hayat
Department of Biological Sciences,
Kean University, Union, NJ, USA
Editor
M.A. Hayat
Department of Biological Sciences
Kean University
Union, NJ, USA
springer.com
New technology, for better or for worse, will be used, as that is our nature.
Lewis Thomas
You have been given the key that opens the gates of heaven; the same key opens the
gates of hell.
Writing at the entrance to a Buddhist temple
Authors and Co-Authors of Volume 5
vii
viii Authors and Co-Authors of Volume 5
Suguru Yonezawa D. Yu
Department of Human Pathology, Field Royal Free Hospital, Liver
of Oncology, Kagoshima University Transplantation and Hepatobiliary
Graduate School of Medical and Medicine, Pond Street,
Dental Sciences, 8-35-1, Sakuragaoka, London NW3 2QG, UK
Kagoshima, 890-8544, Japan
Alex I. Zaika
Kenya Yoshida Department of Surgery and Cancer
Division of Heptobiliary and Pancreatic Biology, Vanderbilt University Medical
Medicine, Department of Internal Center, 1255 Light Hall, 2215 Garland
Medicine, Hyogo College of Medicine, Avenue, Nashville,
Mukogawa-cho, 1-Ishinomiya, Hyogo TN 37232, USA
663-8501, Japan
Yoh Zen
Junji Yoshino Division of Pathology,
Department of Internal Medicine, Second Kanazawa University Hospital,
University Hospital, Fujita Health Kanazawa, Japan
University School of Medicine, 3-6-10,
Otobashi, Nakagawa-ku, Nagoya
454-8509, Japan
Preface
xv
xvi Preface
The hepatitis B and C viruses constitute their writings, which should build and fur-
a major global health problem. It is esti- ther the endeavors of the readers. The text
mated that 180 million people, ∼3% of is divided basically into seven subheadings
the world’s population are infected with for the convenience of the readers. It is my
these viruses (WHO). Most cases result in hope that the current volume will join the
chronic hepatitis, causing liver cirrhosis and preceding volumes of this series for assist-
HCC. Spontaneous eradication of the virus, ing in the more complete understanding of
however, does occur in ∼50% of acute globally relevant cancer syndromes. There
infections, which is associated with specific exists a tremendous public demand on
immune response to the virus. Because the scientific community to address cancer
HCC is the most frequent form of liver can- prevention, diagnosis, treatment and, hope-
cer, it is discussed extensively here. fully, cures.
Biliary tract carcinomas are not very I am grateful to the contributors for their
common tumors, but middle-aged or eld- promptness in accepting my suggestions,
erly people have poor prognosis. Although and I appreciate their dedication and dil-
palliative chemotherapy is commonly used igent work in sharing their invaluable
for these patients, there is no standard knowledge with the public through this
nor very effective therapy for advanced series. Each chapter provides unique indi-
disease. Commonly used therapy includes vidual, practical knowledge based on the
gemcitabine, mitomycin, fluoropyrimi- expertise and experience of the authors.
dines, and platinum compounds, used alone The chapters contain the most up-to-date
or in combination. Response rates with practical and theoretical information. It is
these treatments range from 10–35%, and my hope that these handbooks will assist
median overall survival times vary between in the more complete understanding of
5 and 12 months. Clinicopathological fea- at least come of the globally-encountered
tures of biliary cystic tumors are discussed cancer problems.
in this volume. The role of Mucin 4 as a I am thankful to Dr. Dawood Farahi
prognostic factor for extrahepatic bile duct and Dr. Kristie Reilly for recognizing
carcinoma is also presented. the importance of scholarship in an
As in the previous volumes of this series, institution of higher education and for
each chapter is written by distinguished, providing the resources for completing
practicing clinicians/surgeons/pathologists this project. I appreciate receiving expert
who provide detailed methodologies for help from Myrna Ortiz, Erin McNally
diagnosis and treatment of various forms and Lina Builes in preparing this volume.
of liver cancer. This volume was written
by 78 oncologists representing 16 coun- M.A. Hayat
tries. Their practical experience highlights November 2008
Introduction
M.A. Hayat
The enormous burden of liver cancer on society becomes clear by considering the fact
that approximately 625,000 new cases of this cancer are diagnosed globally each year.
Distressingly, the number of deaths is approximately the same at 598,000 per year. Liver
cancer, therefore, is the third most common cause of death from cancer. Survival rates
for liver cancer are only 3–5% globally. In the United States, 19,160 new cases of liver
cancer and 16,780 deaths were reported for 2007. The major risk factors for this cancer
include prior infection with hepatitis B and C viruses, with the former more prevalent.
Dietary exposure to fungus Aspergillus fumigatus (aflatoxins) also contributes to the
incidence of liver cancer in many parts of the world. Tobacco use is the most serious
preventable cause of cancer, as its use causes cancer of the lung, throat, mouth, liver,
pancreas, urinary bladder, stomach, kidney, as well as other types. Alcohol-induced liver
injury is another major risk factor for hepatocellular carcinoma (HCC).
In view of these devastating statistics, the urgency of deciphering the molecular mech-
anism underlying this disease, perfecting reliable diagnostic methods, understanding
risk factors, developing effective targeted drugs, improving other treatments, assessing
the effectiveness of therapies, and providing improved care for post-treatment patients,
becomes apparent. This volume provides up-to-date information on the above-mentioned
aspects of liver cancer; specifically, details of the methodologies used are included. The
other seven volumes in this series provide similar information on other types of can-
cers.
This series of handbooks has taken the unique approach of discussing cancer diagno-
sis, treatment, and prognosis in the same volume. It is pointed out that this vast subject
cannot be fully discussed by only one author. This is the primary reason for inviting a
large number of oncologists/clinicians/surgeons to write each of the eight volumes of
this series of handbooks. Another advantage of involving more than one author is to
present different points of view on a specific controversial aspect of cancer. I hope these
goals were accomplished in this and other published volumes of this series.
xvii
Contents of Volumes 1, 2, 3, and 4
Volume 1
1. Breast Cancer: An Introduction
xix
xx Contents of Volumes 1, 2, 3, and 4
34. Radiation Therapy for Older Women with Early Breast Cancer
Volume 2
Part I General Methods and Overviews
Volume 3
Index
Contents of Volumes 1, 2, 3, and 4 xxvii
Volume 4
Part I Colorectal Cancer
16. Detection of Tumor Cells in Lymph Nodes of Colon Cancer Patients Using
Real-Time Quantitative Reverse Transcription-Polymerase Chain Reaction
Preface ..................................................................................................................... xv
xxix
xxx Contents
B. Treatment
6. Ultrasonography During Liver Surgery ....................................................... 63
Guido Torzilli
Introduction ................................................................................................... 63
Technical Aspects ..................................................................................... 63
Liver Exploration ...................................................................................... 64
Resection Guidance ...................................................................................... 69
Hepatocellular Carcinoma ........................................................................ 69
Liver Metastases ....................................................................................... 73
References ..................................................................................................... 74
B. Prognosis
9. Unresectable Liver Metastases from Colorectal Cancer:
Methodology and Prognosis with Radiofrequency Ablation ...................... 117
Junji Machi
Introduction ................................................................................................... 117
Methodology of Radiofrequency Ablation ................................................... 118
Prognosis Using Radiofrequency Ablation: Our Study ................................ 120
Patients and Methods ................................................................................ 120
Contents xxxiii
B. Treatment
21. Treatment of Hepatocellular Carcinoma
with Thalidomide: Assessment with Power Doppler Ultrasound............. 277
Chiun Hsu, Chiung-Nien Chen, and Ann-Lii Cheng
Summary ..................................................................................................... 277
Introduction ................................................................................................. 277
Evaluation of Tumor Vascularity Using Power
Doppler Sonography ............................................................................... 279
Evaluation of Vascular Response of HCC to Thalidomide
by Power Doppler Ultrasound: A Prospective Study .............................. 279
Future Perspectives of Imaging Studies for Evaluation
of Anti-angiogenesis Therapy ................................................................. 282
References ................................................................................................... 284
C. Prognosis
24. Hepatocellular Carcinoma: Overexpression
of Homeoprotein Six 1 as a Marker for Predicting Survival ..................... 313
Kevin Tak-Pan Ng and Kwan Man
Introduction ................................................................................................. 313
Materials and Methods................................................................................ 314
Cell Lines ................................................................................................ 314
Clinical Samples ..................................................................................... 314
Contents xxxix
Part V Metastases
A. Diagnosis
28. Liver Metastases from Colorectal Cancer:
Ultrasound Imaging ...................................................................................... 355
Søren R. Rafaelsen
Introduction ................................................................................................. 355
Ultrasound Scanning Technique ................................................................. 356
Advantages of Diagnostic Hepatic Ultrasound ........................................... 357
Grayscale Echo Pattern ............................................................................... 357
Detection of Liver Metastases..................................................................... 358
Doppler Flow Pattern .................................................................................. 359
Ultrasound in Postoperative Follow-Up...................................................... 363
Future Potential Advancements .................................................................. 363
References ................................................................................................... 364
Contents xli
B. Prognosis
33. Extrahepatic Bile Duct Carcinoma: Role of the p53 Protein Family ....... 441
Alexander I. Zaika and Seung-Mo Hong
Introduction ................................................................................................. 441
Epidemiology .......................................................................................... 441
Etiology ................................................................................................... 441
Diagnosis................................................................................................. 442
Pathology ................................................................................................ 442
Prognosis ................................................................................................. 443
Contents xliii
C. Treatment
35. Hilar Cholangiocarcinoma: Photodynamic
Therapy and Stenting ................................................................................... 463
Marcus Wiedmann, Joachim Mössner, and Helmut Witzigmann
Summary ..................................................................................................... 463
Introduction ................................................................................................. 463
Preoperative Biliary Drainage ..................................................................... 464
Palliative Bile Duct Stenting ....................................................................... 464
Palliative Stenting of Malignant Duodenal Obstruction ............................. 469
Photodynamic Therapy (PDT) .................................................................... 470
Mechanism of Action.............................................................................. 470
Contraindications for Photodynamic Therapy ........................................ 471
Photosensitizers....................................................................................... 472
Photodynamic Therapy for Palliation
of Hilar Cholangiocarcinoma .............................................................. 472
Photodynamic Therapy for Neoadjuvant
and Adjuvant Treatment of Hilar Cholangiocarcinoma ...................... 478
Future Directions of Photodynamic Therapy
for Hilar Cholangiocarcinoma ............................................................ 478
References ................................................................................................... 479
xliv Contents
5
6 A.H. Khandani
of glucose-6-phosphatase is high, which the base of the skull through the upper
causes rapid clearance of FDG from the thighs. The CT portion (transmission scan)
liver. This may account for the mild inten- of the PET CT is acquired within seconds
sity of the normal liver on FDG-PET. while the PET acquisition time for each
Furthermore, the interindividual as well as bed position (~ 15 cm) is several minutes;
the intraindividual variation of the hepatic the total PET acquisition time in newer
uptake appears to be minimal (Rydberg machines is 10–25 min.
et al., 2005). Therefore, while comparing In addition to delivering anatomic infor-
PET scans of different patients or of the mation, the CT portion of PET CT is used
same patient at different occasions, liver to measure the attenuation of the X-ray
uptake can be used as the reference; first, photons traveling through the patient in
the PET scans to be compared should be order to produce the so-called “attenuation
displayed with the liver uptake on all the map” and correct the PET data for tissue
scans being the same, then the areas of attenuation (attenuation corrected PET).
interest can be compared to each other. During PET acquisition, photons from
The radioisotope portion of the molecule structures deep in the abdomen or pelvis
used in PET imaging emits a positron (a are attenuated more compared to those
positively charged electron), which travels from the superficial structures and the
a distance of a few millimeters in tissue chest. Therefore, the intensity of uptake
before it collides with a negatively charged in the deeper structures may be underesti-
electron. This collision annihilates the mated on the “non-attenuation corrected”
entire mass of the positron and electron into PET images. The intensity of uptake in
two photons (gamma rays) with energy of the deeper structures is normalized to the
511 KeV each. These two photons travel at intensity of uptake in the superficial struc-
the speed of light in exactly opposite direc- tures on the attenuation-corrected PET
tions (180° apart). Coincidental detection images. While correction of the PET data
of these two photons by two oppositely posi- for tissue attenuation is indispensable,
tioned detectors in the PET scanner results misalignment between PET and CT can
in images with a much higher resolution cause mislocalization of lesions on the
compared to the conventional, single pho- fused PET-CT images. This may be due
ton nuclear medicine studies and presents to changes in position of a body part (e.g.,
the possibility of quantitative measurement neck or legs) or due to physiologic changes
of the tracer uptake in a lesion of interest. in the position of an organ (i.e., respiratory
PET CT allows the fusion of the metabolic movement) between the transmission (CT)
information on PET with the anatomic and the emission scan (non-attenuation
information on CT. PET CT has been corrected PET). As the degree of misalign-
proven to increase the diagnostic accuracy ment and resulting mislocalization can be
compare to stand-alone PET. In PET CT, significant, one has to be cautious when
the patient undergoes a CT scan, followed interpreting the attenuation-corrected PET
by a PET scan without changing the patient CT images or using PET CT images for
position. PET for most oncologic indica- radiation therapy planning. The magnitude
tions (whole body PET) is acquired from of this misalignment can be assessed by
1. Applications of Positron Emission Tomography in Liver Imaging: An Overview 7
a b c
Figure 1.1. Patient with a history of ovarian cancer and rising ca-125 level. T2 (A) and postcontrast
fat-saturated T1 (B) MRI images were without any abnormal findings. PET was requested for further
evaluation and demonstrated focal uptake in the segment IV of the right hepatic lobe consistent with
metastatic disease (C, arrow head)
1. Applications of Positron Emission Tomography in Liver Imaging: An Overview 9
False-negative PET for hepatic metas- ment of the liver during the emission
tases can be due to limitations of spatial scan. The liver is an upper abdominal
resolution. It should be considered, how- organ that moves with respiratory move-
ever, that many of the PET publications ment of the diaphragm. Emission scans
were based on non-attenuation corrected are acquired over several minutes during
images, which may have lower sensitivity, which hepatic lesions are in a repetitive
especially for lesions deeper in the liver; cranio-caudal pendulous movement. The
such lesions generally appear much less respiratory excursion of the liver is
intense on non-attenuation corrected than ~ 10–30 mm. Therefore, it is conceivable
on attenuation corrected PET images. that cranial and caudal portions of small
Generally, the role of PET in detecting lesions are registered only during a portion
subcentimeter lesions should be redefined, of the acquisition time, and their uptake is
considering the higher spatial resolution underestimated, so that they appear less
and sophisticated image correction and intense on images than they really are.
reconstruction algorithms of current PET The degree of this underestimation is vari-
machines. Further improvements in image able, and particularly in the case of a sub-
resolution from currently 0.6–0.8 cm to a centimeter lesion, this may even lead to
few millimeters can be expected with the non-visualization of the lesion. One way
development of small surface area crystal to overcome this problem is to increase
elements in combination with alternative the target-to-background count ratio by
position-sensitive photomultiplier tubes, increasing the acquisition time. This can
and the implementation of depth of inter- be done by increasing the acquisition time
action measurements. in the mid- and upper-abdomen while
False-negative FDG-PET can occur also acquiring the whole body scan, if this is
due to underestimation of uptake, mislo- possible on the PET machine. Another
calization of foci as well as recent comple- solution would be to acquire a second
tion of chemotherapy. The latter is likely PET scan from mid- and upper-abdomen,
associated with significant decrease in the consisting of a transmission scan followed
number of the cells or their glucose metab- by an emission scan with longer acquisi-
olism after chemotherapy. No definitive tion time, once the whole body scan is
information is available in the literature on completed (“Liver View”). A different and
the time interval after completion of chem- preferable approach to solve this problem
otherapy during which PET can be false- is Respiratory Gating, in which case only
negative. Based on experience at large emission data collected in certain parts of
centers, this time interval is ~ 4–6 weeks the respiratory cycle are used for image
for high-grade lymphoma. However, no reconstruction, resulting in better visuali-
definite statement can be made regard- zation of small lesions with the disadvan-
ing the time interval after chemotherapy tage of longer acquisition time.
during which liver metastases may not be A changed position of the liver between
visualized on PET. the emission and transmission scan can cause
Underestimation of uptake of malignant misalignment between these two images,
lesions causing false-negative findings on resulting in mislocalization of pulmonary
PET can occur due to physiologic move- or colonic foci into the liver or vice versa.
10 A.H. Khandani
HEPATOCELLULAR
CARCINOMA
Hepatocellular carcinoma (HCC) or
hepatoma develops through malignant Figure 1.2. Patient status post laparoscopic chole-
transformation of hepatocytes and is com- cystectomy with T2 gallbladder cancer. PET was
mon in the setting of chronic liver changes requested for staging and indicated metastases in
and cirrhosis. This cancer is the most the porta hepatis, paraaortic (arrow heads) and par-
atracheal lymph nodes (arrow)
common primary epithelial malignancy of
the liver, accounting for ~ 80% of malig-
nant epithelial neoplasms of the liver. Facilitative glucose transporters do not
Hepatocellular carcinoma has been showing appear to be overexpressed in HCC as often
an increasing incidence in Western coun- as in other malignant tumors. Zimmerman
tries due to increasing frequency of hepa- et al. (2002b) and Roh et al. (2004) studied
titis B and C viral infection. The diagnosis expression of Glut1 in 35 and 22 cases of
is based on screening risk populations with HCC, respectively; the first group reported
measurements of serum AFP and liver two positive cases and the second group
ultrasound. Magnetic resonance imaging, reported one positive case. Delbeke et al.
CT, and lipiodol angiography with follow- (1998) examined a series of 23 patients
up CT are used in inconclusive cases to with HCC. In visual assessment, differ-
establish the diagnosis. Biopsy is only entiation of tumor from normal liver was
performed on patients in whom the radio- definitive in 13 patients, equivocal (mildly
logical diagnosis cannot be made. increased compared to normal liver) in
1. Applications of Positron Emission Tomography in Liver Imaging: An Overview 13
3 patients, and poor (same or less intense experience, whole body FDG-PET is very
compared to normal liver) in 7 patients. helpful to assess the malignant poten-
The sensitivity of FDG-PET for HCC is tial of a hepatic lesion of unknown pri-
~ 50–60%. There appears to be some asso- mary because FDG avidity of the lesion
ciation between histologic differentiation increases the likelihood of its malignant
of HCC and FDG uptake, with poorly dif- nature. Furthermore, simultaneous visu-
ferentiated tumors being more intense on alization of FDG avid foci outside of the
FDG-PET, likely explained by enzymology liver can be viewed as almost consistent
of HCC. Concentration of glucose-6- with malignant etiologyof the liver lesion.
phospatase is high in normal liver, causing Positron emission tomography may also
rapid clearance of glucose-6-phospate or visualize an easily accessible biopsy site
FDG-6-phosphate from hepatocytes with outside of the liver. Also in cases of known
consequent mildly intense appearance of HCC with clinically suspected extrahe-
liver on PET. The enzymology of well- patic metastasis, FDG-PET can be used
differentiated HCC resembles that of the to confirm this suspicion. Nonetheless, a
normal liver, possibly explaining mild FDG negative FDG-PET scan in such patients
uptake or non-visualization of these tumors does not exclude the possibility of HCC.
on PET. Less differentiated HCC tumors [1-11C] acetate (11-C Acetate) is a
have lower levels of glucose-6-phosphatase short-lived PET tracer with higher uptake
and higher levels of hexokinase, possi- in well-differentiated hepatomas compared
bly causing intense FDG uptake of these to poorly differentiated ones. The physical
tumors on PET (Torizuka et al., 1995). half-life of 11-C is 20 min. Several possible
Also, there appears to be some association pathways of acetate incorporation into
between FDG uptake and tumor-volume tumor metabolism have been postulated;
doubling time as well as between FDG the dominant one appears to be partici-
uptake and tumor size. Trojan et al. (1999) pation in free fatty acid synthesis. Ho
visualized all six tumors larger than 5 cm et al. (2003) studied a series of 57 patients
in a series of 14 HCC tumors. Therefore, with liver lesions, consisting of 39 cases
FDG PET could possibly be used to assess of HCC (group I), 13 cases of non-HCC
the effect of treatment in larger and less malignancies (three cases of cholangi-
differen-tiated HCC. High FDG uptake in ocarcinomas and ten cases of hepatic
HCC has been associated with overexpres- metastases from colon, breast, lung and
sion of mRNA levels for several markers, carcinoid primaries, group II) as well as
such as VEGF of aggressive tumor behavior 5 cases of benign liver lesions (2 cases of
(Lee et al., 2004). cavernous hemangioma, 2 cases of FNH
Detection of extrahepatic FDG avid and 1 case of adenoma, group III). In group
metastases originating from HCC has also I, there was a “complementing” sensitivity
been reported; especially in cases of less- for HCC on FDG and 11-C Acetate-PET,
differentiated HCC, metastases appear as these two tracers together gave a sensi-
to be more FDG avid (Ho et al., 2003; tivity of 100% for detecting HCC. Most of
Trojan et al., 1999). Although more data the well-differentiated HCC lesions were
are needed to establish the clinical role intense on 11-C Acetate scan with mild
of FDG-PET in HCC, based on author’s tracer uptake or not visualized on FDG
14 A.H. Khandani
scan, while this relationship was reversedproven to be more accurate than CT and
for poorly differentiated tumors. In groupMRI in assessing the effect of chemo-
II, all lesions showed intense FDG uptake therapy. In many malignancies such as
but no abnormal 11-C Acetate uptake. In high-grade lymphoma, early decrease of
group III, the uptake patterns were as fol-
uptake after the start of chemotherapy
lows: hemangiomas were isointense on indicates better prognosis and longer sur-
FDG and hypointense on 11-C Acetate vival. The role of FDG-PET in assessing
scan. Both FNH cases showed mild inten- the efficacy of chemotherapy in metastatic
sity uptake on 11-C Acetate scan, while liver disease, cholangiocarcinoma, and
one of them had mild and another one no other FDG avid hepatic malignancies has
visualized uptake on FDG scan. The only not yet been established. While increased
adenoma case showed no abnormal FDG intensity or number of liver foci under
or 11-C Acetate uptake. With an appar- ongoing chemotherapy indicates progres-
ently high specificity of 11-C Acetate forsion of disease, the pathologic and clini-
HCC, the authors concluded that when a cal correlates of decreased uptake in this
lesion is positive for both tracers or only
setting are not clearly understood. One
for 11-C Acetate, the likelihood of HCC has also to consider that in most cases of
is very high. On the other hand, when a metastatic disease to the liver and cholan-
liver lesion is positive only for FDG but giocarcinoma, chemotherapy has only a
is negative for 11-C Acetate, the possi- palliative effect and, therefore, although
bility of non-HCC malignancy or poorly disappearance of tumor on PET may indi-
differentiated HCC should be considered. cate effectiveness of the treatment and
In case both tracers are negative, benign better prognosis, it should not be under-
pathology is very likely. These conclu- stood as a sign of cure.
sions can be particularly helpful for eval- Due to the general assumption that tis-
uation of tumors < 2 cm in patients with sue injury causes inflammatory uptake,
low or intermediate likelihood of having FDG-PET has never been used to monitor
HCC (negative serum status for hepatitis the effect of invasive procedures such as
B or C, borderline or normal AFP). It radiofrequency ablation (RFA) of the liver.
must be considered that these very prom- However, there is no significant accumula-
ising data have not been reproduced yet tion of reactive cells at the ablation site in
in a large series and also that the tumorsthe liver after RFA (Tsuda et al., 2003)
in Asia, where this study was performed, and RFA of the liver does not appear to
may differ somewhat from the tumors cause false-positive findings on PET. This
seen in the US. is of great importance since capillaries
are leaky in the weeks and months after
RFA and, therefore, CT and MRI can-
THERAPY MONITORING not be used for several months after this
procedure to assess for residual tumor.
Effective chemotherapy is associated with In author’s experience, PET can be used
a decrease in the number and/or glucose to demonstrate complete ablation within
metabolism of cancer cells and decreased hours to days after RFA of liver metastases
FDG uptake on PET. FDG-PET has been (Figure 1.3).
1. Applications of Positron Emission Tomography in Liver Imaging: An Overview 15
a b c
Figure 1.3. Patient with solitary metastasis to the right lobe of the liver from an ocular malignant
melanoma, diagnosed on an outside PET scan (a, arrow head). PET was performed at our institution 16 h
after laparoscopic RFA and revealed complete ablation without any inflammatory uptake (b, arrow head).
This was confirmed on the 3-month follow-up PET scan (c, arrow head)
extrahepatic bile duct cancer. Eur. J. Nucl. Med. Rydberg, J., Khandani, A.H., Ivanovic, M., and
Mol. Imaging 29: 1047–1054. McCartney, W.H. 2005. Comparison of Liver,
Kim, Y.J., Yun, M., Lee, W.J., Kim, K.S., and Lee, Mediastinum, and Muscle as References for
J.D. 2003. Usefulness of 18F-FDG PET in intra- Background Activity in 18-F FDG PET-CT
hepatic cholangiocarcinoma. Eur. J. Nucl. Med. Imaging. RSNA Annual Meeting: Abstract # LPL
Mol. Imaging 30: 1467–1472. 15-04-P.
Kim, Y.W., Park, Y.K., Yoon, T.Y., and Lee, Torizuka, T., Tamaki, N., Inokuma, T., Magata, Y.,
S.M. 2002. Expression of the GLUT1 glu- Sasayama, S., Yonekura, Y., Tanaka, A.,
cose transporter in gallbladder carcinomas. Yamaoka, Y., Yamamoto, K., and Konishi, J.
Hepatogastroenterology 49: 907–911. 1995. In vivo assessment of glucose metabolism
Kinkel, K., Lu, Y., Both, M., Warren, R.S., and in hepatocellular carcinoma with FDG-PET.
Thoeni, R.F. 2002. Detection of hepatic metas- J. Nucl. Med. 36: 1811–1817.
tases from cancers of the gastrointestinal tract Trojan, J., Schroeder, O., Raedle, J., Baum, R.P.,
by using noninvasive imaging methods (US, CT, Herrmann, G., Jacobi, V., and Zeuzem, S. 1999.
MR imaging, PET): a meta-analysis. Radiology Fluorine-18 FDG positron emission tomography
224: 748–756. for imaging of Hepatocellular carcinoma. Am. J.
Koh, T., Taniguchi, H., Yamaguchi, A., Kunishima, S., Gastroenterol. 94: 3314–3319.
and Yamagishi, H. 2003. Differential diagnosis Tsuda, M., Rikimaru, H., Majima, K., Yamada, T.,
of gallbladder cancer using positron emission Saito, H., Ishibashi, T., Takahashi, S., Miyachi,
tomography with fluorine-18-labeled fluoro- H., Endoh, M., and Yamada, S. 2003. Time-
deoxyglucose (FDG-PET). J. Surg. Oncol. 84: related changes of radiofrequency ablation lesion
74–81. in the normal rabbit liver: findings of magnetic
Lee, J.D., Yun, M., Lee, J.M., Choi, Y., Choi, Y.H., resonance imaging and histopathology. Invest.
Kim, J.S., Kim, S.J., Kim, K.S., Yang, W.I., Radiol. 38: 525–531.
Park, Y.N., Han, K.H., Lee, W.J., Yoo, N., Lim, Yang, M., Martin, D.R., Karabulut, N., and Frick,
S.M., and Park, J.H. 2004. Analysis of gene M.P. 2003. Comparison of MR and PET imaging
expression profiles of hepatocellular carcinomas for the evaluation of liver metastases. J. Magn.
with regard to 18F-fluorodeoxyglucose uptake Reson. Imaging 17: 343–349.
pattern on positron emission tomography. Eur. J. Zimmerman, R.L., Burke, M., Young, N.A.,
Nucl. Med. Mol. Imaging 3: 1621–1630. Solomides, C.C., and Bibbo, M. 2002a.
Marom, E.M., McAdams, H.P., Erasmus, J.J., Diagnostic utility of Glut-1 and CA 15-3 in
Goodman, P.C., Culhane, D.K., Coleman, R.E., discriminating adenocarcinoma from hepatocel-
Herndon, J.E., and Patz, E.F. Jr. 1999. Staging lular carcinoma in liver tumors biopsied by fine-
non-small cell lung cancer with whole-body needle aspiration. Cancer 96: 53–57.
PET. Radiology 212: 803–809. Zimmerman, R.L., Fogt, F., Burke, M., and
Roh, M.S., Jeong, J.S., Kim, Y.H., Kim, M.C., and Murakata, L.A. 2002b. Assessment of Glut1
Hong, S.H. 2004. Diagnostic utility of GLUT1 expression in cholangiocarcinoma, benign
in the differential diagnosis of liver carcinomas. biliary lesions and hepatocellular carcinoma.
Hepatogastroenterology 5: 1315–1318. Oncol. Rep. 9: 689–692.
2
Localized Fibrous Tumor
of the Liver: Imaging Features
Thomas Moser and Tereza S. Nogueira
17
18 T. Moser and T.S. Nogueira
In their observations, Fuksbrumer et al. portal phases and became marked on both
(2000) depicted large solitary lesions dem- CT and MRI delayed images. This large
onstrating early contrast uptake and het- tumor also showed multiple areas of cystic
erogeneous enhancement with areas of degeneration (Figure 2.1).
differential uptake and washout. These These two separate imaging patterns
findings suggestive of a hypervascu- are well correlated with histological find-
lar tumor are also consistent with most ings. Localized fibrous tumors are com-
descriptions of pleural LFTs. In the series posed of spindle cells intermixed with
of the Armed Force Institute of Pathology, collagen bundles. Most commonly, tumor
heterogeneous contrast enhancement was cells are disposed haphazardly in an
observed in 62% of cases (Rosado-de- arrangement known as the storiform pat-
Christenson et al., 2003). Conversely, we tern. Alternating hyper- and hypocellular
recently reported the case of a LFT of the areas are present in the same tumor.
liver presenting with delayed enhancement Prominent vessels are generally associated
pattern (Moser et al., 2005). After intrave- with the cellular proliferation, determining
nous contrast administration, a progressive the hemangiopericytoma-like vascular pat-
enhancement began during arterial and tern. Such an arrangement was observed
a b c
d e f
Figure 2.1. Hepatic LFT revealed by a hypoglycemic coma in a 73 year old woman. Transverse CT
images obtained before (a), 30 s (b), 1 min (c) and 5 min (d) after contrast injection. A well-demarcated,
hypodense lesion containing numerous prominent cystic areas is shown. The lesion enhances faintly at
arterial and portal phases. However, a marked contrast enhancement is seen 5 min after injection account-
ing for the fibrous content of the lesion. Delayed sagittal MR image obtained 8 min after contrast injection
also confirms the delayed enhancement pattern (e). View of the resected specimen (f). (Reprinted with
permission from the American Journal of Roentgenology.)
2. Localized Fibrous Tumor of the Liver: Imaging Features 19
in all nine cases reported by Moran et al. nosis of other rare mesenchymal hepatic
(1998) as well as in those cases reported by tumors, particularly fibrous tissue contain-
Fuksbrumer et al. (2000). In fact, it repre- ing sarcomas such as fibrosarcoma and
sents the most common microscopic pattern malignant fibrous histiocytoma. Hepatic
described so far and corresponds to the most fibrosarcoma also manifests as a large
common hypervascular pattern of contrast mass and possible hypoglycemia. It con-
uptake where heterogeneous enhancement tains an abundant collagenous stroma with
seems to result from the varying cellularity spindle cells, which are in this case char-
observed throughout the tumor. However, acteristically arranged in a herringbone
the hemangiopericytoma-like vascular pat- pattern. Moreover, immunological labe-
tern was absent in our study. Rather, it was ling for CD34 is negative.
a poorly vascularized tumor with an inter- To summarize, imaging can only sug-
stitium almost exclusively composed of gest the diagnosis of LFT in rare cases,
collagen fibers. Therefore, its enhancement but thorough analysis of vascularity often
dynamic at CT and MR was consistent with brings clues about histological content.
a collagen rich interstitium where contrast
gradually diffuses between the fibers. REFERENCES
Most hepatic LFTs are benign lesions.
However, a malignant variant with a more Chan, J.K. 1997. Solitary fibrous tumour–every-
where, and a diagnosis in vogue. Histopathology
aggressive behavior such as recurrence 31:568–576.
and distant metastasis has been described. Fuksbrumer, M.S., Klimstra, D., and Panicek, D.M.
Malignancy can be suspected solely in 2000. Solitary fibrous tumor of the liver: imaging
the presence of numerous cellular atypias findings. Am. J. Roentgenol. 175:1683–1687.
on histology, or with the identification of Moran, C.A., Ishak, K.G., and Goodman, Z.D.
distant metastases (Chan, 1997). Thus, 1998. Solitary fibrous tumor of the liver: a clin-
icopathologic and immunohistochemical study
from the imaging viewpoint, LFTs should of nine cases. Ann. Diagn. Pathol. 2:19–24.
be included in the differential diagnosis Moser, T., Nogueira, T.S., Neuville, A., Riehm, S.,
of any extensive and well limited hepatic Averous, G., Weber, J.C., and Veillon, F. 2005.
mass. A delayed enhancement pattern has Delayed enhancement pattern in a localized
been demonstrated for a wide range of fibrous tumor of the liver. Am. J. Roentgenol.
fibrous containing hepatic lesions, such 184:1578–1580.
Rosado-de-Christenson, M.L., Abbott, G.F.,
as fibrolamellar hepatocellular carcinoma, McAdams, H.P., Franks, T.J., and Galvin, J.R.
cholangiocarcinoma, or sclerosing heman- 2003. From the archives of the AFIP: local-
gioma. Localized fibrous tumor should ized fibrous tumor of the pleura. Radiographics
also be included in the differential diag- 23:759–783.
3
A Radial Magnetic Resonance Imaging
Method for Imaging Abdominal
Neoplasms
Maria I. Altbach
21
22 M.I. Altbach
techniques and imaging parameters was 100% accuracy. The SSEPI T2 measure-
investigated using qualitative approaches ment used data at five TE time points to
(i.e., observers visually comparing signal improve the accuracy of the T2 estimate,
intensities between a moderate and heavily but required the acquistion of data over
T2-weighted images) as well as quantita- several breath hold periods. This made
tive approaches (i.e., observers measuring the experiment both time-consuming and
signal intensity and/or CNR of lesions dependent on the ability of the subject
relative to liver). Although reports indi- to perform reproducible breath holds in
cated that the quantitative measurements order to avoid errors due to misregistra-
yielded a better characterization of liver tion between TE data sets. The low spatial
lesions, these measurements require resolution of SSEPI and its sensitivity to
significant operator interaction and are susceptibility artifacts limit the technique
not practical for clinical use (Ito et al., even further.
1997). Consequently, in current clinic Abe et al. (2000) used SSFSE to cal-
practice, lesion characerization based on culate T2 at two TE time points within a
T2-weighted imaging is performed by breath hold, but showed significant overlap
visually comparing the signal intensity between malignancies and benign lesions.
of lesions between moderate and heavily Olcott et al. (1999) measured T2 values in
T2-weighted images. the liver with a pulse sequence designed
Several studies have demonstrated that to acquire data from four TE time points
a better quantitative approach for liver with reduced errors due to radiofrequency
lesion charaterization is to directly meas- pulse imperfections. Although they dem-
ure the T2 value by acquiring data at two onstrated a good separation between
or more echo time (TE) points and fitting malignant and benign lesions for a set
the signal intensities to a single exponen- of 24 lesions, their method is also time-
tial decay. Initially, T2 measurements in consuming. Two recent studies based dual-
the upper abdomen were mainly obtained echo respiratory-trigerred 2DFT-FSE, a
with dual-echo 2DFT-SE methods with technique that is also sensitive to motion,
the drawback of long acquisition times reported that the accuracy of separating
(10–20 min) and errors due to motion malignancies from hemangiomas based
(i.e., signal displacement along the phase- on T2 is 92–94% (Cieszanowski et al.,
encoding direction and motion-induced 2002; Kim et al., 2005). In summary,
volume averaging). Thus, although one of the current methods for measuring T2
these studies reported a 97% accuracy for values in the body have at least one of
discriminating malignancies from heman- the following problems: long imaging
giomas (McFarland et al., 1994), the others times, motion-induced errors, misregistra-
show significant overlap between these two tion of images acquired in different breath
type of lesions (Ohtomo et al., 1988). To holds, low spatial resolution, and/or low
reduce the effect of motion in T2 measure- number of measured points on the relaxa-
ments, Goldberg et al. (1993) used SSEPI tion curve. These technical difficulties not
to measure T2 in 46 liver neoplasms and only reduce the accuracy of these methods
show that hemangiomas and cysts can but also make them impractical for routine
be discriminated from malignancies with clinical use.
24 M.I. Altbach
Figure 3.1. A comparison of images acquired with the conventional 2DFT-FSE and the radial FSE method.
(a) Images acquired during free breathing using respiratory triggering. The two metastatic lesions (arrows)
seen in the radial FSE image are masked by flow artifacts in the 2DFT-FSE image. (b) Images acquired in
a breath hold. The image acquired with 2DFT-FSE has ghosting artifacts caused by flow. The signal from
flow is well suppressed in the radial FSE image by using a small amount of diffusion weighting (b = 1.2 s/
mm2), so the metastatic lesions (arrow heads) are clearly seen. Note that the small lesion in the lower right
portion of the liver is only seen in the radial FSE image
in the SSFSE image. An example showing These qualitative findings were cor-
the advantage of flow suppression via dif- roborated by a statistical evaluation of
fusion weighting in radial FSE is shown in image quality for a set of 16 patients with
Figure 3.2c for a patient with very small various types of liver neoplasms (Altbach
metastatic lesions (as reference the lesion et al., 2002). In the study, images were
highlighted by the arrow is 5 mm in diam- acquired with radial FSE and the conven-
eter). In the SSFSE image the signal from tional 2DFT-FSE and SSFSE methods.
lesions cannot be distinguished from the The results of the study indicated that
signal from blood vessels. This is not the radial FSE performed significantly better
case in the radial FSE image where the in terms of organ and lesion definition
signal from flow is suppressed leaving the than SSFSE or 2DFT-FSE, In particular,
bright signal from lesions present. it was demostrated that radial FSE was
26 M.I. Altbach
Figure 3.2. Comparison of images acquired with the conventional SSFSE and the radial FSE method. (a)
Images of a subject with two small metastatic lesions at the bottom right of the liver (arrows). The lesions
are clearly seen in the radial FSE image but not in the SSFSE image due to the intrinsic blurriness of the
latter. (b) Images of a patient with extensive metastatic invasion showing the higher lesion conspicuity
yielded by the radial FSE method. (c) Images of a subject with very small metastatic lesions. As a refer-
ence the lesion highlighted by the arrow is 5 mm in diameter. In the image acquired with SSFSE both
blurriness and the bright signal from blood vessels impairs the assignment of lesions. In the radial FSE
image the signal from blood vessels is suppressed via diffusion weighting which helps identifying the
signal from metastatic lesions
3. A Radial Magnetic Resonance Imaging Method for Imaging Abdominal Neoplasms 27
significantly better than the conventional equals the distance between k-space sam-
methods for malignant lesion definition. ples in the radial direction). In this man-
ner, high-resolution images at various TEeff
values as well as a T2 map are obtained
RADIAL FAST SPIN-ECHO, from a single radial FSE k-space data (Song
and Dougherty, 2000; Altbach et al., 2002,
A NEW ALTERNATIVE FOR 2005). This novel methodology provides a
T2 MAPPING OF THE LIVER new way to measure T2 without the prob-
lems typically encountered in the body. The
In conventional FT-MRI the measurement
T2 maps are generated from motion-insen-
of T2 requires the acquisition of two
sitive data acquired in a single breath hold.
or more k-space data sets with different
Thus, the errors due to motion or misreg-
T2 weighting. In SE acqusitions all the
istration of TE data sets are minimized.
k-space lines on each T2-weighted data Data have goodeffspatial resolution as well as
set are collected at a specific TE. In faster adequate TE time resolution (T2 values can
acqusition methods such as FSE, SSFSE, be calculated from 8 or 16 values of TE ).
eff
and EPI, the k-space lines on each data Moreover, the acquisition of data is fast (for
set have mixed TE values and the T2 abdominal imaging applications, it takes
weighting of the data set is approximated one breath hold to acquire data for several
to the TE of the lines in the central part of slices) and the post-processing algorithm
k-space (i.e., the so-called TEeff). For esti- is simple. Thus, the method can be easily
mating T2, it is necessary to acquire vari-
implemented as part of a regular clinical
ous data sets, each with a different TEeff.
MRI examination.
T2 is then calculated by fitting the MRI
Figure 3.4 shows the anatomical image
signal intensity, I, to:
(reconstructed from the full k-space data
−
TEeff set) as well as the T2-weighted images
I = Ioe T2
(3.1) and the T2 map (generated from the par-
tial k-space data sets) for a patient with
where Io, the equilibrium magnetization, is metastatic liver lesions and a cyst in the
a constant. kidney. All these information is obtained
In radial FSE, data at all TEs sample from a single k-space data set collected
the center of k-space (Figure 3.3a). Since in a breath hold. Note the differences in
the center of k-space is oversampled, contrast in the T2-weighted images and
T2-weighted contrast can be generated the T2 map for the various lesions. The
by partitioning a full k-space data set into signal intensity of the cyst (arrow head)
subsets that have in the center only the stays bright as TEeff increases, whereas the
data from lines acquired at a specific TE. signal intensity of the malignant lesions
Data at other TEs are added progressively (arrows) follows the signal intensity of the
from the center to the outer part of k-space kidney. The T2-weighted images and the
as indicated in Figure 3.3b. The radius on T2 map also reveal areas within two of
each tier of data satisfies the Nyquist crite- the metastatic lesions (long arrows) that
ria (i.e., the distance between two adjacent have a high T2. These correspond to
k-space samples in the azimuthal direction necrotic areas within the tumors. A study
28 M.I. Altbach
Figure 3.3. Schematic representation of the method used to generate multiple high-resolution images
with different T2-weighting and a T2 map from a single radial FSE data set. The method is demonstrated
for data acquired with an echo train length of four (i.e., four TE values)
on a set of 43 neoplasms which included 24 sented above, is the radius at which all TE
malignancies (22 metastases, 1 hepatoma, data are included. This radius determines
1 hepatocellular carcinoma) and 19 benign the accuracy of T2 estimates; the larger
lesions (6 hemangiomas and 12 cysts) cor- the radius the more accurate the estimates,
responding to 28 patients showed that the in particular for small lesions. This radius
T2 value determined from radial FSE data depends on the total number of radial lines
alone can discrimate hemangiomas and collected, and recently it was demon-
cysts from malignancies (Figure 3.5). strated that for a full data set of 512 radial
An important parameter in the determina- lines the error in T2 estimates for lesions
tion of T2, based on the radial method pre- with a diameter of 0.5–0.6 cm is only 7%.
3. A Radial Magnetic Resonance Imaging Method for Imaging Abdominal Neoplasms 29
Figure 3.4. Anatomical images at different TEeff, and the T2 map for a patient with metastatic liver lesions
(arrows) and a cyst (arrow head) in the kidney. All the images were reconstructed from a single radial FSE data
set acquired in a breath hold. The anatomical image is reconstructed using the full k-space data set. The images
at various TEeff were reconstructed using data at selected TE values in the center of k-space as illustrated in
Figure 3.3. The T2 map was obtained by fitting the pixel intensities to equation (3.1)
For lesions with a diameter greater than the other half is calculated and (Barger
0.8–1.0 cm the error is less than 4% (Altbach et al., 2002; Bydder et al., 2002) can be
et al., 2005). It should be pointed out that used to render the data needed for accurate
if the collection of sufficient radial lines is T2 estimation.
by the breath hold period, techniques were As any other technique that uses a single
only half of the radial lines collected and exponential fit to estimate T2, the method
30 M.I. Altbach
Altbach, M.I., Bilgin, A., Li, Z., Clarkson, E.W., Herborn, C.U., Vogt, F., Lauenstein, T.C., Goyen, M.,
and Gmitro, A.F. 2005. Processing of radial fast- Debatin, J.F., and Ruehm, S.G. 2003. MRI
spin echo data for obtaining T2 estimates from of the liver: can True FISP replace HASTE?
a single K-space data set. Magn. Reson. Med. J. Magn. Reson. Imaging 17:190–196.
54:549–559. Ichikawa, T., and Araki, T. 1999. Fast magnetic
Bangerter, N.K., Hargreaves, B.A., Vasanawala, resonance imaging of liver. Eur. J. Radiol.
S.S., Pauly, J.M., Gold, G.E., and Nishimura, 29:186–210.
D.G. 2004. Analysis of multiple-acquisition Ito, K., Mitchell, D.G., Outwater, E.K., Szklaruk, J.,
SSFP. Magn. Reson. Med. 51:1038–1047. and Sadek, A.G. 1997. Hepatic lesions: discrimi-
Barger, A.V., Block, W.F., Toropov, Y., Grist, T.M., nation of nonsolid, benign lesions from solid,
and Mistretta, C.A. 2002. Time resolved con- malignant lesions with heavily T2-weighted
trast-enhanced imaging with isotropic resolu- fast spin-echo MR imaging. Radiology 204:
tion and broad coverage using an undersampled 729–737.
3D projection trajectory. Magn. Reson. Med. Jung, B., Krombach, G.A., Gunther, R.W., and
48:297–305. Buecker, A. 2004. Is postcontrast trueFISP imag-
Bydder, M., Larkman, D.J., and Hajnal, J.V. 2002. ing advantageous? Invest. Radiol. 39:517–523.
Generalized SMASH imaging. Magn. Reson. Kauczor, H.U., and Kreitner, K.F. 1999. MRI
Med. 47:160–170. of the pulmonary parenchyma. Eur. Radiol.
Cieszanowski, A., Szeszkowski, W., Golebiowski, 9:1755–1764.
M., Bielecki, D.K., Grodzicki, M., and Keogan, M.T., Spritzer, C.E., Paulson, E.K., Paine,
Pruszynski, B. 2002. Discrimination of benign S.S., Harris, L., Dahlke, J.L., and MacFall, J.R.
from malignant hepatic lesions based on their 1996. Liver MR imaging: comparison of respira-
T2-relaxation times calculated from moderately tory triggered fast spin echo with T2-weighted
T2-weighted turbo SE sequence. Eur. Radiol. spin-echo and inversion recovery. Abdom.
12:2273–2279. Imaging 21:433–439.
Coates, G.G., Borrello, J.A., McFarland, E.G., Kim, Y.H., Saini, S., Blake, M.A., Harisinghani, M.,
Mirowitz, S.A., and Brown, J.J. 1998. Hepatic Chiou, Y.Y., Lee, W.J., Yu, J.S., and Hahn, P.F.
T2-weighted MRI: a prospective comparison of 2005. Distinguishing hepatic metastases from
sequences, including breath-hold, half-Fourier hemangiomas: qualitative and quantitative
turbo spin echo (HASTE). J. Magn. Reson. diagnostic performance through dual echo res-
Imaging 8:642–649. piratory-triggered fast spin echo magnetic reso-
Constable, R.T., and Gore, J.C. 1992. The loss of nance imaging. J. Comput. Assist. Tomogr. 29:
small objects in variable TE imaging: implica- 571–579.
tions for FSE, RARE, and EPI. Magn. Reson. McFarland, E.G., Mayo-Smith, W.W., Saini, S.,
Med. 28:9–24. Hahn, P.F., Goldberg, M.A., and Lee, M.J.
Gaa, J., Hatabu, H., Jenkins, R.L., Finn J.P., and 1994. Hepatic hemangiomas and malignant
Edelman, R.R. 1996. Liver masses: replace- tumors: improved differentiation with heavily
ment of conventional T2-weighted spin-echo T2-weighted conventional spin-echo MR imag-
MR imaging with breath-hold MR imaging. ing. Radiology 193:43–47.
Radiology 200:459–464. Namimoto, T., Yamashita, Y., Sumi, S., Tang, Y.,
Glover, G.H., and Pauly, J.M. 1992. Projection and Takahashi, M. 1997. Focal liver masses:
reconstruction techniques for reduction of characterization with diffusion-weighted echo-
motion effects in MRI. Magn. Reson. Med. planar MR imaging. Radiology 204:739–744.
28:275–289. Ohtomo, K., Itai, Y., Yoshikawa, K., Kokubo, T.,
Goldberg, M.A., Hahn, P.F., Saini, S., Cohen, and Iio M. 1988. Hepatocellular carcinoma and
M.S., Reimer, P., Brady, T.J., and Mueller, P.R. cavernous hemangioma: differentiation with MR
1993. Value of T1 and T2 relaxation times from imaging. Efficacy of T2 values at 0.35 and 1.5 T.
echoplanar MR imaging in the characterization Radiology 168:621–623.
of focal hepatic lesions. Am. J. Roentgenol. 160: Olcott, E.W., Li, K.C., Wright, G.A., Pattarelli,
1011–1017. P.P., Katz, D.S., Ch’en, I.Y., and Daniel, B.L.
32 M.I. Altbach
1999. Differentiation of hepatic malignancies Semelka, R.C., Kelekis, N.L., Thomasson, D.,
from hemangiomas and cysts by T2 relaxation Brown, M.A., and Laub, G.A. 1996. HASTE
times: early experience with multiply refocused MR imaging: description of technique and pre-
four-echo imaging at 1.5 T. J. Magn. Reson. liminary results in the abdomen. J. Magn. Reson.
Imaging 9:81–86. Imaging 6:698–699.
Outwater, E.K., Blasbalg, R., Siegelman, E.S., and Song, H.K., and Dougherty, L. 2000. K-space
Vala, M. 1998. Detection of lipid in abdominal weighted image contrast (KWIC) for contrast
tissues with opposed-phase gradient-echo images manipulation in projection reconstruction MRI.
at 1.5 T: techniques and diagnostic importance. Magn. Reson. Med. 44:825–832.
Radiographics 18:1465–1480. Tang, Y., Yamashita, Y., Namimoto, T., Abe, Y.,
Pawluk, R.S., Tummala, S., Brown, J.J., and and Takahashi, M. 1997. Liver T2-weighted
Borrello, J.A. 1999. A retrospective analysis MR imaging: comparison of fast and conven-
of the accuracy of T2-weighted images and tional half-Fourier single-shot turbo spin-echo,
dynamic gadolinium-enhanced sequences in the breath-hold turbo spin-echo, and respiratory-
detection and characterization of focal hepatic triggered turbo spin-echo sequences. Radiology
lesions. J. Magn. Reson. Imaging 9:266–273. 203:766–772.
Rasche, V., Holz, D., and Schepper, W. 1994. Trouard, T.P., Theilmann, R.J., Altbach, M.I., and
Radial turbo spin echo imaging. Magn. Reson. Gmitro, A.F. 1999. High-resolution diffusion
Med. 32:629–638. imaging with DIFRAD-FSE (diffusion-weighted
Reimer, P., Rummeny, E.J., Wissing, M., Bongartz, radial acquisition of data with fast spin-echo)
G.M., Schuierer, G., and Peters, P.E. 1996. MRI. Magn. Reson. Med. 42:11–18.
Hepatic MR imaging: comparison of RARE Ward, J., Robinson, P.J., Guthrie, J.A., Downing, S.,
derived sequences with conventional sequences Wilson, D., Lodge, J.P., Prasad, K.R., Toogood,
for detection and characterization of focal liver G.J., and Wyatt, J.I. 2005. Liver metastases in
lesions. Abdom. Imaging 21:427–432. candidates for hepatic resection: comparison of
Robinson, P.J. 2000. Imaging liver metastases: helical CT and gadolinium- and SPIO-enhanced
current limitations and future prospects, Br. J. MR imaging. Radiology 237:170–180.
Radiol. 73:234–241. Yu, J.S., Kim, K.W., Kim, Y.H., Jeong, E.K.,
Schaeffter, T., Weiss, S., Eggers, H., and Rasche, V. and Chien, D. 1998. Comparison of multishot
2001. Projection reconstruction balanced fast turbo spin echo and HASTE sequences for
field echo for interactive real-time cardiac imag- T2-weighted MRI of liver lesions. J. Magn.
ing. Magn. Reson. Med. 46:1238–1241. Reson. Imaging 8:1079–1084.
4
Liver: Helical Computed Tomography
and Magnetic Resonance Imaging
Yuji Baba, Yasuyuki Yamashita, Kazuo Awai, and Koichi Kawanaka
33
34 Y. Baba et al.
Figure 4.1. A case of undifferentiated hepatocellular carcinoma. Upper left: Precontrast CT. Upper right:
Arterial-phase image. Lower left: Equilibrium-phase image. Lower right: Reconstructed coronal image.
Tumor is iso to hypointense relative to the surrounding liver on precontrast CT. In arterial-phase images,
irregularly enhanced tumor is seen. Tumor enhancement is not obvious compared to well-differentiated
tumor (Figure 4.3). In the equilibrium phase, the tumor is seen as iso to hypoattenuating mass relative
to the surrounding liver. The relationship between tumor and vascular/biliary structure is clearly demon-
strated in reconstructed coronal image
4. Liver: Helical Computed Tomography and Magnetic Resonance Imaging 35
Figure 4.2. A case of hepatocellular carcinoma arising in a cirrhotic liver. Left: Arterial-phase image.
Right: Equilibrium-phase image. In arterial-phase image, well-enhanced tumor is seen. In the equilibrium
phase, the tumor is seen as hypoattenuating mass relative to the surrounding liver. A mosaic pattern is
well visualized in arterial phase image
36 Y. Baba et al.
Figure 4.3. A case of multiple metastatic liver tumors from colon cancer. Left: Portal-phase image.
Right: Equilibrium-phase image. Tumors are hypointense relative to the surrounding liver on portal-phase
CT. In the equilibrium phase, contrast enhancement is seen within tumors, resulting from poor contrast
between the tumor and surrounding liver
patients who are under consideration for agent to the portal venous system through
resection or interventional treatment of pri-
a catheter in the superior mesenteric or
mary or secondary liver neoplasms. There splenic artery. Although 70% of the hepatic
are two types of angiography-assisted CT blood supply is from the portal system,
techniques for the evaluation of liver neo- hepatic neoplasms are largely supplied
plasms: CT during hepatic arteriography by the hepatic artery. As a result, hepatic
(CTHA) and CT during arterial portogra- neoplasms are detected as hypoattenuat-
phy (CTAP). Combined use of CTHA with ing lesions compared with the normal
CTAP may be of value in differentiating contrast-enhanced parenchyma.
malignant from benign hepatic defects but Careful pretreatment evaluation of
does not appear to improve lesion detection.patients who are going to receive hepatic
resection or interventional treatment is
essential to determine the precise loca-
tion and number of metastatic lesions for
COMPUTED TOMOGRAPHY accurate surgical planning (Matsui et al.,
DURING ARTERIAL 1987) or embolization therapy. Computed
PORTOGRAPHY tomography during arterial portography is
a highly sensitive technique for detecting
Computed tomography during arterial liver lesions and is superior to intravenous
portography involves CT scanning during contrast-enhanced CT for detecting lesions
selective delivery of an iodinated contrast less than 2 cm in diameter (Heiken et al.,
38 Y. Baba et al.
Figure 4.4. CT angiography of a AV shunt in the liver. Three-dimensional vascular images can be recon-
structed from the data of multislice helical CT. Left: Arterial-phase image shows large abnormal vascular
structure in the liver. Upper right: CT arteriography reconstructed from the arterial-phase CT images.
Lower right: Celiac angiography. Enlarged celiac artery and dilated hepatic vein is demonstrated
1989; Matsui et al., 1988; Deflandre et al., scanning. Multi detector computed tomog-
1994). The reported sensitivity of CTAP raphy is an excellent method for perform-
for lesion detection is among the highest ing rapid hepatic scanning. This method
of the available radiologic techniques, results in high levels of liver parenchymal
ranging from 81% to 91% (Heiken et al., and vascular enhancement that are readily
1989; Deflandre et al., 1994; Bluemke maintained over the duration of the scan.
et al., 1995). Rapid scanning during injec- Computed tomography during arterial
tion is important because recirculation of portography may be particularly problem-
contrast to the systemic circulation results atic in patients with cirrhosis, not only
in diminished contrast differences between because of fibrosis and altered hepatic
the neoplasm and normal parenchyma. architecture but also due to altered hemo-
Recirculation of contrast to the aorta is dynamics.
evident 35-s following the start of the con- In patients with portal hypertension, portal
trast injection, indicating the need for rapid venous blood may be shunted away from the
4. Liver: Helical Computed Tomography and Magnetic Resonance Imaging 39
liver through portosystemic collateral ves- imaging capability. The presence of vari-
sels, limiting hepatic parenchymal enhance- ous kinds of contrast agents, including
ment. In such patients, CTHA may provide extracellular agents, hepatocyte-specific
more useful information than CTAP. agents, and reticuloendothelial system-
specific agents, also has some advantages.
Because of the numerous choices of imag-
ing sequences and parameters, selection
COMPUTED TOMOGRAPHY
and optimization of pulse sequences is a
DURING HEPATIC critical issue in MR imaging of the liver.
ANGIOGRAPHY The current standard MR imaging exam-
ination of the liver includes a T1-weighted
Computed tomography during hepatic sequence, a T2-weighted sequence and a
angiography is performed with the cath- contrast-enhanced dynamic sequence.
eter in the hepatic artery [13, 14]. This The most frequent contrast administration
technique is superb in analyzing neovascu- approach is the use of a nonspecific extra-
larity of hepatic neoplasms. On CT images cellular contrast agent, gadolinium chelate,
obtained with this technique, malignant as a rapid-bolus injection with serial imag-
hepatic neoplasms appear as peripherally ing using a gradient echo sequence as
or homogeneously enhancing hyperdense dynamic MR imaging. In several circum-
masses compared with the normal hepatic stances, tissue-specific contrast agents such
parenchyma (Irie et al., 1995; Chezmar as superparamagnetic iron oxide (SPIO)
et al., 1993), although these tumors may are used not only for tumor detection but
appear hypoattenuating on conventional also for tumor characterization.
liver CT. In early hepatocellular carcino-
mas developed in a precancerous adenom-
atous hyperplasia, enhancement is seen in
DYNAMIC MAGNETIC
most areas of the mass, indicating portal-
dominant supply of the lesion, consistent RESONANCE IMAGING
with precancerous atypical adenomatous
With recent advances in fast MR imaging
hyperplasia on CTAP images. Small foci
technology, the entire liver can be imaged
with absence of portal blood flow show
in a breath-hold with a spoiled gradient
arterial-dominant blood supply on CTHA.
echo sequence. The sequence is repeated
several times after administration of gado-
linium chelates (Figure 4.5). It is usually
MAGNETIC RESONANCE advisable to use an in-phase echo time on
IMAGING spoiled gradient echo images for contrast-
enhanced studies in order to avoid confus-
With the advent of the fast magnetic reso- ing variations in signal intensity based on
nance (MR) imaging technique, this has fatty infiltration of the liver and black-ring
become the modality of choice in some phase cancellation artifacts that will mask
patients for imaging of the liver. An advan- capsular-based disease.
tage of MR imaging compared with CT is In current clinical practice, gadolinium
its excellent tissue contrast and multiplanar chelates are frequently used for dynamic
40 Y. Baba et al.
Figure 4.5. A case of hepatocellular carcinoma after lipiodol TAE. Upper left: Precontrast CT. Upper right:
T1-weighted gradient echo image. Lower left: Arterial-phase MR imaging. Lower right: Equilibrium-
phase MR image. Tumor is visualized as significant hyperintense relative to the surrounding liver because
of lipiodol accumulation in the tumor. It is difficult to evaluate tumor vascularity using dynamic CT. The
tumor is seen as hypointense relative to the surrounding liver on T1-weighted MR images. Lipiodol in the
tumor does not affect signal intensity of the tumor. On arterial-phase image, well-enhanced tumor is seen.
Tumor is not as conspicuous in the equilibrium-phase image as in the arterial-phase image. Peripherally
enhancing pseudocapsule is well visualized
from normal liver tissue. Because the agent hemangioma, which also shows a decrease
is not taken up by tumors, hepatic masses in signal intensity on T2-weighted images.
stand out as hyperintense foci in the low- Some well-differentiated hepatocellular
signal intensity liver on proton density carcinomas and adenomatous hyperplasias
or T2/T2*-weighted images (Figure 4.6). may also show accumulation of this agent
Studies performed at 0.6 and 1.5 T have within the tumor. Therefore, in early hepa-
demonstrated an increased tumor detec- tocellular carcinomas that developed in a
tion rate (Tsang et al., 1988; Ros et al., precancerous adenomatous hyperplasia,
1995), and other studies have shown a signal decrease is seen in most areas of
increased lesion-to-liver contrast (Stark atypical adenomatous hyperplasia, indicat-
et al., 1988; Winter et al., 1993) compared ing the presence of Kupffer cells in this
with unenhanced imaging. Although most area. Small foci of signal hyperintensity
people may agree that this agent is effec- within the lesion may be present, indicat-
tive in detecting metastatic tumors, its role ing the absence of Kupffer cells, consistent
in detecting hepatocellular carcinomas is with malignant foci developed in atypical
questionable. The role of SPIO in char- adenomatous hyperplasia. Delineation of
acterizing focal liver lesions is limited. the clinical role of iron-containing hepatic
Some studies reported SPIO uptake in contrast agents for tumor characterization
lesions containing Kupffer cells, such as requires further testing.
focal nodular hyperplasia, and in lesions In conclusion, with modern imaging
in which blood pooling occurs, such as technologies, including MDCT and fast
Figure 4.6. A case of metastatic liver tumor from colon carcinoma. Left: Pre-SPIO T2-weighted fast spin
echo image. Right: Post-SPIO T2-weighted fast spin echo image. Tumor is hyperintense relative to the
surrounding liver on precontrast T2 image. After administration of SPIO, the signal intensity of the liver
parenchyma is decreased, resulting in better lesion-to-liver contrast on the postcontrast images
42 Y. Baba et al.
Winter, T.C., Freeny, P.C., Nghiem, H.V., Mack, L.A., Yamashita, Y., Yamamoto, H., Arakawa, A.,
Patten, R.M., Thomas, C.R., and Elliott, S. 1993. Matsukawa, T., Hatanaka, Y., and Takahashi, M.
MR imaging with i.v. superparamagnetic iron 1994. Differential diagnosis of liver neoplasms:
oxide: Efficacy in the detection of focal hepatic Role of dynamic MR imaging. Radiology
lesions. Am. J. Roentgenol. 161:1191–1198. 193:59–65.
Winter, T.C., Freeny, P.C., Nghiem, H.V., Yamashita, Y., Mitsuzaki, K., Yi, T., Ogata, I.,
Hommeyer, S.C., Barr, D., Croghan, A.M., Nishiharu, T., Urata, J., and Takahashi, M. 1996.
Coldwell, D.M., Althans, S.J., and Mack, L.A. Small hepatocellular carcinoma in patients with
1995. Hepatic arterial anatomy in transplantation chronic liver damage: Prospective comparison of
candidates: Evaluation with three-dimensional detection with dynamic MR imaging and helical
CT arteriography. Radiology 195:363–370. CT of the whole liver. Radiology 200:79–84.
A. Diagnosis
5
Selection of Patients for Resection
of Hepatic Colorectal Metastases:
18F-Fluorodeoxyglucose/Positron
Emission Tomography
Rebecca Auer and Yuman Fong
49
50 R. Auer and Y. Fong
and specificity of PET scans to detect scanning (i.e., blinded to the CT result)
intrahepatic and extrahepatic lesions. The or the incremental value of PET scan fol-
correlation of PET scans with the preop- lowing CT. These studies have attempted
erative clinical risk score (CRS), and the to define a role for PET scanning in this
impact on surgical management. These specific cohort of patients. A meta-analysis
data will be used to argue that the ability of these studies has also been published
to detect and eliminate patients with unre- providing a concise summary of the data
sectable metastatic colorectal cancer may available to date (Wiering et al., 2005).
also translate into improved survival post- A more general overview on the use of PET
hepatic resection. The benefit of PET dur- scanning in screening patients with colon
ing follow-up, particularly in patients with cancer for local and distant recurrences can
an undiagnosed elevation in carcinoem- be found in other chapters in this volume.
bryonic antigen (CEA) or a suspected
recurrence post-hepatectomy will also be
discussed. Finally, a preoperative staging POSITRON EMISSION
algorithm that incorporates the available
TOMOGRAPHY SCAN
information regarding PET scanning will
be presented. FOR THE DETECTION OF
EXTRAHEPATIC DISEASE
Extrahepatic disease is the most significant
POSITRON EMISSION
operative finding related to adverse long-
TOMOGRAPHY SCANNING term survival (Rosen et al., 1992). The
AS A STAGING MODALITY existence of extrahepatic disease should
TO COMPLEMENT be considered a contraindication for hepa-
CONVENTIONAL IMAGING tectomy, aside from isolated portal lymph
node metastases (Jaeck et al., 2002) or
Although there are no published rand- limited resectable pulmonary metastases
omized controlled trials on the use of FDG- (DeMatteo et al., 1999).
PET in patients with metastatic colorectal The ability of PET scans to detect previ-
cancer to the liver, a number of prospective ously unrecognized extrahepatic disease is
and retrospective studies have attempted to the most compelling reason for use in the
evaluate the ability of PET scan to improve preoperative setting for patients with
the staging of patients in the preoperative hepatic metastases. Numerous studies have
workup for hepatic resection of colorectal demonstrated the superiority of PET scans
metastases. All of these studies included a (Wiering et al., 2005) as compared to
cohort of patients who presented for evalu- conventional CT scans in this setting. The
ation of suspected colorectal cancer recur- sensitivity of a PET scan is 91.5% (range
rence to the liver based on conventional from 63% to 100%) compared to 60.9%
imaging, such as computed tomography (range 25–86%) (Wiering et al., 2005).
(CT) or, in some cases, an elevated CEA Whereas CT scans have a specificity of
level. These studies compare the relative 95.4% as compared to 91.9% with PET
value of PET scans in staging a patient for (Wiering et al., 2005). [18F]fluorodeoxy-
hepatic resection; in the replacement of CT glucose-positron emission tomography
5. Selection of Patients for Resection of Hepatic Colorectal Metastases 51
These parameters include: (1) the presence evidence of additional disease detected
of extrahepatic disease at laparotomy, (2) by PET (Schussler-Fiorenza et al., 2004).
a positive resection margin, (3) nodal The authors concluded that patients with a
metastases from primary cancer, (4) a short CRS of 0 should not undergo preoperative
disease-free interval, (5) largest tumor evaluation with a PET scan because the
> 5 cm, (6) more than one liver metasta- yield is too low to justify the expense and
sis, and (7) a CEA over 200 ng/ml. Using morbidity associated with false-positive
the last five criteria, a preoperative CRS results (Schussler-Fiorenza et al., 2004).
system was created with each positive
criterion counting as one point. This CRS
is a simple, easily remembered staging POSITIVE IMPACT OF
system for classifying patients with liver- POSITRON EMISSION
exclusive metastatic colorectal cancer. TOMOGRAPHY ON THE
The presence of any one of these charac- MANAGEMENT OF PATIENTS
teristics was still associated with a 5-year WITH LIVER METASTASES
survival of 24–34%, and therefore no single
criterion can be considered a contraindica- Positron emission tomography (PET)
tion to resection (Fong et al., 1999a). It is scans have repeatedly shown the ability to
the total score out of 5 that is highly predic- alter clinical treatment plans when used
tive of outcome. A score of 2 or less places for preoperative staging in patients with
a patient in a good prognostic group, for hepatic colorectal metastases. In a well
whom resection is ideal. For scores of 3 or designed prospective study evaluating the
4, outcome is less favorable and patients added benefit of CT/PET scan to CT scan
should be considered for aggressive tri- alone, 76 patients with hepatic colorectal
als of adjuvant therapy. For a score of 5, metastases were imaged preoperatively
long-term disease-free survivors are rarely (Selzner et al., 2004). In this study a pre-
encountered and resections in this high risk PET treatment plan was prospectively
group should be accompanied by trials of defined and compared to the treatment
adjuvant therapy (Fong et al., 1999a). plan outlined after consideration of the
This CRS is also now being used to results of the CT/PET scan (Selzner et al.,
help select the extent and sophistication 2004). In 60 patients (79%) PET/CT did
of preoperative assessment, including the not change the therapeutic strategy, and
use of FDG-PET. The ability of the CRS in 10 patients (13%) considered resect-
to predict the yield of a preoperative PET able by CT alone, the PET/CT demon-
scan was demonstrated in a subset of 63 strated extensive extrahepatic disease,
patients presenting with a first occur- representing a contraindication to surgery
rence of hepatic colorectal metastases (Selzner et al., 2004). In the remaining
(Schussler-Fiorenza et al., 2004). Among six patients (8%) the surgical strategy
patients with a CRS of 0, no patient and extent of resection was changed
had extrahepatic disease detected by PET because the PET/CT demonstrated pre-
and 57% showed a false-positive reading viously unrecognized positive nodes in
(Schussler-Fiorenza et al., 2004). Among the hepatoduodenal ligament (Selzner
patients with a CRS of ≥ 1, 14% had et al., 2004). Taken together, PET/CT
54 R. Auer and Y. Fong
survival estimate at 3 years was 40%, likely to have extrahepatic disease detected
higher than the previously reported results by FDG-PET and therefore considered
of 15–28% (Strasberg et al., 2001). unresectable (Strasberg et al., 2001). In
In a follow-up study, 100 patients who the subsequent analysis of survival follow-
underwent complete hepatic resection for ing hepatic resection among this selected
curative intent were evaluated by preoperative PET negative population, there is loss of
FDG-PET and CT scans, and were followed significance for the prognostic indicators,
for a median of 31 months (Fernandez et al., reflecting the added information obtained
2004). This study did not include patients by preoperative PET scanning.
who underwent an exploratory laparotomy
but were found to be unresectable; therefore,
it is not possible to comment on the ability of STANDARD UPTAKE
PET scans to prevent unnecessary laparoto- VALUE AND PREDICTING
mies from this study. The Kaplan-Meier esti- PROGNOSIS OR RESPONSE
mate of 5-year overall survival in this group TO THERAPY
of patients was 58.6% (95% CI of 46–72%)
which compares favorably to the results A fundamental metabolic difference
of 19 previously reported series where the between normal and malignant cells
estimated 5-year overall survival was 30% relates to the cells primary energy source.
(range 12–41%) (Fernandez et al., 2004). Whereas a normal cell derives most of its
One criticism of this study is that in com- energy from oxidative phosphorylation, a
paring to historical controls, the influences rapidly growing tumor cell derives more
of improved surgical techniques and new than half of its energy from glycolysis,
chemotherapeutic agents which might have with only a minor amount from the citric
contributed to the improved results are not acid cycle (Nakashima et al., 1984). The
considered. The authors acknowledge the glycolytic intermediates are used not only
potential for these confounding variables but to provide energy, but also to intermedi-
state that only 4% of the patients received ate in the synthesis of DNA/RNA and
irinotecan as adjuvant treatment and that membrane lipids. The glycolytic capacity
oxaliplatin was not yet available by the close of tumor cells has been demonstrated in a
of the study (Fernandez et al., 2004). number of studies to correlate with lack
It is interesting to note that in this study of differentiation of these cells (Pauwels
none of the previously established prog- et al., 2000), and when cells undergo
nostic factors relating to the hepatic metas- oncogenic transformation, their glycolytic
tasis, including the components of the activity increases. Many human tumors
CRS, were significant predictors of worse including colon cancers, have demon-
outcome (Fernandez et al., 2004). This strated increased expression and activity
suggests that PET scanning represents a of several glucose transporters (GLUT 1
surrogate for previously identified clinical and 3) and the key enzyme hexokinase
variables that make up the CRS. This (reviewed in Pauwels et al., 2000).
observation may be related to the fact that Positron emission tomography scan-
PET scans preferentially exclude patients ning has the ability to give functional
with multiple, bilateral, and synchronous information about a tumor and may be
tumors because such patients are more able to predict the prognosis or response
56 R. Auer and Y. Fong
a c
Figure 5.3. A patient with bilobar, multiple hepatic metastases treated with neoadjuvant chemotherapy
of FOLFOX with good response and shrinking of all the lesions. The lesions almost completely disap-
peared by PET scan (white arrow) suggesting a good response to chemotherapy. One lesion still took
up a significant amount of FDG and this lesion was considered to be chemoresistant (black arrow). The
prognostic value of a decrease in SUV following chemotherapy is still being evaluated
5. Selection of Patients for Resection of Hepatic Colorectal Metastases 57
however, fail to document the number of only one patient had a false-positive (PET
invasive procedures, such as biopsies that specificity 89%) (Selzner et al., 2004). In
PET scans either avoid or mandate, and the another blinded study, PET had a diag-
associated morbidity and costs. Moreover, nostic accuracy of 86% in differentiat-
these studies cannot claim that a PET scan ing intrahepatic postsurgical change from
is a superior test to evaluate for recurrent malignant recurrence, as compared to 45%
disease in the setting of an elevated CEA for CT portography (Delbeke et al., 1997).
during surveillance, because these studies The significant difference is the result of
evaluated PET with full knowledge of the more false-positives with CT portography.
results of conventional imaging.
ALGORITHM TO INCLUDE
THE ROLE OF POSITRON POSITRON EMISSION
EMISSION TOMOGRAPHY TOMOGRAPHY SCANNING
IN FOLLOW-UP POST IN WORK-UP OF HEPATIC
HEPATIC RESECTION COLORECTAL METASTASES
The most common site for recurrence after The preoperative staging prior to hepatic
resection of hepatic colorectal metastases resection of colorectal metastases must be
is the liver, and it is the sole site of recur- meticulous to avoid unexpected findings or
rence in up to 40% of cases (Taylor, 1996). obstacles at the time of laparotomy. A sys-
In the absence of extrahepatic disease and tematic approach to the colorectal cancer
in a patient with adequate hepatic reserve, patient who is referred for consideration of
a repeat hepatectomy may be considered. hepatic resection can facilitate a thorough
In approximately one third of patients the patient work-up and minimize unnecessary
recurrence will be amenable to further investigations. Patients should initially be
resection with a 5-year overall survival evaluated with a high quality CT scan of the
range from 31–41% (Bozzetti et al., 1992). chest abdomen and pelvis to identify hepatic
Positron emission tomography/com- metastases as well as local recurrence and
puted tomography (PET/CT) imaging may lung metastases. In patients with an equiv-
be superior to conventional CT in estab- ocal finding on CT scan or a persistently
lishing the diagnosis of intrahepatic recur- elevated CEA and a normal CT scan, PET
rence in patients with prior hepatectomy. scanning is warranted and may delineate
In a prospective study of the impact of the site of metastasis and/or demonstrate
FDG-PET on treatment in patients with additional extrahepatic disease. In patients
colorectal liver metastases, a subset of with a documented suspicious lesion in
18 patients with previous hepatic resec- the liver on a CT scan, a PET scan is
tion had a suspected local recurrence by best reserved for patients with a CRS ≥ 2
CT scan (Selzner et al., 2004). A total of because these patients will have the highest
nine patients had an intrahepatic recur- incidence of extrahepatic disease that may
rence, resulting in specificity of 50% for impact on the clinical management.
CT scans (Selzner et al., 2004). By com- As previously mentioned, PET scans may
parison, all nine patients were positive by have a false-negative rate as high as 17%
FDG-PET (PET sensitivity 100%) and (Truant et al., 2005), and this is frequently
5. Selection of Patients for Resection of Hepatic Colorectal Metastases 59
related to small peritoneal metastases. tion for recurrent metastases from colorectal
Laparoscopy is, therefore, still indicated in cancer. Br. J. Surg. 79:146–148.
Delbeke, D., Vitola, J.V., Sandler, M.P., Arildsen,
patients with a CRS of > 2 because these
R.C., Powers, T.A., Wright, J.K. Jr., Chapman,
patients have a high incidence of peritoneal W.C., and Pinson, C.W. 1997. Staging recurrent
metastases that would represent a contrain- metastatic colorectal carcinoma with PET. J.
dication to resection (Jarnagin et al., 2001). Nucl. Med. 38:1196–1201.
Finally, because false-positives can occur, DeMatteo, R.P., Minnard, E.A., Kemeny, N.,
a biopsy is warranted to confirm the pres- Downey, R., Burt, B.M., Fong, Y., and Blumgart,
L.H. 1999. Outcome after resection of both liver
ence of metastatic malignancy in patients
and lung metastases in patients with colorectal
with PET evidence of extrahepatic metas- cancer. ASCO Annual Meeting, abstract.
tases that would render them inoperable. Dimitrakopoulou-Strauss, A., Strauss, L.G., and
Rudi, J. 2003. PET-FDG as predictor of therapy
response in patients with colorectal carcinoma.
CONCLUSION Q. J. Nucl. Med. 47:8–13.
Fernandez, F.G., Drebin, J.A., Linehan, D.C.,
Dehdashti, B., Siegel, B.A., and Strasberg,
There is no doubt that FDG-PET has been
S.M. 2004. Five-year survival after resection
an important addition to our staging modal- of hepatic metastases from colorectal cancer in
ities for metastatic colorectal cancer. For patients screened by positron emission tomogra-
selection of patients for liver resection, this phy with F-18 fluorodeoxyglucose (FDG-PET).
test has become indispensable. The most Ann. Surg. 240:438–447.
exciting current area of research includes Findlay, M., Young, H., Cunningham, D., Iveson,
A., Cronin, B., Hickish, T., Pratt, B., Husband,
studies to correlate FDG uptake with tissue
J., Flower, M., and Ott, R. 1996. Noninvasive
markers of prognosis. Ongoing studies of monitoring of tumor metabolism using fluorode-
cost-effectiveness will likely decipher the oxyglucose and positron emission tomography
clinical parameters that should be used for in colorectal cancer liver metastases: correlation
selection of PET scanning in various clinical with tumor response to fluorouracil. J. Clin.
scenarios. Studies of the use of this test Oncol. 14:700–708.
Flamen, P., Hoekstra, O.S., Homans, F., Flamen,
as an early predictor of response to neo-
P., Hoekstra, O.S., Homans, F., Van Cutsem, E.,
adjuvant chemotherapy will likely further Maes, A., Stroobants, S., Peeters, M., Penninckx,
streamline and optimize patient care. F., Filez, L., Bleichrodt, R.P., and Mortelmans, L.
2001. Unexplained rising carcinoembryonic
antigen (CEA) in the postoperative surveillance
REFERENCES
of colorectal cancer: the utility of positron tom-
Akhurst, T., Kates, T.J., Mazumdar, M., Yeing, ography (PET). Eur. J. Cancer 37:862–869.
H., Riedel, E.R., Burt, B.M., Blumgart, L.H., Flanagan, F.L., Dehdashti, F., Ogunbiyi, O.A.,
Jarnagin, W., Larson, S.M., and Fong, Y. 2005. Kodner, I.J., and Siegel, B.A. 1998. Utility of
Recent chemotherapy reduces the sensitivity FDG-PET for investigating unexplained plasma
of [18F]fluorodeoxyglucose positron emission CEA elevation in patients with colorectal cancer.
tomography in the detection of colorectal metas- Ann. Surg. 227:319–323.
tases. J. Clin. Oncol. 23:8713–8716. Fong, Y., Fortner, J., Sun, R.L., Brennan, M.F.,
Bines, S.D., Doolas, L., Jenkins, L., Millikan, K., and Blumgart, L.H. 1999a. Clinical score for
and Roseman, D.L. 1996. Survival after repeat predicting recurrence after hepatic resection for
hepatic resection for recurrent colorectal hepatic metastatic colorectal cancer: analysis of 1001
metastases. Surgery 120:591–596. consecutive cases. Ann. Surg. 230:309–318.
Bozzetti, F., Bignami, R., Montalto, F., Doci, R., Fong, Y., Saldinger, P.F., Akhurst, T., Macaplinac,
and Gennari, L. 1992. Repeated hepatic resec- H., Yeung, H., Finn, R.D., Cohen, A., Kemeny,
60 R. Auer and Y. Fong
N., Blumgart, L.H., and Larson, S.M. 1999b. Oyen, W.J. 2002. Value of positron emission
Utility of 18F-FDG positron emission tom- tomography with [F-18]fluorodeoxyglucose in
ography scanning on selection of patients for patients with colorectal liver metastases: a pro-
resection of hepatic colorectal metastases. Am. spective study. J. Clin. Oncol. 20:388–395.
J. Surg. 178:282–287. Rydzewski, B., Dehdashti, F., Gordon, B.A.,
Guillem, J.G., Moore, H.G., Akhurst, T., Klimstra, Teefey, S.A., Strasberg, S.M., and Siegel, B.A.
D.S., Ruo, L., Mazumdar, M., Minsky, B.D., 2002. Usefulness of intraoperative sonography
Gollub, M.J., Saltz, L., Wong, W.D., and Larson, for revealing hepatic metastases from colorec-
S. 2004. Sequential preoperative fluorodeoxyglu- tal cancer in patients selected for surgery after
cose-positron emission tomography assessment undergoing FDG PET. AJR. Am. J. Roentgenol.
of response to preoperative chemoradiation: a 178:353–358.
means for determining long term outcomes of Schussler-Fiorenza, C.M., Mahvi, D.M.,
rectal cancer. J. Am. Coll. Surg. 199:1–7. Niederhuber, J., Rikkers, L.F., and Weber,
Jaeck, D., Nakano, H., Bachellier, P., Inoue, K., S.M. 2004. Clinical risk score correlates with
Weber, J.C., Oussoultzoglou, E., Wolf, P., and yield of PET scan in patients with colorec-
Chenard-Neu, M.P. 2002. Significance of hepatic tal hepatic metastases. J. Gastrointest. Surg.
pedicle lymph node involvement in patients with 8:150–157.
colorectal liver metastases: a prospective study. Selzner, M., Hany, T.F., Wildbrett, P., McCormack,
Ann Surg. Oncol. 9:430–438. L., Kadry, Z., and Clavien, P.A. 2004. Does the
Jarnagin, W.R., Conlon, J., Bodniewicz, J., novel PET/CT imaging modality impact on the
Dougherty, E., DeMatteo, R.P., Blumgart, L.H., treatment of patients with metastatic colorectal
and Fong, Y. 2001. A clinical scoring system cancer to the liver? Ann. Surg. 204:1027–1034.
predicts the yield of diagnostic laparoscopy Strasberg, S.M., Dehdashti, F., Siegel, B.A.,
in patients with potentially resectable hepatic Drebin, J.A., and Linehan, D. 2001. Survival
colorectal metastases. Cancer 91:1121–1128. of patients evaluated by FDG-PET before
Nakashima, R.A., Paggi, M.G., and Pederson, P.L. hepatic resection for metastatic colorectal car-
1984. Contributions of glycolysis and oxidative cinoma: a prospective database study. Ann.
phosphorylation to adenosine 5′-triphosphate Surg. 233:293–299.
production in AS-30D hepatoma cells. Cancer Taylor, I. 1996. Liver metastases from colorectal
Res. 44:5702–5706. cancer: lessons from past and present clinical
Pauwels, E.K., Sturm, E.J., Bombardieri, E., studies. Br. J. Surg. 83:456–460.
Cleton, F.J., and Stokkel, M.O. 2000. Positron- Truant, S., Huglo, D., Hebbar, M., Ernst, O.,
emission tomography with [18F]fluorodeoxy- Steinling, M., and Pruvot, F.R. 2005. Prospective
glucose. Part I. Biochemical uptake mechanism evaluation of the impact of [18F]fluoro-2-deoxy-
and its implications for clinical studies. J. Can. D-glucose positron emission tomography of
Res. Clin. Oncol. 126:549–559. resectable colorectal liver metastases. Br. J.
Rohren, E.M., Paulson, E.K., Hagge, R., Wong, Surg. 92:362–369.
T.Z., Killius, J., Clavien, P.A., and Nelson, R.C. Wiering, B., Krabbe, P.F., Jager, G.J., Oyen, W.J.,
2002. The role of F-18 FDG positron emission and Ruers, T.J. 2005. The impact of fluor-18-
tomography in preoperative assessment of the deoxyglucose-positron emission tomography in
liver in patients being considered for curative the management of colorectal liver metastases.
resection of hepatic metastases from colorectal Cancer 104:2658–2670.
cancer. Clin. Nucl. Med. 27:550–555. Young, H., Baum, R., Cremerius, U., Herholz,
Rosen, C.B., Nagorney, D.M., Taswell, H.F., K., Hoekstra, O., Lammertsma, A., Pruim, J.,
Helgeson, S.L., Illstrup, D.M., Van Heerden, and Price, P. 1999. Measurement of clinical
J.A., and Adson, M. 1992. Perioperative blood and subclinical tumour response using [18F]-
transfusion and determinants of survival after fluorodeoxyglucose and positron emission
liver resection for metastatic colorectal carci- tomography: review and 1999 ORTC recommen-
noma. Ann. Surg. 216:492–505. dations. European Organization for Research
Ruers, T.J., Langenhoff, B.S., Neeleman, N., Jager, and Treatment of Cancer (EORTC) PET Study
G.J., Strijk, S., Wobbes, T., Corstens, F.H., and Group. Eur. J. Cancer 35:1773–1782.
B. Treatment
6
Ultrasonography During Liver Surgery
Guido Torzilli
63
64 G. Torzilli
adopted. In all patients 4.8 ml sulphur- detection of new lesions in around 30%
hexafluoride microbubbles (SonoVue®, of cases (Kokudo et al., 1996) (Figure
Bracco Imaging, Italy) was injected intra- 6.1). Conversely, in the case of CRC liver
venously through a peripheral vein by metastases 10–40% of cases do not have
the anaesthesiologist. Intraoperative high- palpable fore sites (Machi et al., 1991)
frequency probes have limited value for (Figure 6.2). As a consequence of this
CE-IOUS although they keep the advan- IOUS exploration of the liver has a great
tages in terms of volume, and stability; impact on the surgical strategy: Kane et al.
indeed, the effects of contrast enhancement (1994) experienced 51% of changed pro-
are less evident using these probes. More
recently, a micro-convex prototype with
the same frequency pattern of the per-
cutaneous probe used for CEIOUS was
released, and its adequacy during surgery
is under evaluation.
A background of perfect knowledge of
the liver anatomy surgically and ultra-
sonographically is requested to proceed.
For surgical anatomy, the Couinaud seg-
ments are considered here (Couinaud,
1957). After entering the abdominal cav-
ity, liver mobilisation dividing the round
and falciform ligaments, and division of
eventual adhesions to free the antero-
superior and inferior surfaces of the liver
are the steps which should precede the Figure 6.1. Millimetric hypoechogenic nodule
liver exploration with IOUS. Pulling the (arrow)
round ligament, the liver surface is widely
exposed and following the portal branches
and the hepatic veins, the liver can be
entirely studied.
The use of IOUS in liver resections can
be schematically divided into two prin-
cipal phases: the liver exploration for the
staging of the disease and planning of
the surgical strategy, and the guidance
of the surgical maneuvers.
Liver Exploration
The hard and irregular surface of cirrhotic
liver makes the detection of small nodules Figure 6.2. Millimetric hyperechogenic nodule
by palpation difficult; IOUS allows the (arrow)
6. Ultrasonography During Liver Surgery 65
cedures by IOUS in their series. However, even biopsy does not seem to be adequate.
the impact of IOUS on the operative deci- The only nodule which can be easily dif-
sion making, when compared with those ferentiated intraoperatively from a HCC
of preoperative imaging techniques, has or a liver metastases is the small hemangi-
decreased to 4–7% (Jarnagin et al., 2001; oma, which is often discovered primarily
Cerwenka et al., 2003; Sahani et al., 2004). at IOUS: it has a typical ultrasonographic
However, the problem of the impact of pattern and, moreover, when compressed
IOUS on the operative decision making changes its size and appearance. Further
depends on two main factors: the surgical improvement in differential diagnosis of
policy of each specific team, and the type liver nodules with IOUS may be expected
of tumor. Indeed, the low rates shown in by the introduction and diffusion of the
the latest reports (Jarnagin et al., 2001; intraoperative use of the last generation
Cerwenka et al., 2003; Sahani et al., contrast agents.
2004) are also partially motivated by the
surgeon’s surgical policy: in fact, since a Contrast Enhanced Intraoperative
considerable number of patients under- Ultrasonography
went major hepatectomies, new nodules More recently the introduction of CE-IOUS
detected by IOUS in the same hemiliver has set the rate of modified operative deci-
would not have modified the surgical strat- sion-making in 30–40% of cases (Torzilli
egy. In our experience major hepatectomies et al., 2004a, b, 2005b).
are carried out in the minority of patients Tumor vascularity as a criterion for dif-
(Torzilli et al., 2005a, 2006) because of the ferentiating the regenerative or dysplastic
extensive use of the IOUS-guidance for nodules from the HCC correlate well with
achieving parenchymal sparing resections, the histological evidence of a progressive
so that detection of new nodules is more increase in unpaired arteries from dys-
suitable for changing the surgical strategy. plastic to neoplastic nodules in a cirrhotic
Conversely, depending on the type of the liver (Roncalli et al., 1999). Certainly, the
tumor, liver exploration is focused mainly pattern of vascular enhancement is not suf-
on the detection of new lesions as in the ficient for differentiating malignant from
case of CRC liver metastases rather than non malignant nodules in a cirrhotic liver
both on the detection and on the differen- with 100% specificity. However, CE-US
tiation, as in the case of HCC. Indeed, for provides differential diagnosis of FLL
this last tumor, which generally develops with a 95% specificity (Quaia et al.,
in the cirrhotic liver the risk nowadays is 2004); of course, it must be considered
to overestimate the tumor stage with IOUS. that this last rate was referred to another
In fact, in the case of patients with HCC type of lesion in comparison to the target
in a cirrhotic liver, except for those nod- of the present study. Indeed, the intraop-
ules with mosaic ultrasonographic pattern erative exploration takes profit from the
which are malignant in 84% of cases, only higher resolution of the ultrasonography in
24–30% of hypoechoic nodules (Figure direct contact with the liver. Therefore, the
6.1), and 0–18% of those hyperechoic are need of differentiating nodules detected at
neoplasm (Kokudo et al., 1996; Takigawa IOUS is mostly focused on lesions smaller
et al., 2001). To overcome this problem than 1 cm: for these nodules, the vascularity
66 G. Torzilli
as criterion for differential diagnosis is These last were not removed, and
less specific. However, some improve- their benignity was confirmed after at
ments compared with conventional IOUS least 6 months of follow-up if there
could be expected. In our preliminary was no tumor growth or they were
experience, CE-IOUS provided remark- still not detectable at spiral CT.
able findings, either by adding information (C) If a new nodule showed a mosaic pat-
on nodular vascularity in patients with tern at IOUS, it was considered malig-
HCC, or by detecting nodules that were nant (Takigawa et al., 2001), and then
not visible at IOUS, in patients with CRC removed, independently from the pat-
liver metastases (Torzilli et al., 2004b, tern of enhancement at CE-IOUS.
2005b). Focusing the attention on patients
Specificity of this classification criteria
operated for HCC our preliminary experi-
is in our experience 69% (Torzilli et al.,
ence showed encouraging results (Torzilli
2007) which is certainly closer to the
et al., 2004a): however, they were related
true performance of CEIOUS, rather than
to a very limited number of patients, and
the results obtained at the initial experi-
data did not meet the expectancies shown
ence (Torzilli et al., 2004a). This value is
by optimistic findings with 100% specifi-
probably not that high, especially when
city. Therefore, we classified the CEIOUS
compared with that reported for CE-US
pattern based on our preliminary experi-
(Quaia et al., 2004). However, as we men-
ence and on what was reported in literature
tioned earlier the small size of the lesions
about the CE-US (Quaia et al., 2004).
targeted for CEIOUS study could explain
Classification was as follows:
this discrepancy: for these tiny nodules
(A) If an hypoechoic or hyperechoic nod- the neovascularity as criterion for differ-
ule showed at CEIOUS a full enhance- entiation between malignant and benign
ment (hyperechoic) in the arterial lesions has limits which are independ-
phase and became hypoechoic in the ent from the method we use for studying
delayed portal and late phases (A1) them. Therefore, CEIOUS can be helpful
or it remained hypoechoic with thin in a certain percentage of nodules but not
vessels detectable into the nodule in the in all. In this perspective the rate of 69%
arterial phase and delayed phase (A2) of specificity is encouraging as it means
or did not show any early enhance- that we can provide proper information
ment but remained hypoechoic in the with this new technique in seven out of ten
delayed phases with peripheral and/or lesions we detect at the time of laparotomy.
intralesional neovascularization (A3), Further improvements in this sense could
it was considered malignant and then come from the introduction of hepto-
removed. specific contrast spents for CE-IOUS.
(B) If an hypoechoic or hyperechoic
Colorectal Cancer Liver Metastases
nodule had no early enhancement at
CEIOUS and showed a hysoechoic Intraoperative ultrasonography is still
or slightly hypoechoic pattern in all the most accurate diagnostic technique
the phases in comparison with the for detecting focal liver lesions (FLLs)
surrounding liver parenchyma, it was (Sahani et al., 2004). However, during
considered a non neoplastic lesion. surgery for CRC liver metastases, IOUS
6. Ultrasonography During Liver Surgery 67
has a sensitivity of just 82% (Machi et al., out the risk of tumor exposure, and thus
1991), and as a consequence may miss enhancing the treatment radicality.
nodules less than 1 cm in diameter: this is
particularly evident in those patients who Tumor Location
undergo surgery after chemotherapy. In IOUS allows an accurate three-dimen-
these patients, liver fore sites have similar sional reconstruction of the relation-
echo-pattern with the surrounding liver ship between the tumor and the portal
parenchyma. branches, and hepatic veins: this is a fun-
In our experience, CE-IOUS confirmed damental step in the definition of the
its improvement of IOUS sensitivity in proper surgical strategy. Indeed, surgical
detecting new small fore sites reducing decision making should be obtained hav-
the risk of down-staging the disease and ing portal branches and hepatic veins as
enhancing the rate of treatment with cura- landmarks to reduce the risk of major
tive intent: indeed 9% of patients had morbidity and mortality.
new lesions discovered only by CEIOUS The definition of tumor-vessels relation-
(Figure 6.3a, b) (Torzilli et al., 2008a). ship is relevant for planning the type of
The association of IOUS and CEIOUS resection and for this purpose we classi-
increased the sensitivity from 88 to 93%. fied its appearance at IOUS. Based on this
Additionally, the better visualization of classification each category corresponds to
the margins of the main lesion in 19% of a specific operation.
patients with CRC liver metastases helps Based on IOUS findings, relationship with
the surgeon to better define the resection intrahepatic vascular structures were classi-
area and the proper dissection plane ruling fied as follows (Torzilli et al., 2005a):
a b
Figure 6.3. (a) The liver shows a inhomogeneous pattern at IOUS; (b) at CEIOUS is well evident black-
hole (arrow) which corresponds to a metastatic nodule
68 G. Torzilli
a b
HCC
RHV
HCC
RHV
IVC
Figure 6.4. (a) This patient was a carrier of HCC in contact without sign of infiltration with the right
hepatic vein (RHV) indicated by the arrows (pattern 1): operation consisted of a limited resection without
resection of the RHV; (b) this patient was a carrier of HCC in contact with sign of infiltration with the
right hepatic vein (RHV) indicated by the arrows (pattern 5): operation consisted of an extended Segment
7 resection including the RHV. IVC=Inferior Vena Cava
6. Ultrasonography During Liver Surgery 69
in more conservative but radical treatments portal branch of segment 5 instead of the
and in a lower rate of major hepatectomies. planned subsegmental branch of segment
Furthermore, in those patients in which 8 and then, necrosis of the segment 5 may
major resections are needed, IOUS allows occur. The hooking technique under IOUS
better achievement of the proper dissection control enables the identification of the
plane, which should run along the hepatic branch, which was encircled, and then the
vein to be fully anatomic. surgeon can decide with certainty whether
The artifacts which allow the dissec- to ligate it. This is useful also in the case of
tion plan at IOUS can sometimes mask tumor thrombus in major portal branches.
structures such as portal branches which In this situation, once the portal branch is
should be ligated or conversely respected. skeletonized, it is encircled with a stitch
For this reason, to better visualise the and, under IOUS control, the stitch, is
targeted point where the portal branch gently pulled up: this traction stretches
should be divided, the so-called hooking the portal branch slightly and the traction
technique has been devised (Torzilli et al., point is demonstrated clearly by IOUS. If
1999b). When the Glissonian sheath is the traction point is not at the level of the
exposed and skeletonized, it is encircled tumor thrombus it is possible to ligate the
with a stitch, which is visualized by IOUS portal branch and proceed with the liver
as an echogenic spot with a posterior resection being sure that thrombus will not
shadow. Then under sonographic control, migrate because of surgical manipulation
the stitch, hooking the exposed vessel, is (Torzilli et al., 2005c).
gently pulled up, which stretches the portal During liver dissection the back-flow
branch slightly and the traction point is bleeding from the hepatic veins is an
demonstrated clearly by IOUS. If the important source of blood loss, and it
exposed portal branch is not clearly visible is one of the most important factors in
because it has collapsed, the portal triad is determining the short-term and long-term
unclamped to enable it to fill with blood patient’s outcome. Therefore, limiting
and then it is visualised better by IOUS. If the backflow bleeding from the hepatic
the target site is correct, the portal branch veins is a priority in liver resections. We
is ligated and divided and segmentec- described a ultrasound-guided technique
tomy is completed under IOUS guidance; for backflow bleeding control from the
conversely, if the exposed vessel was not RHV during right-sided liver resection.
the targeted one, it is spared and useless The technique is very simple: once the
sacrifice of further liver parenchyma is right surface of the extrahepatic RHV is
avoided. exposed to allow its compression by the
A practical example in which the hook- surgeon’s finger-tips, the effectiveness of
ing technique is used is during ventral or finger compression is checked by IOUS,
dorsal subsegmentectomy of segment 8. and color-Doppler.
The portal trunk to this segment may show
bifurcation in its dorsal branch and ventral Post-resectional Control
trunk just close to the origin of the portal There are possibilities given by IOUS after
vessel to segment 5. In this situation, there nodule removal: one is the “water bath”
is the risk of ligating and dividing the technique which consists of the real-time
6. Ultrasonography During Liver Surgery 73
control of the proper resection of the tar- resections, but it is mandatory to obtain the
geted nodule verifying its complete inclu- complete tumor clearance: for this purpose
sion in the specimen once removed from IOUS is essential.
the liver (Makuuchi, 1987); the second is
done by checking the cut surface refilled Definition of the Resection Area
with saline to avoid the artifacts generated Definition of the resection area once a non-
by the residual air bubbles and clots. anatomical resection has been chosen is
The aforementioned methods guarantee, similar for both resection for HCC and CRC
whenever possible, anatomical and limited liver metastases. In both these conditions,
resection: this has consequences either for the surgeon’s target should be to determine
the effectiveness of the surgical treatment a resection volume which could result in a
or for its safety. Patients with HCCs who flat and regular cut surface to minimize the
underwent IOUS-guided subsegmentec- risk of postoperative collections.
tomies had better survivals (Makuuchi
et al., 1991; Hasegawa et al., 2005), and Liver Dissection and Post-resectional Control
those with HCCs in contact with intrahe- The same concepts discussed for HCC
patic vessels, once removed under IOUS- regarding IOUS-guided dissection plane
guidance sparing the involved veins did maintenance and IOUS post-resectional
not show local recurrences (Torzilli et al., control are applied in the case of resection
2005a, 2006). Additionally, this kind of for liver metastases.
surgery allows radical surgical treatment The complete surgical clearance of the
of HCC without mortality (Torzilli et al., tumor tissue, even for multiple liver metas-
1999a). Procedures, which are not IOUS- tases, could be justified assuming that
guided, led to dangerous and useless major every single nodule is secondary to the
resection or incomplete operations. primary tumor rather than an expression
of the intrahepatic diffusion of the first
metastatic lesion. For this purpose several
Liver Metastases
authors showed that the occurrence of
At present, tumor clearance can be achieved satellite nodules around the main meta-
in most cases of patients undergoing liver static lesion are rare and wedge resection
resection for metastatic disease with sur- even with tumor-free margin less than
vival benefits even in the presence of 0.5–1 cm in thickness but without expo-
multiple fore sites or vascular infiltration sure of the tumor on the cut surface is jus-
(Minagawa et al., 2000). To achieve this tified (Kokudo et al., 2002; Pawlik et al.,
standard, most of the merits of conserva- 2005). Therefore, the goal to be achieved
tive and radical resections should prob- is the margin clearance: in this sense IOUS
ably be addressed to the extensive use of is fundamental to guarantee the complete
IOUS. In fact, similar to the case of HCC, removal of the lesions. The other goal
IOUS is not only a simple, accurate diag- is the conservativeness of the treatment
nostic modality, but it is a real surgical which should avoid useless parenchyma
device which allows a proper hepatic sacrifice and then safer procedures. Also
resection. In the case of CRC liver in this sense, IOUS with the techniques
metastases there is no need for anatomical described above allows the accomplishment
74 G. Torzilli
of radical but limited resection introduc- The American College of Surgeons has
ing the aforementioned concept of the recently recognized the need for surgeons
safe rounded dissection plane. With this of a specific training in US, meanwhile
approach, it has been possible to carry dedicated monographs have been pub-
out resections without mortality, with no lished almost simultaneously in America
residual gross tumor on the surgical stump and in Europe (Machi and Staren, 2004;
and with a relatively low rate of major Torzilli et al., 1997). The way for a wider
hepatectomies (21.3%) (Minagawa et al., diffusion of ultrasound in surgeons’ prac-
2000). These results were obtained in tice has definitely been opened.
spite of the high mean number of nodules
resected per patient (3.2) and its relatively
large mean size (4.5 cm). REFERENCES
This is a further confirmation that liver Cerwenka, H., Raith, J., Bacher, H., Werkgartner,
resection is an imaging-guided procedure G., el-Shabrawi, A., Kornprat, P., and Mischinger,
and as every interventional imaging-guided H.J. 2003. Is intraoperative ultrasonography
during partial hepatectomy still necessary
procedure, its features are the highest ther- in the age of magnetic resonance imaging?
apeutic efficacy combined with the mini- Hepatogastroenterology 50: 1539–1541.
mal invasiveness. With IOUS aid, it is now Couinaud, C. 1957. Le foie, etude anatomiques et
possible to carry out surgical procedures chirurgicales. Paris, Masson.
which are safe, oncologically radical and Hasegawa, K., Kokudo, N., Imamura, H.,
conservative for the liver function. Mostly Matsuyama, Y., Aoki, T., Minagawa, M., Sano,
K., Sugawara, Y., Takayama, T., and Makuuchi,
because of these reasons, surgery can be M. 2005. Prognostic impact of anatomic resec-
still considered the treatment of choice for tion for hepatocellular carcinoma. Ann. Surg.
most liver metastases, even in the case of 242: 252–259.
the diffusion of other treatments such as Jarnagin, W.R., Bach, A.M., Winston, C.B., Hann,
the thermal ablation which is impaired by L.E., Heffernan, N., Loumeau, T., DeMatteo,
a 39% tumor relapse rate in the first year of R.P., Fong, Y., and Blumgart, L.H. 2001. What is
the yield of intraoperative ultrasonography dur-
follow-up (Solbiati et al., 2001). ing partial hepatectomy for malignant disease? J.
In conclusion, IOUS has acquired pro- Am. Coll. Surg. 192: 577–583.
gressive importance in the field of sur- Kane, R.A., Hughes, L.A., Cua, E.J., Steele,
gical resection in cirrhotic and normal G.D., Jenkins, R.L., and Cady, B. 1994. The
liver, either in the diagnosis and staging impact of intraoperative ultrasonography on
or strictly in the surgical technique, and surgery for liver neoplasms. J. Ultrasound
Med. 13: 1–6.
now it can be considered a fundamental Kojiro, M. 2004. Focus on dysplastic nodules and
tool for hepatobiliary and other surgi- early hepatocellular carcinoma: an Eastern point
cal procedures (Makuuchi et al., 1998). of view. Liver Transpl. 10 (2 Suppl 1): 3–8.
Presently, we can affirm that liver resec- Kokudo, N., Bandai, Y., Imanishi, H., Minagawa,
tion is an imaging-guided procedure and M., Uedera, Y., Harihara, Y., and Makuuchi,
as every interventional imaging-guided M. 1996. Management of new hepatic nodules
detected by intraoperative ultrasonography during
procedure, its features include the highest hepatic resection for hepatocellular carcinoma.
therapeutic efficacy combined with mini- Surgery 119: 634–640.
mal invasiveness. IOUS should be a famil- Kokudo, N., Miki, Y., Sugai, S., Yanagisawa, A., Kato,
iar instrument for the hepatic surgeon. Y., Sakamoto, Y., Yamamoto, J., Yamaguchi, T.,
6. Ultrasonography During Liver Surgery 75
Muto, T., and Makuuchi, M. 2002. Genetic and Quaia, E., Calliada, F., Bertolotto, M., Rossi, S.,
histological assessment of surgical margins in Garioni, L., Rosa, L., and Pozzi-Mucelli, R.
resected liver metastases from colorectal carci- 2004. Characterization of focal liver lesions with
noma: minimum surgical margins for successful contrast-specific US modes and a sulfur hexaflu-
resection. Arch. Surg. 137: 833–840. oride-filled microbubble contrast agent: diag-
Lin, T.Y. 1979. Resectional therapy for primary nostic performance and confidence. Radiology
malignant hepatic tumors. In: International 232: 420–430.
Advances in Surgical Oncology. Murphy, GP. Roncalli, M., Roz, E., Coggi, G., Di Rocco, M.G.,
ed., New York, Alan R Liss, pp. 25–54. Bossi, P., Minola, E., Gambacorta, M., and
Liu, C.L., Fan, S.T., Lo, C.M., Wong, Y., Ng, I.O., Borio, M. 1999. The vascular profile of regen-
Lam, C.M., Poon, R.T., and Wong, J. 2004. erative and dysplastic nodules of the cirrhotic
Abdominal drainage after hepatic resection is liver: implications for diagnosis and classifica-
contraindicated in patients with chronic liver tion. Hepatology 30: 1174–1178.
diseases. Ann. Surg. 239: 194–201. Sahani, D.V., Kalva, S.P., Tanabe, K.K., Hayat,
Machi, J., and Staren, E.D. 2004. Ultrasound for S.M., O’Neill, M.J., Halpern, E.F., Saini, S.,
Surgeons. New York, Lippincott-Williams & and Mueller, P.R. 2004. Intraoperative US in
Wilkins, 2nd Edition. patients undergoing surgery for liver neoplasms:
Machi, J., Isomoto, H., Kurohiji, T., Yamashita, Y., comparison with MR imaging. Radiology 232:
Shirouzu, K., Kakegawa, T., Sigel, B., Zaren, 810–814.
H.A., and Sariego, J. 1991. Accuracy of intra- Solbiati, L., Livraghi, T., Goldberg, S.N., Ierace,
operative ultrasonography in diagnosing liver T., Meloni, F., Dellanoce, M., Cova, L., Halpern,
metastasis from colorectal cancer: evaluation E.F., and Gazelle, G.S. 2001. Percutaneous
with postoperative follow-up results. World J. radiofrequency ablation of hepatic metastases
Surg. 15: 551–556. from colorectal cancer in 117 patients. Radiology
Makuuchi, M. 1987. Abdominal Intraoperative 221: 159–166.
Utrasonography. Tokyo/New York, lgaku-Shoin. Takayama, T., Makuuchi, M., Watanabe, K.,
Makuuchi, M., Yamazaki, S., and Hasegawa, H. Kosuge, T., Takayasu, K., Yamazaki, S., and
1980. Ultrasonically guided liver surgery. Jpn. J. Hasegawa, H. 1991. A new method for mapping
Ultrasonics Med. 7: 45–49. hepatic subsegment: counterstaining identifica-
Makuuchi, M., Takayama, T., Kosuge, T., Yamazaki, tion technique. Surgery 109: 226–229.
S., Yamamoto, J., Hasegawa, H., and Takayasu, Takigawa, Y., Sugawara, Y., Yamamoto, J., Shimada,
K. 1991. The value of ultrasonography for hepatic K., Yamasaki, S., Kosuge, T., and Makuuchi,
surgery. Hepatogastroenterology 38: 64–70. M. 2001. New lesions detected by intraoperative
Makuuchi, M., Torzilli, G., and Machi, J. 1998. ultrasound during liver resection for hepatocellular
History of intraoperative ultrasonography. carcinoma. Ultrasound Med. Biol. 27: 151–156.
Ultrasound Med. Biol. 24: 1229–1242. Torzilli, G., and Makuuchi, M. 2001. Ultrasound-
Minagawa, M., Makuuchi, M., Torzilli, G., guided liver subsegmentectomy: the peculiarity
Takayama, T., Kawasaki, S., Kosuge, T., of segment 4. J. Am. Coll. Surg. 193: 706–708.
Yamamoto, J., and Imamura, H. 2000. Extension Torzilli, G., and Makuuchi, M. 2004. Ultrasound-
of the frontiers of surgical indications in the guided finger compression in liver subsegmen-
treatment of liver metastases from colorectal tectomy for hepatocellular carcinoma. Surg.
cancer: long-term results of our experience. Ann. Endosc. 18: 136–139.
Surg. 231: 487–499. Torzilli, G., Olivari, N., Livraghi, T., and Di Candio,
Pawlik, T.M., Scoggins, C.R., Zorzi, D., Abdalla, G. 1997. Ecografia in chirurgia: modalità diag-
E.K., Andres, A., Eng, C., Curley, S.A., Loyer, nostiche e terapeutiche. Milano, Poletto Editore.
E.M., Muratore, A., Mentha, G., Capussotti, L., Torzilli, G., Makuuchi, M., Inoue, K., Takayama,
and Vauthey, J.N. 2005. Effect of surgical mar- T., Sakamoto, Y., Sugawara, Y., Kubota, K.,
gin status on survival and site of recurrence after and Zucchi, A. 1999a. No-mortality liver resec-
hepatic resection for colorectal metastases. Ann. tion for hepatocellular carcinoma in cirrhotic
Surg. 241: 715–722. and non-cirrhotic patients: is there a way?
76 G. Torzilli
A prospective analysis of our approach. Arch. N., and Makuuchi, M. 2005c. Utility of the
Surg. 134: 984–992. hooking technique for cases of major hepatec-
Torzilli, G., Takayama, T., Hui, A.M., Kubota, K., tomy. Surg. Endosc. 19: 1156–1157.
Harihara, Y., and Makuuchi, M. 1999b. A new Torzilli, G., Montorsi, M., Del Fabbro, D.,
technical aspect of ultrasound-guided liver sur- Palmisano, A., Donadon, M., and Makuuchi
gery. Am. J. Surg. 178: 341–343. M. 2006. Ultrasonographically guided surgical
Torzilli, G., Olivari, N., Moroni, E., Del Fabbro, approach to liver tumours involving the hepatic
D., Gambetti, A., Leoni, P., Montorsi, M., and veins close to the caval confluence. Br. J. Surg.
Makuuchi, M. 2004a. Contrast-enhanced intra- 93: (10)1238–1246.
operative ultrasonography in surgery for hepato- Torzilli, G., Palmisano, A., Del Fabbro, D.,
cellular carcinoma in cirrhosis. Liver Transpl. 10 Marconi, M., Donadon, M., Spinelli, A., Bianchi,
(2 Suppl 1): 34–38. PP., Montorsi, M., 2007. Contrast-enhanced
Torzilli, G., Del Fabbro, D., Olivari, N., Calliada, intraoperative ultrasonograpy during surgery for
F., Montorsi, M., and Makuuchi, M. 2004b. hepatocellular carcinoma in liver cirrhosis: is it
Contrast-enhanced ultrasonography during liver useful or useless? A prospective cohort study
surgery. Br. J. Surg. 91: 1165–1167. of our experience. Ann. Surg. Oncol. 14(4):
Torzilli, G., Montorsi, M., Donadon, M., Palmisano, 1347–1355.
A., Del Fabbro, D., Gambetti, A., Olivari, N., Torzilli, G., Botes, F., Procopio, F., Donadon, M.,
and Makuuchi, M. 2005a. “Radical but con- Balzarini, L., Lutman, F., Calliada, F., Montorsi,
servative” is the main goal for ultrasonography- M. 2008a. Use of contrast-enhanced intraop-
guided liver resection: prospective validation of erative ultrasonograpy during liver surgery for
this approach. J. Am. Coll. Surg. 201: 517–528. colorectal cancer liver metastases. Its impact
Torzilli, G., Del Fabbro, D., Palmisano, A., on operative outcome. Analysis of prospective
Donadon, M., Bianchi, P., Roncalli, M., Balzarini, Cohort study. Eu. J. Cancer 6: 16–23.
L., and Montorsi, M. 2005b. Contrast-enhanced Torzilli, G., Donadon, M., Marconi, M., Botes,
intraoperative ultrasonography during hepatec- F., Palmisano, A., Del Fabbro, D., Procopio,
tomies for colorectal cancer liver metastases. J. F., Montorsi, M. 2008b. Systematic extended
Gastrointest. Surg. 9: 1148–1153. right posterior sectionectomy: a safe and effec-
Torzilli, G., Montorsi, M., Gambetti, A., Del tive alternative to right hepatectomy. Ann. Surg.
Fabbro, D., Donadon, M., Bianchi, P., Olivari, 247(4): 603–611.
A. Treatment
7
Intraoperative Magnetic Resonance
Imaging for Radiofrequency Ablation
of Hepatic Tumors
Oliver F. Bathe and Houman Mahallati
81
82 O.F. Bathe and H. Mahallati
Figure 7.1. To ablate larger lesions, the largest available RFA probe should be utilized (a). However, if
the lesion is larger than the probe, multiple deployments are required (b). Even when this approach is
taken, small regions of the tumor may still be insufficiently ablated (arrow, c)
destruction. During the application of radio- of the alternating current, causing fric-
frequency energy, a high frequency alter- tional heating of the tissue. At tempera-
nating current is passed through the tip of tures above 60°C, intracellular proteins
an electrode into the tissue surrounding become denatured and cell death occurs.
the electrode. The ions within the tissue Radiofrequency ablation treatments typi-
are compelled to move in the direction cally result in local tissue temperatures in
7. Intraoperative Magnetic Resonance Imaging for Radiofrequency Ablation of Hepatic Tumors 83
excess of 100°C, which produces coagu- RFA is typically best for smaller lesions
lative necrosis of the tumor tissue and sur- (i.e., ≤ 3 cm), in the absence of coagulopathy
rounding hepatic parenchyma. or thrombocytopenia.
The temperature is greatest near the tips In patients who can tolerate a general
of the electrode array and falls rapidly anesthetic, operative (open) RFA is cho-
with increasing distance from the elec- sen in some circumstances. Larger lesions
trode. Only small cylindrical regions of typically require multiple deployments
coagulative necrosis are possible when to ablate the entire tumor area and they
using simple monopolar needle electrodes. are frequently best treated by open RFA.
Modern RFA systems typically consist of Lesions that cannot be easily accessed
an array of electrodes that deploy in an percutaneously may be treated by an open
umbrella formation. The result is a sphere procedure. Finally, RFA may be utilized
of tissue destruction with some viable in conjunction with resection in instances
cells closer to the periphery of the sphere. of bilateral hepatic disease (Abdalla et al.,
Because the electrode shaft is insulated, 2004; Elias et al., 2005).
the tract is spared of thermal injury, mak- Hepatocellular carcinoma most fre-
ing transcutaneous administration of RFA quently occurs in patients with underlying
possible. cirrhosis. Therefore, in this situation, treat-
ment options available to any individual
depend on the degree of liver dysfunction.
Indications
The best outcomes have been reported in
In general, RFA is indicated in patients patients undergoing liver transplantation
with primary hepatic malignancies and (Llovet, 2005), but the limited supply of
metastatic liver tumors in the absence donor livers precludes its widespread use.
of extrahepatic disease, in whom resec- Therefore, liver transplantation is typi-
tion is not feasible. Reasons for ineligi- cally reserved for patients with impaired
bility for resection might include poor liver function (i.e., Child B or C) in
general medical condition (precluding a whom there is a limited disease burden
major operation attendant with potentially (Llovet, 2005). In practical terms, there
considerable blood loss, under general is frequently a delay before an organ is
anesthesia) or insufficient hepatic func- available, so (percutaneous) RFA can be
tional reserve. utilized as a temporizing measure, to avoid
Radiofrequency ablation can be admin- tumor growth beyond the limits of criteria
istered percutaneously if lesions are clearly for transplantation (Lu et al., 2005). Other
visible by transcutaneous ultrasound, ena- treatments that must also be considered
bling one to position the probe into the include transarterial chemoembolization
lesion. Another prerequisite for this approach and ethanol ablation.
is that the lesion is easily accessible by Colorectal liver metastases infreque-
that route. Lesions that are poorly situated ntly occur in the setting of cirrhosis,
for the percutaneous approach include so hepatic functional reserve is a lesser
those directly adjacent to the gallbladder or concern. However, in patients pre-
another hollow viscus, or those lying high treated with chemotherapy, some degree of
up near the dome of the liver. Percutaneous hepatic dysfunction may occur secondary
84 O.F. Bathe and H. Mahallati
Table 7.2. Outcomes following radiofrequency ablation for colorectal liver metastases.
# Patients Incomplete Local
Series (# lesions) ablation (%) recurrence (%) Survival
Pearson et al., 1999 46 4.3
Solbiati et al., 2001 117 (179) 2 39.1 46% (3 years)
Machi et al., 2001 (130) 9.2
Bowles et al., 2001 39 30.8 25 months (median)
Kosari et al., 2002 76 6.6
Bleicher et al., 2003 59 18.3
Aloia et al., 2006 30 37 57% (3 years)
(all solitary) 27% (5 years)
Abdalla et al., 2004 57 (110) 9 22% (4 years)
ablated via a percutaneous route (Mulier et al., 2003). There are currently two
et al., 2005). In a meta-analysis, local con- competing thermal ablation technologies:
trol was best for neuroendocrine metastases microwave coagulation therapy (MCT)
(Mulier et al., 2005). In that same study, and laser-induced thermotherapy (LITT).
local recurrence rates for HCC and colorec- Microwave coagulation therapy involves
tal metastases were similar. While several ultrasound-guided placement of an elec-
investigators have demonstrated that prox- trode into a lesion, followed by microwave
imity of a lesion next to a large vessel is treatment. Laser-induced thermotherapy
associated with an increased risk of local involves placement of a laser catheter into
recurrence (Bowles et al., 2001; Machi et al., a lesion and heat photocoagulation using a
2001) due to a heat sink effect, a meta- Nd-YAG laser. Each of these technologies
analysis failed to demonstrate the predic- can be applied during laparotomy or by
tive significance of this factor (Mulier percutaneous techniques.
et al., 2005). As a whole, these data sug- In general, MCT and LITT are advanta-
gest that a percutaneous approach should geous in that multiple electrodes or cath-
be reserved for small, easily accessible eters can be simultaneously placed to ablate
lesions. Moreover, adjunctive techniques larger or more complex liver lesions. Success
should be considered when treating large rates (i.e., complete ablation rates and local
lesions near major blood vessels by RFA. recurrence rates in primary or metastatic
tumors) are reportedly similar with each of
these technologies and with RFA. However,
Alternative Ablative Techniques
there are few direct comparisons between
Cryoablation and ethanol injection are two RFA and LITT or MCT. In one randomized
alternatives that are becoming less fre- controlled trial, RFA and MCT were found
quently used. RFA is better tolerated than to be equally safe; RFA and microwave
cryoablation, as there are few if any sys- ablation were approximately equivalent in
temic side effects, and most complications their ability to completely ablate tumor, but
are minor (de Baere et al., 2003; Pearson microwave ablation on average took less
et al., 1999). In HCC, RFA outcomes time to administer (Shibata et al., 2002).
appear slightly superior to those seen in More such comparisons will be essential
percutaneous ethanol injection (Lencioni for progress in this field.
86 O.F. Bathe and H. Mahallati
imaging may also be useful in supporting recurrence rates should approach those
the lack of significant residual tumor after seen after resection.
ablation. The technology exists to monitor Lesions are often difficult to visualize
intratumoral temperature by MRI (Dick in their entirety with diagnostic modali-
et al., 2003; Puls et al., 2003; Weidens- ties other than MRI. This is especially
teiner et al., 2004). While functional MR true for some neuroendocrine tumors and
techniques have been used in neuroimaging some HCCs. Monitoring the extent of
for some time, it is only relatively recently the ablation is similarly difficult, as the
that these techniques have become appli- ablated region is immediately filled with
cable in abdominal imaging. Magnetic hyperechoic bubbles. The difficulty is
resonance spectroscopy has been used to compounded in large lesions, where mul-
evaluate patients with prostate cancer and tiple overlapping RFA applications must
renal cell carcinoma (Hyslop et al., 2005; be utilized to cover a lesion larger than
Katz-Brull et al., 2005). Kamel et al. the probe (Figure 7.1). In this instance,
(2006) used functional MR techniques as treatment is planned such that the zones
an adjunct to conventional MR imaging to of necrosis overlap, to ensure complete
assess response to chemotherapy (TACE) destruction of the tumor. However, during
in patients with HCC. Early results are the latter RFA applications, the bubbles
promising, and MR imaging has the unique disappear in the initially ablated regions,
advantage of offering a combination of making it impossible to keep track of the
anatomic as well biochemical informa- total extent of the burn. The area of the
tion in assessing tumor response. These burn is not always consistent, particularly
combined approaches may offer improved if there is a nearby large vessel which
sensitivity and specificity in assessing induces a heat sink. Finally, the surgeon
response. cannot always accurately reach the desired
target with the RFA probe. All of these
obstacles can limit the surgeon’s capability
INTRAOPERATIVE to completely ablate any lesion, particu-
MAGNETIC RESONANCE larly with adequate margins (Mulier et al.,
IMAGING AS AN ADJUNCT 2005). Improved intraoperative imaging
may help to promptly identify local treat-
TO RADIOFREQUENCY ment failures, reducing the need for repeat
ABLATION treatments and, perhaps, reducing local
recurrence rates.
Rationale
As described earlier, RFA is a promis-
ing technology that enables ablation of Technical Developments
lesions that may otherwise not be resect- The first intraoperative MRI systems were
able. However, technical limitations limit relatively recently described (Jolesz and
the efficacy of the treatment. Local failure Blumenfeld, 1994; Schenck et al., 1995),
rates and local recurrence rates are unac- and operative interventions guided by
ceptably high, particularly in larger lesions these systems were pioneered by the neu-
that are difficult to access. Ideally, local rosurgical community. The first systems
7. Intraoperative Magnetic Resonance Imaging for Radiofrequency Ablation of Hepatic Tumors 91
consisted of 2 weak magnets (i.e., ≤ 0.5 the 1.5T magnet remains retracted except
tesla (T) ), which were arranged in a verti- when it is utilized for imaging (Sutherland
cal biplanar donut configuration (Schenck et al., 2002). The use of a movable magnet
et al., 1995). The more recent iterations of is more “patient focused” than previously
these systems permitted excellent vertical developed systems. That is, the operating
and lateral exposure between magnets. table is fixed, the patient is positioned
Subsequently, horizontal low-field sys- once, and the magnet moves to the patient,
tems were developed (Steinmeier et al., as opposed to transporting the patient to
1998), but patients had to be transported the magnet. When the magnet is retracted,
between an operating theater and an imag- conventional surgical instruments can be
ing room. In this way, MR incompatibility utilized behind the 5-Gauss line. This
of surgical instruments was no longer an patient-focused approach is safer in that
issue, but patient safety was suboptimal it minimizes the anesthetic risk associated
due to the need for transportation during with patient transport, and it minimizes
the procedure. Using such systems, virtu- disruption of the surgical procedure.
ally any anatomical region could be easily Recently, we have demonstrated the fea-
reached for interventional procedures, as sibility of doing complex open abdominal
the patient could be placed supine, prone operations using this system (Bathe et al.,
or lateral along the longitudinal axis of 2006). Major hepatic surgery would not be
the magnets. However, such systems had readily feasible in a high-field open magnet
limited imaging capabilities and resolution system, given the equipment requirements
due to the low magnetic field strength. and access constraints. Our ability to do
Stronger magnetic field strength increa- surgery with a 1.5T magnet was facilitated
ses signal-to-noise ratio, enhances image by the portability of the magnet. The one
resolution, and accelerates image acquisi- drawback of a closed unit such as the one
tion. Therefore, “high-field” imaging sys- we use is that it does not allow real-time
tems incorporating stronger magnets were monitoring of the ablation or changes in
developed (Sutherland et al., 2002; Truwit intralesional temperature, which is fea-
and Hall, 2001). These 1.5T systems pro- sible for percutaneous procedures (Dick
duced image quality that was equivalent to et al., 2003; Puls et al., 2003). Therefore,
diagnostic scanners. Moreover, the higher a moveable, retractable system is essential
magnetic field strength enabled physi- for safely performing high-field imaging
ologic MR imaging, such as perfusion in the setting of open intracavitary surgery,
MR, spectroscopic MR, and functional but real-time information is not available.
MR. However, because of the enhanced Others have described minimally inva-
magnetic field strength, it was not possible sive intraabdominal and hepatic proce-
to perform procedures that were moni- dures in low magnetic field systems (i.e.,
tored “real-time” by MRI. Rather, the sys- 0.2–0.5T) (Dick et al., 2003; Morikawa
tems were limited to taking static images et al., 2002). Typically, these systems
before, during and after surgery. employ open magnets that enable access
At our institution, the neurosurgeons for percutaneous procedures. The imag-
developed a unique iMRI system that is ing limitations imposed by such low-field
comprised of an operating theatre in which systems limit their use in hepatic surgery.
92 O.F. Bathe and H. Mahallati
Moreover, with a 1.5T magnet temperature MRI are similarly not eligible for this type
mapping of the thermal burn is possible of approach. Finally, because intraopera-
(Weidensteiner et al., 2004), which poten- tive cardiac monitoring is impaired by the
tially provides additional important intra- iMRI system, patients with significant cor-
operative information. Recently, a report onary artery disease are currently excluded
demonstrated that percutaneous RFA was from these procedures.
possible through a closed high-field (1.5T)
system (Mahnken et al., 2004), but open
intracavitary surgery with such a strong Conduct of Intraoperative
magnet has not so far been reported out- Magnetic Resonance Imaging
side of our own experience. for Radiofrequency Ablation
The iMRI system that is used at the
University of Calgary has been described
Potential Indications
in previous reports (Bathe et al., 2006;
Patients currently selected for iMRI as an Sutherland et al., 2002). Briefly, the sys-
adjunct to surgery are those in whom it is tem is based on a moveable 1.5T mag-
anticipated that RFA (with or instead of net designed and constructed by Magnex
resection) is a likelihood and that iMRI UK. The system enables the positioning
might be helpful in making intraopera- of the patient and, because the magnet
tive decisions. Specific indications include enters and exits the room as required,
the following situations: (1) large lesions the patient remains the focal point. The
(≥ 3 cm) that will be treated by RFA; (2) intubated patient is positioned on a MR
recurrent lesions in which RFA will play compatible table that is bolted to the
an essential role; (3) lesions that are vis- floor, with the anesthetic machine at his/
ible on MRI; (4) lesions located at difficult her foot. The hands must be tucked to
sites, such as near potential heat sinks, the side to accommodate the patient in the
close to the gallbladder or any other hol- magnet. An inflatable positioning device
low viscus, or up near the diaphragm. is inserted under the right flank to elevate
The present closed iMRI systems utilize the patient’s right side slightly, enabling
magnets with a small bore. Therefore, a more complete right subcostal incision,
large patients who did not fit into a closed which is essential for open hepatic surgery.
magnet system cannot undergo this proce- The anesthetic tubes are passed along the
dure, until larger bore magnets are devel- left flank to the feet, where the anesthetic
oped. Open systems may be less limited in equipment resides as far from the magnet
this regard, although it is essential that any as feasible.
lesions targeted by this approach are vis- Following induction of anesthesia, base-
ible in the therapeutic magnet. Frequently, line images are obtained and compared to
in patients with a larger body habitus, a diagnostic MRI taken before the date of
lesions may be suboptimally visible if surgery. Sequence parameters are set to
they lie outside of the magnetic “sweet minimize breath-hold time and maximize
spot”. Patients who had ferrous prostheses coverage. Time to repetition (TR), echo
or other ferrous foreign bodies that would time (TE), flip angle, echo train lengths
normally be a contraindication to diagnostic (ETL), and slice thickness are all adjusted
7. Intraoperative Magnetic Resonance Imaging for Radiofrequency Ablation of Hepatic Tumors 93
to optimize imaging of the liver in each check of probe position is made with ultra-
patient. All patients are imaged with res- sound. The protocol for increases in elec-
piration suspended and monitored by the trical power delivered to the probe is set
attending anesthesiologist. Axial precon- by the manufacturer, and this is dependent
trast 2D T1 weighted gradient echo (GRE) on probe size. Larger probes require more
images are obtained with or without fat power to achieve the target temperature
saturation, and axial 2D fast spin echo throughout the volume of the burn region.
(FSE) T2 weighted images with fat satura- Where multiple deployments are required
tion are obtained. An intravenous bolus to cover the entire tumor volume, it is best
of 0.15 mmol/kg gadopentetate dimeg- to do the first deployments at the least
lumine (Gd-DTPA; Magnevist; Berlex accessible part of the tumor. Usually, that
Laboratories, Wayne, NJ) is then admin- means targeting the most posterior and
istered, and postcontrast T1 weighted superior aspects of the lesion. Thereafter,
fat saturated images are acquired, with the probe is pulled back 2 cm, the tines
parameters identical to the precontrast T1 are redeployed, and another thermal abla-
weighted fat saturated images. All target tion cycle is applied. This is repeated
lesions are visualized, measured, and until it is estimated that the entire lesion
localized relative to nonmobile extrahe- has been ablated. It is best if the regions
patic landmarks (e.g., vertebral body and ablated overlap, to ensure that no areas
liver capsule). In particular, it is important are exposed to insufficient heat to induce
to measure the distances from these non- coagulative necrosis. The latter ablations
mobile landmarks to the tumor edges. The are more difficult to target accurately with
mobile magnet is then retracted, allowing ultrasound, as the lesion fills with bubbles
the safe introduction of normal surgical after the first ablation. Therefore, if it is
instruments to the surgical field. anticipated that multiple deployments will
Following a laparotomy, the liver is be required, it is essential to plan probe
mobilized as necessary to permit access placements prior to initiating the ablation
to the targeted hepatic segments. An intra- for each lesion.
operative ultrasound is performed, and In general, we take an aggressive
decisions regarding the final operative approach to thoroughly burn each of the
approach are made. It is feasible using target lesions. For lesions that are close to
this system to do a resection, if that is the gallbladder, it is essential to remove
considered appropriate. RFA is performed the gallbladder to avoid perforation sec-
for any remaining lesions, using as large a ondary to thermal injury. For lesions that
probe as required to cover as much of the are located near the center of the liver, care
tumor field as possible. Currently, probes must be taken to avoid a thermal injury
up to 5 cm are available. If it is expected to the bile ducts. Injury to the bile ducts
that the probe will cover the entire diam- may be manifested postoperatively by a
eter of the tumor in a single deployment, bile leak, a bile duct stricture, or complete
the probe is initially guided with the aid of disruption of the bile ducts; such compli-
ultrasound to the center of the tumor. Once cations add considerable morbidity to the
the probe tip is at the desired region of the procedure. If necessary, the bile ducts can
tumor, the tines are deployed and a final be infused with cooled saline via the cystic
94 O.F. Bathe and H. Mahallati
Figure 7.5. Intraoperative MRI using a hepatocyte-specific contrast agent. (a) A pre-ablation T1 weighted
image after Gd-EOB-DTPA (Primovist) shows tumor as hypointense against a hyperintense liver. (b) The
intradissectional image demonstrates the ablation zone relative to tumor. The thermal ablation zone is
more T1 hyperintense peripherally, compared to surrounding liver. The central hypointensity represents
tumor bed, which is excessively close to the ablative margin at the anterior aspect (white arrow). Further
deployments were required to improve this margin. (c, d) Images taken 2 weeks later confirmed that the
final ablation margin was sufficient. Following administration of Gd-EOB-DTPA, the T1 weighted image
(c) demonstrates the extent of the ablation bed. The subtraction image confirms the absence of residual
tumor. (e, f) Images taken 8 months later demonstrate no tumor recurrence. Post-contrast (Gd-DTPA)
3D T1 weighted images (VIBE) demonstrate the ablation bed (e). The subtraction images confirm no
enhancement in the RFA bed (f)
96 O.F. Bathe and H. Mahallati
Developments in Intraoperative
Magnetic Resonance Imaging
The magnet itself must be reengineered to
enhance the utility of MRI in a wide variety
of clinical circumstances. It will be essen-
tial to engineer such a magnet so that the
magnetic “sweet spot” is large enough to
encompass the targeted organ. Finally, the
magnet bore must be built so that it is large
enough to accommodate larger patients, as
the prevalence of obesity is increasing.
The ideal iMRI system would allow
Figure 7.6. Lesions in segment IV of the liver rapid, high quality imaging with combined
(white arrow) can be obscured by the aortic “ghost”
artefact. If the lesion is veiled by aortic ghost, then
morphologic and functional information.
tilting the patient effectively separates the target Systems must be built with a stronger
lesion from the artefact. magnetic field (e.g., ≥ 3.0T), as this ena-
bles faster image acquisition, greater reso-
lution, and better functional imaging. On
by aortic ghost, then tilting the patient the other hand, MRI using stronger mag-
effectively separates the target lesion from nets is associated with some disadvantages
the artefact (Figure 7.6). For this reason, such as increased susceptibility artefacts,
we routinely place air bags beneath each increased energy deposition, and altered
flank, enabling repositioning with less T1 properties of tissues (Hyslop et al.,
trouble. Third, we have found that image 2005; Lauzon et al., 2006). In the case of
quality is unacceptable in large patients. intraoperative MR, the increased suscep-
Not only is it difficult to fit these patients tibility artefacts warrant special attention.
into the magnet, it is impossible to image The presence of air/tissue interfaces in the
the anterior aspect of their liver. Each of surgical field can lead to signal loss from
these limitations can be addressed with susceptibility and may decrease the image
specific design changes to the magnet (see quality, negating the advantages of high
below). field imaging. Sequences and systems will
need to be optimized to minimize artefacts
in order to gain the maximum potential
Future Developments
from higher field intra-operative systems.
Image-guided ablation is an important Systems should ideally offer targeting
alternative to resection. However, to capabilities and real-time monitoring of
ensure that its therapeutic effectiveness therapies such as RFA. Magnetic resonance
better resembles that of resection, specific has been used to monitor RF ablation in
technological developments will be neces- real-time in animal models (Lepetit-Coiffe
sary. This is particularly important for the et al., 2006), but this is not yet feasible in
treatment of larger lesions, which have so high-field systems so far described in clin-
far infrequently been treated using these ical use. Real-time MR monitoring of RFA
techniques. would be greatly enhanced by robotics,
7. Intraoperative Magnetic Resonance Imaging for Radiofrequency Ablation of Hepatic Tumors 97
which would allow the remote insertion noise, impairing the acquisition of quality
of instruments into the lesion (McBeth images when monitoring the progression
et al., 2004). Robots are indefatigable, of the burn. To address the latter issue,
have superior spatial resolution and excel- Zhang et al. (1998) have described a
lent geometric accuracy. The development switching circuit to reduce this noise.
of an MR-compatible robot would there- While we have only had experience with
fore facilitate real-time monitoring of RFA RFA, MCT and LITT could be explored
of liver lesions. in the context of MR-guided ablation of
Finally, iMRI could benefit from the liver lesions. The MRI characteristics of
development of novel contrast agents. ablation zones produced by each of these
While conventional extracellular space methods are unlikely to differ significantly
agents are useful in determining enhance- from those produced by RFA (Braga and
ment of post-RFA beds and potential tumor Semelka, 2005). MR-guided insertion of
recurrence (Dromain et al., 2002; Kuszyk microwave electrodes and real-time moni-
et al., 2000), in our experience, the tem- toring of the ablation by MRI has success-
poral window offered by such agents and fully been performed (Morikawa et al.,
the dose limitations (Kirchin and Runge, 2002). Laser-induced thermotherapy may
2003) limit their utility in the intraopera- be particularly suited for MR imaging
tive setting. Ideally, a contrast agent that because laser fibers and catheter systems
has greater retention and better definition are not affected by and do not interfere with
of the target lesion relative to the extent of MR signal (Puls et al., 2003). Puls et al.
thermal injury would be useful. Gd-EOB- (2003) recently described a new miniatur-
DTPA fulfills these requirements, but it ized irrigated laser catheter system that
may not be as good for lesions containing can be inserted percutaneously under MR
functioning hepatocytes, such as hepatic guidance and that allows monitoring of
adenoma or well differentiated hepatocel- therapy in a closed high-field MR imager.
lular carcinoma. Tumor-specific contrast In conclusion, radiofrequency ablation
agents may be useful. Necrosis-avid agents may occasionally cause incomplete tumor
and high molecular weight intravascular necrosis. Moreover, the local recurrence
agents have been used in animal models rates following RFA of liver tumors are
to assess response to RFA therapy (Kim high, especially in larger tumors that are
et al., 2005; Ni et al., 2006), but we have difficult to access. One contributing fac-
not yet tested them in the clinical context. tor to this poor local control rate is the
difficulty in monitoring thermal ablations
Development of Alternative Ablative of larger tumors in which multiple RFA
Techniques deployments are required. Additionally,
Radiofrequency ablation has some dis- the margins of ablation may be inadequate.
advantages, when utilized in the context Intraoperative MRI enables the surgeon
of MRI. If real-time monitoring of the to better estimate the technical success of
ablation is to be done using MRI, then it RFA as well as the size of the ablative mar-
is essential to utilize MR-compatible tita- gins. However, it is still unknown whether
nium probes, which are quite expensive. iMRI changes local recurrence rates.
Moreover, RFA produces electromagnetic Complex hepatic surgeries (i.e., resections
98 O.F. Bathe and H. Mahallati
and/or RFA) are feasible and safe in the Radiofrequency ablation in 447 complex unre-
University of Calgary iMRI suite. The 1.5T sectable liver tumors: lessons learned. Ann. Surg.
Oncol. 10:52–58.
system enables high resolution imaging.
Bowles, B.J., Machi, J., Limm, W.M., Severino,
Our initial experience has identified some R., Oishi, A.J., Furumoto, N.L., Wong, L.L., and
technical limitations to our system. This Oishi, R.H. 2001. Safety and efficacy of radi-
experience will point the way to future ofrequency thermal ablation in advanced liver
system modifications. tumors. Arch. Surg. 136:864–869.
Braga, L., and Semelka, R. 2005. Magnetic reso-
nance imaging features of focal liver lesions
REFERENCES
after intervention. Top Magn. Reson. Imaging
Abdalla, E., Vauthey, J., Ellis, L., Ellis, V., Pollock, 16:99–106.
R., Broglio, K., Hess, K., and Curley, S. 2004. deBaere, T., Risse, O., Kuoch, V., Dromain, C.,
Recurrence and outcomes following hepatic Sengel, C., Smayra, T., Din, M.G.E., Letoublon,
resection, radiofrequency ablation, and com- C., and Elias, D. 2003. Adverse events during
bined resection/ablation for colorectal liver radiofrequency treatment of 582 hepatic tumors.
metastases. Ann. Surg. 239:818–825. Am. J. Roentgenol. 181:695–700.
Aliberti, C., Soriani, M., Tilli, M., Benea, G., Giorgi, Dick, E., Wragg, P., Joarder, R., Jode, M.D., Lamb,
U.D., and Fiorentini, G. 2004. Radiofrequency G., Gould, S., and Gedroyc, W. 2003. Feasibility
ablation of liver malignancies: MRI for evalua- of abdomino-pelvic T1-weighted real-time ther-
tion of response. J. Chemother. 16:79–81. mal mapping of laser ablation. J. Magn. Reson.
Aloia, T.A., Vauthey, J.N., Loyer, E.M., Ribero, D., Imaging 17:197–205.
Pawlik, T.M., Wei, S.H., Curley, S.A., Zorzi, D., Dromain, C., de Baere, T., Elias, D., Kuoch, V.,
and Abdalla, E.K. 2006. Solitary colorectal liver Ducreux, M., Boige, V., Petrow, P., Roche, A.,
metastasis: resection determines outcome. Arch. and Sigal, R. 2002. Hepatic tumors treated with
Surg. 141:460–466. percutaneous radio-frequency ablation: CT and
Arata, M., Nisenbaum, H., Clark, T., and Soulen, MR imaging follow-up. Radiology 223:255–262.
M. 2001. Percutaneous radiofrequency ablation Elias, D., Baton, O., Sideris, L., Boige, V., Malka,
of liver tumors with the LeVeen probe: is roll-off D., Liberale, G., Pocard, M., and Lasser, P. 2005.
predictive of response? J. Vasc. Interv. Radiol. Hepatectomy plus intraoperative radiofrequency
12:455–458. ablation and chemotherapy to treat technically
Asbach, P., Klessen, C., Kroencke, T.J., Kluner, C., unresectable multiple colorectal liver metas-
Stemmer, A., Hamm, B., and Taupitz, M. 2005. tases. J. Surg. Oncol. 90:36–42.
Magnetic resonance cholangiopancreatography Frich, L., Bjørnerud, A., Fossheim, S., Tilung, T.,
using a free-breathing T2-weighted turbo spin- and Gladhaug, I. 2004. Experimental application
echo sequence with navigator-triggered pro- of thermosensitive paramagnetic liposomes for
spective acquisition correction. Magn. Reson. monitoring magnetic resonance imaging guided
Imaging 23:939–945. thermal ablation. Magn. Reson. Imaging 52:
Bathe, O., Mahallati, H., Sutherland, F., Dixon, E., 1302–1309.
Pasieka, J., and Sutherland, G. 2006. Complex Goldberg, S.N., Gazelle, G.S., Compton, C.C.,
hepatic surgery aided by a 1.5-tesla moveable Mueller, P.R., and Tanabe, K.K. 2000. Treatment
magnetic resonance imaging system. Am. J. of intrahepatic malignancy with radiofrequency
Surg. 191:598–603. ablation: radiologic-pathologic correlation.
Berber, E., Pelley, R., and Siperstein, A. 2005. Cancer 88:2452–2463.
Predictors of survival after radiofrequency ther- Huppertz, A., Balzer, T., Blakeborough, A., Breuer,
mal ablation of colorectal cancer metastases to J., Giovagnoni, A., Heinz-Peer, G., Laniado, M.,
the liver: a prospective study. J. Clin. Oncol. Manfredi, R.M., Mathieu, D.G., Mueller, D.,
23:1358–1364. Reimer, P., Robinson, P.J., Strotzer, M., Taupitz,
Bleicher, R.J., Allegra, D.P., Nora, D.T., Wood, M., and Vogl, T.J. 2004. Improved detection of
T.F., Foshag, L.J., and Bilchik, A.J. 2003. focal liver lesions at MR imaging: multicenter
7. Intraoperative Magnetic Resonance Imaging for Radiofrequency Ablation of Hepatic Tumors 99
and Oishi, R.H. 2001. Ultrasound-guided radi- Puls, R., Stroszczynski, C., Gaffke, G., Hosten, N.,
ofrequency thermal ablation of liver tumors: Felix, R., and Speck, U. 2003. Laser-induced
percutaneous, laparoscopic, and open surgical thermotherapy (LITT) of liver metastases:
approaches. J. Gastrointest. Surg. 5:477–489. MR-guided percutaneous insertion of an MRI-
Mahnken, A., Buecker, A., Spuentrup, E., compatible irrigated microcatheter system using
Krombach, G., Henzler, D., Gunther, R., and a closed high-field unit. J. Magn. Reson. Imaging
Tacke, J. 2004. MR-guided radiofrequency abla- 17:663–670.
tion of hepatic malignancies at 1.5 T: initial Schenck, J.F., Jolesz, F.A., Roemer, P.B., Cline,
results. J. Magn. Reson. Imaging 19:342–348. H.E., Lorensen, W.E., Kikinis, R., Silverman,
Martin, D.R., and Semelka, R.C. 2001. Imaging of S.G., Hardy, C.J., Barber, W.D., and Laskaris,
benign and malignant focal liver lesions. Magn. E.T. 1995. Superconducting open-configuration
Reson. Imaging Clin. N. Am. 9:785–802. MR imaging system for image-guided therapy.
McBeth, P.B., Louw, D.F., Rizun, P.R., and Radiology 195:805–814.
Sutherland, G.R. 2004. Robotics in neurosur- Shibata, T., Iimuro, Y., and Yamamoto, Y. 2002. Small
gery. Am. J. Surg. 188:68S–75S. hepatocellular carcinoma: comparison of radio-
Morikawa, S., Inubushi, T., Kurumi, Y., Naka, S., frequency ablation and percutaneous microwave
Sato, K., Tani, T., Yamamoto, I., and Fujimura, coagulation therapy. Radiology 223:331–337.
M. 2002. MR-guided microwave thermocoagu- Solbiati, L., Livraghi, T., Goldberg, S.N., Ierace,
lation therapy of liver tumors: initial clinical T., Meloni, F., Dellanoce, M., Cova, L., Halpern,
experiences using a 0.5 T open MR system. J. E.F., and Gazelle, G.S. 2001. Percutaneous
Magn. Reson. Imaging 16:576–583. radio-frequency ablation of hepatic metastases
Mulier, S., Ni, Y., Jamart, J., Ruers, T., Marchal, from colorectal cancer: long-term results in 117
G., and Michel, L. 2005. Local recurrence after patients. Radiology 221:159–166.
hepatic radiofrequency coagulation: multivariate Solbiati, L., Ierace, T., Tonolini, M., and Cova, L.
meta-analysis and review of contributing factors. 2004. Guidance and monitoring of radiofrequency
Ann. Surg. 242:158–171. liver tumor ablation with contrast-enhanced
Ni, Y., Chen, F., Mulier, S., Sun, X., Yu, J., ultrasound. Eur. J. Radiol. 51:S19–S23.
Landuyt, W., Marchal, G., and Verbruggen, Steinmeier, R., Fahlbusch, R., Ganslandt, O.,
A. 2006. Magnetic resonance imaging after Nimsky, C., Buchfelder, M., Kaus, M., Heigl,
radiofrequency ablation in a rodent model of T., Lenz, G., Kuth, R., and Huk, W. 1998.
liver tumor: tissue characterization using a Intraoperative magnetic resonance imaging with
novel necrosis-avid contrast agent. Eur. Radiol. the magnetom open scanner: concepts, neurosur-
16:1031–1040. gical indications, and procedures: a preliminary
Noone, T.C., Semelka, R.C., Chaney, and D.M., report. Neurosurgery 43:739–747.
Reinhold, C. 2004. Abdominal imaging studies: Sutherland, G., Kaibara, T., and Louw, D. 2002.
comparison of diagnostic accuracies resulting Intraoperative MR at 1.5 tesla - experience and
from ultrasound, computed tomography, and future directions. Acta Neurochir. 85:21–28.
magnetic resonance imaging in the same indi- Therasse, P., Arbuck, S.G., Eisenhauer, E.A.,
vidual. Magn. Reson. Imaging 22:19–24. Wanders, J., Kaplan, R.S., Rubinstein, L.,
Pawlik, T.M., Scoggins, C.R., Zorzi, D., Abdalla, Verweij, J., vanGlabbeke, M., van Oosterom, A.T.,
E.K., Andres, A., Eng, C., Curley, S.A., Loyer, Christian, M.C., and Gwyther, S.G. 2000. New
E.M., Muratore, A., Mentha, G., Capussotti, L., guidelines to evaluate the response to treatment in
and Vauthey, J.N. 2005. Effect of surgical mar- solid tumors. J. Natl. Cancer Inst. 92:205–216.
gin status on survival and site of recurrence after Truwit, C., and Hall, W. 2001. Intraoperative MR
hepatic resection for colorectal metastases. Ann. systems: high-field approaches. Neuroimaging
Surg. 241:715–724. Clin. N. Am. 11:645–650.
Pearson, A.S., Izzo, F., Fleming, R.Y., Ellis, L.M., Veenendaal, L.M., Rinkes, I.H., Lips, C.J., and van
Delrio, P., Roh, M.S., Granchi, J., and Curley, Hillegersberg, R. 2006. Liver metastases of neu-
S.A. 1999. Intraoperative radiofrequency abla- roendocrine tumours; early reduction of tumour
tion or cryoablation for hepatic malignancies. load to improve life expectancy. World J. Surg.
Am. J. Surg. 178:592–599. Oncol. 4:35.
7. Intraoperative Magnetic Resonance Imaging for Radiofrequency Ablation of Hepatic Tumors 101
Weidensteiner, C., Kerioui, N., Quesson, B., Yu, H., Cheng, J., Lai, K., Lin, C., Lo, G., Lin, C.,
Senneville, B.D., Trillaud, H., and Moonen, C. Hsu, P., Chan, H., Lo, C., Tsai, W., and Chen,
2004. Stability of real-time MR temperature W. 2005. Factors for early tumor recurrence
mapping in healthy and diseased human liver. J. of single small hepatocellular carcinoma after
Magn. Reson. Imaging 19:438–446. percutaneous radiofrequency ablation therapy.
Weinmann, H., Schuhmann-Giampieri, G., Schmitt- World J. Gastroenterol. 11:1439–1444.
Willich, H., Vogler, H., Frenzel, T., and Gries, H. Zhang, Q., Cheung, Y.-C., Lewin, J., and
1991. A new lipophilic gadolinium chelate as a Duerk, J. 1998. A method for simultaneous
tissue-specific contrast medium for MRI. Magn. RF ablation and MRI. J. Magn. Reson. Imaging
Reson. Med. 22:233–237. 8:110–114.
8
Surgically Unresectable and
Chemotherapy-Refractory Metastatic Liver
Carcinoma: Treatment with Yttrium-90
Microsphere Followed by Assessment
with Positron Emission Tomography
Ching-Yee Oliver Wong
103
104 C.-Y.O. Wong
particles and to evaluate the amount of shunting fraction (F) for the given physi-
pulmonary and gastrointestinal flow. ology and anatomy of the hepatic tumors.
All treated patients are then evaluated Prior to treatment, planar anterior 99mTc
to confirm an acceptable risk of shunting MAA scans (Figure 8.1) are obtained
of Yttrium-90 microsphere to the lungs within 30 min after intrahepatic arterial
(< 30 Gy) (Wong et al., 2004, 2005). In injection of 150 MBq 99mTc MAA during
patients with hepatic tumors, a portion of angiography to include the lungs, liver,
the arterial supply may bypass the capil- and stomach (Figure 8.2). Recently, single
lary bed and drain directly into venous photon emission computed tomography
system. The microspheres that are not (SPECT) or combined SPECT/CT is also
trapped in the liver will be shunted via the utilized after planar scintigraphic imag-
heart and deposited in the lungs. Thus, a ing to define more precisely the hepatic
lung shunting measurement is essential distribution of MAA (Figure 8.3), which
before the treatment to obtain a unique mimics the subsequent Y-90 microsphere
Figure 8.1. Planar Tc-99m MAA scan (left panel = anterior view; right panel = posterior view), showing
distribution to the right lobe only (expected from the injection into the right hepatic artery) and hetero-
geneity due to tumor uptake
Figure 8.2. The regions of interest for calculating lung shunting fraction (F)
8. Surgically Unresectable and Chemotherapy-Refractory Metastatic Liver Carcinoma 107
Figure 8.3. Tc-99m MAA SPECT images showing the localization of tracer in the tumor, which mirrors
the subsequent targeted distribution of Y-90 microspheres
distribution within the normal liver and exposure to lung to be < 30 Gy. The treat-
metastatic tumor lesions. Lung shunting ment dose needs to be further verified by
(F) is assessed using intra-hepatic arterial the experienced physicists according to
99m
Tc MAA images [F = 100% × lung the formulae outlined later. The method
counts/(lung + liver + stomach counts)] of lobar treatment for Y-90 microsphere
(Wong et al., 2004). Lung activity < 10% radioembolization has recently been used
is considered insignificant for any dose by some investigators (Wong et al., 2004,
or activity adjustment (Figure 8.4). The 2005), which differs from the whole liver
lung fraction, even in the setting of F approach reported for hepatocellular
being > 10%, is used to adjust treatment carcinoma (Dancey et al., 2000; Herba
dose or activity to optimize the radiation et al., 1988) and metastatic colorectal
dose given to each lobe of liver or the cancer (Gray et al., 1992) by other inves-
whole liver while maintaining radiation tigators. In the lobar method, each lobe of
108 C.-Y.O. Wong
the liver is treated separately using a lobar Activity (GBq) = (BSA − 0.2) + (Tumor
arterial injection. volume/Total liver volume)
If there is no observable and uncor-
where BSA (m2) = 0.20247[height (m)]0.725
rectable gastrointestinal flow and if liver
[weight (kg)]0.425
reserve and tumor vascular anatomy are
With the calculated activity, the dose
judged to be suitable, a treatment plan
can be deduced from the activity – dose
employing either a lobar or whole liver
formula. Despite the different approach
treatment strategy, is constructed for each
that is used between Y-90 resin-bound
patient. As the treatment is a form of
and glass matrix microspheres, a level of
brachytherapy, the radiation is directly
radioactivity ranging from 0.5 to 2.5 GBq
imposed into the tumor by intraarterial
of Y-90 microspheres is usually pre-
delivery. Unlike conventional radiother-
scribed for left or right lobe or whole liver
apy, only target tumor lesions are identi-
based on the mass (Wong et al., 2004,
fied by CT and PET without the need of
2005; Stubbs et al., 2001). The activity is
drawing gross tumor volumes and their
generally higher for glass microspheres
margin. The targeted liver volume is,
than resin based microspheres due to
however, measured by CT. The amount
different activity uptake fractions to the
of activity required for the target dose of
tumor (TUF) (Wong et al., 2004, 2005).
Yttrium-90 microsphere is calculated (if
Using this technique, normal liver tis-
ignoring lung shunt) using the following
sue exposure will be less than the level
formula (Ho et al., 1997):
above which complications have been
Activity (GBq) = Tumor Dose (Gy) x Tumor Mass (kg) reported using external beam radiother-
50 x Tumor uptake fraction
apy (Campbell et al., 1997).
The prescribed dose of Y-90 will take
The tumor uptake fraction (TUF) depends
the amount of lung shunting into consider-
on the lung shunting fraction and tumor
ation. It has been previously estimated that
vascularity factor. For Y-90 glass micro-
the radiation dose to the lungs assuming a
spheres, TUF is assigned to be 1 while for
uniform microsphere distribution using the
Y-90 resin microspheres, TUF is adjusted
formula (Berger, 1971):
with lung shunting fraction (F) as: TUF = 1
− F. The mass of the targeted lobe was Lung Dose (Gy) = Activity (GBq) x F x 50
Mass of lungs (kg)
determined using CT images for volume
calculation with a conversion factor of
1.03 g/cm3. The target dose level for Y-90 when the total lung mass, including blood,
glass microsphere is set ~100–120 Gy to was assumed to be 1 kg (Snyder et al.,
balance between tumor response rates and 1975–1976), the lung radiation dose is
the risk of normal hepatocellular necrosis simply:
(Kennedy et al., 2007). The required activ- Lung Dose (Gy/kg) = Activity (GBq) x F x 50
ity of Y-90 can be calculated directly from
the above formula. An alternative approach, Patients are considered eligible for
which has been proposed for Y-90 resin Yttrium-90 microsphere treatment if the
microsphere, is to calculate the activity first lung shunt fraction can only result in
based on the body surface area: an estimated dose < 30 Gy to the lungs.
8. Surgically Unresectable and Chemotherapy-Refractory Metastatic Liver Carcinoma 109
To prevent any possible radiation pneu- be scheduled for treatment of the oppo-
monitis, the cumulative radiation dose to site lobe 30–60 days following initial
lungs is also limited to a maximum of treatment. To determine eligibility for
30 Gy. The activity prescribed may also retreatment, patients have repeat labora-
be reduced if the hepatic function is com- tory tests including hepatic function panel
promised by as much as 30% (Kennedy and cross-sectional imaging, similar to the
et al., 2007). All these considerations first treatment. At the time of retreatment
are balanced against the assessment of evaluation, consideration of the total dose
whether the final adjusted radiation dose to the lungs is based on the cumulative
or activity will be enough to deliver to dose over all prior treatments so as not to
treatment response on the target lesions exceed a total estimated cumulative dose
based on the following formula (Ho et al., of 30 Gy to the lungs.
1997; Wong et al., 2004).
Delivered Dose (Gy) =Delivered Activity (GBq) x (1 - F) x 50
Liver Lobe Mass (kg)
CLINICAL RESULTS
where the lung shunt fraction (F) is
obtained as the ratio of counts in the Prior data have suggested that PET is
lungs to the total counts in the abdo- useful in assessing therapy for hepatic
men and lungs, as measured by intra- colorectal metastases (Ho et al., 1997;
hepatic arterial 99mTc-MAA scan. After Selzner and Hany, 2004). Positron
the appropriate activity has drawn into emission tomography has changed the
the delivery device, Yttrium-90 micro- conventional management of patients with
spheres are injected into a percutaneous liver metastases due to its enhanced ability
catheter inserted via the femoral artery that to detect recurrent or metastatic lesions
directed to the targeted liver lobe or the over CT (Meta et al., 2001; Rohren et al.,
entire liver. The administration of Y-90 is 2002). The former has also been known to
done by slow intraarterial infusion at low be more accurate than the latter in detect-
pressure under strict aseptic conditions, ing liver metastases (Wong et al., 2002;
fluoroscopic guidance, institutional radia- Valk et al., 1999; Ogunbiyi et al., 1997).
tion safety guidelines, and the manufac- A prior prospective study on Y-90 glass
turer’s instructions. Treatment is usually microsphere has shown that it is feasible
delivered on an outpatient basis. Patients to quantify reduction of hepatic tumor
are observed in recovery room for 4–6 h metabolism objectively after Y-90 glass
for stability of vital signs, bleeding com- microsphere treatment for unresectable
plications or signs or symptoms of acute metastatic colorectal cancer to the liver
post-embolization syndrome. Patients can (Wong et al., 2004). The total SUV of the
be discharged home when they are clini- entire axial slices of liver agrees well with
cally stable and without complaints. Liver visual evaluation. The PET findings are
function tests are monitored on all patients, also correlated with serum tumor marker,
before and after treatment. suggesting that the reduction of metabo-
Patients with bilobar disease, if not lism is visually, quantitatively, and bio-
yet treated by whole liver approach, can logically related to the tumoricidal effects
110 C.-Y.O. Wong
Figure 8.5. Pre-treatment PET-CT scan showing liver metastasis. CT shows one of the two hypodense
lesions in the posterior segment of the right lobe of the liver. PET shows the two foci of high metabolic
activity within the liver metastases
112 C.-Y.O. Wong
Figure 8.6. Post-treatment PET-CT showing response to Y-90 therapy in the liver (with persistent
hypodensity in the posterior segment on CT but resolution of the foci with high metabolic activity seen
on the pre-treatment PET within the liver)
2.3 × 2.3 and 3.9 × 3.4 cm. But PET/CT Dancey, J.E., Shepherd, F.A., Paul, K., Sniderman,
showed drastic improvement in metabolism K.W., Houle, S., Gabrys, J., Hendler, A.L., and
Goin, J.E. 2000. Treatment of nonresectable
(Figure 8.6). She had transient twofold ele-
hepatocellular carcinoma with intrahepatic 90Y-
vation of transaminases. Her CEA decreased microspheres. J. Nucl. Med. 41:1673–1681.
to 5.8 ng/dl 3 months after Y-90 and the liver Drbe, E.M., and Day, D.E. 1993. Chemical durabil-
function tests returned to normal levels. ity of Y2O3-Al2O3-SiO2 glasses for the in vivo
delivery of beta radiation. J. Biomed. Mat. Res.
27:1301–1308.
Eubank, W.B., Mankoff, D.A., Vesselle, H.J., Eary,
REFERENCES
J.F., Schubert, E.K., Dunnwald, L.K., Lindsley,
Arulampalam, T.H.A., Costa, D.C., Loizidou, M., S.K., Gralow, J.R., Austin-Seymour, M.M., Ellis,
Visvikis, D., Ell, P.J., and Taylor, I. 2001. Positron G.K., and Livingston, R.B. 2002. Detection of
emission tomography and colorectal cancer. Br. J. locoregional and distant recurrences in breast can-
Surg. 88:176–189. cer patients by using FDG PET. Radiographics
Barker, D.W., Zagoria, R.J., Morton, K.A., 22:5–17.
Kavanagh, P.V., and Shen, P. 2005. Evaluation Fernandez, F.G., Drebin, J.A., Linehan, D.C.,
of liver metastases after radiofrequency ablation: Dehdashti, F., Siegel, B.A., and Strasberg,
Utility of 18F-FDG PET and PET/CT. Am. J. S.M. 2004. Five-year survival after resection
Roentgenol. 184:1096–1102. of hepatic metastases from colorectal cancer in
Berger, M.J. 1971. Distribution of absorbed dose patients screened by positron emission tomogra-
around point sources of electrons and beta par- phy with F-18 fluorodeoxyglucose (FDG- PET).
ticles in water and other media. J. Nucl. Med. Ann. Surg. 240:438–447.
12(Suppl 5):5–23. NM/MIRD Pamphlet No. 7. Gray, B.N., Anderson, J.E., Burton, M.A., van
Campbell, A.M., Bailey, I.H., and Burton, M.A. Hazel, G., Codde, J., Morgan, C., and Klemp, P.
1997. Tumour dosimetry in human liver follow- 1992. Regression of liver metastases following
ing hepatic Yttrium-90 microsphere therapy. treatment with Yttrium-90 microspheres. Aust.
Eur. J. Nucl. Med. 24:293–298. N. Z. J. Surg. 62:105–110.
8. Surgically Unresectable and Chemotherapy-Refractory Metastatic Liver Carcinoma 113
Herba, M.J., Illescas, F.F., Thirlwell, M.P., Boos, rate in man with (F-18) fluorodeoxyglucose:
G.J., Rosenthall, L., Atri, M., and Bret, P.M. Validation of method. Ann. Neurol. 6:371–388.
1988. Hepatic malignancies: Improved treat- Rilling, W.S., and Drooz, A. 2002. Multidisciplinary
ment with intraarterial Yttrium-90. Radiology management of hepatocellular carcinoma. J.
169:311–314. Vasc. Interv. Radiol. 13(Suppl 9):S259–S263.
Ho, S., Lau, W.Y., Leung, T.W.T., Chan, M., Rohren, E.M., Paulson, E.K., Hagge, R., Wong,
Johnson, P.J., and Li, A.K.C. 1997. Clinical T.Z., Killius, J., Clavien, P.A., and Nelson, R.C.
evaluation of the partition model for estimating 2002. The role of F-18 FDG positron emission
radiation doses from Yttrium-90 microspheres tomography in preoperative assessment of the
in the treatment of hepatic cancer. Eur. J. Nucl. liver in patients being considered for curative
Med. 24:293–298. resection of hepatic metastases from colorectal
Huang, S.C. 2000. Anatomy of SUV. Nucl. Med. cancer. Clin. Nucl. Med. 27:550–555.
Biol. 27:643–646. Salem, R., Lewandowski, R.J., Atassi, B., Gordon,
Huang, S.C., Phelps, M.E., Hoffman, E.J., Sideris, S.C., Gates, V.L., Barakat, O., Sergie, Z., Wong,
K., Selin, C.J., and Kuhl, D.E. 1980. Noninvasive C.Y., and Thurston, K.G. 2005. Treatment of
determination of local cerebral metabolic rate of unresectable hepatocellular carcinoma with use
glucose in man. Am. J. Physiol. 238:E69–E82. of 90Y microspheres (TheraSphere): Safety,
Kennedy, A., Nag, S., Salem, R., Murthy, R., McEwan, tumor response, and survival. J. Vasc. Interv.
A.J., Nutting, C., Benson, A., Espat, J., Bilbao, Radiol. 16:1627–1639.
J.I., Sharma, R.A., Thomas, J.P., and Coldwell, Selzner, M., and Hany, T.F. 2004. Does the novel
D. 2007. Recommendations for radioemboliza- PET/CT imaging modality impact on the treat-
tion of hepatic malignancies using Yttrium-90 ment of patients with metastatic colorectal cancer
microsphere brachytherapy: A consensus panel of the liver? Ann. Surg. 240:1027–1036.
report from the radioembolization brachytherapy Silvestri, G.A., Tanoue, L.T., Margolis, M.L., Barker,
oncology consortium. 1. Int. J. Radiat. Oncol. J., and Detterbeck, F. 2003. The noninvasive
Biol. Phys. 68:13–23. staging of non-small cell lung cancer: The guide-
Koukourakis, M.I. 2001. Tumour angiogenesis lines. Chest 123(Suppl 1):147S–156S.
and response to radiotherapy. Anticancer Res. Snyder, W.S., Ford, M.R., Warner, G.G., and
21:4285–4300. Watson, S.B. 1975–1976. Absorbed dose per unit
Meta, J., Seltzer, M., Schiepers, C., Silverman, D.H., cumulated activity for selected radionuclides and
Ariannejad, M., Gambhir, S.S., Phelps, M.E., organs. MIRD Pamphlet No. 11:S. New York:
Valk, P., and Czernin, J. 2001. Impact of 18F-FDG Society of Nuclear Medicine.
PET on managing patients with colorectal cancer: Sokoloff, L., Reivich, M., Kennedy, C., Des Rosiers,
The referring physician’s perspective. J. Nucl. M.H., Patlak, C.S., Pettigrew, K.D., Sakurada,
Med. 42:586–590. O., and Shinohara, M. 1977. The [C-14] deoxy-
Ogunbiyi, O.A., Flanagan, F.L., Dehdashti, F., glucose method for the measurement of local
Siegel, B.A., Trask, D.D., Birnbaum, E.H., cerebral glucose utilization: Theory, procedure,
Fleshman, J.W., Read, T.E., Philpott, G.W., and and normal values in the conscious and anesthe-
Kodner, I.J. 1997. Detection of recurrent and tized albino rat. J. Neurochem. 28:897–916.
metastatic colorectal cancer: Comparison of Stubbs, R.S., Cannan, R.J., and Mitchell, A.W.
positron emission tomography and computed 2001. Selective internal radiation therapy with
tomography. Ann. Surg. Oncol. 4:613–620. 90-Yttrium microspheres for extensive color-
Park, K.C., Schwimmer, J., Shepherd, J.E., Phelps, ectal liver metastases. J. Gastrointest. Surg.
M.E., Czernin, J.R., Schiepers, C., and Gambhir, 5:294–302.
S.S. 2001. Decision analysis for the cost-effective Valk, P.E., Abella-Columna, E., Haseman, M.K.,
management of recurrent colorectal cancer. Ann. Pounds, T.R., Tesar, R.D., Myers, R.W., Greiss,
Surg. 233:310–319. H.B., and Hofer, G.E. 1999. Whole-body PET
Phelps, M.E., Huang, S.C., Hoffman, E.J., Selin, imaging with F-18 fluorodeoxyglucose in man-
C.E., and Kuhl, D.E. 1979. Tomographic meas- agement of recurrent colorectal cancer. Arch.
urement of regional cerebral glucose metabolic Surg. 134:503–511.
114 C.-Y.O. Wong
Van Hazel, G., Blackwell, A., Anderson, J., Price, Nagle, C., and Salem, R. 2005. Reduction of
D., Moroz, P., Bower, G., Cardaci, G., and Gray, B. metastatic load to liver after intraarterial hepatic
2004. Randomised phase 2 trial of SIR-Spheres Yttrium-90 radioembolization as evaluated by
plus Fluorouracil/Leucovorin chemotherapy ver- [18F]fluorodeoxyglucose positron emission
sus Fluorouracil/Leucovorin chemotherapy alone tomographic imaging. J. Vasc. Interv. Radiol.
in advanced colorectal cancer. J. Surg. Oncol. 16:1101–1106.
88:78–85. Wong, C.Y., Savin, M., Sherpa, K.M., Qing, F.,
Wong, C.Y., Salem, R., Raman, S., Gates, V.L., and Campbell, J., Gates, V.L., Lewandowski, R.J.,
Dworkin, H.J. 2002. Evaluating 1 2 9 0 Y- g l a s s Cheng, V., Thie, J., Fink-Bennett, D., Nagle,
microsphere treatment response of unresectable C., and Salem, R. 2006. Regional yttrium-90
colorectal liver metastases by [18F]FDG PET: A microsphere treatment of surgically unresectable
comparison with CT or MRI. Eur. J. Nucl. Med. and chemotherapy refractory metastatic liver
Mol. Imaging 29:815–820. carcinoma: A quantitative comparison of tumor
Wong, C.Y., Salem, R., Qing, F., Wong, K.T., Barker, response with glass and resin based agents.
D., Gates, V., Lewandowski, R., Hill, E.A., Cancer Biother. Radiopharm. 21:305–313.
Dworkin, H.J., and Nagle, C. 2004. Metabolic Wong, C.Y., Schmidt, J., Bong, J.S., Chundru, S.,
response after intraarterial 90Y-glass micro- Kestin, L., Yan, D., Grills, I., Gaskill, M., Cheng,
sphere treatment for colorectal liver metastases: V., Martinez, A.A., and Fink-Bennett, D. 2007.
Comparison of quantitative and visual analyses Correlating metabolic and anatomic responses
by 18F-FDG PET. J. Nucl. Med. 45:1892–1897. of primary lung cancers to radiotherapy by
Wong, C.Y., Qing, F., Savin, M., Campbell, J., combined F-18 FDG PET-CT imaging. Radiat.
Gates, V.L., Sherpa, K.M., Lewandowski, R.J., Oncol. 2:18.
B. Prognosis
9
Unresectable Liver Metastases from
Colorectal Cancer: Methodology and
Prognosis with Radiofrequency Ablation
Junji Machi
117
118 J. Machi
1999; Bowles et al., 2001; Curley et al., opposite first- and second-line chemo-
1999; Hellman et al., 2002; Iannitti et al., therapy, and to identify prognostic factors
2002; Jiao et al., 1999; Machi et al., 2001; associated with survival. We focused on
Parikh et al., 2002; Siperstein et al., 2000; colorectal cancer metastases, although
Solbiati et al., 1997). However, as described noncolorectal metastases were also treated
by Seidenfeld et al. (2002), many ques- by RFA during the study period. Finally,
tions remain unanswered regarding the I describe the current role and the future
current and future role of RFA in the treat- perspective of RFA.
ment of liver metastases: a comparison of
RFA to alternative methods, efficacy of
combining RFA with other methods such
METHODOLOGY OF
as chemotherapy, comparison of different RADIOFREQUENCY
RFA approaches (percutaneous, laparo- ABLATION
scopic, and open surgical), and long-term
outcome after RFA. To determine the role Most of RFA therapies are performed
of RFA for metastatic liver tumors, long- under ultrasound guidance. Computed
term follow-up data are essential but have tomography-guided RFA is possible, but
rarely been reported until recently. it is much cumbersome. Because of real-
Moreover, while series published to date time feature, ultrasound is a much better
have generally included patients with prog- imaging method to guide and monitor
ressive disease after chemotherapy, no one RFA. The basic ultrasound guidance tech-
has described and compared survival after nique for needle, cannula, or probe place-
RFA given prior to the first chemotherapy ment for RFA is the same as other needle
regimen for metastatic disease (i.e., “first- guidance under ultrasound.
line”), and survival after RFA when given Radiofrequency ablation treatment of
following disease progression subsequent abdominal organs such as tumors of the
to chemotherapy (i.e., “second-line”). liver can be performed percutaneously,
Failure to describe the impact of RFA in laparoscopically, or open surgically (Machi
relation to chemotherapy in distinct stages et al., 2001). Advantages and disadvan-
of disease progression has hampered the tages of each approach are as follows:
ability to estimate incremental survival 1. Percutaneous Advantages: least invasive,
benefit gained by RFA over the expected possible out-patient procedure, needle
survival on chemotherapy alone. guidance system (+), possible local
We have prospectively acquired data of anesthesia
patients undergoing RFA for unresectable
Disadvantages: less accurate cancer
metastatic liver tumors since 1997. In this
staging, some areas not accessible, risk of
chapter, first I describe the methodology
thermal injury to adjacent organs (bowel,
of RFA. Next, I describe our study, a part
diaphragm, gallbladder, etc.)
of which was recently published (Machi
et al., 2006). The purpose of this study 2. Laparoscopic Advantages: less invasive,
is to assess the long-term outcome of better cancer staging with laparoscopic
the patients based on a specific grade of ultrasound, short recovery, possible con-
unresectability and in relation to its timing current colorectal surgery
9. Unresectable Liver Metastases from Colorectal Cancer 119
Also included in this group were patients of metastatic tumors, visualization or acces-
who were judged not to be able to tolerate sibility on transabdominal ultrasound, need
resection. For example, some patients had for concurrent operations (i.e., synchronous
too many tumors and/or tumors too large colorectal or other organ resection), need
to resect by anatomical major hepatectomy for precise staging by laparoscopy or open
and/or by multiple metastasectomy, unless surgery, newly diagnosed tumors versus
two-stage hepatectomies combined with recurrent tumors following previous resec-
portal vein embolization were performed tion or ablation, and patient condition and
(Jaeck et al., 2004). Other patients had too liver function with regard to surgical and
high (prohibitive) surgical risks such as anesthesia risks (Bowles et al., 2001; Machi
significant comorbidity or liver dysfunction et al., 2000, 2001, 2002, 2006). The selec-
to undergo resection. Many of the patients tion of RFA approach was individualized,
in this group also had multiple risk factors and informed consent was obtained from
for recurrence as patients in grade 1. each patient after full explanation of ben-
Grade 3: Extrahepatic metastasis efits and risks of approaches.
Patients who had extrahepatic metastases During RFA operation, final number,
in addition to liver metastases were included size, and location of the tumors, vascular or
in this group. Extrahepatic metastases bile duct invasion were determined. When
(usually limited lung or peritoneal metas- laparoscopic or open surgical RFA was
tases) in these patients were not extensive performed, high-frequency (5–10 MHz)
as compared to liver metastases so that RFA laparoscopic or open intraoperative ultra-
would be considered to possibly provide sound was routinely performed to screen
benefits to patients. the entire liver for occult tumors. All occult
tumors were ablated when detected.
Preoperative and Intraoperative
Many patients had multiple tumors of
Evaluation
different sizes. In order to estimate the total
Patients had preRFA laboratory tests includ- (cumulative) tumor volume (sum of the
ing liver function tests and CEA. Patients volumes of all tumors to be ablated in a
had computed tomography (or magnetic given RFA session), “total tumor volume”
resonance imaging) and transabdominal was approximated by “total tumor size
ultrasound preoperatively. Positron emis- (sum diameter of tumors)”, which was
sion tomography (PET) was introduced calculated by adding the largest dimension
in October 1999; it was selectively used (in millimeter) of each tumor for all tumors
mainly to diagnose or exclude extrahepatic ablated in each treatment session.
metastases. The timing of offering RFA to
each patient in relation to chemotherapy Radiofrequency Ablation Methods
for metastases was individualized, and was The majority of patients had general anes-
determined by medical oncologists and thesia, although several patients underwent
surgeons together. RFA under local anesthesia. Radiofrequency
Radiofrequency ablation approaches ablation was performed under ultrasound
(percutaneous, laparoscopic, or open sur- guidance using a RFA system (RITA
gical) were determined by considering Medical System, Mountain View, CA) as
factors including number, size, and location described in the previous section. Initially, a
122 J. Machi
survival rate after resection has been relative safety, less invasive nature and
reported in a range of 25–40% (Fong et al., patient willingness to undergo repeated
1999; Nordlinger et al., 1996; Registry procedures, as compared to liver resec-
of hepatic metastases, 1988; Scheele et al., tion. Approximately, 30% of treatments in
1995): the median 5-year survival of 19 our study were repeated RFA for recurrent
study series is 30% (Fernandez et al., metastatic tumors. Local tumor recurrence
2004). Recently, even higher 5-year sur- at the site of RFA occurred in 6.7% of
vival rates of 58% were achieved in three tumors, which is similar or superior to
studies (Abdalla et al., 2004; Choti et al., other studies (~ 5–15%). The diagnosis of
2002; Fernandez et al., 2004). This excel- local recurrence was made mostly within
lent outcome was attributed to improve- 1 year after RFA (mean 8.7 months). With
ment in patient selection (e.g., use of longer follow-up, new tumor recurrence
PET), perioperative care, multidisciplinary occurred much more frequently (higher
treatment including newer chemotherapy, than 80%) in other areas of the liver or
and aggressive surgical management. extrahepatic organs. The diagnosis of new
However, resection can be performed only tumor recurrence was made slightly later
in 10–20% of patients with colorectal (mean 10.4 months) than local tumor
metastases. Even with aggressive surgical recurrence, even as late as 40 months after
approaches such as a two-stage hepate- RFA.
ctomy (Jaeck et al., 2004), resectability In order to evaluate the effect of RFA as a
rates would still be ~ 30%. potentially curative treatment, a long-term
For unresectable metastatic liver tumors, follow-up of patients after RFA is essen-
ablation therapies including RFA were tial. However, the long-term outcome for
initially used for local control, but more liver metastases treated by RFA has rarely
recently its use was also aimed for curative been documented until recently. Most of
purpose. A number of reports documented the data on long-term survival have been
and confirmed the safety and local control reported by investigators in Europe, par-
ability of RFA for metastatic liver tumors ticularly Italy, where RFA was introduced
(Berber et al., 2002; Bilchik et al., 1999; in the early to mid-1990s. Initial survival
Bowles et al., 2001; Curley et al., 1999; outcomes for colorectal metastases were
Hellman et al., 2002; Iannitti et al., 2002; reported by Solbiati et al. (1997) and
Jiao et al., 1999; Machi et al., 2001; by Lencioni et al. (1998). Solbiati et al.
Parikh et al., 2002; Siperstein et al., (2003) updated their data, and Lencioni et
2000; Solbiati et al., 1997). Our study (as al. (2004) also updated data, which were
described above) demonstrated the safety, included in a long-term multicenter study
repeatability and local control efficacy of of 423 patients. Table 9.1 summarizes
RFA consistent with other studies. Major studies showing long-term prognosis of
complications occurred in fewer than 10% RFA. In these Italian and other European
of RFA procedures and were similar to studies with long-term follow-up (3–5-
those noted by others including Curley year survival), RFA was performed by the
et al. (2004) who recently reported early percutaneous approach by radiologists (De
complications in 7.1% and late complica- Baere et al., 2003; De Meijer et al., 2006;
tions in 2.4% of 608 patients. The repeat- Gillams and Lees, 2004; Lencioni et al.,
ability is one advantage of RFA due to its 2004; Solbiati et al., 2003).
126 J. Machi
Table 9.1. Literature review: long-term prognosis of radiofrequency ablation for colorectal cancer liver metastases.
Overall survival
Number of RFAa Follow-up 1-year 4-year 5-year
Author Year patients approach (mean) Median (%) 3-year (%) (%) (%)
(Italy/Europe)
Solbiati et al. 2003 166 P 41.6 Mo 33 Mo 96 46 37 22
Gillams and Lees 2003 167 P 22 Mo 71 21 – 14
De Baere et al. 2003 155 P 18 Mo 27 Mo 82 31
Oshowo et al. 2003 25 P 37 Mo 25 Mo 37 22
Lencioni et al. 2004 423 P 19 Mo 86 47 29 24
White et al. 2004 30 P 17 Mo 22 Mo 75 45
(United States)
Siperstein et al. 2000d 110 P, L, O 30 Mo 88 49 – –
Abdalla et al. 2004 57 O 21 Mob 37 22 –
Abdalla et al. 2004 101c O 21 Mob 43 36 –
Berber et al. 2005 135 L 28.9 Mo
Machi et al. 2006 100 P, L, O 29 Mo 28 Mo 90 42 31 31
(our study) 24.5 Mob
a
P = Percutaneous, L = Laparoscopic, O = Open surgical approach.
b
Median follow-up.
c
Patients underwent RFA combined with liver resection.
d
Study presented but not published.
In the United States, RFA was intro- 24.5-month median and 29-month mean
duced around 1996–1997. The first long- follow-up, which is the longest period ever
term follow-up outcome (3-year survival) reported in the United States (Machi et al.,
of a United States multicenter study was 2006). We used percutaneous, laparoscopic
presented by Siperstein et al. (2000) at and open surgical approaches to optimize
the congress of the American College of the potential benefit to each patient, reco-
Surgeons in October 2002; however, the gnizing the limitation of the percutaneous
data were not published. Curley et al. approach alone (Bowles et al., 2001; Machi
(1999), working with Italian investigators, et al., 2000, 2001, 2002, 2006).
have been demonstrating excellent results The long-term follow-up in our study has
of RFA for unresectable primary and met- demonstrated that RFA can contribute to
astatic hepatic malignancies, and their encouraging survival in patients with unre-
most recent report by Abdalla et al. (2004) sectable colorectal cancer liver metastases.
described long-term outcomes of open The survival outcome was similar or supe-
surgical RFA and combined RFA plus rior to those of Italian/European studies
resection (4-year survival) as compared to and similar to that reported by Berber et al.
resection for colorectal metastases (Table (2005) (Table 9.1). This may be due to
9.1). Berber et al. (2005), using exclusively our aggressive performance of RFA using
a laparoscopic RFA approach, reported laparoscopic and open approaches when
a median survival of 28.9 months in 135 the percutaneous approach is considered
patients with colorectal metastases. In our suboptimal. Additionally, the inclusion of
study (as described in the previous section), the unresectability grade 1 patients, who
we have reported follow-up results of 100 might have otherwise undergone hepatic
patients undergoing RFA for colorectal resection although not optimally, may have
metastatic tumors, up to 84 months, with skewed our survival rates. However, the
9. Unresectable Liver Metastases from Colorectal Cancer 127
overall 5-year survival rate of 30.5% was Radiofrequency ablation following pre-
comparable to those historically achieved vious chemotherapy for metastatic dis-
by hepatic resection (Fong et al., 1999; ease was associated with worse survival
Nordlinger et al., 1996; Registry of hepatic likely because patients who had pre-RFA
metastases, 1988; Scheele et al., 1995), treatment generally had more extensive
although lower than the 58% achieved in tumors, chemotherapy-refractory tumors
recent three reports of resection (Abdalla et or extrahepatic metastasis. However, the
al., 2004; Choti et al., 2002; Fernandez et median survival of 48 months among
al., 2004). patients receiving RFA prior to first-line
Prognostic factors such as the clinical chemotherapy compares favorably with
risk score for recurrence after hepatic resec- survival reported from various trials of
tion have been identified and are useful for first-line chemotherapy alone, which have
predicting post-resection survival outcome reported survival in the range of 19–22
in patients with colorectal metastatic liver months (Grothey et al., 2004; Hurwitz
tumors (Fong et al., 1999; Nordlinger et al., 2004). Similarly, the median survival
et al., 1996). In our study, we have iden- of 22 months in this study for RFA when
tified factors predictive of survival after given in the second-line setting compares
RFA for unresectable metastatic tumors. favorably with survival of 11–15 months
Among the factors not associated with reported from recent trials of second-line
overall and recurrent-free survivals was chemotherapy alone (Giantonio et al.,
the presence or absence of repeated RFA. 2005; Pitot et al., 2005; Rougier et al.,
Repeated RFA did not result in worse sur- 2002). In order to preserve the opportu-
vival, suggesting that RFA should be con- nity to measure secondary end points of
sidered for recurrent metastatic tumors. tumor response and response duration,
On the other hand, the factors associated many chemotherapy trial protocols pro-
with survival were CEA, total tumor size, hibit patients from receiving a potentially
unresectability grade including extrahe- curative resection or ablation after the
patic metastasis, RFA approach, previous start of chemotherapy, even though initial
therapeutic chemotherapy, complications, response to such therapy may move many
local tumor recurrence and new recur- of these initially unresectable tumors to a
rence. Many of these factors are related potentially resectable or ablatable status.
to one another. Complications and local Berber et al. (2005) have reported pre-
tumor recurrence are more frequent when dictors of survival after RFA: significant
total tumor size (volume) to be ablated is factors associated with overall survival
larger (Bowles et al., 2001; Machi et al., were CEA and tumor size, which are simi-
2001; Siperstein et al., 2000). Differences lar to our study. However, the presence of
in survival outcome among three RFA limited extrahepatic disease did not appear
approaches in our study are likely due to affect survival significantly in their study.
to selection of patients and tumors (e.g., This is in contradiction to patients with
unresectability grade) to be ablated: the extrahepatic metastasis in our study, prob-
laparoscopic approach was superior to ably due to differences in inclusion crite-
other approaches because patients in lower ria of patients for RFA depending on the
unresectability grade were selected. extent of extrahepatic metastases.
128 J. Machi
In our study, we have attempted to classify cancer recurrence. The precise monitoring
patients with unresectable liver metastatic of the size or extent of ablation intraopera-
tumors into groups in the hope of pre- tively by ultrasound is presently not pos-
dicting prognosis or outcome after RFA sible, and local cancer recurrence is higher
treatment. As compared to resectable meta- than resection. New ultrasound or other
static tumors for which easily applicable intraoperative technology such as 3-D
and valuable patient stratification methods (three-dimensional) ultrasound or contrast
such as the clinical risk score are avail- enhancement may improve the accuracy
able (Fong et al., 1999; Nordlinger et al., of monitoring RFA with improvement of
1996), unresectable liver tumors encom- ablation negative margin.
pass a wider variety of patient and tumor Hepatic resection is considered as the
characteristics, and thus stratification of gold standard for resectable colorectal
such heterogeneous patients by a straight- metastatic liver tumors. On the other hand,
forward scale is difficult. Unresectability the true role of RFA has not been determined.
is generally determined based on oncolo- Our study has demonstrated that RFA and
gical, technical and medical factors and, in repeated RFA can contribute to a relatively
a certain extent, the unresectability criteria good survival for patients with unresectable
differ among surgeons. Radiofrequency metastatic tumors, and has characterized
ablation can be used for the purpose survival following RFA in relation to first- and
of cure, palliation or life prolongation. second-line chemotherapy (Machi et al.,
Taking these factors and RFA purposes 2006). However, new liver or extrahepatic
into account, we have presented an “unre- recurrence is common after RFA; new liver
sectability grade” by defining three groups: tumors probably recur more frequently
grade 1 (relatively unresectable or not after RFA than after hepatic resection, as
optimally resectable), grade 2 (absolutely noted by Abdalla et al. (2004), probably
unresectable), and grade 3 (extrahepatic because larger liver parenchyma fields,
metastasis). This unresectability grade sys- which are at risk for recurrence remain
tem has demonstrated significant survival after RFA compared to resection.
differences among patients with colorec- To prevent and control such recurrences
tal metastatic liver tumors after RFA in and to achieve a long survival, multimo-
this study, and therefore may be a useful dality therapies, including locoregional
predictor in future studies. and systemic methods, are definitively
required. Because there is a limitation in
hepatic resection as a locoregional treat-
FUTURE PERSPECTIVE ment method, RFA is an additionally
valuable modality, particularly for unresect-
The current methodology of RFA is able metastatic tumors. Our study sug-
advancing, and a 5 cm ablation and even gests greater survival benefit when RFA
a 7 cm ablation can be performed. It is is used in the first-line setting than when
anticipated that further advances of RFA used in the second-line or salvage setting.
technology will improve the ablation size Further investigation is needed to evaluate
larger and the speed of ablation faster the role and timing of RFA in combination
with lower complications and lower local with systemic treatment, especially with
9. Unresectable Liver Metastases from Colorectal Cancer 129
J.L. 2002. Trends in long-term survival following Grothey, A., Sargent, D., Goldberg R.M., and
liver resection for hepatic colorectal metastases. Schmell, H.J. 2004. Survival of patients with
Ann. Surg. 235: 759–766. advanced colorectal cancer improves with avail-
Curley, S.A., Izzo, F., Delrio, P., Ellis, L.M., Granchi, ability of fluorouracil-leucovorin, irinotecan,
J., Vallone, P., Fiore, F., Pignata, S., Daniele, B., and oxaliplatin in the course of treatment. J. Clin.
and Cremona, F. 1999. Radiofrequency ablation Oncol. 22: 1209–1214.
of unresectable primary and metastatic hepatic Hellman, P., Ladjevardi, S., Skogseid, B., Akerstrom,
malignancies. Results in 123 patients. Ann. Surg. G., and Elvin, A. 2002. Radiofrequency tissue
230:1–8. ablation using cooled tip for liver metastases
Curley, S.A., Marra, P., Beaty, K., Ellis, L.M., of endocrine tumors. World J. Surg. 26: 1052–
Vauthey, J.N., Abdalla, E.K., Scaite, C., Rant, 1056.
C., Wolff, R., Choi, H., Loyer, E., Vallone, P., Hurwitz, H., Fehrenbacher, L., Novotny, W.,
Fiore, F., Scardino, F., DeRosa, V., Orlando, Cartwright, T., Hainsworth, J., Heim, W., Berlin,
R., Pignata, S., Daniele, B., and Izzo, F. 2004. J., Baron, A., Griffing, S., Holmgren, E., Ferrara,
Early and late complications after radiofre- N., Fyfe, G., Rogers, B., Ross, R., and Kabbinavar,
quency ablation of malignant liver tumors in 608 F. 2004. Bevacizumab plus irinotecan, fluorour-
patients. Ann. Surg. 239: 450–458. acil, and leucovorin for metastatic colorectal
De Baere, T.J., Elias, D., Ducreux, M., Kuoch, V., cancer. N. Engl. J. Med. 350: 2335–2342.
Boige, V., Dromain, C., Lasser, P., and Tursz, Iannitti, D.A., Dupuy, D.E., Mayo-Smith, W.W.,
T. 2003. Percutaneous radiofrequency of liver and Murphy, B. 2002. Hepatic radiofrequency
metastases: single center experience over a five ablation. Arch. Surg. 137: 422–427.
year period. Proc. Am. Soc. Clin. Oncol. 22: 345 Jaeck, D., Oussoultzoglou, E., Rosso, E., Greget,
(Abstract 1387). M., Weber, J.C., and Bachellier, P. 2004. A two-
De Meijer, V.E., Verhoef, C., Kuiper, J.W., Alwayn, stage hepatectomy procedure combined with
I.P.J., Kazemier, G., and Ijzermans, J.N.M. 2006. portal vein embolization to achieve curative
Radiofrequency ablation in patients with pri- resection for initially unresectable multiple and
mary and secondary hepatic malignancies. J. bilobar colorectal liver metastases. Ann. Surg.
Gastrointest. Surg. 10: 960–973. 240: 1037–1051.
Fernandez, F.G., Drebin, J.A., Linehan, D.C., Jiao, L.R., Hansen, P.D., Havlik, R., Mitry, R.R.,
Dehdashti, F., Siegel, B.A., and Strasberg, Pignatelli, M., and Habib, N. 1999. Clinical
S.M. 2004. Five-year survival after resection short-term results of radiofrequency ablation in
of hepatic metastases from colorectal cancer in primary and secondary liver tumors. Am. J. Surg.
patients screened by positron emission tomogra- 177: 303–306.
phy with F-18 fluorodeoxyglucose (FDG-PET). Lencioni, R., Crocetti, L., Cioni, D., Dellapina, C.,
Ann. Surg. 240: 438–450. and Bartolozzi, C. 2004. Percutaneous radiofre-
Fong, Y., Fortner, J., Sun, R.L., Brennan, M.F., quency ablation of hepatic colorectal metastases.
and Blumgart, L.H. 1999. Clinical score for Technique, indications, results, and new prom-
predicting recurrence after hepatic resection for ises. Invest. Radiol. 39: 689–697.
metastatic colorectal cancer. Analysis of 1001 Lencioni, R., Goletti, O., Armillotta, N., Paolicchi,
consecutive cases. Ann. Surg. 230: 309–321. A., Moretti, M., Cioni, D., and Donati, F. 1998.
Giantonio, B.J., Catalano, P.J., and Meropol, N.J. Radio-frequency thermal ablation of liver metas-
2005. High-dose bevacizumab improves survival tases with cooled-tip electrode needle: results of
when combined with FOLFOX 4 in previously a pilot clinical trial. Eur. Radiol. 8: 1205–1211.
treated advanced colorectal cancer: results from Machi, J., Oishi, A.J., Morioka, W.K., Yu, M.,
ECOG study E3200. Proc. Am. Soc. Clin. Oncol. Hundahl, S.A., Furumoto, N.L., and Oishi, R.H.
(Abstract 2). 23: Part 1. 2000. Radiofrequency thermal ablation of syn-
Gillams, A.R., and Lees, W.R. 2004. Radio- chronous metastatic liver tumors can be performed
frequency ablation of colorectal liver metastases safely in conjunction with colorectal cancer resec-
in 167 patients. Eur. Radiol. 14: 2261–2267. tion. Cancer J. 6(Suppl 4): S344–S350.
9. Unresectable Liver Metastases from Colorectal Cancer 131
Machi, J., Uchida, S., Sumida, K., Limm, W.M.L., (FOLFOX4) in patients with advanced colorectal
Hundahl, S.A., Oishi, A.J., Furumoto, N.L., cancer previously treated with 5FU. Proc. Am.
and Oishi, R.H. 2001. Ultrasound-guided radi- Soc. Clin. Oncol. 24 (Abstract 3506), page 3506.
ofrequency thermal ablation of liver tumors: Registry of hepatic metastases. 1988. Resection of
percutaneous, laparoscopic, and open surgical the liver for colorectal carcinoma metastases: a
approaches. J. Gastrointest. Surg. 5:477–489. multi-institutional study of indications for resec-
Machi, J., Oishi, A.J., Mossing, A.J., Furumoto, tion. Surgery 103: 278–288.
N.L., and Oishi, R.H. 2002. Hand-assisted laparo- Rougier, P., Lepille, D., Bennouna, J., Marre, A.,
scopic ultrasound-guided radiofrequency thermal Ducreux, M., Mignot, L., Hua, A., and Mery-
ablation of liver tumors: a technical report. Surg. Mignard, D. 2002. Antitumour activity of three
Laparosc. Endosc. Percutan. Tech. 12: 160–164. second-line treatment combinations in patients
Machi, J., Oishi, A.J., Sumida, K., Sakamoto, K., with metastatic colorectal cancer after optimal
Furumoto, N.L., Oishi, R.H., and Kylstra, J.W. 5-FU regimen failure: a randomized, multicentre
2006. Long-term outcome of radiofrequency phase II study. Ann. Oncol. 13: 1558–1567.
ablation for unresectable liver metastases from Scheele, J., Stang, R., Altendorf-Hofmann, A.,
colorectal cancer: evaluation of prognostic fac- and Paul, M. 1995. Resection of colorectal liver
tors and effectiveness in first- and second-line metastases. World J. Surg. 19: 59–71.
management. Cancer J. 12: 318–326. Seidenfeld, J., Korn, A., and Aronson, N. 2002.
Nordlinger, B., Guiguet, M., Vaillant, J.C., Balladur, Radiofrequency ablation of unresectable liver
P., Boudjema, K., Bachellier, P., and Jaeck, D. metastases. J. Am. Coll. Surg. 195: 378–386.
1996. Surgical resection of colorectal carcinoma Siperstein, A., Garland, A., Engle, K., Rogers,
metastases to the liver. A prognostic scoring S., Barbar, E., Foroutani, A., String, A., Ryan,
system to improve case selection, based on 1568 T., and Ituarte, P. 2000. Local recurrence after
patients. Cancer 77: 1254–1262. laparoscopic radiofrequency thermal ablation of
Oshowo, A., Gillams, A., Harrison, E., Lees, W.R., hepatic tumors. Ann. Surg. Oncol. 7: 106–113.
and Taylor, I. 2003. Comparison of resection Solbiati, L., Goldberg, S.N., Ierace, T., Livraghi, T.,
and radiofrequency ablation for treatment of Meloni, F., Dellanoce, M., Sironi, S., and Gazelle,
solitary colorectal liver metastases. Br. J. Surg. G.S. 1997. Hepatic metastases: Percutaneous
90: 1240–1243. radio-frequency ablation with cooled-tip elec-
Parikh, A.A., Curley, S.A., Fornage, B.D., and trodes. Radiology 205: 367–373.
Ellis, L.M. 2002. Radiofrequency ablation of Solbiati, L., Ierace, T., Tonolini, M., Bellobuono,
hepatic metastases. Semin. Oncol. 29: 168–182. A., and Cova, L. 2003. Long-term survival of
Pawlik, T.M., Izzo, F., Cohen, D.S., Morris, J.S., patients treated with radiofrequency ablation
and Curley, S.A. 2003. Combined resection and for liver colorectal metastases: improved out-
radiofrequency ablation for advanced hepatic come with increasing experience. Radiology
malignancies: results in 172 patients. Ann. Surg. 229: 411–412.
Oncol. 10: 1059–1069. White, T.J., Roy-Choudhury, S.H., Breen, D.J.,
Pitot, H.C., Rowland, K.M., Sargent, D.J., Philip, Cast, J., Maraveyas, A., Smyth, E.F., Hartley,
P.A., Mitchell, E.P., Goldberg, R.M., and Alberts, J.E., and Monson, J.R.T. 2004. Percutaneous
S.R. 2005. N9841: a randomized, phase III radiofrequency ablation of colorectal hepatic
equivalence trial of irinotecan (CPT-11) ver- metastases – initial experience. Dig. Surg.
sus oxaliplatin/5-fluorouracil (FU)/leucovorin 21: 314–320.
A. Diagnosis
10
Screening with Ultrasonography
of Patients at High-Risk for Hepatocellular
Carcinoma: Thrombocytopenia as a Valid
Surrogate of Cirrhosis
Sheng-Nan Lu, Jing-Houng Wang, Kwong-Ming Kee, and Po-Lin Tseng
137
138 S.-N. Lu et al.
However, HCV-related HCC is increasing Toyoda et al., 2006; Leykum et al., 2007;
in some countries such as Taiwan (Lu et al., Chen et al., 2002). A community study
2006a), Japan (Kiyosawa et al., 2004), and with a cohort of 4,843 Taiwanese at risk
the United States (Wong et al., 2000). In for HCC has demonstrated that the morta-
the United States, after incidents of acute lity was 41% lower in the screening group
HCV infection peaked in the mid-1980s, compared to the control group (Chen et al.,
increasing numbers of future cases of 2002). To assess the value of screening,
HCV-related HCC can be expected. Until one randomized controlled study with
an HCV vaccination is developed and 18,816 Chinese at risk for HCC indicated
become readily available, chronic HCV that biannual screening reduces HCC mor-
infection will remain a major cause of LC tality by 37%. Compared with the control
and HCC. group, the screened group had a higher
number of HCC patients diagnosed at an
earlier stage, with better overall survival
Benefit of HCC Screening
rates (Zhang et al., 2004). It is obvious that
Individuals at risk for developing HCC surveillance of high risk groups for HCC
usually present with clinical symptoms has a sound scientific basis.
when HCC is in a far advanced stage, Alpha-fetoprotein (AFP) and/or ultra-
without effective therapeutic options. sonography (US) are the most widely used
There is clearly a need for early diagnosis tools for surveillance. For HCC screening,
of HCC. Thus, screening and surveillance AFP has a sensitivity of 17–65% and a
appear to be very appropriate. Surveillance specificity of 80–99%, depending on the
programs are ideal for a disease such as prevalence of HCC in the screened popula-
HCC, because it is relatively common tion as well as the cut-off level for the diag-
in patients with chronic liver disease, nosis (Daniele et al., 2004). The results for
has readily identifiable risk factors, high US examination depend on the expertise of
mortality rates, and potentially curative operator and the equipment available. In a
therapy (Di Bisceglie, 2004). However, systemic review, the pooled estimated sen-
there is considerable controversy regard- sitivity and specificity for US in diagnosing
ing the role of screening and surveillance HCC is 60% and 97%, respectively (Colli
in HCC management. The value of screen- et al., 2006). Although neither AFP nor US
ing and surveillance for HCC should be alone is an effective surveillance tool, their
assessed by its impact on detecting earlier combination increases the sensitivity of
stage of the disease, increasing the effec- screening by 5–10% (Daniele et al., 2004;
tive application of curative treatment, and Colli et al., 2006). The optimal surveil-
thus reducing mortality. Although there lance interval for HCC has yet to be deter-
are inconsistent conclusions, most cohort mined. However, based on the estimated
studies from Western and Eastern countries growth rate of HCC, the suggested interval
have shown that surveillance of patients for surveillance in patients has been set at
with viral hepatitis or cirrhosis increases 6 months. Although there are no firm data
early detection and effective application to support the hypothesis that 6 months
of curative treatments (Zoli et al., 1996; is better than 12 months, recent practice
Yuen et al., 2000; Bolondi et al., 2001; guideline on HCC management recommend
10. Screening with Ultrasonography of Patients at High-Risk for Hepatocellular Carcinoma 139
Table 10.1. Non-invasive diagnostic surrogate tests for liver fibrosis and cirrhosis.
Sensitivity Specificity
Tests Cutoff Disease (%) (%) Comments References
Direct
markers
HA 30 ug/l HCV 90 73 Boeker et al., 2002
100 ug/l Viral 84 87 Murawaki et al., 1996
110 ug/l HCV 79.2 89.4 Guechot et al., 1996
183.5 ug/l HCV 80 80 Saitou et al., 2005
Laminin 1.26 U/ml HCV 80 83 Severe liver injury Walsh et al., 2000
Type IV C 148 ng/ml HCV 73 85 Severe liver injury Walsh et al., 2000
160 ng/ml Viral 80 81 Murawaki et al., 1996
6.55 ng/ml HCV 60 61 Saitou et al., 2005
PIIINP 1.0 U/ml HCV 60 74 Guechot et al., 1996
0.995 U/ml HCV 77 66 Saitou et al., 2005
Pro 1,500 ug/l HCV 74 100 Boeker et al., 2002
MMP-2
TIMP1 950 ug/l HCV 100 75 Boeker et al., 2002
YKL-40 284.8 ng/ml HCV 80 71 Saitou et al., 2005
Indirect
markers
AST/ALT ratio ≥1 HCV 77.8 96.9 Giannini and Testa, 2003
APRI ≤1 HCV 89 75 Wai et al., 2003
≤2 HCV 75 93 Wai et al., 2003
Platelet <130 × 109/l HCV 91.1 88.3 Giannini and Testa, 2003
<150 × 109/l HCV 68.2 76.4 Lu et al., 2006b
<150 × 109/l HBV 52.5 78 Lu et al., 2006b
API ≥6 HCV 67 87 Borroni et al., 2006
CDS ≥7 HCV 17 100 Borroni et al., 2006
PGA index >3 Mixed 91 81 Liver fibrosis Teare et al., 1993
Forns <4.2 HCV 94 51 Significant fibrosis Forns et al., 2002
index
Fibrotest <0.1 HCV 100 22 Imbert-Bismut et al., 2001
>0.6 HCV 70 95 Imbert-Bismut et al., 2001
<0.1 HCV 92 29 Rossi et al., 2003
>0.6 HV 42 94 Rossi et al., 2003
Abbreviations: AST: aspartate transaminase; ALT: alanine transaminase; HA: hyaluronic acid; PIIINP: procollagen type III aminot-
erminal peptide; MMP: matrix metalloproteinase TIMP: tissue inhibitor of metalloproteinase; APRI: AST to platelet ratio index;
API: age-platelet index; CDS: cirrhosis discriminate score.
The efficacy is not appropriate for all forms A blood test for platelet count is the
of liver diseases, especially alcoholic liver simplest, most inexpensive, and easily
disease and autoimmune hepatitis. APRI is available serum marker for predicting LC.
a simple test for predicting cirrhosis with The possible causes of thrombocytopenia
an AUC of 0.94 (Wai et al., 2003). When are decreased thrombopoietin production,
using upper (1.0) and lower (2.0) as cut-off antibody mediated platelet destruction,
values, 81% of the patients with or without hypersplenism, and myelotoxic effects.
cirrhosis were characterized. Its perform- Confounding factors are infectious disease
ance for evaluating cirrhosis is similar and hematological disease. The study by
to other, more complex indexes, such as Lu et al. (2006b) investigated the diagnos-
Forns index and Fibrotest. However, APRI tic performance of platelet counts in pre-
cannot be completely standardized due to dicting pathological LC in chronic HBV
the variability of AST ranges considered and HCV patients and US LC in HCV
normal in different laboratories. endemic community residents. Platelet
10. Screening with Ultrasonography of Patients at High-Risk for Hepatocellular Carcinoma 141
counts decrease while pathological fibro- percent of all participants were identified
sis stages increase in chronic HCV, but as belonging to the high risk group, and
not in HBV patients. AUC was larger were scheduled for US (Chen et al., 2002).
in HCV (0.799) than HBV (0.672) and Although more than 95% of HCC cases
the validity was acceptable when cutoff could be identified in this manner, the cost
was set at 150 × 109/l in cases of chronic of screening was high, and it is a daunting
hepatitis B. At this cutoff level, sensitivity task to perform US on 29% of the adult
and specificity were 52.5% and 78% in population in a community.
HBV patients, and 68.2% and 76.4% in Among 4,042 proven HCC cases at a
HCV patients, respectively. In a small medical center in Taiwan, 2,032 (50.3%)
scale HCV endemic community (Lu et al., were HBVAg positive, 1,151 (28.5%)
2006b), the prevalence of HBsAg, anti- were anti-HCV positive, 449 (11.1%)
HCV, both, and neither were 9.0%, 37.3%, were positive for both, and 410 (10.1%)
3.5% and 50.2%; the patients’ prevalence were negative for both. Using 150 × 109/l
of thrombocytopenia were 5.6%, 26.7%, as the cutoff, 1,926 (48%) of the HCC
42.9%, and 3.0%, respectively. The mean cases were found to be thrombocyto-
platelet counts decreased as the US paren- penic. Thrombocytopenic patients made
chyma score increased. Using this cutoff up 42% (860/2,032) of the HBV-related
to identify US cirrhosis, sensitivity was HCC cases, 63% (720/1,151) of the HCV-
76.2%, specificity was 87.8%, and accu- related cases, 50% (225/449) of cases
racy was 86.6%. For cases with chronic positive to both, and 30% (121/410) of the
HCV infection, platelet counts inversely cases tested negative for both. In short,
correlated to both histological and US HCV-related HCC cases had a higher rate
scores for hepatic fibrosis. Platelet count of thrombocytopenia than HBV-related
is cheap, easily interpreted and not inferior cases (Lu et al., 2006b).
to others tests, and should be one of most In an HBsAg (12.4%) and HCV
feasible surrogates. Platelet counts can be (40.8%)-endemic area, 825 (17.9%) of
employed not only for the diagnosis of 4,616 participants were thrombocyto-
hepatic fibrosis and cirrhosis, but also for penic. Of 586 (71%) participants in HCC
identification of the population at high risk screening by US and AFP, 116 (19.8%)
of developing HCC. were diagnosed as LC. Forty-two cases
had suspected liver nodules, and 25 (4.3%)
cases were confirmed as HCC. Only 13
Ultrasonographic HCC Screening on the
of 25 HCC cases had elevated AFP. In
Thrombocytopenic Adult
the low HCV prevalence area (HBsAg
Two-staged community-based HCC 12%, anti-HCV 2%), only 104 throm-
screening seems reasonable. The first bocytopenic cases (6.1%) were found
stage is identification of high risk groups, among 1,694 participants. Forty-four
and the second stage is US screening of patients (42.3%) participated in US and
the high risk group only. In a community AFP screening, but only two (4.5%)
study in Taiwan, HBsAg, anti-HCV, AST, were diagnosed at an early stage of liver
ALT, AFP, and a family history of HCC cirrhosis and no incidence of HCC was
were selected as markers for the identifi- found (Lu et al., 2006b). Among the 25
cation of high risk groups. Twenty-nine detected HCC cases, tumor diameters
142 S.-N. Lu et al.
and mortality in compensated cirrhosis type C: high-risk for hepatocellular carcinoma. Cancer
a retrospective follow-up study of 384 patients. 107: 2212–2222.
Gastroenterology 112: 463–472. Lu, S.N., Wang, J.H., Chen, P.F., Tung, H.D.,
Forns, X., Ampurdanes, S., Llovet, J.M., Aponte, Tseng, P.L., Hung, C.H., Kee, K,M., Chen,
J., Quintó, L., Martínez-Bauer, E., Bruguera, C.H., Chang, K.C., Lee, C.M., Changchien,
M., Sánchez-Tapias, J.M., and Rodés, J. 2002. C.S., Chen, Y.D., Tsai, L.S., and Chen, T.H.
Identification of chronic hepatitis C patients 2008. Community-based mass ultrasonographic
without hepatic fibrosis by a simple predictive screening of hepatocellular carcinoma among
model. Hepatology 36: 986–992. thrombocytopenic adults. Cancer Epidemiol.
Giannini, E., and Testa, R. 2003. The metabolic Biomarkers Prev. 17: 1813–1821.
syndrome: all criteria are equal, but some criteria Murawaki, Y., Ikuta, Y., Koda, M., Yamada, S., and
are more equal than others. Arch. Intern. Med. Kawasaki, H. 1996. Comparison of serum 7S
163: 2787–2788. fragment of type IV collagen and serum central
Guechot, J., Laudat, A., Loria, A., Serfaty, L., triple-helix of type IV collagen for assessment of
Poupon, R., and Giboudeau, J. 1996. Diagnostic liver fibrosis in patients with chronic viral liver
accuracy of hyaluronan and type III procollagen disease. J. Hepatol. 24: 148–154.
amino-terminal peptide serum assays as markers of Rossi, E., Adams, L., Prins, A., Bulsara, M., de
liver fibrosis in chronic viral hepatitis C evalu- Boer, B., Garas, G., MacQuillan, G., Speers, D.,
ated by ROC curve analysis. Clin. Chem. 42: and Jeffrey, G. 2003. Validation of the FibroTest
558–563. biochemical markers score in assessing liver
Imbert-Bismut, F., Ratziu, V., Pieroni, L., Charlotte, fibrosis in hepatitis C patients. Clin. Chem. 49:
F., Benhamou, Y., Poynard, T. and MULTIVIRC 450–454.
Group. 2001. Biochemical markers of liver fibro- Saitou, Y., Shiraki, K., Yamanaka, Y.,
sis in patients with hepatitis C virus infection: a Yamaguchi, Y., Kawakita, T., Yamamoto, N.,
prospective study. Lancet 357: 1069–1075. Sugimoto, K., Murata, K., and Nakano, T.
Kelleher, T.B., and Afdhal, N. 2005. Noninvasive 2005. Noninvasive estimation of liver fibrosis
assessment of liver fibrosis. Clin. Liver Dis. 9: and response to interferon therapy by a serum
667–683. fibrogenesis marker, YKL-40, in patients
Kiyosawa, K., Umemura, T., Ichijo, T., Matsumoto, with HCV-associated liver disease. World J.
A., Yoshizawa, K., Gad, A., and Tanaka, E. Gastroenterol. 11: 476–481.
2004. Hepatocellular carcinoma: recent trends in Teare, J.P., Sherman, D., Greenfield, S.M.,
Japan. Gastroenterology 127: S17–26. Simpson, J., Bray, G., Catterall, A.P., Murray-
Leykum, L.K., El-Serg, H.B., Cornell, J., and Lyon, I.M., Peters, T.J., Williams, R., and
Papadopoulos, K.P. 2007. Screening for hepato- Thompson, R.P. 1993. Comparison of serum
cellular carcinoma veterans with hepatitis C on procollagen III peptide concentrations and PGA
disease stage, treatment received, and survival. index for assessment of hepatic fibrosis. Lancet
Clin. Gastroenterol. Hepatol. 5: 508–512. 342: 895–898.
Lu, S.N., Su, W.W., Yang, S.S., Chang, T.T., Toyoda, H., Kumada, T., Kiriyama, S., Sone, Y.,
Cheng, K.S., Wu, J.C., Lin, H.H., Wu, S.S., Tanikawa, M., Hisanaga, Y., Yamaguchi, A.,
Lee, C.M., Changchien, C.S., Chen, C.J., Sheu, Isogai, M., Kaneoka, Y., and Washizu, J. 2006.
J.C., Chen, D.S., and Chen, C.H. 2006a. Secular Impact of surveillance on survival of patients
trends and geographic variations of hepatitis B with initial hepatocellular carcinoma: a study
virus and hepatitis C virus-associated hepatocel- from Japan. Clin. Gastroenterol. Hepatol. 4:
lular carcinoma in Taiwan. Int. J. Cancer 119: 1170–1176.
1946–1952. Wai, C.T., Greenson, J.K., Fontana, R.J.,
Lu, S.N., Wang, J.H., Liu, S.L., Hung, C.H., Chen, Kalbfleisch, J.D., Marrero, J.A., Conjeevaram,
C.H., Tung, H.D., Chen, T.M., Huang, W.S., Lee, H.S., and Lok, A.S. 2003. A simple noninvasive
C.M., Chen, C.C., and Changchien, C.S. 2006b. index can predict both significant fibrosis and
Thrombocytopenia as a surrogate for cirrhosis cirrhosis in patients with chronic hepatitis C.
and a marker for the identification of patients at Hepatology 38: 518–526.
144 S.-N. Lu et al.
Walsh, K.M., Fletcher, A., MacSween, R.N., and treatment: Hong Kong experience. Hepatology
Morris, A.J. 2000. Basement membrane peptides 31: 330–335.
as markers of liver disease in chronic hepatitis C. Zhang, B.H., Yang, B.H., and Tang, Z.Y. 2004.
J. Hepatol. 32: 325–330. Randomized controlled trial of screening for
Wong, J.B., McQuillan, G.M., McHutchison, J.G., hepatocellular carcinoma. J. Cancer. Res. Clin.
and Poynard, T. 2000. Estimating future hepatitis Oncol. 130: 417–422.
C morbidity, mortality, and costs in the United Zoli, M., Magalotti, D., Bianchi, G., Gueli, C.,
States. Am. J. Public Health 90: 1562–1569. Marchesini, G., and Pisi, E. 1996. Efficacy
Yuen, M.F., Cheng, C.C., Lauder, I.J., Lam, S.K., of surveillance program for early detection
Ooi, C.G., and Lai, C.L. 2000. Early detection of of hepato-cellular carcinoma. Cancer 78:
hepatocellular carcinoma increases the chance of 977–985.
11
Hepatocellular Carcinoma:
Contrast-Enhanced Sonography
Byung Ihn Choi and Se Hyung Kim
145
146 B.I. Choi and S.H. Kim
coded harmonic sonography, contrast- and the subsequent ones are the second,
enhanced agent detection imaging, and low third, and fourth harmonics.
mechanical index imaging. Finally, the Bubbles in a liquid tend to diffuse and
usefulness of contrast-enhanced sonogra- disappear unless they are stabilized by
phy in assessing the therapeutic response some form of a shell. Once the shell is dis-
of HCCs is reviewed. rupted, the gas inside will diffuse into the
surrounding fluid. The mechanical index
(MI), defined originally to predict the onset
PHYSICS OF MICROBUBBLES of cavitation in fluids, also gives an indica-
tion of the likelihood of bubble destruction.
Microbubble contrast agents with diagnos- It has been well established that the acous-
tic ultrasound have been an active area of tic power level used during routine exami-
research since Gramiak and Shah (1968) nations destroys the contrast microbubbles.
observed opacification of the right ventri- The blood flow in a normal capillary bed is
cle after an injection of saline. The earliest
on the order of 1 mm/s, and a typical capil-
microbubbles were unable to pass through lary is ~ 1 mm long. Thus, if the contrast
the lungs, and so were only able to opacify within a capillary is destroyed, it will take
the right ventricle. But, during the past 2 ~ 1 s or more to refill the capillary. Given the
decades, stabilized microbubbles capable of branching structure of the microvasculature
transpulmonary passage for left-side blood and the thickness of a typical scan plane,
pool enhancement were developed by sev- it can take several seconds to replenish the
eral major pharmaceutical companies. contrast in the scan plane, depending on
An acoustic wave generated by an the flow rate to the organ. During real-time
ultrasonic system consists of alternating scanning at normal output power levels, the
high and low pressures at frequencies contrast is never given a chance to fill the
of 1.5–10 MHz. When an acoustic wave microvasculature. This was first observed
encounters a microbubble, it alternately by Porter et al. (1996) when they found
compresses the microbubble on the posi- that triggered imaging allows much better
tive pressure and expands it on the nega- visualization of contrast within the myo-
tive pressure. On the positive portion of cardium. This led to the widespread use
the wave the microbubbles are compressed of electrocardiographic triggering during
in a different fashion than the way they myocardial contrast echo, users often trig-
expand in the negative portion. This resultsgering only once every four or more cardiac
in an asymmetric, nonlinear bubble oscil- cycles. Similar techniques have been used
lation. Instead of producing a sinusoidal to image flow in the hepatic parenchyma.
echo with a clean frequency spectrum
like the transmitted signal, it produces an
odd-looking echo with an asymmetric top CONTRAST-ENHANCED
and bottom. This asymmetry produces COLOR DOPPLER
harmonics and can be used to enhance the SONOGRAPHY
signals from the bubbles. In the frequency
spectrum of the bubble echoes, the first Doppler sonography can provide informa-
major hump is the fundamental component tion on blood flow that is useful in the
11. Hepatocellular Carcinoma: Contrast-Enhanced Sonography 147
differential diagnosis of a focal hepatic of several studies have shown that power
lesion. Tanaka et al. (1990) have described Doppler sonography is more sensitive than
a color Doppler pattern of intra-lesional color Doppler US in the depiction of vas-
flow which may be helpful in differen- cular flow in focal hepatic lesions, includ-
tiating HCC from other hepatic tumors. ing HCC (Choi B.I. et al., 1996; Lencioni
A basket pattern, fine blood flow net- et al., 1996). Even using the optimized
work around the HCC nodule, and vessels parameters for power Doppler sonography,
within the tumor are seen only in HCC. however, the ability to detect low-velocity
This contrasts with the detour or dot pat- blood flow in very small intratumoral
tern seen in metastases or hemangioma, vessels, particularly in small HCCs and
respectively. However, on conventional hemangioma, is still limited.
color Doppler sonography, the ability to According to the previous report by Kim
evaluate low-velocity blood flow in very et al. (1998), intra- and peritumoral flow
small intratumoral vessels, in particular in signals increased in most HCCs (95%)
small hepatic lesions and lesions located after administration of contrast agent and
deep within hepatic parenchyma, is less enhanced vascular flow signals in the tumor
impressive. Attempts to improve color changed from an internal flow pattern or a
Doppler imaging led to the development of basket pattern to a prominent mixed pattern.
ultrasound contrast agents. Consequently, Contrast enhancement of the Doppler signal
contrast-enhanced color Doppler sono- in the tumor began ~ 15–20 s after injection
graphy demonstrated more intratumoral of contrast agent and reached a maximum
vascularity in HCCs than did conventional at 30–60 s. This study concluded that con-
color Doppler sonography (Ernst et al., trast-enhanced power Doppler sonography
1996; Strobel et al., 2000). Tanaka et al. was superior to unenhanced power Doppler
(1998) have also shown an improvement sonography in the demonstration and char-
in the detection rate of intratumoral vascu- acterization of tumor vascularity in nodular
larity. In a study on contrast-enhanced or massive HCC.
color Doppler sonography, the majority of
HCCs showed extensive or moderate intra-
tumoral vascularity (Strobel et al., 2000). CONTRAST-ENHANCED
HARMONIC POWER
DOPPLER SONOGRAPHY
CONTRAST-ENHANCED
POWER DOPPLER Even though contrast-enhanced power
SONOGRAPHY Doppler sonography showed a good per-
formance to depict tumor vascularity in
Unlike color Doppler sonography, which HCC, it still has several limitations related
emphasizes velocity, variance informa- to intrinsic weaknesses of power Doppler
tion, or both, power Doppler sonography technique. First, high susceptibility to tissue
is a technique based on the total integrated motion limits the application of this tech-
power of the Doppler spectrum, which is nique in patients with poor breath-holding
related to the number of red blood cells and those with hepatic masses near the heart
that produce the Doppler shift. Findings or pulsatile great vessels. Second, color
148 B.I. Choi and S.H. Kim
blooming artifact, which can degrade image sonography. Similarly, the sizes of intrahe-
quality, is frequently encountered immedi- patic and intratumoral vessels likely were
ately after bolus injection of contrast agent overestimated. Enhanced vessels adjacent
(Kim et al., 1998). In one study, the size to a liver tumor can be seen as if they are
of the vessel enhanced with contrast agent located within the tumor on conventional
frequently appeared larger than its actual power Doppler sonography. This can be an
size, even though reduction of Doppler important problem when a physician needs
gain was done until artifacts outside the to know whether any residual viable tumor
vessel nearly disappeared (Choi et al., is present after local treatment for a malig-
2000a). Third, power Doppler artifacts pro- nant tumor. In that case, harmonic power
duced from areas with high echogenicity Doppler sonography might be superior to
frequently mimic intratumoral vessels and conventional power Doppler sonography
make assessing intratumoral vascularity (Choi et al., 2000b).
difficult (Kim et al., 1999). To circumvent
such limitations, the harmonic techniques
have been introduced and merged into con- CONTRAST-ENHANCED
trast-enhanced power Doppler sonography. PULSE-INVERSION
The use of harmonic mode can effectively HARMONIC SONOGRAPHY
reduce artifacts including blooming and
motion-related artifacts. Like tissue harmonic imaging, pulse-
In our previous study, we performed a inversion harmonic imaging (PIHI) displays
prospective study to compare contrast- the amplitude of the harmonic signals result-
enhanced harmonic power Doppler sonog- ing from non-linear echoes. However, in
raphy with conventional power Doppler PIHI, two identical pulses with reversed
sonography in the assessment of tumor polarity which are transmitted down each
vascularity in HCC (Choi et al., 2000b). ray line were used. When these result-
In most cases, the intensity of enhanced ant returned waveforms are added, the
signals in the tumor and hepatic paren- harmonic component gives the multiplied
chyma in conventional mode was stronger harmonic level of a single waveform,
than that in harmonic mode (Choi et al., while all linear fundamental components
2000b). This is not surprising when consid- are canceled out more effectively than
ering the relatively low intensity of second those on tissue harmonic imaging using a
harmonic echoes compared with funda- simple bandpass filter (Hohl et al., 2004;
mental echoes (Burns, 1996). However, Jang et al., 2000). This technique, instead
harmonic mode was superior to conven- of using a narrow receiver filter tuned
tional mode in terms of power Doppler around harmonic frequency components
artifacts at any imaging time. In harmonic which is used in a simple bandpass fil-
power Doppler mode, continuous scan- tering method, allows the use of broader
ning during a patient’s quiet breathing was transmit and receive bandwidths to achieve
possible without producing considerable improved resolution. Therefore, PIHI can
motion-related artifacts. Overestimation of overcome the bandwidth limitations of
the size of contrast-enhanced vessel was filtering harmonic imaging (Hohl et al.,
evident on conventional power Doppler 2004; Jang et al., 2000).
11. Hepatocellular Carcinoma: Contrast-Enhanced Sonography 149
Although PIHI produces significantly bet- motion artifacts are seen when a transducer
ter image quality for focal hepatic lesions, moves between frames.
it has several limitations that should be In contrast to power Doppler sonography,
overcome for it to replace conventional contrast-enhanced PIHI with interval delay
sonography in routine practice. Because scanning can provide strong gray-scale
multiple pulses are used, the frame rate is enhancement by microbubble contrast agents
relatively low compared with that of con- and detect signals from microbubbles in very
ventional sonography or tissue harmonic slow flow without Doppler-related arti-
imaging. Although we often found that PIHI facts (Figure 11.1). It has been shown that
improves the conspicuousness of a HCC this technique can depict the enhancement
barely recognizable on conventional US, patterns of various hepatic tumors includ-
we cannot use PIHI during biopsy because ing HCC. Kim et al. (2000) performed a
of its slow frame rate. In addition, the fact prospective study with contrast-enhanced
that multiple pulses are used to form each PIHI in various liver tumors. Hepatocellular
ray line, makes it susceptible to motion arti- carcinomas with or without irregular intra-
facts. Anything that moves between pulses tumoral enhancing vessels show findings
is not completely cancelled, so some tissue similar to those on hepatic angiography
(Figure 11.1). The pattern of enhancement (Lee et al., 2002). It combines PIHI with
on PIHI in that study was thought to be a coded excitation technique, which trans-
useful in the differential diagnosis of liver mits coded pulse sequence, decoding them
tumors (Kim et al., 2000). As compared on receipt. Coded technology uses a code
with computed tomography (CT) or mag- and decode pulse sequence system, so-
netic resonance imaging (MRI), there were called digitally encoded sonographic tech-
a few limitations of the sonographic tech- nology, to enhance the signals from the
nique used in that study. The first is that contrast agent and, at the same time, sup-
dynamic sonographic scanning is possible press unwanted background, fundamental,
in only one scanning plane; thus, it is not and harmonic signals from the tissue. This
possible to characterize all lesions simul- technique can be paired with the benefits
taneously in patients with multiple lesions. of B-flow. B-flow optimizes direct visuali-
Second, interval delay scanning in the same zation of blood cells on gray-scale imag-
area is not easy for unskilled examiners. ing. Because blood echoes are very weak,
It may be difficult even for experienced it uses codes to boost weak blood echoes
examiners to scan the lesion in the liver and equalize non-moving tissue signal.
region that is visible only during a patient’s For that reason, B-flow technique does
deep inspiration. Sufficient practice of inter- not cause Doppler artifact. Ultimately,
val delay scanning at the same area, includ- contrast-enhanced CHA can provide
ing the tumor, must be done before injection extremely high quality and detailed vas-
of the contrast agent in such cases. However, cular information without the artifacts
sonography is superior to CT or MRI in associated with contrast agent as well as
that immediate, accurate characterization preservation of wideband resolution of
of a newly detected focal hepatic lesion on harmonic signals.
sonographic examination is possible (Kim In a previous study using contrast-
et al., 2000). enhanced CHA, the typical hemodynamic
pattern of HCCs was shown as abundant
tumor vessels supplied from the periphery
CONTRAST-ENHANCED to the center of the tumor and dense paren-
CODED HARMONIC chymal tumor staining with fast washout
SONOGRAPHY (Figure 11.2) and the sensitivity and
specificity of those findings were 92.8%
In general, transmitted sonographic pulse and 92.3%, respectively (Wen et al., 2004).
is distorted in shape and loses energy Our previous study showed similar results
as it travels through tissue. It is true for that contrast-enhanced CHA was an effec-
contrast-enhanced signals as well as for tive sonographic technique in evaluating
fundamental sonic waves. To improve the the vascularity of HCCs because it has the
detectability of the enhanced microbub- advantages such as the depiction of excel-
ble signals, researchers have studied the lent dynamic enhancement of tumors, high-
behavior of microbubbles in blood flow resolution imaging of detailed intratumoral
and developed newer sonographic tech- vessels, and no sonographic artifacts (Lee et
nology. Coded harmonic angio (CHA) al., 2003) (Figure 11.2). However, despite
is one of the new imaging techniques many advantages of CHA, CHA had several
11. Hepatocellular Carcinoma: Contrast-Enhanced Sonography 151
used with ADI, it may effectively reveal 2005b; Youk et al., 2003a). According to
the vascularity of focal liver lesions and our recent report, contrast-enhanced ADI
may also help in lesion characterization. sonography can be used to characterize
Furthermore, ADI may separate the tis- focal hepatic lesions with a higher diag-
sue image from the contrast-only image nostic confidence than baseline sonogra-
(Figure 11.3). Because the tissue image phy and furthermore, it reduced the need
and contrast-only image can be shown sep- for further diagnostic workups for focal
arately, the positional relationship between hepatic lesion characterization (Kim et al.,
the tumor and other structures can be accu- 2005b) (Figure 11.3). Our another recent
rately assessed (Figure 11.3). report showed that single-level dynamic
Several reports have suggested that sonography with ADI was potentially
contrast-enhanced ADI sonography has useful in diagnosing HCC, with excellent
improved the detection and characteriza- visibility of intra- and peri-tumoral vascu-
tion of focal liver lesions (Kim et al., larity, early tumor staining, and sometimes
Figure 11.3. Contrast-enhanced agent detection imaging of hepatocellular carcinoma. Images obtained
at 20 s (upper) and 30 s (lower) after injection of the contrast agent shows a small hepatocellular carci-
noma (arrows) with hyperechoic contrast enhancement compared with hepatic parenchyma. Comparison
of combined (left), contrast (middle), and tissue (right) modes of display enables precise location of the
distribution of the microbubble; therefore, the hypervascularity of the tumor during the arterial phase can
be assessed confidently
11. Hepatocellular Carcinoma: Contrast-Enhanced Sonography 153
with Doppler sonography, there have been tic response to RF ablation therapy (Choi
a lot of researches underway in the use- et al., 2003; Kim et al., 2005b) (Figures
fulness of contrast-enhanced sonography 11.5 and 11.6). This technique allows
for assessing the therapeutic responses of detection of residual tumor after ablation
HCC after treatment. Several reports have without the blooming and motion artifacts
suggested that contrast-enhanced color and seen at contrast-enhanced color or power
power Doppler sonography can be useful Doppler sonography. Another advantage
in detecting residual or recurrent tumors of contrast-enhanced gray-scale harmonic
after local ablation of HCC (Goldberg sonography over CT and MR imaging
et al., 2000) (Figure 11.4). According to is that it can directly show persistent
a preliminary report, contrast-enhanced fine vascular flow signals within the
power Doppler sonography can be used tumor during a period of effective con-
as an alternative to immediate follow-up trast enhancement that lasts for more than
CT for early assessment of the therapeu- 10 min. Consequently, contrast-enhanced
tic efficacy of RF ablation of HCC (Choi gray-scale harmonic sonography could be
et al., 2000b) (Figure 11.4). helpful in detecting residual or recurrent
More recently, contrast-enhanced gray- tumor after local ablation of HCC. These
scale harmonic sonography has been modalities may also be helpful in identify-
developed for the evaluation of therapeu- ing residual tumor when performed during
Figure 11.5. Contrast-enhanced coded harmonic sonography for evaluation of local recurrence after
radiofrequency ablation (RFA) of HCC. Unenhanced, oblique coronal sonogram (upper left) obtained 1
year after RFA shows a heterogeneous high echoic nodule (arrow) in right lobe of the liver. On contrast-
enhanced coded harmonic sonogram (upper middle) obtained 42 s after injection of the contrast agent, a
nodular enhancing focus (arrow) is noted in the anterior side of the lesion. This represents the recurred
tumor. During RFA (upper right), a large echogenic lesion (arrow) suggesting ablated zone is seen. Portal
phase contrast CT scan (lower left) before second RFA shows a nodular enhancing focus (arrow) of
the residual tumor. On CT scan (lower right) obtained immediate after second RFA, the ablated lesion
remains unenhanced, which indicates a successful RFA
repeat ablation, when accurate localization to the risk of toxicity. However, real-time
of viable tumor is needed. This real-time sonography and the ability to administer
confirmation of the accurate localization repeated doses of sonographic contrast
of viable tumor and of correct placement material may, in some cases, allow better
of the electrode within the tumor cannot detection of residual tumor.
be expected at CT, and may reduce the In conclusion, many recent advances
number of treatment sessions without fur- in contrast-enhanced sonography imag-
ther use of contrast-enhanced CT or MR ing have occurred, making it a strong
imaging. Furthermore, contrast-enhanced contender for several problem-solving
CT is less than ideal because errors in applications. Admittedly, CT and MR
contrast bolus timing cannot always be cor- imaging continue to have complimentary
rected immediately with repeat scanning due roles. However, sonographic examinations
156 B.I. Choi and S.H. Kim
Figure 11.6. Contrast-enhanced agent detection imaging for evaluation of therapeutic response after
transarterial chemoembolization (TACE) of HCC. Unenhanced, sagittal sonogram (upper left) shows a
1.5 cm-size, hypoechoic nodule (arrow) in the left hepatic lobe. Contrast-enhanced sonograms obtained
at 30 s (upper right), 40 s (lower left), and 120 s (lower right) after injection of the contrast agent reveal
focal enhancing portion (arrows) that represents residual viable tumor. The residual tumor was treated
with second TACE later on the same day
with reproducibly excellent quality are in the depiction of vasculature in liver tumors.
expected with the aid of newer imaging Radiology 200:55–58.
Choi, B.I., Kim, T.K., Han, J.K., Kim, A.Y., Seong,
techniques combined with the use of more
C.K., and Park, S.J. 2000a. Vascularity of hepa-
stable, second generation contrast agents. tocellular carcinoma: assessment with contrast-
Contrast-enhanced sonography will con- enhanced second-harmonic versus conventional
tinue to have a major role in liver imaging power Doppler US. Radiology 214:381–386.
into the foreseeable future. Choi, D., Lim, H.K., Kim, S.H., Lee, W.J., Jang,
H.J., Lee, J.Y., Paik, S.W., Koh, K.C., and Lee,
J.H. 2000b. Hepatocellular carcinoma treated
REFERENCES
with percutaneous radio-frequency ablation: use-
Bartolozzi, C., Lencioni, R., Ricci, P., Paolicchi, A., fulness of power Doppler US with a microbubble
Rossi, P., and Passariello, R. 1998. Hepatocellular contrast agent in evaluating therapeutic response-
carcinoma treatment with percutaneous etha- preliminary results. Radiology 217:558–563.
nol injection: evaluation with contrast-enhanced Choi, D., Lim, H.K., Lee, W.J., Kim, S.H., Kim,
color Doppler US. Radiology 209:387–393. Y.H., Kim, S.H., and Lim, J.H. 2003. Early
Burns, P.N. 1996. Harmonic imaging with ultrasound assessment of the therapeutic response to radio
contrast agents. Clin. Radiol. 51: Suppl. 1:50–55. frequency ablation for hepatocellular carcinoma:
Choi, B.I., Kim, T.K., Han, J.K., Chung, J.W., Park, utility of gray scale harmonic ultrasonography
J.H., and Han, M.C. 1996. Power versus conven- with a microbubble contrast agent. J. Ultrasound
tional color Doppler sonography: comparison Med. 22:1163–1172.
11. Hepatocellular Carcinoma: Contrast-Enhanced Sonography 157
Dill-Macky, M.J., Burns, P.N., Khalili, K., and utility of contrast-enhanced agent detection
Wilson, S.R. 2002. Focal hepatic masses: imaging. Eur. J. Radiol. 56:66–73.
enhancement patterns with SH U 508A and Kim, K.W., Choi, B.I., Park, S.H., Kim, H.C.,
pulse-inversion US. Radiology 222:95–102. Lee, M.W., Kim, S.H., Lee, K.H., Park, C.H.,
Ding, H., Kudo, M., Onda, H., Suetomi, Y., Kim, J.S., Won, H.J., and Han, J.K. 2003.
Minami, Y., Chung, H., Kawasaki, T., and Hepatocellular carcinoma: assessment of vascu-
Maekawa, K. 2001. Evaluation of posttreatment larity with single-level dynamic ultrasonography
response of hepatocellular carcinoma with con- during the arterial phase. J. Ultrasound Med.
trast-enhanced coded phase-inversion harmonic 22:887–896.
US: comparison with dynamic CT. Radiology Kim, S.H., Lee, J.M., Lee, J.Y., Han, J.K., An,
221:721–730. S.K., Han, C.J., Lee, K.H., Hwang, S.S., and
Dodd, G.D. III, Miller, W.J., Baron, R.L., Skolnick, Choi, B.I. 2005b. Value of contrast-enhanced
M.L., and Campbell, W.L. 1992. Detection of sonography for the characterization of focal
malignant tumors in end-stage cirrhotic livers: hepatic lesions in patients with diffuse liver dis-
efficacy of sonography as a screening technique. ease: receiver operating characteristic analysis.
AJR Am. J. Roentgenol. 159:727–733. AJR Am. J. Roentgenol. 184:1077–1084.
Ernst, H., Hahn, E.G., Balzer, T., Schlief, R., and Kim, T.K., Han, J.K., Kim, A.Y., Park, S.J., and
Heyder, N. 1996. Color Doppler ultrasound of Choi, B.I. 1999. Signal from hepatic hemangi-
liver lesions: signal enhancement after intrave- omas on power Doppler US: real or artefactual?
nous injection of the ultrasound contrast agent Ultrasound Med. Biol. 25:1055–1061.
Levovist. J. Clin. Ultrasound 24:31–35. Kim, T.K., Choi, B.I., Han, J.K., Hong, H.S.,
Goldberg, S.N., Gazelle, G.S., Compton, C.C., Park, S.H., and Moon, S.G. 2000. Hepatic
Mueller, P.R., and Tanabe, K.K. 2000. Treatment tumors: contrast agent-enhancement patterns
of intrahepatic malignancy with radiofrequency with pulse-inversion harmonic US. Radiology
ablation: radiologic-pathologic correlation. 216:411–417.
Cancer 88:2452–2463. Lee, J.Y., Choi, B.I., Han, J.K., Kim, A.Y., Shin,
Gramiak, R., and Shah, P.M. 1968. Echocardiography S.H., and Moon, S.G. 2002. Improved sono-
of the aortic root. Invest. Radiol. 3:356–366. graphic imaging of hepatic hemangioma with
Hohl, C., Schmidt, T., Haage, P., Honnef, D., contrast-enhanced coded harmonic angiography:
Blaum, M., Staatz, G., and Guenther, R.W. 2004. comparison with MR imaging. Ultrasound Med.
Phase-inversion tissue harmonic imaging com- Biol. 28:287–295.
pared with conventional B-mode ultrasound in Lee, J.Y., Choi, B.I., Han, J.K., Kim, A.Y., Shin,
the evaluation of pancreatic lesions. Eur. Radiol. S.H., and Park, C.M. 2003. Assessment of
14:1109–1117. hepatocellular carcinoma vascularity: compari-
Jang, H.J., Lim, H.K., Lee, W.J., Kim, S.H., Kim, son of contrast-enhanced coded harmonic ultra-
K.A., and Kim, E.Y. 2000. Ultrasonographic sound with harmonic power Doppler ultrasound.
evaluation of focal hepatic lesions: comparison J. Med. Ultrasound 11:147–155.
of pulse inversion harmonic, tissue harmonic, and Lencioni, R., Caramella, D., and Bartolozzi, C.
conventional imaging techniques. J. Ultrasound 1995. Hepatocellular carcinoma: use of color
Med. 19:293–299. Doppler US to evaluate response to treatment
Kim, A.Y., Choi, B.I., Kim, T.K., Han, J.K., with percutaneous ethanol injection. Radiology
Yun, E.J., Lee, K.Y., and Han, M.C. 1998. 194:113–118.
Hepatocellular carcinoma: power Doppler Lencioni, R., Pinto, F., Armillotta, N., and
US with a contrast agent–preliminary results. Bartolozzi, C. 1996. Assessment of tumor vascu-
Radiology 209:135–140. larity in hepatocellular carcinoma: comparison
Kim, C.K., Choi, D., Lim, H.K., Kim, S.H., Lee, of power Doppler US and color Doppler US.
W.J., Kim, M.J., Lee, J.Y., Jeon, Y.H., Lee, J., Radiology 201:353–358.
Lee, S.J., and Lim, J.H. 2005a. Therapeutic Meloni, M.F., Goldberg, S.N., Livraghi, T., Calliada,
response assessment of percutaneous radiofre- F., Ricci, P., Rossi, M., Pallavicini, D., and
quency ablation for hepatocellular carcinoma: Campani, R. 2001. Hepatocellular carcinoma
158 B.I. Choi and S.H. Kim
159
160 V. Migaleddu and G. Virgilio
Ultrasound contrast media are sus- US beam, depends on the density of the
pensions containing air or gas micro- tissue. The compression part of the US
bubbles that increase reflection of the wave increases by fractions the density of
US beam used during the investigation. the tissue, and thus this part of the wave
Contrast agents were previously used as travels marginally faster than the rarefac-
echo-enhancers as they amplified blood tion part of the wave where the density is
reflectivity and, therefore, increased the lower. Thus, over distance, the shape of
signal-to-noise ratio for Doppler signals the waveform becomes distorted and its
amplifying blood reflectivity and increas- angular components represent the over-
ing the signal-to-noise ratio. This enhances tones or harmonics. These can be filtered
the echogenicity of the blood but creates and used for imaging (Ward et al., 1997).
more artifacts due to blooming and tissue The advantage is that they depend on a
motion (Powers et al., 1997). Another, high acoustic pressure for their formation,
more efficient way to avoid artifacts would and therefore are not produced as much
be to obtain a contrast B-mode image by the weaker unwanted side and grat-
that suppresses Doppler artifacts without ing lobes or by reverberations than by the
the use of a velocity-dependent filter. main lobe. Thus, tissue harmonic images
Contrast- enhanced ultrasound, defined as are cleaner with fewer side lobe and rever-
nonlinear or harmonic imaging, aims to beration artifacts.
provide such a method. The harmonics are generated from within
the tissue and only at the center portion
of the main US beam. This results in higher
HARMONIC IMAGING: tissue contrast. The disadvantage is that
BASIC PRINCIPLES the harmonic echoes are some 20 dB less
intense than the fundamental echoes with
Harmonics are additional frequencies that the risk of reducing the signal-to-noise
are multiples of the fundamental (transmit- ratio. The disadvantage is that the harmonic
ted) frequency and are commonly found in echoes are some 20 dB less intense than
acoustics. Double and fourfold overtones fundamental echoes with a corresponding
(harmonics) give character to the different reduction of the signal-to-noise ratio. This
musical instruments. The same pheno- is why harmonics are better handled by
menon occurs in ultrasound (Hamilton and newer generation systems powered with a
Blackstock, 1998). We need to distinguish high dynamic range than those with older
two different modes of harmonic gene- technology-designed scanners. When it
ration: those produced by transmission works well, harmonic imaging is spec-
through the tissue and those produced tacularly useful in sharpening up images
by interaction with microbubbles. Tissue since it increases contrast by eliminating
harmonics are derived from the nonlin- noise, especially in the evaluation of the
ear distortion of the incident beam as it fluid-filled structures such as the biliary
passes through the tissues. The pheno- tree, gallbladder and vascular structures
menon occurs for US beams of a sufficiently (Ortega et al., 2001). In clinical practice,
high power. A well-known basic principle for most scanners the benefit in abdomi-
of US, the speed of propagation of an nal and general work is so great that most
12. Focal Liver Lesion: Nonlinear Contrast-Enhanced Ultrasound Imaging 161
users now set the systems to default in a cardiac evaluations, are effectively one-
tissue harmonic mode; however, in our pass-only agents and have been replaced
opinion, harmonic mode is more useful by new-generation agents such as Optison,
in the evaluation of fluid-filled structures Levovist and SonoVue.
positioned in the middle field because For radiology, the two most important
a penetration loss occurs in the far field are Levovist (Schering) and SonoVue
with the poor echo intensity of deeper (Bracco). Levovist is a first-generation
tissue second harmonics (Migaleddu contrast medium of air containing micro-
et al., 2002). bubbles made by shaking galactose micro-
Microbubble nonlinear (harmonic) particles with water. The galactose
responses can be produced in two dif- micro-particles contain micro-defects,
ferent ways: microbubble oscillations or which force the attached air microbubbles
disruption. Microbubbles are air or gas into the required size. They are stabilized
bodies ranging from 1 to 7 μm in diameter, by a monomolecular layer of a surfactant,
enclosed by some sort of membrane. Two palmitic acid (Nanda and Carstensen,
important characteristics of microbubbles, 1997). The second-generation contrast
stability and diameter, must be optimized medium SonoVue uses an inert high-
in order to have efficient contrast media. density gas (sulfur hexafluoride), which
Stability means that microbubbles of con- improves the longevity of the microbub-
trast medium must dwell in the blood bles because of its high molecular weight,
stream long enough to allow the comple- which slows diffusion. Microbubble shell
tion of an US investigation with contrast is constituted by a phospholipid layer
enhancement. Microbubbles must also be which is similar to that of cell membranes
of a diameter necessary to obtain contrast (Schneider et al., 1995; Morel et al.,
medium that can pass through the pul- 2000). The essence of their function
monary circulation and that, once in the as contrast agents is that microbubbles
systemic circulation, can act as a contrast- behave quite differently from solid or
enhancing medium in all organs (http:// watery tissues in that they can be comp-
www.smirg.org/.contrast.php). Several ressed and expanded much more readily.
cardiac and vascular US contrast agents Thus, they change their diameter from
are available on the American, Canadian, two to tenfold, depending on the power
European or worldwide market; Echovist of the incident US and, like all reac-
and Levovist (Schering), SonoVue (Bracco) tive processes, have a natural resonant
are available in a number of European frequency at which they respond most
countries. Other agents that underwent or actively. It happens that the resonant fre-
are still undergoing American regulatory quency for 1–7 μm microbubbles lies in
approval include Optison (Amersham/GE the 2–10 MHz range that is used for diag-
Healthcare), Definity (BMS), Imagent nostic imaging. This lucky coincidence
(Alliance Pharmaceutical), Sonazoid explains the remarkable reflectivity of
(Amersham/GE Healthcare), Quantison microbubbles which are many more times
(Quadrant Healthcare), SonoVue (Bracco), echogenic than comparable tissue elements
Biosphere (Ponit Biomedical), and AI-700. such as red blood cells (Cosgrove in
Albunex and Echovist, used mainly in http://www.smirg.org/lectures.php).
162 V. Migaleddu and G. Virgilio
the second pulse undergoes a 180° phase signal intensity that is greater than that of
change so to be a mirror image of the first. the second harmonics alone of the contrast
The fundamental response produced in agent. The identification of this nonlinear
the tissues which will be phase-inverted response requires the use of an algorithm
compared with the fundamental pulses is capable of modulating the pulses so that
erased because the sum of two inverted fundamental responses are elicited from
pulse is zero. In this way the echoes from the tissue and from the contrast agent.
the microbubbles are enhanced so as to Such an algorithm requires perfect pulse
differentiate them from those coming from modulation both in phase and in ampli-
stationary tissues (Burns et al., 2000). tude. It has been marketed with the name
A great spatial and details resolution (capa- cadence Contrast Pulse Sequencing (CPS
bility to identify macro and micro vessels) by Acuson Sequoia, Siemens) and it works
is achieved by intermittent imaging and spe- in real-time with continuous observation,
cific software – Agent Detection Imaging using second-generation contrast agents.
(ADI by Acuson Sequoia, Siemens) – with Previous algorithms exploit only the non-
a high MI and first-generation contrast linear harmonic response of contrast agent.
agent. The ADI can be used for a short These require cancellation of the funda-
time either with static techniques or in mental echoes coming from the tissue
real-time (Migaleddu et al., 2004). forcing examination to be done blindly
Compared with the intermittent imaging (without an image) until the contrast agent
characteristic of the reconstruction algo- arrives. Contrast Pulse Sequencing software
rithms used in breaking first-generation has the immediate advantage of enabling
contrast agents, the algorithms based on observation of the fundamental signals
the use of a low mechanical index enable coming from the tissue at the same time
signals from the resonance of second- as signals coming from the nonlinear
generation microbubbles to be detected fundamental response of the contrast agent
in real-time; this permits enhancement (Phillips and Gardner, 2004).
of the contrast resolution of the vascular
phenomena present in an organ or in the
lesions it contains. These methods, however, NONLINEAR CONTRAST
show low spatial and details resolution,
useful for detecting peri- and intra-lesional
ENHACEMENT OF FOCAL
macro- and micro-circulation due to the LIVER LESIONS
low intensity of signal coming from micro-
(A) Characterization of Focal Liver
bubble resonance.
Lesions
The observation that there is also a non-
linear response coming from the contrast Due to the dual blood supply of the liver
agent in the proximity of the fundamental tissue by the hepatic artery (25–30%) and
frequency has enabled a new algorithm the portal vein (70–75%), three vascular
to be developed, which permits the non- phases can be defined in all the contrast
linear fundamental echo to be detected. imaging modalities and even in CEUS.
The advantage of detecting this nonlinear The arterial phase starts from 10–20 s after
fundamental component is an increase in contrast administration into a peripheral
164 V. Migaleddu and G. Virgilio
vein and persists for ∼10–15 s; followed by related to the perfusion behavior of the
the portal vein phase which usually lasts lesions which shows the evolution of lesion
2 min after i.v. contrast injection. The last vascular volume over time.
phase (parenchymal or sinusoidal) persists
until clearance of the contrast medium Enhancement Patterns of Benign Focal
from the hepatic parenchyma. It has two Lesions
different lengths: 4–6 min for SonoVue
Hemangioma
and 15–20 min for Levovist, which sug-
gests several mechanisms including sinu- Hemangioma is the most frequent solid
soid pooling and RES/Kupffer’s cells focal lesion in the liver from 1% to
uptake. This differs from the extracellular 20% in the general population (Ros
equilibrium phase observed in contrast et al., 2001). These can be divided into
agent CT and MR imaging (EFSUMB two different types: capillary and cavern-
Study Group. Guidelines for the Use of ous hemangioma. The first are relatively
Contrast Agents in Ultrasound 2004). In small networks of vascular spaces
our opinion the evaluation criteria of the (< 3 cm). Some of them can show an
US and CEUS examinations should be early hypervascular pattern in the arterial
specified; meaning that the terms Hypo-, phase due to concentrated arterial micro-
Iso- and Hyper-echogenic should be used vessels – high-flow hemangioma (Jang
in the description of the conventional US et al., 1998). These are generally charac-
examinations; basket patterns, peripheral terized by the presence of small arterial
or central, regular or irregular vessels for branches, venous lakes and fibrous areas,
color Doppler examinations. and absence of capsule. The latter, cav-
Terms to be used for CEUS should be: ernous hemangiomas (> 3 cm) and giant
hemangiomas (> 10 cm) may show an
1. Type of enhancement: Rim, Marginal, inhomogeneous structure composed of
Nodular; Centripetal, Centrifugal filling venous lakes, dense connective tissue and
2. Degree of enhancement or enhancing: thrombosed areas (Unal et al., 2002).
Hypo-iso- and Hyper-enhancement or Arterial portal micro-fistulas can be
enhancing – Lower, Equal or Higher identified with CEUS (http://www.smirg.
than surrounding parenchymal it/research/avi-shunt.php?sl=13). On con-
3. Time of washout of contrast in the differ- ventional US, hemangiomas are often inci-
ent phases – portal and parenchymal dental findings since they are asymptomatic
4. Lack of enhancement in the different except for the relatively rare giant type.
phases – arterial portal and parenchymal Conventional US capillary hemangiomas
5. Sustained enhancement in the different have a typical homogeneous hyper-echoic
phases – arterial portal and parenchymal appearance in comparison with surround-
Hypo-, Iso- and Hyper-echoic terms might ing hepatic parenchyma with well defined
be useful to describe static phenomena margins; in a fatty liver it may be found as
underlying the conventional US imaging a hypo-echoic lesion. Cavernous heman-
but are ambiguous for explaining dynamic giomas can present an inhomogeneous
phenomena underlying the CEUS imaging. aspect with hyper-echoic and hypo-echoic
It might be appropriate to uniform terms areas (Moody and Wilson, 1993). In all
12. Focal Liver Lesion: Nonlinear Contrast-Enhanced Ultrasound Imaging 165
Adenoma
Regenerative Nodule
Adenoma is an uncommon primary benign
Regenerative nodules are found in liver
tumor that has hepatocellular origin. Young
cirrhosis surrounded by fibrous septae.
females who use hormonal medication
The nodular dimension characterizes the
are the most frequently affected patients
micro-macro or mixed type of cirrhosis
(Karhunen, 1986). An adenoma may be
evolution. Large regenerative nodules from
asymptomatic or may present abdominal
5 to 20 mm can be detected on conven-
pain. This tumor has a marked tendency
tional US as hypo-echoic nodular lesions;
to intra-lesional or intra-abdominal high
hyper-echoic patterns due to fatty change
risk bleeding, although very rare potential
are rarer. Color Doppler or CEUS real-
malignant transformation could be consid-
time vascular evaluation show no apparent
ered. Hepatic adenoma presents (80% of
vessels in lesions (Migaleddu et al., 2004;
cases) as a solitary well-defined or encap-
Nicolau et al., 2006).
sulated tumor. This is a hypervascular
Typical features: lesion composed of cords of proliferating
Arterial phase: iso-enhancing hepatocytes without portal venous tracts,
Portal vein phase: iso-enhancing terminal hepatic veins, and bile ductules,
Parenchymal phase: iso-enhancing but with typical large sub-capsular tribu-
Additional features: tary arteries originating from the hepatic
Arterial phase: hypo- or hyper-enhancing arterial system. In larger masses necrotic
12. Focal Liver Lesion: Nonlinear Contrast-Enhanced Ultrasound Imaging 167
(Nam et al., 1996). In our experience small vascular enhancement behavior is related
encapsulated nodules may be observed. to the evolution of the lesion. Thus, dur-
Color Doppler presents some flow sig- ing the arterial phase dysplastic nodules
nals in the periphery of the lesion in large or early HCC appear usually hypo- iso-
masses and no flow signal in the small vascularized, while advanced HCCs are
nodules. Contrast-enhanced ultrasound hyper-vascularized. On conventional US,
shows marginal enhancement in the arte- findings are non-specific; usually small
rial phase, while hypo enhancement areas HCCs (< 3 cm) are hypo-echoic and homo-
are demonstrated in the portal and paren- geneous but may appear hyper-echoic due
chymal phases. Small nodules present to fatty content (Kim et al., 2001; Ward and
hypo or lack of enhancement especially in Robinson, 2002). Color Doppler or power
the late phase. This ambiguous behavior Doppler may detect characteristic hyper-
may mimic metastases. vascularity; the “basket” pattern of vessels
surrounding and penetrating or “vessel-
Enhancement Patterns of Malignant Focal within the tumor” pattern may be visual-
Lesions
ized in HCC as has been reported (Tanaka
Dysplastic Nodule, Early and Advanced et al., 1990). Contrast-enhanced ultrasound
Hepatocellular Carcinoma may show any selective enhancement in
Hepatocellular carcinoma (HCC) is the the arterial phase differentiating HCCs
most common primary hepatic malignancy. from regenerative nodules and dysplastic
Virus infections (B and particularly C), nodules (Migaleddu et al., 2004; Nicolau
genetic hemochromatosis, primary bil- et al., 2006) (Figure 12.1a–d). The most
iary cirrhosis, environmental and dietary common features are described below:
factors such as persistent organic pol-
Typical features:
lutant substances, alcohol and aflotoxins
Arterial phase: hyper-enhancing, com-
are the most important causes. Regional
plete, non-enhancing areas (necrosis)
distribution worldwide is related to the
Portal vein phase: iso-, hypo-enhancing,
different etiologies. Chronic liver diseases
non-enhancing areas (necrosis)
correlated with C virus infection and alco-
Parenchymal phase: hypo-enhancing or
hol abuse are becoming the most diffuse
wash-out
causes of HCC. Dysplastic nodules occupy
an intermediate position between regenera- Additional features:
tive nodules and well-differentiated HCC. Arterial phase: “chaotic” vessels, enhanc-
Dysplastic nodules are characterized by ing tumor thrombus in Portal Vein + HCC/
cellular atypia without the development portal vein: hypo-enhancing with intral-
of small arteries with a size ranging from esional vessels
8–10 to 15 mm. Their pre-malignant nature Portal vein phase: hypo-iso-enhancing
has been well demonstrated. Hepatic car- Parenchymal phase: hypo-iso-enhancing
cinogenesis is described as a multi-step
process where the progressive arterializa- Liver Metastases
tion and gradual loss of portal vessels are Metastatic liver disease is the most fre-
the principal features (Ueda et al., 1992; quent malignant hepatic affliction, the liver
Matsui et al., 1991). It is evident that the being the most common site for secondary
12. Focal Liver Lesion: Nonlinear Contrast-Enhanced Ultrasound Imaging 169
Figure 12.1. (a–d): Well-differentiated hepatocellular carcinoma in 80-year-old man. (a) On base line
conventional US, hyper-echoic lesion is identified (arrow). (b) On CEUS in the early arterial phase, “cha-
otic” micro-vessels within the lesion are observed (arrow). (c) On CEUS in arterial phase, hyper-enhanced
feature is evident (arrow). (d) On CEUS in late phase, a poor wash-out is demonstrated (arrow)
Figure 12.2. (a–d): Multiple Hypo-vascular liver metastases from pancreas carcinoma in 75-year-old
man. (a) On base line conventional US: a large hypo-echoic mass is well detectable. (b) On CEUS in the
early arterial phase, minimal vessels within the lesions are visible. (c) On CEUS in the arterial phase, mul-
tiple Hypo-vascular lesions show a prevalent marginal enhancement. (d) On CEUS in late phase, multiple
lacks of enhancement are evident; metastases less than 10 mm are proved (arrows)
12. Focal Liver Lesion: Nonlinear Contrast-Enhanced Ultrasound Imaging 171
Data related to the use of intra-operative blood-pool contrast agent without any
CEUS are still scarce (Siösteen and interstitial equilibrium phase. The appear-
Elvin, 2004). ance of metastases in the arterial phase is
In rectal-colon cancer, patients who variable. Hypo-vascular metastases show
underwent liver resection range between up in CEUS as hypo-enhanced lesions
30–40% in terms of 5 years survival at with and without an additional marginal
best because other metastases at the time enhancement, while hyper-vascular metas-
of resection were undetectable. To detect tases appear as brightly hyper-enhanced
metachronous metastases, a 3-monthly more or less homogeneous lesions. Hyper-
conventional US and yearly CT follow-up vascular metastases occur most often from
program may identify 88% of patients who primary tumors of neuro-endocrine origin
will develop liver metastases in an asympto- or from renal or breast cancer.
matic stage in order to plan liver resection, A common pitfall is that small cysts are
chemotherapy or percutaneous treatments sometimes detected on late phase scanning.
(Howell et al., 1999). Additionally, it has These can usually be distinguished from
been reported that patients with diffuse metastases as they characteristically show
metastatic diseases receiving systemic a higher US beam transmission (Bleuzen
chemotherapy at an asymptomatic stage and Tranquart, 2004). Differential diag-
have higher response rates, better quality nosis may be difficult for metastases with
of life, and survival, than those who begin dominant cystic component, as in ovarian
the chemotherapy at a symptomatic stage metastases. Contrast-enhanced ultrasound
(Nordic Gastrointestinal Tumor Adjuvant could help to observe enhanced septae
Therapy Group, 1992). In these follow-up or nodules that are very frequent in such
programs CEUS results as the more cost- metastases. Thrombosed hemangiomas
effectiveness imaging procedure in the may mimic the hypo enhancing behav-
detection of liver metastases. ior of the metastases. Contrast-enhanced
Evaluation in the late parenchymal phase ultrasound follow-up could identify the
with high M.I., using a first-generation modification of lesion vascularity due to
contrast agent (Levovist) has been recom- intralesional thrombus changes over time.
mended in some reports (Albrecht et al.,
Surveillance for Hepatocellular Carcinoma
2003b). A long liver-specific late phase
of up to 20 min has been described due Surveillance means repeated application
to adherence of the microbubbles to the of screening tests. Conventional US and
hepatic sinusoid or to phagocytosis by serum alpha-fetoprotein are the most
Kupffer’s cells (Albrecht et al., 2001). widespread screening tests for hepatocel-
The late phase is the most useful time lular carcinoma particularly in European
to detect metastases that usually appear countries; a biannual interval between US
as hypo-enhanced or lack enhancement examinations is generally accepted. A ran-
areas. In comparison, most benign lesions domized study demonstrated the benefit of
show uptake at this time and are therefore such a program considering that in the sur-
not likely to be confused with metastases. veillance arm the HCC-related mortality
Second-generation gas filled microbub- was reduced by 37% (Zhang et al., 2004).
bles as in SonoVue are considered a truly During the Consensus Conference of
174 V. Migaleddu and G. Virgilio
European Association for the study of the low M.I., 100 primary tumors (96.2%)
liver (held in Barcelona in the year 2000), demonstrated hyper-enhancement in the
a biannual US and AFL level evaluations arterial phase but 4 did not (3.8%) (Nicolau
were recommended as the minimal follow up. et al., 2004). One study did not find arterial
A spiral CT investigation is suggested for enhancement in three well differentiated
patients who do not have nodules on US HCCs out of 60 cases consequently high-
but an increasing AFP level. For patients lighted the relationship between absence
who have a US finding of a nodule of of enhancement in the arterial phase and
10 mm or less in diameter, 3 monthly US histological differentiation (Tanaka et al.,
is suggested taking into account that these 2001). After diagnosis of HCC, the pres-
lesions are too small to be accurately char- ence and degree of cirrhosis, as well as
acterized and at least 50% of these lesions the tumor extension, need to be evalu-
will not be HCC. In patients with a nod- ated for staging and management. In fact,
ule of > 20 mm in diameter HCC can be the presence of satellite nodules, biliary
confirmed if the AFP level is over 400 ng/ extension, vascular, and extra-hepatic
ml and there is contrast-enhanced CT, involvement modify therapeutic planning.
MR, or angiographic proof of the hyper- Entire hepatic evaluation during the arte-
vascularity of the lesion. At least two pieces rial phase (20–30 s) might be difficult, also
of positive imaging evidence are required because the sweeping scanning might fail
for HCC diagnosis (Bruix et al., 2001). to detect some lesions. The administration
Biopsy is a controversial option espe- of several doses of contrast agents and
cially for small nodules as it has a poten- scanning of different hepatic segments
tial risk of bleeding and tumor seeding in are suggested as well as the use of a high
the needle track, whereas it is not always M.I. at intervals to evaluate the destruction
useful due to the heterogeneity of vari- and parenchyma replenishment in order to
ous types of hepatic nodular lesions (Brú detect hyper-enhancing foci. To date, there
et al., 1989). Microbubble contrast as echo- are no data enabling comparison between
enhancer permits useful information to be CEUS and spiral CT MRI in this chal-
obtained for diagnosis if prevalent hyper- lenging field. The extra-hepatic staging
vascularity of the lesion is demonstrated favors total body cross-sectional imaging.
(Lencioni et al., 1996; Fracanzani et al., The evaluation of portal tumor invasion is
2001). These imaging modalities are asso- obviously easier with US examination.
ciated with motion and blooming artifacts.
With first and second-generation con-
trast media, CEUS reveals a well defined (C) Monitoring of Percutaneous Ablative
hyper-enhancement in arterial phase and Treatment
prevalent washout in the late phase. For In the last 10 years numerous changes
real-time evaluation of all three phases, the have occurred in therapeutic approach
use of second-generation contrast medium for primary and metastatic hepatic
and low M.I., is more suitable (Nicolau malignancies. Many minimally invasive
et al., 2006). therapies as the percutaneous ethanol
In a series of 104 patients, using a second- injection (PEI) and the thermal ablation
generation of contrast media SonoVue and using different energy sources (radio
12. Focal Liver Lesion: Nonlinear Contrast-Enhanced Ultrasound Imaging 175
Figure 12.3. (a–d): Hepatocellular carcinoma: pre and post radio-frequency ablative treatment features.
(a) and (b): Pre; on CEUS in the early and arterial phase: typical hyper-vascular pattern is identified
(arrows). (c) and (d): Post; on CEUS in the early and arterial phase: no vessels within the lesion and no
enhancement are present. 100% of necrosis is achieved (arrows)
176 V. Migaleddu and G. Virgilio
Earlier we have described the guide- useful to read together fundamental echoes
lines for primary tumor management pro- from tissue as well as nonlinear fundamental
duced by the European Association for the echoes from the contrast agent; the CPS
Study of Liver Disease (EASLD) (Bruix (the most recent technological approach)
et al., 2001). The American Association for permits this also with second-generation
the Study of Liver Disease (AASLD) has contrast media in real-time. The possibility
recently published “Practice Guidelines of aiming at the lesion, and in particular
for the management of HCCs”. The use seeing both extra- and intra-lesional vascular
of only one contrast enhanced imaging contribution in the arterial phase, consid-
modality including CEUS to demonstrate erably assists early phase characteriza-
hyper-vascularization and diagnose a HCC tion of HCC. Using second-generation
was suggested for a lesion larger than of contrast media, nonlinear fundamental
20 mm (Bruix and Sherman, 2005). (CPS) and “conventional” harmonic imag-
The characterization of the focal liver ing findings proves to be equivalent in
lesions by evaluation of hemodynamic the study of the late phase useful for
contrast enhancement behavior is con- metastases diagnosis. This is true except
firmed by an important number of studies, for the lesions positioned in the distal
using both first or second-generation con- field because there is poor intensity of the
trast media (Leen et al., 2002; Migaleddu contrast harmonic echoes in comparison
et al., 2004; Wen et al., 2004; Quaia et al., with contrast nonlinear fundamental ones
2004; Catalano et al., 2005). (Migaleddu et al., 2005). Excellent details
Whereas some authors have emphasized resolution especially in the arterial phase
the importance of observing enhancement is also valuable in outcome assessment of
at a late time, in our experience it is also ablative therapy.
important to observe arterial and early
portal enhancement. Other experiences REFERENCES
seem to confirm this (Nicolau et al.,
Abbey-Toby, A., Cazals-Hatem, D., Colombat, M.,
2006). In particular, detection through Belghiti, J., Vilgrain, V., and Degott, C. 2003.
good details resolution of peri- and intra- Inflammatory pseudo-tumor of the liver: is pre-
lesional macro- and micro-vessels con- operative diagnosis possible? Gastroenterol.
tributes to the diagnosis of focal liver Clin. Biol. 27: 883–890.
lesions, above all in the cirrhotic liver and Akahn, O., and Pringot, J. 2002. Imaging of abdom-
inal tubercolosis. Eur. Radiol. 12: 321–323.
therefore in differential diagnosis between
Albrecht, T., Hoffmann, C.W., Schmitz, S.A.,
Regenerative macro-nodules, Dysplastic Schettler, S., Ovemberg, A., Germen, C.T., and
nodules in the transformation phase and Wolf, K.J. 2001. Phase-inversione sonography
well-differentiated HCC; as well, early during the liver-specific late phase of contrast
rim enhancement can suggest diagnosis enhancement: improved detection of liver metas-
of high flow hemangioma. The study of tases. Am. J. Roentgenol. 176 (5): 1191–1198.
Albrecht, T., Blomley, M.J., Burns, P.N., Wilson, S.,
the only late phase is more useful in the
Harvey, C.J., Leen, E., Claudon, M., Calliada,
diagnosis of metastases. F., Correas, J.M., LaFortune, M., Campani,
In order to obtain excellent details reso- R., Hoffmann, C.W., Cosgrove, D.O., and
lution, especially in the arterial phase, it is LeFevre, F. 2003a. Improved detecion of hepatic
178 V. Migaleddu and G. Virgilio
metastases with pulse-inversion US during the Chin C.T., and Burns P.N. 1997. Predicting the
liver-specific phase of SHU 505A: multicenter acoustic response of a microbubble population
study. Radiology 227: 361–370. for contrast imaging. In: IEEE Ultrason. Symp.
Albrecht, T., Oldenburg, A., Hohmann, J., Skrok, J., 22: 1557–1560.
Hoffmann, C.W., Schettler, S., and Wolf, K.J. Clarke, M.P., Kane, R.A., Steele, G. Jr., Hamilton,
2003b. Imaging of liver metastases with contrast- E.S., Ravikumar, T.S., Onik, G., and Clouse,
specific low-MI real-time ultrasound and SonoVue. M.E. 1989. Prospective comparision of pre-
Eur. Radiol. 13 (Suppl 3): 79–86. operative imaging and intraoperative ultrasonog-
Albrecht, T., Hohmann, J., Oldenburg, A., Skrok, J., raphy in the detecion of liver tumours. Surgery
and Wolf, K.J. 2004. Detection and characterisa- 106: 849–855.
tion of liver metastases. Eur. Radiol. 14 (Suppl 8): Conlon, R., Jacobs, M., Dasgupta, D., and Loge,
25–33. J.P. 2003. The value of intraoperative ultra-
Apfel, R.E., and Holland, C.K. 1991. Gauging the sound during hepatic resection compared with
likelihood of cavitation from short-pulse, low improved preoperative magnetic resonace imag-
duty cycle diagnostic ultrasound. Ultrasound ing. Eur. Ultrasound 16: 211–216.
Med. Biol. 17: 175–185. Cosgrove, D.O. 2001. Malignant liver disease. In:
Bartolozzi, C., Cioni, D., Donati, F., and Lencioni, R. Meire, H.B., Cosgrove, D.O., Dewbury, K.C.,
2001. Focal liver lesions: MR imaging-patho- and Farrant, P. (eds). Clinical Ultrasound: A
logic correlation. Eur. Radiol. 11: 1374–1388. Comprehensive Text. Abdominal and General
Bleuzen, A., and Tranquart, F. 2004. Incidental Ultrasound, vol. 1, 2nd ed. Churchill Livingstone,
liver lesions: diagnostic value of cadence con- London, 211–231.
trast pulse sequencing (CPS) and SonoVue. Eur. Cosgrove, D.O. 2005. Harmonic imaging: tissue
Radiol. 14 (Suppl 8): 53–62. and microbubbles in http://www.smirg.org/lec-
Brauner, M., Buffard, M.D., Jaentils, V., Legrand, I., tures.php.
and Gotheil, C. 1989. Sonography and computed De Baere, T., Elias, D., Dromain, C., Din, M.G.,
tomography of mocroscopic tuberculosis of the Kuoch, V., Ducreux, M., Boige, V., Lassau, N.,
liver. J. Clin. Ultrasound 17: 563–568. Marteau, V., Lasser, P., and Roche, A. 2000.
Brú, C., Marotto, A., Bruix, J., Faus, R., Biachi, L., Radiofrequency ablation of 100 hepatic metas-
Calvet, X., Ayuso, C., Vilana, R., Gilalbert, R., tases with a mean follow-up of more than 1 year.
and Rodes, J. 1989. Diagnostic accurancy of Am. J. Roentgenol. 175: 1619–1625.
fine-needle biopsy in patients with hepatocellular Dodd, G.D. 3rd., Soulen, M.C., Kane, R.A.,
carcinoma. Dig. Dis. Sci. 34: 1765–1769. Livraghi, T., Lees, W.R., Yamashita, Y.,
Bruix, J., and Sherman, M., 2005. Management Gillams, A.R., Karahan, O.I., and Rhim, H.
of Hepatocellular carninoma. J. Hepatol. 42: 2000. Minimally invasive treatment of malig-
1208–1236. nant hepatic tumors: at the threshold of a major
Bruix, J., Sherman, M., Llovet, J.M., Beaugrand, M., breakthrough. Radiographics 20: 9–21.
Lencioni, R., Burroughs, A.K., Christensen, E., EFSUMB Study Group Albrecht, T., Blomley, M.,
Pagliaro, L., Colombo, M., and Rodes, J. 2001. Bolondi, L., Claudon, M., Correas, J.M.,
Clinical management of hepatocellular carci- Cosgrove, D., Greiner, L., Jager, K., Jong, N.D.,
noma. Conclusions of the Barcelona. EASL Leen, E., Lencioni, R., Lindsell, D., Martegani,
Conference. J. Hepatol. 35: 421–430. A., Solbiati, L., Thorelius, L., Tranquart,
Burns, P.N., Wilson, S.R., and Simpson, D.H. F., Weskott, H.P., and Whittingham, T.
2000. Pulse inversion imaging of liver blood 2004. Guidelines for the Use of Contrast
flow: an improved method for characterization Agents in Ultrasound. Ultraschall. Med. 25:
of focal masses with microbubble contrast. 249–256.
Invest. Radiol. 35: 58–71. Fracanzani, A.L., Burdick, L., Borzio, M., Roncalli, M.,
Catalano, O., Nunziata, A., Lobianco, R., and Siani, A. Bonelli, N., Borzio, F., Maraschi, A., Fiorelli,
2005. Real-time harmonic contrast material- G., and Fargion, S. 2001. Contrast-enhanced
specific US of focal liver lesions. RadioGraphics Doppler ultrasonography in the diagnosis of
25: 333–349. hepatocellular carcinoma and premalignant
12. Focal Liver Lesion: Nonlinear Contrast-Enhanced Ultrasound Imaging 179
lesions in patients with cirrhosis. Hepatology 34: sign at two-phase spiral CT. Radiology 208:
1109–1112. 543–548.
Goldberg, S.N., Gazelle, C.S., Compton, C.C., Karhunen, P.J. 1986. Benign hepatic tumours and
Mueller, P.R., and Tanabe, K.K. 2000. Treatment tumour-like conditions in men. Clin. Pathol. 39:
of intrahepatic malignancy with radiofrequency 183–188.
ablation: radiologic-pathologic correlation. Kim, C.K., Lim, J.H., and Lee, W.J. 2001. Detection
Cancer 88: 2452–2463. of hepatocellular carcinomas and dysplastic nod-
Golli, M., Van Nhieu, J.T., Mathieu, D., Zafrani, ules in cirrhotic liver: accuracy of ultrasonogra-
E.S., Cherqui, D., Dhumeaux, D., Vasile, N., and phy in transplant patients. J. Ultrasound Med.
Rahmouni, A. 1994. Hepatocellular adenoma: 20: 99–104.
Color Doppler US and pathologic correlations. Kroncke, T.J., Taupitz, M., Kivelitz, D., Scheer,
Radiology 190: 741–744. I., Daberkow, U., Rudolph, B., and Hamm, B.
Grazioli, L., Federle, M.P., Ichikawa, T., Balzano, E., 2000. Multifocal nodular fatty infiltration of the
Nalesnik, M., and Madariaga, J. 2000. Liver liver mimicking metastatic disease on CT: imag-
adenomatosis: clinical, histopathologic and ing findings and diagnosis using MR imaging.
imaging findings in 15 patients. Radiology 216: Eur. Radiol. 10: 1095–1100.
395–402. Leen, E., Angerson, W.J., Yarmenitis, S.,
Greis, C. 2004. Technology overview: SonoVue Bongartz, G., Blomley, M., Del Maschio, A.,
(Bracco, Milan). Eur. Radiol. 14 (Suppl 8): 11–15. Summaria, V., Maresca, G., Pezzoli, C., and
Hamilton, M.F., and Blackstock, D.T. 1998. Llul, J.B. 2002. Multi-centre clinical study
Nonlinear Acoustics. Academic, San Diego, CA. evaluating the efficacy of SonoVue (BR1),
Harvey, C.J., and Albrecht, T. 2001. Ultrasound of a new ultrasound contrast agent in Doppler
focal liver lesions. Eur. Radiol. 11: 1578–1593. investigation of focal hepatic lesions. Eur. J.
Harvey, C.J., Blomley, M.J., Eckersley, R.J., Radiol. 42: 200–206.
Cosgrove, D.O., Patel, N., Heckemann, R.A., and Lencioni, R., Pinto, F., Armillota, N., and
Butler-Barnes, J. 2000. Hepatic malignancies: Bartolozzi, C. 1996. Assessment of tumor vasco-
improved detection with pulse-inversion US in larity in hepatocarcinoma: comparison of power
late phase of enhancement with SH U 508A-early Doppler and Color Doppler US. Radiology 201:
experience. Radiology 216: 903–908. 353–358.
Horiuchi, R., Uchida, T., Kojima, T., and Shikata, T. Lencioni, R.A., Allgaier, H.P., Cioni, D.,
1990. Inflammatory pseudotumor of the liver. Olschewski, M., Deibert, P., Crocetti, L., Frings,
Clinicalpathological study and review of litera- H., Laubenberger, J., Zuber, I., Blum, H.E.,
ture. Cancer 65: 1583–1590. and Bartolozzi, C. 2003. Small hepatocellular
Howell, J.D., Wotherspoon, H., Leen, E., Cooke, carcinoma in cirrhosis: randomized compari-
T.C., and MacArlde, C.S. 1999. Evaluation of son of radiofrequency thermal ablation versus
a follow up programme after curative resec- percutanous ethanol injection. Radiology 228:
tion for colorectal cancer. Br. J. Cancer 79: 235–240.
308–310. Levine, C. 1990. Primary macronodular hepatic
Ishak, K.G., Anthony, P.P., and Sobin, L.H. 1994. tuberculosis: US and CT appearances. Gastro-
World Health Organization: International Histo- intest. Radiol. 15: 307–309.
logic Classification of Tumors, 2nd ed. Berlin, Livraghi, T. 2006. Role of image-guided thera-
Springer. pies in hepatocellular carcinoma in http://www.
Iwase, K., Higaki, J., Yoon, H.E., Mikata, S., smirg.org/lectures.php.
Miyazaki, M., Torikai, K., Shirai,Y., Awai, K., Livraghi, T., Benedini, V., Lazzaron, S., Meloni, F.,
Imakita, M., and Kamiike, W. 2002. Solitary Torzilli, G., and Vettori, C. 1998. Long term
necrotic nodule of the liver. J. Hepatobiliary results of single session PEI in patients with large
Pancreat. Surg. 9: 120–124. hepatocellular carcinoma. Cancer 83: 48–57.
Jang, H.J., Choi, B.I., Kim, T.K., Jun, A.J., Kim., Livraghi, T., Meloni, F., Morabito, A., and Vettori,
K.W., Han, J.K., and Han, M.C. 1998. Atypical C. 2004. Multimodal image-guided tailored
small hemangiomas of the liver: “bright dot” therapy of early and intermediate hepatocellular
180 V. Migaleddu and G. Virgilio
carcinoma: long-term survival in the experi- management of the cirrothic patient and for detec-
ence of a single radiologic referral center. Liver tion of HCC. Eur. Radiol. 14 (Suppl 8): 63–71.
Transpl. 10: 98–106. Nicolau, C., Vilana, R., Catalá, V., Bianchi, L.,
Matsui, O., Kadoja, M., Kamejama, T., Takashima, T., Gilabert, R., Carcía, A., and Brú, C. 2006.
Nakanuma, J., Unoura, M., Kobayashi, K., Importance of evaluating all vascular phases on
Izumi, R., and Ida, M. 1991. Benign and malig- contrast enhanced sonography in the diferentia-
nant nodules in cirrhotic patients: distinction on tion of benign from malignant focal lesions. Am.
blood supply. Radiology 178: 493–497. J. Roentgenol. 186: 158–167.
Migaleddu, V., Virgilio, G., Campisi, G., Sirigu, D., Nordic Gastrointestinal Tumour Adjuvant Therapy
and Canalis, G.C. 2002. Conventional ultra- Group, 1992. Expectancy or primary chemo-
sonography versus tissue harmonic imaging for therapy in patients with advanced asymptomatic
the assessment of the common bile duct in chole- colorectal cancer; a randomisede trial. J. Clin.
cystectomized patients. Radiol. Med. 103: 1–6. Oncol. 10: 904–911.
Migaleddu, V., Virgilio, G., Turilli, D., Conti, M., Numata, K., Tanaka, K., Kiba, T., Norimoto, M.,
Campisi, G., Canu, M., Sirigu, D., and Arata, S., Kondo, M., and Sekihara, H. 1997.
Vincentelli, I. 2004. Charaterization of focal Use of hepatic tumor index on color Doppler
liver lesions in real-time using harmonic sonography for differentiating large hepatic
imaging with high mechanical index and con- tumors. Am. J. Roentgenol. 168: 991–995.
trast agent levovist. Am. J. Roentgenol. 182: Ortega, D., Burns, P.N., Simpson, D.H., and
1505–1512. Wilson, S.R. 2001. Tissue harmionic imaging:
Migaleddu, V., Manca, A., Turilli, D., Luciano, P., is it a benefit for bile duct sonography? Am. J.
Canu, N., Campisi, G., and Virgilio, G. 2005. Roentgenol. 176: 653–659.
Fundamental nonlinear imaging versus harmonic Phillips, P., and Gardner, E. 2004. Contrast-agent
imaging with contrast medium sonovue in the detection and quantification. Eur. Radiol. 14
evaluation of the hepatic focal lesions: prelimi- (Suppl 8): 4–10.
nary results. Eur. Radiol. 14 (Suppl. 1): B-126. Powers, J.E., Burns, P.N., and Souquet, J. 1997.
Mirilas, P., and Skandalakis, J.E. 2002. Benign Imaging instrumentation for ultrasound con-
anatomical mistakes: incidentaloma. Am. Surg. trast agents. In: Nanda, N.C., Schlief, R., and
68: 1026–1028. Goldberg, B.B. (eds). Advances in Echo Imaging
Moody, A.R., and Wilson, S.R. 1993. Atypical Using Contrast Enhancement, 2nd, ed. Kluwer,
hepatic hemangioma: a suggestive sonographic Dordrecht, 139–170.
morphology. Radiology 188: 413–417. Quaia, E., Calliada, F., Bertolotto, M., Rossi, S.,
Morel, D.R., Schwieger, I., Hohn, L., Terrettaz. J., Garioni, L., Rosa, L., and Pozzi-Mucelli, R.
Llull, J.B., Cornioley, Y.A., and Schneider, M. 2004. Characterization of focal liver lesions with
2000. Human pharmacokinectics and safety eval- contrast-specific US modes and a sulfur hexaflu-
uation of SonoVueTM, a new contrast agent for oride-filled microbubble contrast agent: diag-
ultrasound imaging. Invest. Radiol. 35 (Suppl 1): nostic performance and confidence. Radiology
80–85. 232: 420–430.
Nam, K.J., Kang, H.K., and Lim, J.H. 1996. Ros, P., and Davis, G. 1998. The incidental focal liver
Inflammatory pseudotumor of the liver: CT and lesion: photon, proton, or needle? Hepatology
sonographic features. Am. J. Roentgenol. 167: 98: 1183–1190.
485–487. Ros, P.R., Menu, Y., Vilgrain, V., Mortele, K.J., and
Nanda, N.C., and Carstensen, E.L. 1997. Echo- Terris, B. 2001. Liver neoplasms and tumor-like
enhancing agents: safety. In: Nanda N.C., conditions. Eur. Radiol. 11 (Suppl 2): 145–165.
Schlief, R., and Goldberg, B.B. (eds). Advances Schneider, M., Arditi, M., Barrau, M.B., Brochot,
in Echo Imaging Using Contrast Enhancement, J., Broillet, A., Ventron, R., and Yan, F. 1995.
2nd, ed.Kluwer, Dordrecht, 139–170. BR1: a new ultrasonographic contrast agent
Nicolau, C., Vilana, R., and Brù, C. 2004. The based on sulfur hexafluoride-filled microbub-
use of contrast-enhanced ultrasound in the bles. Invest. Radiol. 30: 451–457.
12. Focal Liver Lesion: Nonlinear Contrast-Enhanced Ultrasound Imaging 181
Schwartz, L.H., Gandras, E.J., Colangelo, S.M., Ueda, K., Terada, T., Nakanuma, Y., and Matsui, O.
Ercolani, M.C., and Panicek, D.M. 1999. 1992. Vascular supply in adenomatous hyperplasia
Prevalence and importance of small hepatic of the liver and hepatocellular carcinoma: a mor-
lesions found at CT in patients with cancer. phometric study. Hum. Pathol. 23: 619–626.
Radiology 210: 71–74. Unal, O., Sakarya, M.E., Arslan, H., Tuncer, I.,
Siösteen, A.K., and Elvin, A. 2004. Intra-operative and Etlik, O. 2002. Hepatic cavernous hemangi-
uses of contrast-enhanced ultrasound. Eur. omas: patterns of contrast enhancement on MR
Radiol. 14 (Suppl 8): 87–95. fluoroscopy imaging. Clin. Imaging 26: 39–42.
Solbiati, L., Livraghi, T., Golberg, S.N., Ierace, T., Volk, M., Strotzer, M., Lenhart, M., Techert, J.,
Meloni, F., Dellanoce, M., Cova, L., Halpern, Seitz, J., and Feuerbach, S. 2001. Frequency of
E.F., and Gazelle, G.S. 2001. Percutaneous benign hepatic lesions incidentally detected with
radio-frequency ablation of liver metastase contrast-enhanced section portal venous pahase
from colon rectal cancer: long term risults 117 spiral CT. Acta Radiol. 42: 172–175.
patients. Radiology 221: 159–166. Ward, B., Baker, A.C., and Humprey, V.F. 1997.
Solbiati, L., Tonolini, M., and Cova, L. 2004. Non linear propagation applied to the improve-
Monitoring RF ablation. Eur. Radiol. 14 (Suppl 8): ment of resolution in diagnostic medical ultra-
34–42. sound. J. Acoust. Soc. Am. 101: 143–154.
Tanaka, S., Kitamura, T., Fuijta, M., Nakanishi, K., Ward, J., and Robinson, J. 2002. How to detect
and Okuda, S. 1990. Color Doppler flow imag- hepatocellular carcinoma in chirrhosis. Eur.
ing of liver tumors. Am. J. Roentgenol. 154: Radiol. 12: 2258–2272.
509–514. Wen, Y.L., Kudo, M., Zheng, R.Q., Ding, H., Zhou, P.,
Tanaka, S., Ioka, T., Oshikawa, O., Hamada, Y., and Minami, Y., Chung, H., Kitano, M., Kawasaki, T.,
Yoshioka, F. 2001. Dynamic sonography of hepatic and Maekawa, K. 2004. Characterizazione of
tumors. Am. J. Roentgenol. 177: 799–805. hepatic tumors: value of contrast-enhanced
Torzilli, G., Del Fabro, D., Palmisano, A., Donadon, M., coded phase-inversion harmonic angio. Am. J.
Bianchi, P., Roncalli, M., Balzarini, L., and Roentgenol. 182: 1019–1026.
Montorsi, M. 2005. Contrast-enhanced intraop- Zhang, B.H., Yang, B.H., and Tang, Z.Y. 2004.
erative ultrasonography during hepatectomies Randomized controll trial of screening for hepa-
for Colon rectal metastaes. J. Gastrointest. Surg. tocellular carcinoma. J. Cancer Res. Clin. Oncol.
9: 1148–1154. 130: 417–422.
13
Hepatocellular Carcinoma: Magnetic
Resonance Imaging
Bachir Taouli
183
184 B. Taouli
(whole acquisition time < 30 min). The degree of differentiation (Ebara et al.,
protocol includes: T1-weighted in- and out- 1999; Amano et al., 2003). In-phase and
of-phase gradient-recalled echo (GRE) to opposed-phase images can demonstrate
evaluate for microscopic fat content, Turbo microscopic fat components within HCC,
Spin Echo T2-weighted with spectral fat areas of fat accumulation showing a signal
saturation or T2-fast STIR (short tau inver- drop on opposed-phase images.
sion recovery), T2-weighted single-shot On T2-weighted images, HCCs are
RARE (rapid acquisition with relaxation classically hyperintense, whereas regener-
enhancement) and 3D T1-weighted fat ative nodules (Figure 13.1) and most dys-
suppressed GRE before and after dynamic plastic nodules are hypointense (Earls et al.,
injection of 15–20 ml of gadolinium 1996; Matsui et al., 1989; Krinsky et al.,
chelates using an automatic injector, with 2001) (Figures 13.2 and 13.4). However,
acquisitions at the arterial, portal venous, well-differentiated HCC lesions can be
and equilibrium phases. We use the short- iso- or rarely hypointense on T2-weighted
est possible TR (repetition time) in order images (Earls et al., 1996; Hecht et al.,
to minimize acquisition time and allow 2006). A focus of high signal intensity on
sequential acquisitions through the entire T2 within a hypo- or isointense nodule
liver during breath-holding. Fat suppres- (“nodule within a nodule” sign) is highly
sion improves the conspicuity of contrast suggestive of HCC developing within a
enhancement, and eliminates chemical- dysplastic nodule (Mitchell et al., 1991;
shift artifacts at fat-water interfaces. van den Bos et al., 2006). Ebara et al.
(1999) have shown that the signal inten-
sity of HCC on T1- and T2-weighted
MAGNETIC RESONANCE images is related to the degree of his-
IMAGING APPEARANCE tologic differentiation and intratumoral
OF HEPATOCELLULAR copper content, whereas the signal inten-
CARCINOMA sity on T2-weighted images is related to
the degree of histologic differentiation.
The macronodular architecture of the However, a relationship between copper
cirrhotic liver and the morphology of content and signal intensity of HCC was
HCC are usually well demonstrated on T1- not confirmed by other studies (Nakakoshi
and T2-weighted images. On T1-weighted et al., 1996). Hemosiderin deposition is
images, HCC can be hypo-, iso- or hyper- common in regenerative nodules (called
intense, compared to the surrounding liver siderotic nodules), producing specific imag-
(Aschoff et al., 2001; Honda et al., 1997). ing features on MRI, such as hypointensity
Most well differentiated HCCs and high- on T1-GRE images and T2-weighted GRE
grade dysplastic nodules have high signal images (Krinsky et al., 2001). Dysplastic
intensity on T1-weighted images, and T1 nodules are usually homogeneously hyper-
hyperintensity can be related to copper, intense on T1-weighted images, but small
Fe3+, glycogen deposition, or high pro- HCC nodules may exhibit similar findings
tein or lipid content, and possibly to the (Krinsky and Israel, 2003).
13. Hepatocellular Carcinoma: Magnetic Resonance Imaging 185
Figure 13.1. (a–c): Regenerative nodule and HCC in the same patient. (a): T2-weighted images show
that the regenerative nodule is hypointense (left arrow) and the HCC is hyperintense (right arrow). (b):
Post-contrast images at the arterial phase show that the regenerative nodule is not enhancing (left arrow)
whereas the HCC is hypervascular (right arrow). (c): Post-contrast images at the portal venous phase
show that the regenerative nodule is hypovascular (left arrow) and the HCC has washout with a capsule
(right arrow)
186 B. Taouli
Figure 13.2. Dynamic post-contrast T1-weightedimages at the arterial (left) and portal venous (right)
show a segment 4 hypervascular HCC (arrow) with washout at the portal venous phase
13. Hepatocellular Carcinoma: Magnetic Resonance Imaging 187
Figure 13.3. Dynamic post-contrast T1-weighted images at the arterial phase (left: source image, right:
coronal maximum intensity projection reformat) show multifocal HCCs (arrows). The largest lesion is
located in the posterior right hepatic lobe, with disseminated nodules in the rest of the liver
dropped to 38% for lesions smaller than intense pseudo-lesions and HCCs is
1 cm. However, these authors did not difficult with MRI, and follow-up is
evaluate dysplastic nodules. In a previ- often required. Hyperintense pseudo-
ous study (Ward et al., 2000), using lesions should be recognized by their
SPIO and dynamic gadolinium injection lack of interval growth or frank dis-
in 27 patients (with explant correla- appearance (Shimizu et al., 2003;
tion available in 15 patients), the same Jeong et al., 2002), whereas HCC will
authors were able to detect significantly have a tendency to grow. Jeong et al.
more HCC nodules (particularly HCCs (2002) have calculated the doubling
< 1 cm), although few lesions were vis- time of small enhancing lesions, and
ible only on SPIO-enhanced images. showed that all nodules with interval
Dysplastic nodules were also character- growth and/or change in MR signal are
ized with greater confidence on the basis HCCs (mean doubling time of HCCs,
of their combined enhancement charac- 2 months 2 weeks). The same authors
teristics (no enhancement with gadolin- reported positive and negative predic-
ium, enhancement with SPIO). However, tive values of 100% and 98% respec-
this approach is more complex and adds tively, on the basis of nodule growth,
to the cost and duration of MRI. Another and suggested a close interval follow-up
study found that the benefit of the com- in such patients.
bined use of SPIO and gadolinium is 2. Large HCC: A large HCC can have
negligible (Halavaara et al., 2002). a number of distinctive features, such
as a mosaic pattern, tumor capsule,
satellite nodules, vascular invasion,
ATYPICAL APPEARANCES and metastasis. The mosaic pattern is
OF HEPATOCELLULAR related to confluent nodules separated
CARCINOMA by areas of necrosis and thin septa,
best seen on T2-weighted images and
1. Small HCC: A small HCC is defined post-contrast images (Kadoya et al.,
as a lesion < 2 cm in size. Most small 1992). A large HCC often has a fibrous
HCC will show intense arterial phase tumor capsule (seen in 60–82% of
enhancement. The widespread use of cases) (Kadoya et al., 1992; Choi et
arterial phase CT and MRI in cir- al., 1993), appearing hypointense on
rhotic patients has shown the existence T1- and T2-weighted images, rarely
of enhancing “pseudo-lesions” (small hyperintense on T2-weighted images,
hepatic nodules with arterial phase with delayed enhancement. Vascular
enhancement, not seen on the portal invasion is often present histologically
venous phase images), that may be in the HCC, affecting the portal vein
wedge-shaped, geographic, oval or and the hepatic veins, particularly in a
round in shape (Shimizu et al., 2003). large non-encapsulated HCC. On MRI,
These lesions may represent arterioportal vascular invasion appears as non-
shunts, but the underlying structure is visualization of the portal branches or
still not clear (Yu et al., 1997). The as an intraluminal mass, with enhance-
differentiation between small hyper- ment at the arterial phase, and a filling
13. Hepatocellular Carcinoma: Magnetic Resonance Imaging 189
Figure 13.4. Diffuse type HCC with portal vein invasion. Pre-contrast T2- (left upper row), pre- (right
upper row) and post-contrast (lower image) T1-weighted images show an infiltrative HCC of the right
hepatic lobe (arrows) invading the portal vein (small arrow)
190 B. Taouli
This is likely related to free water motion in In conclusion, although MRI with
benign lesions, and restricted water motion dynamic gadolinium enhancement appears
in the presence of tumor. We showed in to be superior to helical CT for detecting
a previous study (Taouli et al., 2003) a HCC, its sensitivity for the detection of
significant difference between ADCs of small lesions is still low, and improvement
benign and malignant liver lesions (2.45 ± in accuracy is still needed. The use of new
0.96 × 10−3 and 1.08 ± 0.50 × 10−3 mm2/s contrast media, high field (3T) magnets
for b = 0 and 500 s/mm2; respectively, and new MRI techniques such as perfusion
p < 0.001). The mean ± SD ADCs (x10−3 and diffusion MRI will likely improve the
mm2/s) of the different groups of lesions image quality and the accuracy of MRI in
were: HCCs 1.33 ± 0.13, hemangiomas the future.
2.95 ± 0.67 and cysts 3.63 ± 0.56 (Taouli
et al., 2003).
REFERENCES
Perfusion-weighted MRI: Perfusion-weighted
MRI provides the ability to detect regional Amano, S., Ebara, M., Yajima, T., Fukuda, H.,
Yoshikawa, M., Sugiura, N., Kato, K., Kondo, F.,
and global alterations in liver blood flow, Matsumoto, T., and Saisho, H. 2003. Assessment
and could be performed with CT or MRI. of cancer cell differentiation in small hepatocel-
Whole-liver CT imaging has the potential lular carcinoma by computed tomography and
to provide both high-temporal-resolution magnetic resonance imaging. J. Gastroenterol.
and high-spatial-resolution imaging of the Hepatol. 18:273–279.
entire liver for HCC detection; however, Aschoff, A.J., Merkle, E.M., Wong, V., Zhang,
Q., Mendez, M.M., Duerk, J.L., and Lewin,
the radiation dose must be considered. J.S. 2001. How does alteration of hepatic blood
Perfusion-weighted MRI represents a flow affect liver perfusion and radiofrequency-
promising technique for HCC surveillance induced thermal lesion size in rabbit liver? J.
(Pandharipande et al., 2005). The closest Magn. Reson. Imaging 13:57–63.
technique to perfusion-weighted MRI has Bhartia, B., Ward, J., Guthrie, J.A., and Robinson,
been use of a double-arterial phase MR P.J. 2003. Hepatocellular carcinoma in cirrhotic
livers: double-contrast thin-section MR imaging
imaging technique enabled by a rapid par- with pathologic correlation of explanted tissue.
allel imaging method; one group reported AJR Am. J. Roentgenol. 180:577–584.
sensitivity of HCC detection of 91.7% Choi, B.I., Takayasu, K., and Han, M.C. 1993.
using double arterial phase vs. 76.3% with Small hepatocellular carcinomas and associ-
a single arterial phase (Yoshioka et al., ated nodular lesions of the liver: pathology,
2002). These findings suggest the poten- pathogenesis, and imaging findings. AJR Am. J.
Roentgenol. 160:1177–1187.
tial of high-temporal-resolution perfusion- Earls, J.P., Theise, N.D., Weinreb, J.C., DeCorato,
weighted MRI to improve detection of D.R., Krinsky, G.A., Rofsky, N.M., Mizrachi, H.,
HCC. and Teperman, L.W. 1996. Dysplastic nodules
3T MRI: Advantages of imaging at 3T and hepatocellular carcinoma: thin-section MR
include higher signal-to-noise ratio, better imaging of explanted cirrhotic livers with patho-
image contrast, particularly in gadolinium- logic correlation. Radiology 201:207–214.
Ebara, M., Fukuda, H., Kojima, Y., Morimoto, N.,
enhanced applications, which translate to Yoshikawa, M., Sugiura, N., Satoh, T., Kondo, F.,
higher spatial resolution images and faster Yukawa, M., Matsumoto, T., and Saisho, H. 1999.
imaging. This could potentially improve Small hepatocellular carcinoma: relationship of
HCC detection in the future. signal intensity to histopathologic findings and
13. Hepatocellular Carcinoma: Magnetic Resonance Imaging 191
metal content of the tumor and surrounding ing for liver disease. AJR Am. J. Roentgenol.
hepatic parenchyma. Radiology 210:81–88. 173:393–398.
El-Serag, H.B., and Mason, A.C. 1999. Rising inci- Krinsky, G.A., and Israel, G. 2003. Nondysplastic
dence of hepatocellular carcinoma in the United nodules that are hyperintense on T1-weighted
States. N. Engl. J. Med. 340:745–750. gradient-echo MR imaging: frequency in cir-
Fung, J., and Marsh, W. 2002. The quandary over rhotic patients undergoing transplantation. AJR
liver transplantation for hepatocellular carcinoma: Am. J. Roentgenol. 180:1023–1027.
the greater sin? Liver Transpl. 8:775–777. Krinsky, G.A., Lee, V.S., Nguyen, M.T., Rofsky,
Halavaara, J., Tervahartiala, P., Isoniemi, H., and N.M., Theise, N.D., Morgan, G.R., Teperman,
Hockerstedt, K. 2002. Efficacy of sequential use L.W., and Weinreb, J.C. 2001. Siderotic nodules
of superparamagnetic iron oxide and gadolinium in the cirrhotic liver at MR imaging with explant
in liver MR imaging. Acta. Radiol. 43:180–185. correlation: no increased frequency of dys-
Hecht, E.M., Holland, A.E., Israel, G.M., Hahn, W.Y., plastic nodules and hepatocellular carcinoma.
Kim, D.C., West, A.B., Babb, J.S., Taouli, B., Lee, Radiology 218:47–53.
V.S., and Krinsky, G.A. 2006. Hepatocellular car- Matsui, O., Kadoya, M., Kameyama, T., Yoshikawa,
cinoma in the cirrhotic liver: gadolinium-enhanced J., Arai, K., Gabata, T., Takashima, T., Nakan-uma,
3D T1-weighted MR imaging as a stand-alone Y., Terada, T., and Ida, M. 1989. Adenomatous
sequence for diagnosis. Radiology 239:438–447. hyperplastic nodules in the cirrhotic liver: differ-
Honda, H., Kaneko, K., Kanazawa, Y., Hayashi, entiation from hepatocellular carcinoma with MR
T., Fukuya, T., Matsumata, T., Maeda, T., and imaging. Radiology 173:123–126.
Masuda, K. 1997. MR imaging of hepatocellular Matsui, O., Kadoya, M., Kameyama, T., Yoshikawa,
carcinomas: effect of Cu and Fe contents on sig- J., Takashima, T., Nakanuma, Y., Unoura, M.,
nal intensity. Abdom. Imaging 22:60–66. Kobayash, K., Izumi, R., Ida, M., et al. 1991.
Jeong, Y.Y., Mitchell, D.G., and Kamishima, T. Benign and malignant nodules in cirrhotic livers:
2002. Small (<20 mm) enhancing hepatic nod- distinction based on blood supply. Radiology
ules seen on arterial phase MR imaging of the 178:493–497.
cirrhotic liver: clinical implications. AJR Am. J. Mazzaferro, V., Regalia, E., Doci, R., Andreola,
Roentgenol. 178:1327–1334. S., Pulvirenti, A., Bozzetti, F., Montalto, F.,
Kadoya, M., Matsui, O., Takashima, T., and Ammortuna, M., Morabito, A., and Gennari, L.
Nonomura, A. 1992. Hepatocellular carci-noma: 1996. Liver transplantation for the treatment of
correlation of MR imaging and histopathologic small hepatocellular carcinomas in patients
findings. Radiology 183:819–825. with cirrhosis. N. Engl. J. Med. 334:693–699.
Katyal, S., Oliver, J.H. 3rd, Peterson, M.S., Ferris, Mitchell, D.G., Kim, I., Chang, T.S., Vinitski, S.,
J.V., Carr, B.S., and Baron, R.L. 2000. Extrahe- Consigny, P.M., Saponaro, S.A., Ehrlich, S.M.,
patic metastases of hepatocellular carcinoma. Rifkin, M.D., and Rubin, R. 1991. Fatty liver.
Radiology 216:698–703. Chemical shift phase-difference and suppres-
Kanematsu, M., Semelka, R.C., Leonardou, sion magnetic resonance imaging techniques in
P., Mastropasqua, M., and Lee, J.K. 2003. animals, phantoms, and humans. Invest. Radiol.
Hepatocellular carcinoma of diffuse type: MR 26:1041–1052.
imaging findings and clinical manifestations. J. Murakami, T., Kim, T., Oi, H., Nakamura, H.,
Magn. Reson. Imaging 18:189–195. Igarashi, H., Matsushita, M., Okamura, J., and
Kim, T., Murakami, T., Oi, H., Matsushita, M., Kozuka, T. 1995. Detectability of hypervascu-
Kishimoto, H., Igarashi, H., Nakamura, H., and lar hepatocellular carcinoma by arterial phase
Okamura, J. 1995. Detection of hypervascular images of MR and spiral CT. Acta Radiol.
hepatocellular carcinoma by dynamic MRI and 36:372–376.
dynamic spiral CT. J. Comput. Assist. Tomogr. Murakami, T., Baron, R.L., Peterson, M.S., Oliver,
19:948–954. J.H. 3rd, Davis, P.L., Confer, S.R., and Federle,
Kim, T., Murakami, T., Takahashi, S., Hori, M., M.P. 1996. Hepatocellular carcinoma: MR imag-
Tsuda, K., and Nakamura, H. 1999. Diffusion- ing with mangafodipir trisodium (Mn-DPDP).
weighted single-shot echoplanar MR imag- Radiology 200:69–77.
192 B. Taouli
Nakakoshi, T., Kajiyama, M., Fujita, N., Jong- diffusion isotropy and characterization of focal
Hon, K., Takeichi, N., and Miyasaka, K. 1996. hepatic lesions with two single-shot echo-planar
Quantitative analyses of correlations of signal MR imaging sequences: prospective study in 66
intensity on T1-weighted images and T1 relaxa- patients. Radiology 226:71–78.
tion time with copper concentration in the rat van den Bos, I.C., Hussain, S.M., Terkivatan, T.,
liver. Acad. Radiol. 3:36–39. Zondervan, P.E., and de Man, R.A. 2006. Stepwise
Oi, H., Murakami, T., Kim, T., Matsushita, M., carcinogenesis of hepatocellular carcinoma in the
Kishimoto, H., and Nakamura, H. 1996. Dynamic cirrhotic liver: demonstration on serial MR imag-
MR imaging and early-phase helical CT for ing. J. Magn. Reson. Imaging 24:1071–1080.
detecting small intrahepatic metastases of hepa- Ward, J., Guthrie, J.A., Scott, D.J., Atchley, J.,
tocellular carcinoma. AJR Am. J. Roentgenol. Wilson, D., Davies, M.H., Wyatt, J.I., and
166:369–374. Robinson, P.J. 2000. Hepatocellular carcinoma
Pandharipande, P.V., Krinsky, G.A., Rusinek, H., in the cirrhotic liver: double-contrast MR imag-
and Lee, V.S. 2005. Perfusion imaging of the ing for diagnosis. Radiology 216:154–162.
liver: current challenges and future goals. Yamashita, Y., Mitsuzaki, K., Yi, T., Ogata, I.,
Radiology 234:661–673. Nishiharu, T., Urata, J., and Takahashi, M. 1996.
Rofsky, N.M., Weinreb, J.C., Bernardino, M.E., Small hepatocellular carcinoma in patients with
Young, S.W., Lee, J.K., and Noz, M.E. 1993. chronic liver damage: prospective comparison of
Hepatocellular tumors: characterization with detection with dynamic MR imaging and helical
Mn-DPDP-enhanced MR imaging. Radiology CT of the whole liver. Radiology 200:79–84.
188:53–59. Yoshioka, H., Takahashi, N., Yamaguchi, M., Lou,
Shimizu, A., Ito, K., Koike, S., Fujita, T., Shimizu, D., Saida, Y., and Itai, Y. 2002. Double arterial
K., and Matsunaga, N. 2003. Cirrhosis or chronic phase dynamic MRI with sensitivity encoding
hepatitis: evaluation of small (< or =2-cm) early- (SENSE) for hypervascular hepatocellular carci-
enhancing hepatic lesions with serial contrast- nomas. J. Magn. Reson. Imaging 16:259–266.
enhanced dynamic MR imaging. Radiology Yu, J.S., Kim, K.W., Sung, K.B., Lee, J.T., Yoo,
226:550–555. and H.S. 1997. Small arterial-portal venous
Taouli, B., Vilgrain, V., Dumont, E., Daire, J.L., shunts: a cause of pseudolesions at hepatic imag-
Fan, B., and Menu, Y. 2003. Evaluation of liver ing. Radiology 203:737–742.
14
Expression of Vascular Endothelial Growth
Factor in Hepatocellular Carcinoma:
Correlation with Radiologic Findings
Masayuki Kanematsu, Richard C. Semelka, and Shinji Osada
193
194 M. Kanematsu et al.
several advanced cancers, including color- the first or second most common cause
ectal cancer, breast cancer, glioblastoma of cancer death worldwide, and in 1990
multiforme, prostate, and renal cancer the World Health Organization estimated
where VEGF is known to play a role, that there were ∼430,000 new cases of
have been expected. Most interestingly, liver cancer worldwide, and a similar
previous results have demonstrated that number of patients died as a result of this
gadolinium-enhanced MRI findings may disease. The most common risk factors for
be useful biomarkers for defining the HCC include hepatitis B or C viral infec-
pharmacological response of colorectal tion, aflatoxins, iron overload, and alcohol
cancer and liver metastases and for dose abuse. This disease is particularly preva-
of angiogenesis inhibitors, such as PTK/ lent in parts of Asia, Africa, and Europe.
ZK (Morgan et al., 2003). However, rela- Eighty percent of HCC patients have cir-
tions between VEGF activity in tumors rhosis of the liver, and 70% have elevated
and radiologic findings have yet to be serum alpha fetoprotein levels. Treatment
ascertained. options include liver transplantation, par-
We previously assessed the correla- tial hepatectomy, transcatheter arterial
tion between radiologic findings in ang- embolization, percutaneous ablation, and
iographically-assisted CT and MRI and radiation therapy. Reported 5-year survival
angiogenetic activities determined by rates range from 10% to 60%, depend-
immunohistochemistry (Kanematsu et al., ing on the tumor stage and the treatment
2004a) and Western blotting (Kanematsu method. The number of hepatitis virus
et al., 2004b, 2005) for VEGF in vari- carriers in Japan is estimated to be > 3
ous hepatocellular nodules in cirrhosis, million (1.6% of the population), and of
HCCs, and the surrounding liver. They these 1 million develop chronic hepati-
have described the potential usefulness tis and 30,000 HCC. A national survey
of radiologic imaging in the evaluation of (the third National Health and Nutrition
angiogenetic activity of hepatic nodules Examination Survey [NHANES III]) of the
and the surrounding liver. We summa- civilian, non-institutionalized U.S. popula-
rize the results and discuss the relations tion found that 1.8% of Americans (3.9
between radiologic findings and VEGF million) have been infected with hepatitis
expression in this chapter. C virus (HCV), of whom most (2.7 mil-
lion) are chronically infected with HCV.
One interesting footnote is that tattooing,
which is gaining in popularity, increases
HISTORY OF RADIOLOGIC the risk of HCV infection by ninefold.
DIAGNOSIS Hepatocellular carcinoma may become
OF HEPATOCELLULAR the most common malignant tumor in
CARCINOMA North America in a decade or two, and is
the only major cancer with a substantial
Hepatocellular carcinoma is a malignant increase in incidence.
neoplasm that arises from hepatocytes A pathology-based study from 1984
chiefly in patients with chronic liver dam- reported that the common major gross
age. Hepatocellular carcinoma is either patterns were expanding, spreading, and
14. Expression of Vascular Endothelial Growth Factor in Hepatocellular Carcinoma 195
As the HCC grows beyond 3 cm in size, ing in angiograms, and VEGF protein was
the distance between tumor vessels and substantial in HCC cells on immunohis-
cancer cells tends to increase. As shown by tochemical observation. In the same year,
Folkman (1971), oxygen diffuses no more Suzuki et al. (1996) studied vascular per-
than 150 μm in tissue, thus some parts of meability factor (VPF)/VEGF transcript
the enlarged HCC develop hypoxia and expression using reverse-transcription
resultant internal necrosis. The hypoxia polymerase chain reaction (PCR) analy-
induces the generation of VEGF peptides sis, and reported that 16 (69.6%) cases
once again. However, the generation of showed VEGF transcript expression in
tumor vessels is physically limited when the tumor, whereas only 9 (39%) showed
the HCC reaches 5–15 cm in size, due to it in adjacent liver parenchyma. Vascular
tissue pressure and cell crowding, and large endothelial growth factor mRNA expres-
HCCs may require anaerobic metabolism sion in HCCs in their study was associated
with glycolytic pathway with hexokinase with fibrous capsule formation and septal
II to sustain themselves (Gwak et al., formation. Vascular endothelial growth
2005). This is observed on imaging stud- factor activity was not correlated with the
ies as HCCs measuring 5–15 cm in size vascularity of HCCs as determined by con-
often do not enhance well on contrast- ventional angiography.
enhanced hepatic arterial-phase CT or MRI, El-Assal et al. (1998) performed quantita-
despite highly activated VEGF expression. tive reverse-transcription PCR and immu-
Yamaguchi et al. (1998) described that sar- nostaining in 71 patients with HCCs to
comatous HCCs, the final form of dediffer- determine microvessel density (MVD) and
entiation with highly malignant behavior, VEGF expressions in HCCs and in the sur-
showed low VEGF expression in immuno- rounding liver. They reported that tumor vas-
histochemistry studies. cularity on angiography was not correlated
with the MVD, and neither VEGF mRNA
levels nor protein expression in HCCs was
CORRELATION OF VASCULAR correlated with the MVD or any histopatho-
ENDOTHELIAL GROWTH logical features of HCCs, although they did
FACTOR EXPRESSION AND not directly correlate VEGF expression and
angiographic findings.
RADIOLOGIC FINDINGS Kwak et al. (2001) attempted to deter-
IN HEPATOCELLULAR mine whether VEGF is a histopathological
CARCINOMA IN PRIOR factor influencing contrast enhancement of
REPORTS HCC by qualitatively evaluating 22 nodular
HCCs on intravenously contrast-enhanced
Mise et al. (1996) performed Northern CT in correlation with the qualitative
blot analysis and immunohistochemistry to extent of anti-VEGF antibody expres-
semi-quantify and localize the expression sion determined by immunohistochemical
of VEGF in 20 HCCs and nine metastatic staining. They reported that the degree of
tumors, and reported that the degree of enhancement on the hepatic arterial-phase
VEGF mRNA expression was significantly CT had a significant positive correlation
correlated with the intensity of tumor stain- with VEGF expression in HCCs.
200 M. Kanematsu et al.
Magnetic resonance imaging was per- fat deposition in HCCs, the phase-shift
formed in a similar fashion to our prior index was calculated as follows: phase-
study obtaining T1-weighted gradient- shift index = (SI in-phase – SI opposed-
echo, T2-weighted echo-train spin-echo, phase)/SI in-phase, where SI in-phase and
and gadolinium-enhanced multiple-phase SI opposed-phase are the signal intensi-
gradient-echo images in the hepatic arte- ties of HCC on in-phase and opposed-
rial-dominant, portal venous, and equilib- phase T1-weighted gradient-echo images,
rium phases. Computed tomography was respectively.
performed using a combination of unen- To measure CT values in HCCs and in
hanced CT, CT during arterial portogra- the surrounding liver, a circular region
phy, and CT hepatic arteriography. On of interest was drawn to encompass as
CT portography, scan was started 25–35 s much of the HCC as possible, and another
after the initiation of a transcatheter supe- was drawn in the surrounding liver paren-
rior mesenteric arterial injection of 95 ml chyma. Quantitative degree of contrast
of nonionic contrast material containing enhancement was expressed as contrast
150 mg iodine/ml at 3 ml/s, and on CT enhancement index, which was calculated
arteriography scan was started 5 s after by subtracting the CT values on unenhanced
the initiation of a transcatheter common CT from those on contrast-enhanced CT.
hepatic arterial injection of 30 ml of the Two experienced gastrointestinal radi-
same contrast at 1.5 ml/s. ologists independently reviewed MRI in
This study evaluated MRI data using a retrospective manner. They subjectively
a quantitative approach. Our technique evaluated the signal intensity in HCC and
was as follows: to determine MR signal in the surrounding liver using a seven-point
intensity values and their standard devia- scale from “strong hypointensity (−3)” to
tions (SDs) in HCCs and in the surround- “marked hyperintensity (+3)”. Intensity of
ing liver, a circular region of interest was enhancement and degree of heterogeneity
drawn to encompass as much of the lesion in HCCs were determined on postcontrast
as possible, and another circular region of images using four-point scales from “vir-
interest was drawn in a region of the sur- tually no enhancement (0)” to “marked
rounding liver. The SD of the background enhancement (+3)” and from “virtually no
noise, SDB, was measured in the phase- heterogeneity (0)” to “marked heterogene-
encoding direction outside the anterior ity (+3)”, respectively.
abdominal wall to calculate the following: The two radiologists independently
lesion-to-liver contrast-to-noise ratio = (SI reviewed CT findings in a retrospective
lesion – SI liver)/SDB, where SI lesion manner. They evaluated the contrast-
and SI liver are the signal intensities of enhanced CT with reference to unen-
the HCC and of the surrounding liver, hanced CT to determine the degree of
respectively. As a quantitative parameter contrast enhancement in HCC and in the
of signal intensity heterogeneity, the sig- surrounding liver using a four-point scale
nal intensity SD ratio was calculated as from “virtually no enhancement (0)” to
follows: SD ratio = SD lesion/SDB, where “intense enhancement (+3)”. The radiolo-
SD lesion is the signal intensity SD of gists evaluated the degree of heterogeneity
the HCC. As a quantitative parameter of of hepatic arterial contrast enhancement in
204 M. Kanematsu et al.
a b
c d
Figure 14.2. A 57-year-old man with chronic type C viral hepatitis and a poorly differentiated 5.8-cm
HCC showing high VEGF expression. Child-Pugh grade was A. a, Fat-suppressed T2-weighted echo-
train spin-echo (repetition time [TR] of 4,286 ms, echo time [TE] of 80 ms) axial image shows the HCC
(arrow) as a moderately hyperintense lesion with internal areas of higher signal intensity (arrowheads)
due to internal necrosis. b, On gadolinium-enhanced hepatic arterial-phase T1-weighted gradient-echo
(TR of 150 ms, TE of 1.6 ms) axial image, the HCC is totally hypovascular other than its peripheries,
which are slightly enhanced (arrows). Central areas corresponding to hyperintense internal areas on a
remain unenhanced (arrowheads). c, CT hepatic arteriogram shows weak hepatic arterial enhancement
with moderate heterogeneity in the HCC (arrow). Note that central areas of the HCC exhibit water density
with no enhancement due to necrosis. d, Scheme shows electrophoretic bands and corresponding histo-
grams in this patient; 1.25 mg/ml of VEGF solution was used for calibration. Areas of the histogram were
376 pixels for the calibration band, 3,485 pixels for the HCC band, and 1,395 pixels for the surrounding
liver band. The VEGFIND values were 9.27 in the HCC and 3.71 in the surrounding liver. Note that electro-
phoretic peaks adjacent to those of HCC and liver are due to expression of irregular protein. (Permission
to reproduce granted by the American Journal of Roentgenology.)
a b
Figure 14.3. A 74-year-old woman with chronic type C viral hepatitis and a moderately differentiated
2-cm HCC showing weak VEGF expression. Child-Pugh grade was A. a, CT hepatic arteriogram shows
very intense, homogeneous hepatic arterial enhancement in the HCC (arrow). b, Scheme shows elec-
trophoretic bands and their corresponding histograms in this patient; 1.25 mg/ml of VEGF solution was
used for calibration. Areas of histogram were 480 pixels for the calibration band, 761 pixels for the HCC
band, and 2,375 pixels for the surrounding liver band. The VEGFIND values were 1.59 in the HCC and
4.95 in the surrounding liver. Note that electrophoretic peak adjacent to that of liver is due to expression
of irregular protein. (Permission to reproduce granted by the American Journal of Roentgenology)
between VEGF expression in HCCs and level of VEGF mRNA in HCC tissues did
tumor size, Yamaguchi et al. (1998), who not significantly correlate with tumor size,
immunohistochemically investigated the cellular differentiation, capsule, daughter
expression of VEGF in HCCs, found that nodules, vascular permeation, necrosis and
VEGF positivity gradually decreased with hemorrhage of tumors. Ng et al. (2001)
an increase in tumor size. It may be similarly reported that there was no sta-
that because the population they evalu- tistically significant association between
ated included a greater number and per- tumor VEGF mRNA levels and tumor size
centage of small (15 mm or less) HCCs or other pathologic features. Nevertheless,
than other reports, that they might have our opinion is that positive correlation
observed VEGF activity changes during between VEGF expression and tumor
the continuous dedifferentiation in small size is reasonable, because intratumoral
HCCs. Kwak et al. (2001) also observed hypoxia or necrosis induced by tumor
that tumor size correlated negatively with growth may induce VEGF protein synthe-
immunohistochemical VEGF expression. sis (Shweiki et al., 1992; Mukhopadhyay
While, Jeng et al. (2004) found that the et al., 1995; von Marschall et al., 2001).
14. Expression of Vascular Endothelial Growth Factor in Hepatocellular Carcinoma 207
in regions of tumor with decreased oxy- microvascular density and VEGF expres-
gen tension. Previous research indicated sion than noncirrhotic livers. The reason
that the hypoxic regions of solid tumors why in our study the correlations between
produced powerful and directly acting the intensity of hepatic arterial enhance-
angiogenic proteins such as VEGF, or that, ment and of VEGF expression differed
in an experimental model using rat livers, between HCCs and the surrounding liver
VEGF peptides were produced by non- was unclear, this discrepancy in part may
parenchymal and parenchymal cells after be associated with the difference in mag-
necrosis (Mukhopadhyay et al., 1995). nitude between hepatic arterial perfusion
Fat deposition is observed in various in a HCC and in the surrounding liver: the
types of HCCs. Kutami et al. (2000) contrast enhancement indices of HCCs
reported that fatty changes in small HCCs ranged from 33 to 236 Hounsfield unit
are closely related to tumor size, histologic (HU) (mean, 105.3 ± 51.7 HU), whereas
grade, and the insufficient development the contrast enhancement indices of the
of arterial tumor vessels. Previous stud- surrounding liver ranged from 9 to 68 HU
ies have demonstrated that lipid bodies (mean, 28.9 ± 15.7 HU). The increased
in endothelial cells are induced during oxygen supply to the HCC secondary to
hypoxia in any cell type (Scarfo et al., neovascularization was sufficiently great
2001). Based on these previous results, to suppress VEGF expression via VEGF
we suspect that VEGF expression, hepatic downregulation (Figure 14.1B), but the
arterial perfusion, hypoxia, and fat deposi- increased oxygen supply in the surround-
tion in HCCs are closely related. However, ing liver might not be sufficient to cause
in our study, we did not observe a cor- VEGF downregulation (Figure 14.3).
relation between fat deposition of HCCs The most contentious aspect of VEGF
and VEGF expression, where phase shift correlation is the relationship with the con-
imaging accurately reflects intratumoral comitant presence of increased vascularity.
fat deposition. Kwak et al. (2001) correlated tumor atten-
The contrast enhancement index uation qualitatively, and determined on
of the surrounding liver on CT hepatic intravenously contrast-enhanced CT scans
arteriography showed moderate direct with immunoreactivity for anti-VEGF
correlation with the VEGFINDs of the sur- antibody, which was qualitatively deter-
rounding liver with marginal statistical mined by immunohistochemical staining.
significance, which suggested that the They concluded that the degree of VEGF
greater the hepatic arterial enhancement expression in HCC was directly correlated
of the surrounding liver, the stronger is the with the degree of contrast enhancement
VEGF expression in the liver. Rosmorduc during the hepatic arterial phase. The
et al. (1999) described experimental bil- exact reasons why their results differ from
iary cirrhosis, indicating that there was ours (Kanematsu et al., 2004b, 2005) in
evidence that angiogenesis was stimulated terms of the correlation between the inten-
primarily by VEGF in response to hepa- sity of hepatic arterial tumor enhancement
tocellular hypoxia caused by liver dam- and VEGF expression in HCC is unclear.
age. El-Assal et al. (1998) reported that Eighteen (82%) of 22 patients had type-B
cirrhotic livers had a significantly higher hepatitis and 1 (5%) had type-C hepatitis
14. Expression of Vascular Endothelial Growth Factor in Hepatocellular Carcinoma 209
in their study, whereas in our study only histologic tumor grades was uneven.
6 (21%) of 28 patients had type-B hepa- Multi-institutional studies will be needed
titis and 22 (79%) had type-C hepatitis. to confirm our results. Also, although
This substantial difference in the patient we employed Western blotting to semi-
populations and in the underlying hepatic quantify VEGF peptides, this technique is
disease might affect the results. The sensi- limited in terms of its ability to differenti-
tivity for enhancement and timing of data ate the multiple types of VEGF peptides
acquisition of single detector row spiral that exist in cell membranes, cytoplasm,
CT, and subjective ratings of contrast and interstitial spaces.
enhancement on CT images and of VEGF
expression may also affect statistical cor-
relations. SUMMARY
Some researchers have reported that
VEGF activity does not correlate with We have reviewed the historical devel-
the vascularity of HCCs as determined opment of various diagnostic imaging
by conventional angiography (Suzuki modalities for the evaluation of HCC.
et al., 1996; El-Assal et al., 1998), while We further described the research pub-
others have reported that VEGF activity lications on the correlation of VEGF
correlates directly with the intensity of and imaging findings. The apparently
tumor stain on angiography (Mise et al., discrepant observations of the extent of
1996). All these reports focused on the correlation of VEGF activity and HCC,
clinical significance of VEGF expression we believe, reflects that different studies
in HCC and in the surrounding liver, the contained patients with HCCs of different
evaluation of VEGF activity by means evolution. Based on our research studies,
of VEGF mRNA quantification with we have observed that VEGF appears
Northern blotting, and VEGF localiza- to correlate with increased tissue fluid,
tion by immunohistochemical staining, which on MRI appears as low signal on
while the radiological assessment of the T1-weighted images and high signal on
degree of tumor staining by angiogra- T2-weighted images. It is our impres-
phy was performed by subjective rat- sion that VEGF increases as a precursor
ings using a two- or three-point scale. to increased vascularity, as vascularity
We believe that our study methodology increases, seen in small HCCs, VEGF
more accurately clarified the relation- is downregulated and as tumor increases
ship between the vascularity of HCCs to larger size, > 3 cm, VEGF may again
and liver and VEGF activity, as contrast increase in response to hypoxia.
enhancement was quantitatively deter- The future clinical impact of VEGF
mined using region-of-interest measure- measurements and correlation of VEGF
ments and VEGF was determined using values with imaging results have yet to
Western blotting technique. be determined. Our hope is that VEGF
Limitations of our studies have included measurements will provide insight into the
that the study populations were small, angiogenic potential of tumors, and also
reflecting single institution accrual, predict or reflect response to therapeutic
and distribution of varying types of interventions.
210 M. Kanematsu et al.
222584, an inhibitor of the vascular endothe- Shweiki, D., Itin, A., Soffer, D., and Keshet, E.
lial growth factor receptor tyrosine kinases, in 1992. Vascular endothelial growth factor induced
patients with advanced colorectal cancer and by hypoxia may mediate hypoxia-initiated ang-
liver metastases: results from two phase I stud- iogenesis. Nature 359: 843–845.
ies. J. Clin. Oncol. 21: 3955–3964. Suzuki, K., Hayashi, N., Miyamoto, Y., Yamamoto, M.,
Mukhopadhyay, D., Tsiokas, L., Zhou, X.M., Ohkawa, K., Ito, Y., Sasaki, Y., Yamaguchi, Y.,
Foster, D., Brugge, J.S., and Sukhatme, V.P. Nakase, H., Noda, K., Enomoto, N., Arai, K.,
1995. Hypoxic induction of human vascular Yamada, Y., Yoshihara, H., Tujimura, T.,
endothelial growth factor expression through Kawano, K., Yoshikawa, K., and Kamada, T.
c-Src activation. Nature 375: 577–581. 1996. Expression of vascular permeability fac-
Murakami, T., Kim, T., Takamura, M., Hori, M., tor/vascular endothelial growth factor in human
Takahashi, S., Federle, M.P., Tsuda, K., Osuga, hepatocellular carcinoma. Cancer Res. 56:
K., Kawata, S., Nakamura, H., and Kudo, M. 3004–3009.
2001. Hypervascular hepatocellular carcinoma: Ueda, K., Matsui, O., Kawamori, Y., Nakanuma,
detection with double arterial phase multi-detector Y., Kadoya, M., Yoshikawa, J., Gabata, T.,
row helical CT. Radiology 218: 763–767. Nonomura, A., and Takashima, T. 1998.
Ng, I.O., Poon, R.T., Lee, J.M., Fan, S.T., Ng, M., Hypervascular hepatocellular carcinoma: evalu-
and Tso, W.K. 2001. Microvessel density, vas- ation of hemodynamics with dynamic CT during
cular endothelial growth factor and its receptors hepatic arteriography. Radiology 206: 161–166.
Flt-1 and Flk-1/KDR in hepatocellular carci- Viglietto, G., Maglione, D., Rambaldi, M., Cerutti,
noma. Am. J. Clin. Pathol. 116: 838–845. J., Romano, A., Trapasso, F., Fedele, M.,
Rosmorduc, O., Wendum, D., Corpechot, C., Ippolito, P., Chiappetta, G., and Botti, G. 1995.
Galy, B., Sebbagh, N., Raleigh, J., Housset, C., Upregulation of vascular endothelial growth
and Poupon, R. 1999. Hepatocellular hypoxia- factor (VEGF) and downregulation of placenta
induced vascular endothelial growth factor growth factor (PlGF) associated with malig-
expression and angiogenesis in experimental bil- nancy in human thyroid tumors and cell lines.
iary cirrhosis. Am. J. Pathol. 155: 1065–1073. Oncogene 11: 1569–1579.
Scarfo, L.M., Weller, P.F., and Farber, H.W. 2001. von Marschall, Z., Cramer, T., Hocker, M.,
Induction of endothelial cell cytoplasmic lipid Finkenzeller, G., Wiedenmann, B., and Rosewicz S.
bodies during hypoxia. Am. J. Physiol. Heart 2001. Dual mechanism of vascular endothelial
Circ. Physiol. 280: 294–301. growth factor upregulation by hypoxia in human
Semelka, R.C., Shoenut, J.P., Kroeker, M.A., hepatocellular carcinoma. Gut 48: 87–96.
Greenberg, H.M., Simm, F.C., Minuk, G.Y., Yamaguchi, R., Yano, H., Iemura, A., Ogasawara,
Kroeker, R.M., and Micflikier, A.B. 1992. Focal S., Haramaki, M., and Kojiro, M. 1998.
liver disease: comparison of dynamic contrast- Expression of vascular endothelial growth factor
enhanced CT and T2-weighted fat-suppressed, in human hepatocellular carcinoma. Hepatology
FLASH, and dynamic gadolinium-enhanced MR 28: 68–77.
imaging at 1.5 T. Radiology 184: 687–694. Yamane, A., Seetharam, L., Yamaguchi, S., Gotoh, N.,
Shimizu, A., Ito, K., Koike, S., Fujita, T., Shimizu, K., Takahashi, T., Neufeld, G., and Shibuya, M.
and Matsunaga, N. 2003. Cirrhosis or chronic 1994. A new communication system between
hepatitis: evaluation of small (< or = 2-cm) early- hepatocytes and sinusoidal endothelial cells in
enhancing hepatic lesions with serial contrast- liver through vascular endothelial growth factor
enhanced dynamic MR imaging. Radiology 226: and Flt tyrosine kinase receptor family (Flt-1
550–555. and KDR/Flk-1). Oncogene 9: 2683–2690.
15
Detection of Small Hepatic Lesions:
Superparamagnetic Oxide-Enhanced
Diffusion-Weighted T2 FSE Imaging
Shigeru Kiryu and Kuni Ohtomo
INTRODUCTION SUPERPARAMAGNETIC
IRON OXIDE-ENHANCED
The liver is a common site for metastases
MAGNETIC RESONANCE
from colorectal carcinoma and other neo-
plasms. Most colorectal carcinoma deaths IMAGING
are attributable to hepatic metastases, There are several methods such as com-
and the number of these metastases is puted tomography (CT), magnetic reso-
increasing (Ballantyne and Quin, 1993). nance imaging (MRI), and ultrasonography
The prognosis of patients without treat- for preoperative imaging assessment in
ment is dismal because most cases die detecting liver metastases. Presently, super-
within a few years after discovering the paramagnetic iron oxide (SPIO)-enhanced
liver lesion. Hepatic resection is the sole MRI is being widely used to detect liver
treatment that regularly leads to long- metastases. This oxide is taken up by
term survival with a possible chance of phagocytosis of the reticuloendothelial
cure in patients with hepatic metastases, system, including Kupffer cells of the liver,
especially those of colorectal origin. substantially shortening T2 of liver tissue
With techniques such as intraoperative (Stark et al., 1988). The hepatic metastases
ultrasonography (Makuuchi et al., 1991), that contain negligible or a few Kupffer
which improves the management of peri- cells remain largely nonenhanced, while
operative patients, hepatic resection has the normal liver is enhanced and shows
extended the possibility of liver surgery in a low signal intensity on T2-weighted
patients with advanced metastatic tumors images, with the result that the lesion-to-
(Kawasaki et al., 1994). Also, hepatic liver signal intensity ratio is improved on
resection has been proved to improve SPIO-enhanced MRI, as compared with
the prognosis of patients (Penna and nonenhanced T2-weighted images.
Nordlinger, 2002). Therefore, accurately In a multicenter trial, SPIO-enhanced MRI
detecting small lesions is indispensable revealed additional lesions not seen on
in preoperative imaging assessment when unenhanced images in 27% of cases and
planning hepatic resection. additional lesions not seen by conventional
213
214 S. Kiryu and K. Ohtomo
CT in 40% (Ros et al., 1995). In a mul- parenchyma; therefore, the small vessels
tiobserver study that compared the sen- and lesions exhibit similar signal intensities.
sitivities of dual phase spiral CT with Furthermore, SPIO-enhanced MRI dis-
SPIO-enhanced MRI for detecting metas- played more peripheral vascular structures
tases, both techniques showed high sen- than contrast enhanced CT (Fretz et al.,
sitivities for detecting lesions larger than 1990). Distinguishing small lesions from
1 cm at 94% vs. 99%, whereas the sen- the signal intensities from hepatic vessels
sitivities of dual phase spiral CT and is the key to detect small hepatic lesions.
SPIO-enhanced MRI for detecting lesions
smaller than 1 cm were 52% and 47%,
respectively (Ward et al., 1999). Computed
tomography during arterial portography DIFFUSION-WEIGHTED
(CTAP) has been regarded as a technique IMAGING FOR SUPPRESSION
that is more sensitive at detecting small OF SIGNALS FROM HEPATIC
lesions than dual phase spiral CT and VESSELS
SPIO-enhanced MRI, but CTAP is inva-
sive and has a greater risk of providing Diffusion-weighted imaging (DWI) is
false-positive results than other techniques sensitive to molecular diffusion, which
(Soyer et al., 1993). CTAP is also insensi- is the thermally induced motion of water
tive to lesions smaller than 1 cm (Young molecules in biological tissues called
et al., 1997). Therefore, the strategy to Brownian motion. The changes in proton
improve the detection rate of small lesions self-diffusion are an early indicator of
is anticipated. alterations of cellular homeostasis in acute
SPIO-enhanced MRI increases the rate ischemic stroke (Moseley et al., 1990);
of false-positive findings compared to therefore, DWI has a significant impact
spiral-CT (Muller et al., 1998), and these in the detection of early ischemic change
false-positive results are responsible for the and has played an important role in the
difficulty in distinguishing small lesions. clinical setting in the field of neuroradiol-
One of the common false-positive lesions is ogy. DWI is able to suppress the signals
a small hepatic cyst (Ward et al., 1999). Some of vessels and has been used to improve
techniques are recommended for detecting the sensitivity of the detection of small
hepatic cysts, such as T1-weighted breath- hepatic lesions (Okada et al., 1998).
hold GRE imaging after SPIO enhance- In the DWI pulse sequence, motion-
ment and short- and long-TE SSFSE probing gradient (MPG) pulses are
(Oudkerk et al., 1997; Kiryu et al., 2002). placed before and after a 180° RF pulse,
Adding these techniques to routine SPIO- which comprises the spin echo method
enhanced MRI, hepatic cysts can be distin- along with a 90° pulse. The protons
guished from true hepatic lesions. Another that have moved during or after the
cause for false-positives is the signals first MPG randomly change their spin
from hepatic vessels. SPIO-enhanced MRI phase as they move from one region
increases the signal of cross-sectioned of the magnetic field to another. Thus,
vessels in the low signal background liver the signal intensity of diffusion protons
15. Detection of Small Hepatic Lesions: Superparamagnetic Oxide-Enhanced 215
Figure 15.2. A 63-year-old man with hepatic metastases. a: SPIO-enhanced T2-FSE. A small metastasis
(arrow) is indistinguishable from the bright signals from vessels. b: SPIO-enhanced diffusion-weighted
PROPELLER. The lesion is clearly noted (arrow)
Figure 15.3. A 38-year-old man with hepatic metastases. a: SPIO-enhanced T2-FSE. Motion artifacts
are shown. b: SPIO-enhanced diffusion-weighted PROPELLER. Metastases are clearly shown without
motion artifacts
varying phase, which could not be cor- Gaa, J., Hatabu, H., Jenkins, R.L., Finn, J.P., and
rected by the PROPELLER method. Large Edelman, R.R. 1996. Liver masses: replace-
ment of conventional T2-weighted spin-echo
high signal intensity structures caused a
MR imaging with breath-hold MR imaging.
severe streak artifact, and small lesions Radiology 200: 459–464.
near these structures may be obscured. Kawasaki,S.,Makuuchi,M.,Kakazu, T., Miyagawa, S.,
This is a pitfall of PROPELLER imaging. Takayama, T., Kosuge, T., Sugihara, K., and
It is concluded that SPIO-enhanced MRI Moriya, Y. 1994. Resection for multiple meta-
is widely used to detect hepatic metastases; static liver tumors after portal embolization.
Surgery 115: 674–677.
however, a problem arises in the detection
Kiryu, S., Okada, Y., and Ohtomo, K. 2002.
of small lesions. The signal from hepatic Differentiation between hemangiomas and cysts
vessels mimicking small lesions causes of the liver with single-shot fast-spin echo image
false-positive lesions in SPIO-enhanced using short and long TE. J. Comput. Assist.
MRI. MPG with a small b-value is a fea- Tomogr. 26: 687–690.
sible technique to suppress such signals Kiryu, S., Watanabe, M., Kabasawa, H., Akahane, M.,
Aoki, S., and Ohtomo, K. 2006. Evaluation of
from hepatic vessels. The combination of
super paramagnetic iron oxide-enhanced diffu-
MPG and an appropriate pulse sequence, sion-weighted PROPELLER T2-fast spin echo
such as PROPELLER, provides high spa- magnetic resonance imaging: preliminary expe-
tial resolution SPIO images and is useful rience. J. Comput. Assist. Tomogr. 30: 197–200.
for detecting small hepatic lesions. Makuuchi, M., Takayama, T., Kosuge, T.,
Yamazaki, S., Yamamoto, J., Hasegawa, H., and
REFERENCES Takayasu, K. 1991. The value of ultrasonogra-
phy for hepatic surgery. Hepatogastroenterology
Ballantyne, G.H., and Quin, J. 1993. Surgical treat- 38: 64–70.
ment of liver metastases in patients with colorectal Mosley, M.E., and Butts, K. 1999. Diffusion and
cancer. Cancer 71: 4252–4266. perfusion. In: Stark, D. D., Bradley, W. G., Jr.,
Bammer, R. 2003. Basic principles of diffusion- eds. Magnetic Resonance Imaging. St. Louis,
weighted imaging. Eur. J. Radiol. 45: 169–184. MO: Mosby, pp1515–1538.
Bammer, R., Augustin, M., Prokesch, R.W., Moseley, M.E., Cohen, Y., Mintorovitch, J.,
Stollberger, R., and Fazekas, F. 2002. Diffusion- Chileuitt, L., Shimizu, H., Kucharczyk, J.,
weighted imaging of the spinal cord: interleaved Wendland, M.F., and Weinstein, P.R. 1990.
echo-planar imaging is superior to fast spin- Early detection of regional cerebral ischemia in
echo. J. Magn. Reson. Imaging 15: 364–373. cats: comparison of diffusion- and T2-weighted
Bellon, E.M., Haacke, E.M., Coleman, P.E., Sacco, MRI and spectroscopy. Magn. Reson. Med. 14:
D.C., Steiger, D.A., and Gangarosa, R.E. 1986. 330–346.
MR artifacts: a review. Am. J. Roentgenol. 147: Muller, R.D., Vogel, K., Neumann, K., Hirche, H.,
1271–1281. Barkhausen, J., Stoblen, F., Henrich, H., and
Farzaneh, F., Riederer, S.J., and Pelc, N.J. 1990. Langer, R. 1998. MRI with supermagnetic iron
Analysis of T2 limitations and off-resonance particles versus double-spiral CT in identifi-
effects on spatial resolution and artifacts in cation of malignant liver lesions. Rofo 168:
echo-planar imaging. Magn. Reson. Med. 14: 436–443.
123–139. Okada, Y., Ohtomo, K., Kiryu, S., and Sasaki, Y.
Fretz, C.J., Stark, D.D., Metz, C.E., Elizondo, G., 1998. Breath-hold T2-weighted MRI of hepatic
Weissleder, R., Shen, J.H., Wittenberg, J., Simeone, J., tumors: value of echo planar imaging with
and Ferrucci, J.T. 1990. Detection of hepatic diffusion-sensitizing gradient. J. Comput. Assist.
metastases: comparison of contrast-enhanced CT, Tomogr. 22: 364–371.
unenhanced MR imaging, and iron oxide-enhanced Oudkerk, M., van den Heuvel, A.G., Wielopolski,
MR imaging. Am. J. Roentgenol. 155: 763–770. P.A., Schmitz, P.I., Borel Rinkes, I.H., and
15. Detection of Small Hepatic Lesions: Superparamagnetic Oxide-Enhanced 219
Wiggers, T. 1997. Hepatic lesions: detection Soyer, P., Lacheheb, D., and Levesque, M. 1993.
with ferumoxide-enhanced T1-weighted MR False-positive CT portography: correlation with
imaging. Radiology 203: 449–456. pathologic findings. Am. J. Roentgenol. 160:
Penna, C., and Nordlinger, B. 2002. Colorectal 285–289.
metastasis (liver and lung). Surg. Clin. North. Stark, D.D., Weissleder, R., Elizondo, G., Hahn, P.F.,
Am. 82: 1075–1090, x–xi. Saini, S., Todd, L.E., Wittenberg, J., and Ferrucci,
Pipe, J.G. 1999. Motion correction with J.T. 1988. Superparamagnetic iron oxide: clinical
PROPELLER MRI: application to head motion application as a contrast agent for MR imaging of
and free-breathing cardiac imaging. Magn. Reson. the liver. Radiology 168: 297–301.
Med. 42: 963–969. Ward, J., Naik, K.S., Guthrie, J.A., Wilson, D.,
Pipe, J.G., Farthing, V.G., and Forbes, K.P. and Robinson, P.J. 1999. Hepatic lesion detec-
2002. Multishot diffusion-weighted FSE using tion: comparison of MR imaging after the
PROPELLER MRI. Magn. Reson. Med. 47: administration of superparamagnetic iron oxide
42–52. with dual-phase CT by using alternative-free
Ros, P.R., Freeny, P.C., Harms, S.E., Seltzer, S.E., response receiver operating characteristic analy-
Davis, P.L., Chan, T.W., Stillman, A.E., Muroff, sis. Radiology 210: 459–466.
L.R., Runge, V.M., Nissenbaum, M.A., et al. Young, N., Sing, T., Wong, K.P., Hollands, M., and
1995. Hepatic MR imaging with ferumoxides: Tait, N. 1997. Use of spiral and non-spiral com-
a multicenter clinical trial of the safety and puted tomography arterial portography in the
efficacy in the detection of focal hepatic lesions. detection of potentially malignant liver masses.
Radiology 196: 481–488. J. Gastroenterol. Hepatol. 12: 385–391.
16
Diagnosis of Hepatocellular Carcinoma:
Multidetector-Row Computed Tomography
and Magnetic Resonance Imaging
Hiromitsu Onishi, Takamichi Murakami, and Hironobu Nakamura
221
222 H. Onishi et al.
slice thickness transverse images are used imaging techniques are employed. The
for the evaluation of HCC (Kawata et al., main advantage of MR imaging is its
2002), while the thin-slice images are used higher sensitivity for contrast, which plays
for detailed diagnosis or postprocessing. a critically important role in accurate
Truly isotropic voxel data consisting of evaluation of HCC.
submillimeter-slice thickness images can
be acquired easily when using an MDCT
scanner with ³ 16 channels. Multiplanar MAGNETIC RESONANCE
reformatted (MPR) images reconstructed SCANNING TECHNIQUE
from isotropic voxel data have the same
spatial resolution as the original transverse Our standard technique for evaluation
images, and three-dimensional (3D) of HCC using the 1.5-Tesla unit (Signa
CT-angiography and maximum intensity Excite HD; GE Healthcare) employs the
projection (MIP) images can visualize array spatial sensitivity encoding tech-
even small branching arteries. These imaging nique (ASSET; GE Healthcare), which is
techniques are thus useful for acquiring one of the parallel imaging techniques.
the clinical information needed to select For signal reception, an 8-channel phased
the appropriate treatment course and to array coil system for the abdomino-pelvic
plan the treatment procedure including region is used. The parallel imaging tech-
surgical resection or transcatheter arterial nique is based on the fact that receiver sen-
chemoembolization of HCC. To gener- sitivity generally has an encoding effect
ate effective MPR, 3D, and MIP images, complementary to Fourier preparation by
however, we need a clear understanding of linear field gradients. In this technique the
the features and management of the disease. spatial information related to the spatially
varying sensitivity of receiver coils is used
to reduce the number of k-space lines
MAGNETIC RESONANCE needed to form an image. By using mul-
IMAGING tiple receiver coils in parallel, scan time
for Fourier imaging can be considerably
Magnetic resonance imaging has also been reduced.
used for the evaluation of HCC, and faster After localizer scanning, T2-weighted
imaging techniques have been developed, single shot first spin echo (SSFSE)
such as the gradient echo imaging and images are obtained during breath-hold.
three-dimensional Fourier transformation The sequence parameters are: repeti-
techniques. Recently, even faster imaging tion time [TR]: 900 ms, echo time [TE]:
was achieved with the parallel imaging 90 ms, band width: 62.5 kHz, matrix size:
technique. In addition, a sequence dedi- 320 × 160, field of view [FOV]: 320
cated for dynamic studies of the abdomi- × 320 mm, slice thickness: 5.0–6.0 mm,
nal region, especially the liver, was newly slice interval: 2.0–2.5 mm, number of
developed. However, MR imaging used for excitations [NEX]: 0.6. The slice thick-
liver dynamic studies is inferior to MDCT ness and interval are increased accord-
imaging in terms of spatial and temporal ing to the size of the liver to include the
resolution, even when these faster MR whole liver. After calibration scanning
226 H. Onishi et al.
a b c
d e f
Figure 16.1. Hepatocellular carcinoma in a 71-year-old man with severe liver cirrhosis. Precontrast (a),
late arterial phase (b), portal venous phase (c), and equilibrium phase (d) dynamic MDCT images. The
HCC (arrowheads) is visualized as an enhanced nodule in the late arterial phase and as a hypo-attenuated
nodule in the equilibrium phase. Three-dimensional CT angiography of hepatic arteries (e) created with
early arterial phase data by means of the color-coded volume rendering method shows that the right hepatic
artery (arrow in e) branches from the supermesenteric artery. Maximum intensity projection image of the
portal venous system (f) created with portal venous phase data clearly depict the varices (arrow in f)
Because most HCCs are strongly ally hypovascular and are not visualized
enhanced and tumor-to-liver contrast is as enhanced nodules in the arterial phase.
maximal in the late arterial phase, the Computer tomography images obtained
images obtained during this phase are during the portal venous and equilib-
essential to detect the hypervascular HCCs rium phases have proven to be useful for
(Figure 16.1b). The intrahepatic vein the detection of less vascularized tumors
and tumor thrombus are also thoroughly such as well-differentiated or early HCCs.
enhanced during this phase. Iannaccone et al. (2005) reported that
In the portal venous and equilibrium the equilibrium phase images that were
phases, the contrast medium is moving acquired 180 s after initiation of the con-
from the intravascular to the intercellular trast injection were a useful adjunct to the
space. The washout of contrast material biphasic images (arterial and portal venous
from the tumor is rapid because the HCC phase) for the detection of HCC.
contains less intercellular space than the
surrounding liver parenchyma. Although
most HCCs are not supplied by the portal COMPUTED TOMOGRAPHIC
vein, the liver parenchyma is maximally ANGIOGRAPHY
enhanced in the portal venous phase, so
that low attenuation is typically observed in The hepatic arteries are depicted with
these phases (Figures 16.1c, d) (Kim et al., CT angiography using the thin-slice data
1995). Well-differentiated HCCs are usu- obtained from early arterial phase imaging
16. Diagnosis of Hepatocellular Carcinoma: Multidetector-Row Computed Tomography 229
(Figure 16.1e). It has been reported that useful for detection of tumor invasion to
100% of major arterial trunks and ~ 90% the portal or the hepatic vein. Patients
or more of the small branches of the with HCC usually have cirrhosis as an
hepatic arteries were detected on partial underlying disease and have collateral
MIP images and original thin-slice cross- vessels and varices due to portal hyper-
sectional images in the early arterial phase tension. Computed tomographic angi-
(Takahashi et al., 2002). Visualization of ography can also accurately show the
the anatomy of hepatic arteries is essen- anatomy of these vessels.
tial for planning not only surgical resec-
tion but also the transcatheter arterial
chemoembolization or reservoir infusion MAGNETIC RESONANCE
system. The portal and hepatic veins can IMAGING FEATURES
be visualized by CT angiography from OF HEPATOCELLULAR
late arterial and portal venous phases CARCINOMA
imaging (Figure 16.1f). Hepatocellular
carcinomas often invade the portal vein On nonenhanced MR imaging, 35% of HCC
and/or hepatic vein and form the tumor cases show high intensity on T1-weighted
thrombus. Computed tomographic por- images, 25% show isointensity and 40%
tography and CT venography are thus show low intensity (Figure 16.2a), while
a b c
d e f
Figure 16.2. Hepatocellular carcinoma in a 69-year-old man. Precontrast (a), arterial phase (b), portal
venous phase (c), and equilibrium phase (d) dynamic MR images obtained with the LAVA sequence
before therapeutic treatment. The HCC (arrow) is seen adjacent to the hepatic vein as an enhanced nodule
in the late arterial phase and as a hypo-attenuated nodule in the portal venous and equilibrium phases.
T1-weighted gradient echo image (e) and spin echo T2-weighted image (f) 2 days after radiofrequency
ablation therapy for HCC. The completely ablated area shows as hyperintense on T1-weighted images
and as hypointense on T2-weighted images (arrowheads)
230 H. Onishi et al.
~ 90% show high intensity on T2-weighted mal liver parenchyma. In most tumors,
images (Kadoya et al., 1992). However, however, uptake of SPIO declines with a
well-differentiated HCC often shows iso- decrease in Kupffer cells. This contrast
intensity or low intensity on T2-weighted material has a relatively strong T2 and
images. Fatty deposition which is some- moderate T1 shortening effect. Therefore,
times seen in well-differentiated HCC on images obtained with SPIO-enhanced
T2-weighted spin echo and T2*-weighted
results in high-signal intensity on in-phase
T1-weighted images, which decreases gradient echo with long echo time, the
on opposed-phase T1-weighted images. tumor is visualized as an area with a higher
Contrast-enhanced dynamic studies with signal intensity than the surrounding liver
gadolinium chelate materials show HCC as parenchyma (Figure 16.3). Furthermore,
a hypervascular tumor in the arterial phaseHCC which has lost its hepatocellular
function, including its reticuloendothelial
and the washout of contrast materials in the
portal venous phase and the equilibrium function, is visualized in most cases as
phase (Figures 16.2b–d). higher signal intensity. However, SPIO
is sometimes taken up by the tumor of
well-differentiated HCC, adenomatous
SUPERPARAMAGNETIC hyperplasia, regenerating nodules and
focal nodular hyperplasia, which have
IRON OXIDE-ENHANCED maintained some degree of hepatocel-
MAGNETIC RESONANCE lular function (Imai et al., 2000). It may
IMAGING FEATURE thus be possible to differentiate benign
OF HEPATOCELLULAR and malignant tumors by using SPIO-
CARCINOMA enhanced MR imaging. Moreover, since
the SPIO agent ferucarbotran can be
Superparamagnetic iron oxide is taken injected as a bolus, contrast-enhanced
up by reticuloendothelial cell such as dynamic studies can be performed by
Kupffer cells that are present in the nor- using T1-weighted or T2-weighted imaging
a b c
Figure 16.3. Hepatocellular carcinoma in an 85-year-old man. Transverse (a) and sagittal (b) long TE
gradient echo images and transverse fast spin echo T2-weighted image (c) obtained more than 10 min
after SPIO administration. The HCC (arrow) is clearly seen as a hyperintense mass on these images
16. Diagnosis of Hepatocellular Carcinoma: Multidetector-Row Computed Tomography 231
for the evaluation of the tumor hemody- systems can also be evaluated by both of
namics (Hori et al., 2006). Hepatobiliary these imaging techniques.
contrast agents enhance normal liver
parenchyma on T1-weighted images due
to the T1 shortening effect. The tumor is ACCURACY FOR THE
seen as a low-intensity area because the DIAGNOSIS OF
tumor cell does not take up the agents. HEPATOCELLULAR
With certain agents that can be injected
in bolus form, such as Gd-EOB-DTPA
CARCINOMA
and Gd-BOPTA (Kuwatsuru et al., 1999), The accuracy of tumor detection is often
contrast-enhanced dynamic studies can evaluated by means of receiver oper-
also be performed for the evaluation of ating characteristic (ROC) analyses or
tumor hemodynamics. alternative-free response ROC (AFROC)
analyses as well as in terms of sensitiv-
ity and positive predictive values. While
STAGING OF the conventional ROC method allows for
HEPATOCELLULAR only one response per image, the AFROC
CARCINOMA method allows for multiple responses per
image (Chakraborty and Winter, 1990).
Several staging systems for HCC are cur- Several studies have demonstrated the Az
rently in use. The Union Interbationale value, which is the area under the ROC
Contre le Cancer [UICC], the General curve, and sensitivity for the detection of
Rules for the Clinical and Pathological HCC. Az value and sensitivity determined
Study of Primary Liver Cancer [Liver by means of contrast-enhanced dynamic
Cancer Study Group of Japan classifi- MDCT imaging ranged from 0.82 to 0.99
cation system], and the American Joint and from 77% to 93%, respectively (Kim
Committee on Cancer [AJCC] have estab- et al., 2006; Iannaccone et al., 2005).
lished staging systems that are based on When determined by means of gadolin-
the results of pathologic analyses, which ium-enhanced dynamic MR imaging, Az
incorporate anatomic and histologic find- value and sensitivity ranged from 0.68 to
ings obtained at tumor resection. The 0.97 and from 65% to 91%, respectively
number and size of the primary tumors as (Mori et al., 2005; Kim et al., 2004), while
well as the presence or absence of portal the corresponding values ranged from 0.76
or hepatic vein involvement are included to 0.98 and from 56% to 95%, respec-
in the staging of the primary lesion tively, when SPIO-enhanced MR imaging
because HCC is frequently associated was used (Ward et al., 2000; Kang et al.,
with portal vein or hepatic vein tumor 2003).
thrombus. Both MDCT imaging and MR Contrast-enhanced dynamic CT imag-
imaging are also useful for the evalua- ing using a 16-channel MDCT scanner
tion of the tumor thrombus. Lymph node and SPIO-enhanced MR imaging showed
metastasis and distant metastasis which similar diagnostic accuracy, sensitivity,
are taken into consideration in the staging and positive predictive value for the
232 H. Onishi et al.
detection of HCC (Kim et al., 2006). the viable portion of the tumor in the
However, gadolinium-enhanced dynamic arterial phase.
MR imaging proved to be significantly more
accurate than SPIO-enhanced MR imaging
(Pauleit et al., 2002; Kim et al., 2004). HEPATOCELLULAR
CARCINOMA AFTER
PERCUTANEOUS
HEPATOCELLULAR RADIOFREQUENCY
CARCINOMA AFTER ABLATION THERAPY
TRANSCATHERTER ARTERIAL
CHEMOEMBOLIZATION Percutaneous radiofrequency ablation
therapy can be used for local treatment
Transcatheter arterial chemoemboliza- of HCC, with coagulation necrosis gener-
tion (TACE) with iodized oil is often ally occurring after the treatment and the
performed for treatment of HCC, after ablated area often showing a zonal pattern
which the tumor is seen as a nodule with on both T1- and T2-weighted images (Lee
markedly high attenuation on nonen- et al., 2001). The completely ablated area
hanced CT images, since high concentra- appears as hyperintense on T1-weighted
tions of iodized oil remain in the tumor. images and as hypointense on T2-weighted
Because the portion of the tumor where images, reflecting coagulation necrosis or
iodized oil is retained at high density is hemorrhage (Figures 16.2e, f). On the
considered to be necrotic, CT is suit- other hand, the incompletely ablated area
able for the evaluation of the therapeutic that is often seen around the completely
efficacy of TACE with iodized oil for ablated area appears as hypointense on
HCC. However, incomplete retention of T1-weighted images and as hyperin-
iodized oil corresponds to a wide vari- tense on T2-weighted images, reflecting
ety of tumor necroses, from completely fibrous change or sinusoidal congestion
necrotic to viable cancer tissue (Choi (Onishi et al., 2004). These zonal pat-
et al., 1992). Although contrast-enhanced terns on non-enhanced MR imaging may
dynamic CT may be helpful to evaluate be important for the assessment of the
viability for HCC with incomplete reten- effect of ablation therapy on HCC soon
tion after the TACE treatment, the high after completion of the therapy, since
level of attenuation of iodized oil in the these signal changes are seldom affected
tumor can negatively affect the evalua- by the arterio-portal venous shunt or por-
tion by dynamic CT. On the other hand, tal thrombus that is often associated with
as iodized oil does not have much effect percutaneous ablation therapy. In clinical
on the intensity observed on MR images, practice, contrast-enhanced dynamic CT
MR images can clearly show intratu- and contrast-enhanced dynamic MR imag-
moral signal changes (Murakami et al., ing are used as the standard techniques to
1993a, b). Contrast-enhanced dynamic evaluate the therapeutic effect of ablation
MR imaging is thus useful for evaluating therapy, with the necrotic area shown as a
viability for HCC after TACE treatment nonenhanced area and the residual viable
because it shows enhancement only in lesion as an enhanced area. However, it is
16. Diagnosis of Hepatocellular Carcinoma: Multidetector-Row Computed Tomography 233
Choi, B.I., Kim, H.C., Han, J.K., Park, J.H., Kim, with use of noise reduction filters pilot study.
Y.I., Kim, S.T., Lee, H.S., Kim, C.Y., and Han, Radiology 228: 251–256.
M.C. 1992. Therapeutic effect of transcatheter Kalra, M.K., Maher, M.M., Toth, T.L., Schmidt, B.,
oily chemoembolization therapy for encapsu- Westerman, B.L., Morgan, H.T., and Saini, S.
lated nodular hepatocellular carcinoma: CT and 2004. Techniques and applications of automatic
pathologic findings. Radiology 182: 709–713. tube current modulation for CT. Radiology 233:
Dromain, C., de Baere, T., Elias, D., Kuoch, V., 649–657.
Ducreux, M., Boige, V., Petrow, P., Roche, A., Kang, B.K., Lim, J.H., Kim, S.H., Choi, D., Lim,
and Sigal, R. 2002. Hepatic tumors treated H.K., Lee, W.J., and Lee, S.J. 2003. Preoperative
with percutaneous radio-frequency ablation: CT depiction of hepatocellular carcinoma: ferumox-
and MR imaging follow-up. Radiology 223: ides-enhanced MR imaging versus triple-phase
255–262. helical CT. Radiology 226: 79–85.
Heiken, J.P., Brink, J.A., McClennan, B.L., Sagel, Kawata, S., Murakami, T., Kim, T., Hori, M.,
S.S., Crowe, T.M., and Gaines, M.V. 1995. Federle, M.P., Kumano, S., Sugihara, E.,
Dynamic incremental CT: effect of volume and Makino, S., Nakamura, H., and Kudo, M. 2002.
concentration of contrast material and patient Multidetector CT: diagnostic impact of slice
weight on hepatic enhancement. Radiology 195: thickness on detection of hypervascular hepa-
353–357. tocellular carcinoma. AJR. Am. J. Roentgenol.
Hori, M., Murakami, T., Kim, T, Iannaccone, R., 179: 61–66.
Abe, H., Onishi, H., Tomoda, K., Catalano, C, Kim, T., Murakami, T., Oi, H., Matsushita, M.,
Passariello, R., and Nakamura, H. 2006. Kishimoto, H., Igarashi, H., Nakamura, H., and
Hemodynamic characterization of focal hepatic Okamura, J. 1995. Detection of hypervascular
lesions: role of ferucarbotran-enhanced dynamic hepatocellular carcinoma by dynamic MRI and
MR imaging using T2-weighted multishot spin- dynamic spiral CT. J. Comput. Assist. Tomogr.
echo echo planar sequence. J. Magn. Reson. 19: 948–954.
Imag. 23: 509–519. Kim, T., Murakami, T., Hori, M., Takamura, M.,
Iannaccone, R., Laghi, A., Catalano, C., Rossi, P., Takahashi, S., Okada, A., Kawata, S., Cruz, M.,
Mangiapane, F., Murakami, T., Hori, M., Federle, M.P., and Nakamura, H. 2002. Small
Piacentini, F., Nofroni, I., and Passariello, R. hypervascular hepatocellular carcinoma revealed
2005. Hepatocellular carcinoma: role of unen- by double arterial phase CT performed with
hanced and delayed phase multi-detector row single breath-hold scanning and automatic bolus
helical CT in patients with cirrhosis. Radiology tracking. AJR Am. J. Roentgenol. 178: 899–904.
234: 460–467. Kim, Y.K., Kim, C.S., Kwak, H.S., and Lee, J.M.
Imai, Y., Murakami, T., Yoshida, S., Nishikawa, M., 2004. Three-dimensional dynamic liver MR
Ohsawa, M., Tokunaga, K., Murata, M., Shibata, K., imaging using sensitivity encoding for detection
Zushi, S., Kurokawa, M., Yonezawa, T., Kawata, S., of hepatocellular carcinomas: comparison with
Takamura, M., Nagano, H., Sakon, M., superparamagnetic iron oxide-enhanced MR
Monden, M., Wakasa, K., and Nakamura, H. 2000. imaging. J. Magn. Reson. Imag. 20: 826–837.
Superparamagnetic iron oxide-enhanced magnetic Kim, Y.K., Kwak, H.S., Kim, C.S., Chung, G.H.,
resonance images of hepatocellular carcinoma: Han, Y.M., and Lee, J.M. 2006. Hepatocellular
correlation with histological grading. Hepatology Carcinoma in Patients with Chronic Liver
32: 205–212. Disease: Comparison of SPIO-enhanced MR
Kadoya, M., Matsui, O., Takashima, T., and Imaging and 16-Detector Row CT. Radiology
Nonomura, A. 1992. Hepatocellular carcinoma: 238: 531–541.
correlation of MR imaging and histopathologic Kuwatsuru, R., Kadoya, M., Ohtomo, K., Tanimoto, A.,
findings. Radiology 183: 819–825. Hirohashi, S., Murakami, T., Tanaka, Y.,
Kalra, M.K., Maher, M.M., Sahani, D.V., Blake, Yoshikawa, K., and Katayama, H. 1999. Clinical
M.A., Hahn, P.F., Avinash, G.B., Toth, T.L., late phase II trials of MultiHance (Gd-BOPTA) for
Halpern, E., and Saini, S. 2003. Low-dose CT of the magnetic resonance imaging of liver tumors in
the abdomen: evaluation of image improvement Japan. J. Comput. Assist. Tomogr. 23: 65–74.
16. Diagnosis of Hepatocellular Carcinoma: Multidetector-Row Computed Tomography 235
Lee, J.D., Lee, J.M., Kim, S.W., Kim, C.S., and Mun, Murakami, T., Onishi, H., Mikami, K., Iannaccone, R.,
W.S. 2001. MR imaging-histopathologic correla- Federle, M.P., Kim, T., Hori, M., Okada, A.,
tion of radiofrequency thermal ablation lesion in Marukawa, T., Higashihara, H., Passariello, R.,
a rabbit liver model: observation during acute and and Nakamura, H. 2006. Determining the optimal
chronic stages. Korean J. Radiol. 2: 151–158. timing for early arterial phase hepatic CT imaging
Mori, K., Yoshioka, H., Takahashi, N., Yamaguchi, by measuring abdominal aortic enhancement in
M., Ueno, T., Yamaki, T., and Saida, Y. 2005. variable contrast injection protocols. J. Comput.
Triple arterial phase dynamic MRI with sensitiv- Assist. Tomogr. 30: 206–211.
ity encoding for hypervascular hepatocellular car- Onishi, H., Matsushita, M., Murakami, T., Tono, T.,
cinoma: comparison of the diagnostic accuracy Okamoto, S., Aoki, Y., Iannaccone, R., Hori,
among the early, middle, late, and whole triple M., Kim, T., Osuga, K., Tomoda, K., Passariello,
arterial phase imaging. AJR Am. J. Roentgenol. R., and Nakamura, H. 2004. MR appearances
184: 63–69. of radiofrequency thermal ablation region: his-
Murakami, T., Nakamura, H., Hori, S., Tomoda, K., topathologic correlation with dog liver models
Mitani, T., Nakanishi, K., Hashimoto, T., Tsuda, K., and an autopsy case. Acad. Radiol. 11: 1180–
Kozuka, T., Monden, M., and Wakasa, K. 1993a. 1189.
Detection of viable tumor cells in hepatocellular Pauleit, D., Textor, J., Bachmann, R., Conrad, R.,
carcinoma following transcatheter arterial chem- Flacke, S., Layer, G., Kreft, B., and Schild, H.
oembolization with iodized oil. Pathologic cor- 2002. Hepatocellular carcinoma: detection with
relation with dynamic turbo-FLASH MR imaging gadolinium- and ferumoxides-enhanced MR
with Gd-DTPA. Acta. Radiol. 34: 399–403. imaging of the liver. Radiology 222: 73–80.
Murakami, T., Nakamura, H., Tsuda, K., Nakanishi, K., Takahashi, S., Murakami, T., Takamura, M., Kim, T.,
Hori, S., Tomoda, K., Mitani, T., Kozuka, T., Hori, M., Narumi, Y., Nakamura, H., and Kudo, M.
Monden, M., and Wakasa, K. 1993b. Treatment of 2002. Multi-detector row helical CT angiog-
hepatocellular carcinoma by chemoembolization: raphy of hepatic vessels: depiction with dual-
evaluation with 3DFT MR imaging. AJR Am. J. arterial phase acquisition during single breath
Roentgenol. 160: 295–299. hold. Radiology 222: 81–88.
Murakami, T., Baron, R.L., Peterson, M.S., Oliver, Ward, J., Guthrie, J.A., Scott, D.J., Atchley, J.,
J.H. 3rd, Davis, P.L., Confer, S.R., and Federle, Wilson, D., Davies, M.H., Wyatt, J.I., and
M.P. 1996. Hepatocellular carcinoma: MR imag- Robinson, P.J. 2000. Hepatocellular carcinoma
ing with mangafodipir trisodium (Mn-DPDP). in the cirrhotic liver: double-contrast MR imag-
Radiology 200: 69–77. ing for diagnosis. Radiology 216: 154–162.
Murakami, T., Kim, T., Takamura, M., Hori, M., Yamashita, Y., Komohara, Y., Takahashi, M.,
Takahashi, S., Federle, M.P., Tsuda, K., Osuga, K., Uchida, M., Hayabuchi, N., Shimizu, T., and
Kawata, S., Nakamura, H., and Kudo, M. 2001. Narabayashi, I. 2000. Abdominal helical CT:
Hypervascular hepatocellular carcinoma: detec- evaluation of optimal doses of intravenous con-
tion with double arterial phase multi-detector row trast material -a prospective randomized study.
helical CT. Radiology 218: 763–767. Radiology 216: 718–723.
17
Hepatocellular Carcinoma: Effect
of Injection Rate/Injection Duration
of Contrast Material on Computed
Tomography
Tomoaki Ichikawa and Tsutomu Araki
237
238 T. Ichikawa and T. Araki
because of the definite increasing detec- for individual patients. When a fixed injec-
tion rate of the false-positive lesions tion duration is employed in a hepatic
with a high injection rate of contrast CT protocol, injection rates should be
material. Considering the balance of the variable depending on the patient’s body-
sensitivity and the specificity, we finally weight-tailored dose of contrast material.
recommend 4 ml/s as a medium injection Therefore, the concept of an effective
rate for multiphasic contrast-enhanced injection rate (an injection rate applied
hepatic CT imaging in cirrhotic patients. for a patient with mean body weight in
a patients’ group) is needed. An effec-
Fixed Injection Duration of Contrast tive injection rate can be established by
Material refining the optimal fixed injection rate
(4 ml/s) derived from our previous study.
At present, it has started to be recognized
For example, under the assumption of
that a use of a fixed injection duration can
the use of 2.0 ml/kg with 300 mgl/ml
minimize the patients’ variables in the scan
(600 mgI/kg) of contrast material and the
timing for each phase among all patients
mean patient’s body weight of 60 kg in a
(Bae, 2003; Awai et al., 2004). Such effect
patients’ group, a fixed injection duration
of the fixed injection duration depends on
should be established so that the effective
the fact that the peak enhancement time of
injection rate becomes 4 ml/s by following
the aorta (pET-Ao) may be strictly defined
the equation (Ichikawa et al., 2006b);
by the injection duration. Based on our
experience, the mean time delay from the a fixed injection duration = 120 ml (2 ml/
completion of any injection durations to kg × 60 kg)/4 ml/s = 30 s
pET-Ao was ∼10 s regardless of patients’ In conclusion, we recommend the use of
body-weight. In conclusion, pET-Ao, can 30 s as an optimal injection duration in a
be calculated with a simple equation as practical multiphasic contrast-enhanced CT
follows (Ichikawa et al., 2006b); protocol of the liver. This recommendation
pET − Ao = an injection duration + 10 s might be reliable based on our previous
results that the peak enhancement values of
In general, peak enhancement time for
the aorta with injection duration of < 30 s
hypervascular HCC (pET-HCC) might
were significantly higher than those with
occur several seconds (∼5 s based on our
the injection duration of > 30 s (Ichikawa
experience) after pET-Ao because it sup-
et al., 2006b).
plies arterial flow from hepatic artery to
the aorta. Therefore, pET-HCC can be cal-
culated with the following equation; REFERENCES
Awai, K., and Hori, S. 2003. Effect of contrast
pET − HCC = an injection duration + 15 s injection protocol with dose tailored to patient
Thus, use of fixed injection duration can weight and fixed injection duration on aortic and
hepatic enhancement at multidetector-row heli-
minimize the differences in HAP scan
cal CT. Eur. Radiol. 13: 2155–2160.
duration for individual patients and there- Awai, K., Hiraishi, K., and Hori, S. 2004. Effect
fore, this method might become standard of contrast material injection duration and rate
rather than the use of fixed injection rate on aortic peak time and peak enhancement at
that makes a difference in injection duration dynamic CT involving injection protocol with
240 T. Ichikawa and T. Araki
dose tailored to patient weight. Radiology 230: weight-tailored dose of contrast material. Eur. J.
142–150. Radiol. 58: 165–176.
Bae, K.T. 2003. Peak contrast enhancement in CT Ichikawa, T., Nakajima, H., Nanbu, A., Hori, M.,
and MR angiography: when does it occur and and Araki, T. 2006b. Effect of injection rate of
why? Pharmacokinetic study in a porcine model. contrast material on CT of hepatocellular carci-
Radiology 227: 809–816. noma. Am. J. Radiol. 186: 1413–1418.
Bae, K.T., Heiken, J.P., and Brink, J.A. 1998. Aortic Kim, T., Murakami, T., and Oi, H. 1995. Detection
and hepatic contrast medium enhancement at of hypervascular hepatocellular carcinoma by
CT. Part II. Effect of reduced cardiac output in a dynamic MRI and dynamic spiral CT. J. Comput.
porcine model. Radiology 207: 657–662. Assist. Tomogr. 19: 948–954.
Baron, R.L., Oliver, J.H., Dodd, G.D., Nalesnik, Kim, T., Murakami, T., and Takahashi, S. 1998. Effects
M., Holbert, B.L., and Carr, B. 1996. of injection rates of contrast material on arterial
Hepatocellular carcinoma: Evaluation with phase hepatic CT. Am. J. Radiol. 171: 429–432.
biphasic, contrast-enhanced, helical CT. Mitsuzaki, K., Yamashita, Y., Ogata, I., Nishiharu,
Radiology 199: 505–511. T., Urata, J., and Takahashi, M. 1996. Multiple-
Hollett, M.D., Jeffrey, R.B., Nino-Murcia, M., phase helical CT of the liver for detecting small
Jorgensen, M.J., and Harris, D.P. 1995. hepatomas in patients with liver cirrhosis: con-
Dual-phase helical CT of the liver: value of arterial trast-injection protocol and optimal timing. Am.
phase scans in the detection of small (<1.5 cm) J. Radiol. 167: 753–757.
malignant hepatic neoplasms. Am. J. Radiol. Ohashi, I., Hanafusa, K., and Yoshida, T. 1993.
164: 879–884. Small hepatocellular carcinomas: two-phase
Hwang, G.J., Kim, M.J., Yoo, H.S., and Lee, dynamic incremental CT in detection and evalu-
J.T. 1997. Nodular hepatocellular carcinomas: ation. Radiology 189: 851–855.
detection with arterial-, portal-, and delayed- Tublin, M.E., Tessler, F.N., Cheng, S.L., Peters,
phase images at spiral CT. Radiology 202: T.L., and McGovern, P.C. 1999. Effect of injec-
383–388. tion rate of contrast medium on pancreatic and
Ichikawa, T., Erturk, S.M., and Araki, T. 2006a. hepatic helical CT. Radiology 210: 97–101.
Multiphasic contrast-enhanced multidetector- Yu, J.S., Kim, K.W., Sung, K.B., Lee, J.T., and
row CT of liver: contrast-enhancement theory Yoo, H.S. 1997. Small arterial-portal venous
and practical scan protocol with a combination shunts: a cause of pseudolesions at hepatic imag-
of fixed injection duration and patients’ body- ing. Radiology 203: 737–742.
18
Detection of Combined Hepatocellular
and Cholangiocarcinomas: Enhanced
Computed Tomography
Akihiro Nishie and Kengo Yoshimitsu
241
242 A. Nishie and K. Yoshimitsu
preoperative diagnosis of this type of modalities, small HCC and CC can now be
tumor because it is rarely encountered. In diagnosed with relative ease (Honda et al.,
an examination performed by The liver 1993a, b), while a prospective diagnosis of
cancer study group of Japan (1990), the cHCC-CC on enhanced CT remains diffi-
frequency of lymph node metastases in cult, although several reports describe CT
patients with cHCC-CC was 76.2% and findings of this entity (Aoki et al., 1993;
was as high as that in patients with CC, Fukukura et al., 1997; Nishie et al., 2005).
while that in patients with HCC is 30.3%. We herein describe enhanced CT findings
Some of the literature suggests that there of cHCC-CC and discuss the possibility of
was a need for a lymph node resection dur- making a diagnosis of cHCC-CC based on
ing a hepatectomy for cHCC-CC (Uenishi enhanced CT.
et al., 2000; Sasaki et al., 2001; Eguchi
et al., 2002). Therefore, it is important to
make a diagnosis of cHCC-CC preopera- PATHOGENESIS
tively.
Fine needle aspiration biopsy (FNAB) In general, the histopathological definition
is one of the possible methods of diagnos- of cHCC-CC is based on criteria proposed
ing this entity. However, making a reliable by the World Health Organization: a tumor
diagnosis of cHCC-CC based on cytologic that contains unequivocal elements of both
preparations alone is very difficult because HCC and CC. If bile production, intercel-
HCC and CC have a wide spectrum of lular bile canaliculi, trabecular growth
morphologic features (Dusenbery, 1997). pattern or all three are recognized, the
Wee and Nilsson (1999) mentioned that, at epithelium is defined as a hepatocellular
the very least, there should be a high index element; if mucin production, definite
of suspicion, especially when the clinical gland formation, or both are noted, the
history is highly suggestive of HCC, and epithelium is defined as a cholangiocel-
FNAB reveals only an adenocarcinoma or lular component (Gibson, 1978). Allen
tumor cells which could be interpreted as and Lisa (1949) first defined three types
intermediate cells. They also suggested that of cHCC-CC: (1) HCC and CC originating
the FNAB diagnosis of cHCC-CC may be from different sites in the liver, but consist-
possible if the clinical, cytohistologic, and ing of a uniform cell type (double cancer);
immunohistochemical findings support (2) contiguous HCC and CC originating
the presence of two cellular components, independently, but intermingling as they
HCC and adenocarcinoma. However, the grow (combined type); and (3) completely
usefulness of FNAB in diagnosing cHCC- intermingled HCC and CC (mixed type).
CC has been previously discussed in only On the other hand, Goodman et al.
a few studies, and presently we have not (1985) also defined criteria for a patho-
come to a conclusion about it. morphological diagnosis of cHCC-CC
On the other hand, sonography, computed that were different from those proposed
tomography (CT), and magnetic resonance by Allen and Lisa (1949). According
imaging (MRI) are now routinely used to this classification, type I, “collision
to detect liver tumors and evaluate their tumors”, contain separate and apparently
characteristics. With the advent of these coincidental components of HCC and CC,
18. Detection of Combined Hepatocellular and Cholangiocarcinomas 243
a b c
Figure 18.1. A 73-year-old man with the mixed type of cHCC-CC. (a) Early phase of enhanced CT. A
tumor with a diameter of 3.5 cm is located in the right lobe. The left-sided main component shows high
density, and the right-sided component under the hepatic capsule shows low density (arrow). (b) Delayed
phase of enhanced CT. The left-sided component shows low density with mild ring-enhancement, and the
right-sided component shows low density (arrow); they are recognized as an HCC-pattern component and
a CC-pattern component, respectively. (c) Macroscopic finding. The inner part of the tumor is a yellowish
mass with partial necrosis and a capsule, representing HCC, and the outside part is a white mass without
expansion, representing CC (arrow). The enhancing pattern of the two components was consistent with
the histological findings. (Nishie et al., 2005; with permission.)
including tumors in the different sites; (Taguchi et al., 1996; Kim et al., 2004). It
however, no transitional forms between is supposed that the latter type of tumor
HCC and CC are observed; type II, “tran- may originate from hepatic progenitor
sitional tumors”, exhibit hepatocellular cells. Another pathological characteris-
differentiation with a transition area to tic of this tumor is a broad desmoplastic
CC or to a mixed hepatocellular glan- stroma encasing the strands and trabecu-
dular tumor; and type III, “fibrolamellar lae of the tumor cells (Kim et al., 2004).
tumors”, resemble a fibrolamellar vari- It is very important to be familiar with the
ant of HCC but contain mucin-producing known histopathological characteristics
glands. Goodman’s type I tumor is equiva- when radiological imaging of cHCC-CC
lent to Allen and Lisa’s double cancer and is analyzed.
combined type, and Goodman’s type II is
similar to Allen and Lisa’s mixed type.
However, Goodman’s type III tumor does ENHANCED COMPUTED
not correspond to any categories in Allen TOMOGRAPHIC FINDINGS
and Lisa’s classification. Furthermore,
another type of cHCC-CC is also present. Investigations into the imaging character-
This type is composed of intermediate istics of cHCC-CC appear to be somewhat
tumor cells that show morphological fea- limited to case reports and small series using
tures intermediate between HCC and CC imaging techniques that are not up to current
244 A. Nishie and K. Yoshimitsu
standards because of the rarity of this entity. That is to say, these results suggest that
Aoki et al. (1993) reported on the CT find- cHCC-CCs tend to show CT findings charac-
ings of 14 lesions in 12 cases with cHCC- terized by a predominant component identi-
CC. Enhanced CT findings were divided into fied in the mass and that it is difficult to make
three types with reference to differences in a diagnosis of cHCC-CC based on enhanced
the enhancement pattern. Type A (n = 9) was CT alone. However, Aoki et al. (1993) evalu-
defined as a mass in which only the periph- ated the values of alpha-fetoprotein (AFP)
eral portion with a wide to narrow width was and carcinoembryonic antigen (CEA), which
variously enhanced as a lesion that remained are tumor markers characteristic to HCC
isodense or as a lesion in which the central and CC, respectively. The AFP level was
portion was of low density or isodense on positive in 65% of patients with cHCC-CC.
the early phase. On the late phase the central Furthermore, 47% of patients had abnormal
portion showed high density or isodensity. CEA values. They concluded that cHCC-CC
We can guess that viable cancer cells are might be suspected when CT findings and
located in the periphery of a mass, and are AFP or CEA values are at variance.
enhanced on the early phase but fade on Fukukura et al. (1997) also analyzed the
the late phase; the central portion of a mass, CT findings of 15 cases with cHCC-CC.
which mainly comprises fibrous tissue and They compared the CT findings on the
scant tumor cells, is enhanced only on the basis of the gross appearance of resected
late phase. These findings are consistent with specimens. The tumors were classified
those of liver metastasis and CC (Muramatsu into two groups: HCC type which is green-
et al., 1986; Honda et al., 1993b). In fact, ish or yellow in color, has hemorrhagic
there were viable cancer cells consisting of necrosis in varying degrees and is sharply
HCC-predominant components in the mar- demarcated with a fibrous capsule; and CC
ginal portion, and the central portion of the type which is whitish in color and shows
tumor was composed of abundant fibrous an infiltrative growth with an irregular mar-
stroma containing CC-predominant compo- gin. In the HCC type (n = 6), most parts of
nents. In the midzone between the peripheral the masses comprised HCC components,
and central areas, HCC and CC components and CC components were observed only
intermingled. Type B (n = 4) was defined as sporadically or focally within the mass.
a mass in which the entire portion was well On the other hand, in the CC type (n = 9),
enhanced on the early phase and faded to four masses had a large volume of HCC
low density on the late phase. This finding is components. In the two masses the ratio
identical to that for HCC. These masses were of HCC to CC components was almost
mainly composed of HCC with CC compo- equal, and three masses consisted of pre-
nents scattered in the mass or of a main mass dominant CC components. All HCC type
of HCC with a small CC area attached to it. masses were entirely enhanced at the early
Type C (n = 1) showed low density on the late phase and faded into low density at the late
phase as well as on the early phase, and this phase. This result is quite consistent with
finding is one of the characteristic CT find- that reported by Aoki et al. (1993). Two
ings in CC (Honda et al., 1993b). The main of the masses also had a ring enhancement
lesion was histopathologically composed of representing a fibrous capsule, which is
CC and an attached small HCC component. also one of the characteristics of HCC
18. Detection of Combined Hepatocellular and Cholangiocarcinomas 245
(Itai et al., 1986). However, eight of nine Twenty-seven patients with cHCC-CCs his-
CC type masses were enhanced only at the topathologically proven by surgical resec-
peripheral portion at the early phase and tion were enrolled in our study. Before the
changed to low density areas or had only evaluation of enhanced CT findings, the
central portions enhanced at the delayed pathologist identified all areas of HCC or
phase; in other words, they showed CT CC cell type and subclassified cHCC-CC
findings compatible with those of CC, according to the criteria of Allen and Lisa
although a sufficient volume of HCC com- (1949). Then, we focused on the enhanc-
ponents was present in at least six cases. ing pattern of a part or component of the
This was because these HCC components lesions. Based on the previous reports, we
presented remarkable fibrosis along the defined the components showing high den-
blood space, which is also seen in scle- sity on the early phase and low density on
rosing HCC (Yamashita et al., 1993). the late phase (high–low pattern) as HCC-
Therefore, these HCC components were pattern component (Honda et al., 1993a),
considered to show CT findings similar to and we defined the components showing
those seen in CC. The volume of fibrous low density on the early phase and low,
stroma in a mass is one of the factors that iso, or high density on the delayed phase
can change enhanced CT findings, and we (low–low, low-iso, or low–high patterns)
are sometimes puzzled when diagnosing as CC-pattern components (Honda et al.,
a liver mass. Fukukura et al. (1997) con- 1993b). We correlated the radiologically
cluded that this CC type can be diagnosed determined components to the histological
as cHCC-CC by evaluating virus markers areas of each cell type. When a complete
and serum AFP levels, which may suggest concordance between the component
the presence of HCC components. detected radiologically and the histological
Choi et al. (1994) reported iodized-oil finding was obtained in terms of site, size,
CT findings in six cases with cHCC-CC. and enhancing pattern, we determined this
On iodized-oil CT scans, a partial reten- to be a correct concordance (Figure 18.1).
tion of iodized oil was observed in the The diagnostic rate (DR) was defined as the
mass in all six cases. HCC portions of ratio of patients in whom CT (HCC- and
pathologic specimens showed compact CC-pattern components) correctly depicted
retention of iodized oil, whereas CC por- HCC/CC to the total number of patients.
tions did not show compact retention. According to the criteria of Allen and
However, it is usually impossible to make Lisa (1949), histological evaluation deter-
a diagnosis of cHCC-CC based only on mined the presence of double cancer,
an iodized-oil CT result because scleros- combined type, and mixed type in 2, 9 and
ing HCC or necrotic portions of HCC also 16 patients, respectively. In the two cases
show no retention of iodized oil. of double cancer, two distinct masses were
How frequently can we diagnose cHCC- described as HCC-pattern and CC-pattern
CC based on enhanced CT alone? We components, respectively. The DR was
evaluated the performance of enhanced 100% (2/2) in double cancer.
CT in making a diagnosis of cHCC-CC On the other hand, the DR was 44.4%
by closely comparing CT findings with (4/9) in the combined type. Why could we
histologic findings (Nishie et al., 2005). not diagnose at a high rate? The reasons
246 A. Nishie and K. Yoshimitsu
for discordance included the different carcinoma with abundant fibrous stroma.
enhancing pattern from the typical pat- Either component was very small (< 1 cm
terns as defined earlier and the small size in diameter) in nine patients and could not
of each component. The CC in one patient be detected on CT. Another reason for the
showed a high–low (HCC) pattern on CT discordance was the presence of tumors
and was composed of poorly differentiated composed of intermediate cells as men-
adenocarcinoma cells with a small amount tioned earlier (Taguchi et al., 1996; Kim
of desmoplastic stroma. Yoshida et al. et al., 2004). Two different components
(1999) also reported two cases of small int- are not reasonably detected in this type
rahepatic CCs with marked hypervascular- of tumor, which is constructed of single
ity, composed of a large number of tumor tumor cells. In our series, three patients
cells and little interstitial fibrous tissue. had mixed type tumors that were shown
Honda et al. (1993b) reported that only 5% to be totally composed of intermediate
of intrahepatic peripheral cholangiocarci- cells, with one showing an HCC pattern
noma showed the high–low pattern. The and the other two showing a CC pattern
HCC in another patient showed a low–low on CT. The two intermediate cell tumors
(CC) pattern on CT and was composed showing a CC pattern had a large amount
of well-differentiated hepatocellular car- of interstitial fibrosis, which is one of the
cinoma cells. Fujita et al. (1996) reported characteristics of this type (Kim et al.,
that 17.6% of well-differentiated HCC 2004) and is histopathologically similar
showed a low–low pattern. That is to say, to CC. However, the third tumor showing
these results suggest that it becomes dif- the HCC pattern was composed of malig-
ficult to diagnose cHCC-CC on CT when nant cells proliferating in a trabecular,
one component of cHCC-CC shows an solid, and tubular pattern. Some of these
atypical enhancing pattern. Furthermore, cells reacted with CA19-9, and only a
another reason for the discordance is that small amount of fibrous tissue was seen.
either component was very small (<1 cm This result suggests that all tumors com-
in diameter) and could not be detected on posed of intermediate cells do not show
CT. This discordance was observed in the the CC enhancement pattern. As a result,
other three patients. the total DR on enhanced CT was 33.3%
The DR in the mixed type was 18.8% (9/27), which was not very high. Variable
(3/16), and even lower than that for the factors such as an atypical enhancement
combined type. Three masses composed pattern, the small size of the HCC or CC
of intermediate cells were included in this components, and the presence of a tumor
group. The main reasons for discordance composed of intermediate cells led to a
included the different enhancing pattern misreading interpretation.
from the typical patterns as defined ear- The main differential diagnosis of
lier, and the small size of each component. cHCC-CCs on enhanced CT is consid-
In one patient the CC-pattern component ered to be HCC or CC containing various
on CT was histologically proven to be kinds of differentiation. An early advanced
HCC. This HCC showed a low-iso (CC) HCC, which is a representative of such
pattern on CT and was composed of tumors, may be differentiated from cHCC-
moderately differentiated hepatocellular CCs by delayed enhancement of a low
18. Detection of Combined Hepatocellular and Cholangiocarcinomas 247
Jarnagin, W. R., Weber, S., Tickoo, S. K., Koea, hepatocellular and cholangiocellular carcinoma:
J. B., Obiekwe, S., Fong, Y., DeMatteo, R. modes of spreading and choice of surgical treat-
P., Blumgart, L. H., and Klimstra, D. 2001. ment by reference to macroscopic type. J. Surg.
Combined hepatocellular and cholangiocarcinoma: Oncol. 76:37–46.
demographic, clinical, and prognostic factors. Taguchi, J., Nakashima, O., Tanaka, M., Hisaka, T.,
Cancer 94:2040–2046. Takazawa, T., and Kojiro, M. 1996. A clinico-
Kim, H., Park, C., Han, K., Choi, J., Kim, Y. B., pathological study on combined hepatocellu-
Kim, J. K., and Park, Y. N. 2004. Primary liver lar and cholangiocarcinoma. J. Gastroenterol.
carcinoma of intermediate (hepatocyte-cholangi- Hepatol. 11:758–764.
ocyte) phenotype. J. Hepatol. 40:298–304. The liver cancer study group of Japan. 1990.
Koh, K. C., Lee, H., Choi, M. S., Lee, J. H., Paik, Primary liver cancer in Japan: clinicopatho-
S. W., Yoo, B. C., Rhee, J. C., Cho, J. W., Park, logical features and results of surgical treatment.
C. K., and Kim, H. J. 2005. Clinicopathological Ann. Surg. 221:277–287.
features and prognosis of combined hepatocellular Uenishi, T., Hirohashi, K., Shuto, T., Yamamoto, T.,
cholangiocarcinoma. Am. J. Surg. 189:120–125. Kubo, S., Tanaka, H., Ikebe, T., and
Lee, C. C., Wu, C. Y., Chen, J. T., and Chen, Kinoshita, H. 2000. Surgery for mixed hepa-
G. H. 2002. Comparing combined hepatoc- tocellular and cholangiocellular carcinoma.
cellular-cholangiocarcinoma and cholan- Hepatogastroenterology 47:832–834.
giocarcinoma: a clinicopathological study. Wee, A., and Nilsson, B. 1999. Combined hepa-
Hepatogastroenterology 49:1487–1490. tocellular-cholangiocarcinoma: diagnostic chal-
Liu, C. L., Fan, S., Lo, C., Ng, I. O., Lam, C., Poon, lenge in hepatic fine needle aspiration biopsy.
R. T., and Wong J. 2003. Hepatic resection for Acta Cytol. 43:131–138.
combined hepatocellular and cholangiocarci- Yamashita, Y., Fan, Z. M., Yamamoto, H.,
noma. Arch. Surg. 138:86–90. Matsukawa, T., Arakawa, A., Miyazaki, T.,
Muramatsu, Y., Takayasu, K., Moriyama, N., Harada, M., and Takahashi, M. 1993. Sclerosing
Shima, Y., Goto, H., Ushio, K., Yamada T., hepatocellular carcinoma: radiologic findings.
Hasegawa, H., Koyama, Y., and Hirohashi, S. Abdom. Imaging 18:347–351.
1986. Peripheral low-density area of hepatic Yano, Y., Yamamoto, J., Kosuge T., Sakamoto, Y.,
tumors: CT-pathologic correlation. Radiology Yamasaki, S., Shimada, K., Ojima, H.,
160:49–52. Sakamoto, M., Takayama, T., and Makuuchi,
Nishie, A., Yoshimitsu, K., Asayama, Y., Irie, H., M. 2003. Combined hepatocellular and
Aibe, H., Tajima, T., Shinozaki, K., Nakayama, cholangiocarcinoma: a clinicopathologic
T., Kakihara, D., Shimada, M., Aishima, S., study of 26 resected cases. Jpn. J. Clin. Oncol
Yoshida, K., and Honda, H. 2005. Detection of 33:283–287.
combined hepatocellular and cholangiocarcino- Yoshida, Y., Imai, Y., Murakami, T., Nishikawa,
mas on enhanced CT: comparison with histologic M., Kurokawa, M., Yonezawa, T., Tokunaga, K.,
findings. Am. J. Roentgenol. 184:1157–1162. Fukushima, Y., Wakasa, K., Kim, T., Nakamura,
Sasaki, A., Kawano, K., Aramaki, M., Ohno, T., H., Sakon, M., and Monden, M. 1999. Intrahepatic
Tahara, K., Takeuchi, Y., Yoshida, T., and Kitano, cholangiocarcinoma with marked hypervascular-
S. 2001. Clinicopathological study of mixed ity. Abdom. Imaging 24:66–68.
19
Hepatocellular Carcinoma
and Adenomatous Hyperplasia
(Dysplastic Nodules): Dynamic
Computed Tomography and a Combination
of Computed Tomography and Angiography
Kenichi Takayasu
249
250 K. Takayasu
book was published by the World Health that of surrounding hepatic parenchyma
Organization (Hirohashi et al., 2000), (Sakamoto et al., 1991), irregular thin
which mentioned that the morphological trabecular pattern, acinar and/or pseu-
criteria for differential diagnosis among doglandular pattern, fatty change and/or
AH, atypical AH and early HCC was still clear cell changes of cancer cells (Kojiro,
under discussion mainly due to the lack of 1997), and stromal invasion (Kondo et al.,
objective phenotypic or genotypic markers. 1994). Early HCC is deemed to be a well-
Recently, the discrepancy between differentiated HCC, namely cancer in
Japanese (Sakamoto et al., 1991) and non- situ. Meanwhile, AH shows a moderate
Japanese (International Working Party, increase in cell density compared with
1995) investigators regarding dysplastic the surrounding tissue but no structural
nodule and early HCC was reported to abnormalities. The nuclear cytoplasmic
reflect differences in interpretation and ratio is slightly increased due to a slight
applications of nomenclature and not due decrease in the amount of cytoplasm (The
to different biological pathways (Hirohashi liver cancer study group of Japan, 1992,
et al., 2000; Theise et al., 2002): 2003). Atypical AH presents with inter-
Pathology of adenomatous hyperplasia, mediate findings between AH and early
atypical adenomatous hyperplasia, early HCC. Nodule-in-nodule HCC consists of
hepatocellular carcinoma, and nodule-in- progressed HCC within an early HCC;
nodule hepatocellular carcinoma. emergence of less differentiated HCC
Macroscopically, the first three entities within a very well differentiated one.
such as AH (Figure 19.1c), atypical AH, Immunohistochemical examination with
and early HCC (Figure 19.2c) are difficult heat-shock protein 70 was reported to be a
to differentiate. However, mean diameters sensitive molecular marker to differentiate
of AH, atypical AH and early HCC were early HCC from precancerous lesion or
0.8, 1.2 and 1.4 cm, respectively, with a sig- noncancerous liver (Chuma et al., 2003).
nificant difference between AH and atypi- The frequency of vascular invasion and/
cal AH and between AH and early HCC or intrahepatic metastases observed micro-
(Takayasu et al., 1989). Nodule-in-nodule scopically in early HCC, nodule-in-nodule
type HCC, composed of internal develop- type HCC, and progressed HCC (Figure
ment of dedifferentiated HCC within early 19.1c) was 5%, 26%, and 35%, respec-
HCC, can be differentiated from the other tively (Watanabe et al., 1990). In our
three entities when the internal nodule hospital, early HCC and nodule-in-nodule
is large enough to be visualized grossly. HCC accounted for 14% and 5.1%, respec-
The boundary of early HCC as well as tively, of surgically resected 980 HCCs in
nodule-in-nodule type HCC is equivocal 664 patients (Oikawa et al., 2005).
to recognize on both intraoperative and
extracorporeal ultrasonography. Thus, care-
ful attention is needed not to leave positive MULTISTEP PROGRESSION OF
margins especially at surgical resection HEPATOCARCINOGENESIS
and/or local ablation therapy.
Microscopically, early HCC presents Clinical and pathological studies have
focal structural abnormalities of increased elucidated multistep progression of hepa-
cell density which are more than twice tocarcinogenesis (Sakamoto et al., 1991).
19. Hepatocellular Carcinoma and Adenomatous Hyperplasia 251
a b
Figure 19.1. A small advanced HCC associated with adenomatous hyperplasia (dysplastic nodule).
Computed Tomography hepatic arteriography (a) shows a small hyperattenuating nodular lesion (arrow)
close to a simple cyst (arrowhead) in segment 3, the lesion was demonstrated as a hypoattenuating one
(arrow) on CT arterial portography (b). Cut surface of resected specimen (c) disclosing two lesions with
almost the same size measuring 1 cm; a yellow-colored lesion was advanced HCC with satellite lesions
or intrahepatic metastases around it (arrow) and a tan-colored one was adenomatous hyperplasia (arrow-
heads), the latter of which was found only by this cut slice
a b
Figure 19.2. An early HCC located in segment 8. The dynamic CT in the arterial phase (a) discloses a
faintly hypoattenuating lesion (arrow), which was demonstrated as hypoattenuation in the delayed phase
(b) as well as on unenhanced CT (not shown). Cut surface of resected specimen (c) shows a slightly
yellow-colored equivocal lesion (arrows) with several portal tracts (arrowheads). The intratumoral archi-
tecture is preserved and similar to the surrounding parenchyma. Microscopically, the portal tract and
increased cell density with fatty change are recognized (not shown)
aspiration biopsy showed malignant trans- the other was CT hepatic arteriography
formation at 3 years (Kobayashi et al., (CTHA) in which the catheter tip is
2006). This might evoke a critical problem advanced into the superior mesenteric
of whether it suggests limitation of needle artery or the proper hepatic artery, respec-
biopsy or whether regenerative nodule tively. The former was started 20 s after
actually develops to carcinoma. On the the injection of 90 ml of Ioversol (Optiray;
other hand, needle tract seeding and intra- 350 mgI/ml, Yamanouchi, Tokyo) diluted
peritoneal hemorrhage are still possible with saline (1:3 ratio; iodine, 87.5 mgI/
risks of needle biopsy. ml) at a speed of 3 ml/s after the adminis-
tration of 10 μg of prostaglandin E1 into
the superior mesenteric artery. The latter
METHODS OF was commenced 10 s after the injection of
MULTIDETECTOR 60 ml of the same diluted contrast medium
COMPUTED TOMOGRAPHY into the proper, right or left hepatic artery
at a speed of 1.3–2 ml/s based on the
Helical CT with two (arterial and delayed variation of arterial anatomy. Both the
phases) or three phases (arterial, portal, beam collimation and image reconstruc-
and delayed phases) is widely used by tion was 7 mm. Usually CTAP was first
injecting 100–120 ml of nonionic contrast performed to obtain unequivocal contrast
medium into antecubital veins at a speed of image, followed by celiac arteriography
3 ml/s. Computed tomography is started at to know the variation of artery, and next
35–40 s (arterial phase), 70 s (portal phase), CTHA, and superselective arteriography,
and 3 min (delayed phase) after the start of if necessary.
the injection with contrast medium. Entire
livers are scanned within one breath-hold CT IMAGES OF ADVANCED
for ∼ 8–20 s, which depends on liver size. HEPATOCELLULAR
Scanning parameters are as follows: axial
single- or 4-slice mode, 5–10 mm beam
CARCINOMA
collimation, 0.5–1.0 s/rotation, 0.7–3.0 of In general, advanced HCC is demon-
helical pitch (0.7–1.0 of pitch factor), strated as a hyper-attenuating lesion in
120–150 kVp, and 200–250 mAs. Image arterial phase and hypo-attenuating in
reconstructions are 5–7 mm in thickness. portal and/or delayed phases of dynamic
CT (Takayasu et al., 1990). The washout
of contrast medium from a lesion is more
METHODS OF A clearly seen in delayed phase (3–5 min
COMBINATION OF after contrast injection). The small HCC
COMPUTED TOMOGRAPHY (< 3 cm) is homogeneously enhanced by
AND ANGIOGRAPHY contrast medium (Takayasu et al., 1995a),
and faint ring enhancement is recognized
As a further examination of hemodynamics if it is encompassed by fibrous capsule.
of lesions, a combination study of helical Large HCC (∼ 10 cm) is inhomogeneously
CT and angiography was performed. One enhanced mainly due to associated necro-
was CT arterial portography (CTAP) and sis and hemorrhage. Arterioportal shunt is
254 K. Takayasu
also recognized to show a dilated arterial The delayed phase CT showed the highest
branch followed by hepatopetal (periph- detection rate. Only 5% of early HCC pre-
eral) and/or hepatofugal (retrograde) por- sented hyper-attenuation of lesion in the arte-
tal blood flow in the arterial phase. The rial phase CT. Lim et al. (2000) reported that
filling defect of large vessels secondary to the sensitivity of three-phase helical CT for
portal and/or hepatic vein tumor thrombi DN was 39% in transplant patients. Another
could be seen more clearly in the portal study showed that 20% of 55 lesions of DN
phase. On unenhanced CT, small HCC is (13 high grade DNs and 42 low grade DNs),
usually shown as an iso-attenuating lesion which were initially found by ultrasonogra-
and a large HCC is shown as a hypo- phy, could be detected by helical dynamic
attenuating lesion. The hypo-attenuating CT (Kobayashi et al., 2006). However, these
area appears in hepatic lobar, segmental, results do not always suggest that ultra-
or hepatic hilar area due to the obstruc- sonography is superior to CT for detect-
tion of the portal vein on unenhanced CT. ing DN because ultrasonography generally
The hypo-attenuation area changes due shows high sensitivity and low specificity
to hyper-attenuation in the arterial phase rates as compared to dynamic CT.
secondary to increased arterial blood flow The most common CT pattern among
as a compensation. An advanced HCC is detected lesions was iso-iso-hypo-attenuated
shown as hyper-attenuated on CTHA and in 22%, followed by hypo-hypo-hypo pat-
hypo-attenuated on CTAP due to 100% tern in 19% (Figure 19.2A, B), and several
nutritional supply from the arterial blood different patterns in the remaining 13%.
flow (Figure 19.1A, B). For small HCC The reason for hypo- or iso-attenuation
(< 3 cm), 93% of lesions were hyper- of early HCC is presumed to be mainly
attenuated and 7% were iso-attenuated on immature neovascularization (Takayasu
CTHA, whereas 97% were hypo-attenuated et al., 1989, 1995a). The hypoattenuat-
and 3% were iso-attenuated on CTAP ing lesions on unenhanced CT showed
(Takayasu et al., 1995b). microscopically the tendency of moderate
to high grade of fatty change within the
cytoplasm (Takayasu et al., 1995a).
CT IMAGES OF EARLY CT hepatic arteriography of early HCC
HEPATOCELLULAR was hypo-attenuated in 55%, iso- in 30%,
CARCINOMA and hyper- in 15%. Twenty-seven percent
of early HCCs were hypo-attenuated on
The overall sensitivity of conventional CT both CTAP and CTHA, which suggests
for 39 resected early HCCs was 56%; the that they receive less blood flow from both
remaining 44% could not be detected due the portal and arterial vessels than the sur-
to isoattenuation throughout all phases of rounding parenchyma (Takayasu et al.,
dynamic CT and unenhanced CT. Twenty- 1995b). On the other hand, overall sensitiv-
four percent of the lesions were detected as ity of CTAP for early HCC was 66% shown
hypo-attenuation on unenhanced CT, 31% as hypo-attenuation, and the remaining
as hypo-attenuation in the arterial phase of 34% was iso-attenuated. As an unusual
dynamic CT, and 51% as hypo-attenuation in case, hyper-attenuating early HCC and/or
the delayed phase (Takayasu et al., 1995a). AH shown on CTAP has been reported.
19. Hepatocellular Carcinoma and Adenomatous Hyperplasia 255
Chuma, M., Sakamoto, M., Yamazaki, K., Ohta, observation of 795 patients with viral and alco-
T., Ohki, M., Asaka, M., and Hirohashi, S. holic cirrhosis. Hepatology 18: 47–53.
2003. Expression profiling in multistage hepa- Imamura, H., Matsuyama, Y., Tanaka, E., Ohkubo, T.,
tocarcinogenesis: identification of HSP70 as a Hasegawa, K., Miyagawa, S., Sugawara, Y.,
molecular marker of early hepatocellular carci- Minagawa, M., Takayama, T., Kawasaki, S., and
noma. Hepatology 37: 198–207. Makuuchi, M. 2003. Risk factors contributing to
de Baere, T., Risse, O., Kuoch, V., Dromain, C., early and late phase intrahepatic recurrence of
Sengel, C., Smayra, T., Gamal El Din, M., hepatocellular carcinoma after hepatectomy.
Letoublon, C., and Elias, D. 2003. Adverse events J. Hepatol. 38: 200–207.
during radiofrequency treatment of 582 hepatic International Working Party. 1995. Terminology of
tumors. AJR Am. J. Roentgenol. 181: 695–700. nodular hepatocellular lesions. Hepatology 22:
Ding, H., Kudo, M., Onda, H., Suetomi, Y., Minami, Y., 983–993.
and Maekawa, K. 2001. Hepatocellular carci- Ishii, H., Okada, S., Okusaka, T., Yoshimori, M.,
noma: depiction of tumor parenchymal flow Nakasuka, H., Shimada, K., Yamasaki, S.,
with intermittent harmonic power Doppler US Nakanishi, Y., and Sakamoto, M. 1998. Needle
during the early arterial phase in dual-display tract implantation of hepatocellular carcinoma
mode. Radiology 220: 349–356. after percutaneous ethanol injection. Cancer 82:
Hayashi, M., Matsui, O., Ueda, K., Kawamori, Y., 1638–1642.
Kadoya, M., Yoshikawa, J., Gabata, T., Kobayashi, M., Ikeda, K., Hosaka, T., Sezaki, H.,
Takashima, T., Nonomura, A., and Nakanuma, Y. Someya, T., Akuta, N., Suzuki, F., Suzuki, Y.,
1999. Correlation between the blood supply and Saitoh, S., Arase, Y., and Kumada, H. 2006.
grade of malignancy of hepatocellular nodules Dysplastic nodules frequently develop into hepa-
associated with liver cirrhosis: evaluation by CT tocellular carcinoma in patients with chronic viral
during intraarterial injection of contrast medium. hepatitis and cirrhosis. Cancer 106: 636–647.
AJR Am. J. Roentgenol. 172: 969–976. Koda, M., Murawaki, Y., Mitsuda, A., Ohyama, K.,
Hayashi, M., Matsui, O., Ueda, K., Kawamori, Y., Horie, Y., Suou, T., Kawasaki, H., and Ikawa, S.
Gabata, T., and Kadoya, M. 2002. Progression to 2000. Predictive factors for intrahepatic recur-
hypervascular hepatocellular carcinoma: corre- rence after percutaneous ethanol injection ther-
lation with intranodular blood supply evaluated apy for small hepatocellular carcinoma. Cancer
with CT during intraarterial injection of contrast 88: 529–537.
material. Radiology 225: 143–149. Kojiro, M. 1997. Pathology of hepatocellular
Hirohashi, S., Ishak, K., Kojiro, M., Wanless, I., carcinoma. In: Okuda, K., Tabor, E., editors.
Theise, N., Tsukuma, H., Blum, H., Deugnier, Y., Liver Cancer. New York: Churchill Livingstone.
Laurent Puig, P., Fischer, H., and Sakamoto, M. pp. 165–187.
2000. Hepatocellular carcinoma. In: Hamilton, S., Kondo, F., Kondo, Y., Nagato, Y., Tomizawa, M.,
Aaltonen, L., editors. World Health Organization and Wada, K. 1994. Interstitial tumour cell
Classification of Tumours: Pathology and invasion in small hepatocellular carcinoma.
Genetics of Tumours of the Digestive System. Evaluation in microscopic and low magnification
Lyon: International Agency for Research on views. J. Gastroenterol. Hepatol. 9: 604–612.
Cancer Press. pp. 159–172. Lim, J.H., Kim, C.K., Lee, W.J., Park, C.K., Koh,
Ikai, I., Arii, S., Ichida, T., Okita, K., Omata, M., K.C., Paik, S.W., and Joh, J.W. 2000. Detection
Kojiro, M., Takayasu, K., Nakanuma, Y., of hepatocellular carcinomas and dysplastic nod-
Makuuchi, M., Matsuyama, Y., and Yamaoka, Y. ules in cirrhotic livers: accuracy of helical CT in
2005. Report of the 16th follow-up survey of pri- transplant patients. AJR Am. J. Roentgenol. 175:
mary liver cancer. Hepatol. Res. 32: 163–172. 693–698.
Ikeda, K., Saitoh, S., Koida, I., Arase, Y., Tsubota, A., Murakami, T., Kim, T., Takamura, M., Hori, M.,
Chayama, K., Kumada, H., and Kawanishi, M. Takahashi, S., Federle, M.P., Tsuda, K., Osuga,
1993. A multivariate analysis of risk factors K., Kawata, S., Nakamura, H., and Kudo, M.
for hepatocellular carcinogenesis: a prospective 2001. Hypervascular hepatocellular carcinoma:
19. Hepatocellular Carcinoma and Adenomatous Hyperplasia 259
detection with double arterial phase multi-detec- and Hirohashi, S. 1990. The diagnosis of small
tor row helical CT. Radiology 218: 763–767. hepatocellular carcinomas: efficacy of various
Oikawa, T., Ojima, H., Yamasaki, S., Takayama, T., imaging procedures in 100 patients. AJR Am. J.
Hirohashi, S., and Sakamoto, M. 2005. Multistep Roentgenol. 155: 49–54.
and multicentric development of hepatocellular Takayasu, K., Wakao, F., Moriyama, N., Muramatsu, Y.,
carcinoma: histological analysis of 980 resected Sakamoto, M., Hirohashi, S., Makuuchi, M.,
nodules. J. Hepatol. 42: 225–229. Kosuge, T., Takayama, T., and Yamazaki, S.
Parkin, D.M., Bray, F., Ferlay, J., and Pisani, P. 1993. Response of early-stage hepatocellular
2001. Estimating the world cancer burden: carcinoma and borderline lesions to therapeutic
Globocan 2000. Int. J. Cancer 94: 153–156. arterial emboli-zation. AJR Am. J. Roentgenol.
Sadek, A.G., Mitchell, D.G., Siegelman, E.S., 160: 301–306.
Outwater, E.K., Matteucci, T., and Hann, H.W. Takayasu, K., Furukawa, H., Wakao, F., Muramatsu, Y.,
1995. Early hepatocellular carcinoma that devel- Abe, H., Terauchi, T., Winter, T.C., 3rd,
ops within macroregenerative nodules: growth Sakamoto, M., and Hirohashi, S. 1995a. CT
rate depicted at serial MR imaging. Radiology diagnosis of early hepatocellular carcinoma:
195: 753–756. sensitivity, findings, and CT-pathologic correla-
Sakamoto, M., and Hirohashi, S. 1998. Natural tion. AJR Am. J. Roentgenol. 164: 885–890.
history and prognosis of adenomatous hyperpla- Takayasu, K., Muramatsu, Y., Furukawa, H., Wakao, F.,
sia and early hepatocellular carcinoma: multi- Moriyama, N., Takayama, T., Yamasaki, S.,
institutional analysis of 53 nodules followed up Sakamoto, M., and Hirohashi, S. 1995b. Early
for more than 6 months and 141 patients with hepatocellular carcinoma: appearance at CT during
single early hepatocellular carcinoma treated by arterial portography and CT arteriography with
surgical resection or percutaneous ethanol injec- pathologic correlation. Radiology 194: 101–105.
tion. Jpn. J. Clin. Oncol. 28: 604–608. Takayasu, K.M., Mizuguchi, Y., Okusaka, T., Shimada,
Sakamoto, M., Hirohashi, S., and Shimosato, Y. K., Takayama, T., and Sakamoto, M. 2006. CT
1991. Early stages of multistep hepatocarcino- evaluation of the progression of hypoattenuating
genesis: adenomatous hyperplasia and early nodular lesions in virus related chronic liver dis-
hepatocellular carcinoma. Hum. Pathol. 22: ease. AJR Am. J. Roentgenol. 187: 454–463.
172–178. Theise, N.D., Park, Y.N., and Kojiro, M. 2002.
Takayama, T., Makuuchi, M., Hirohashi, S., Dysplastic nodules and hepatocarcinogenesis.
Sakamoto, M., Okazaki, N., Takayasu, K., Clin. Liver Dis. 6: 497–512.
Kosuge, T., Motoo, Y., Yamazaki, S., and The liver cancer study group of Japan. 1992. The
Hasegawa, H. 1990. Malignant transformation General Rules for the Clinical and Pathological
of adenomatous hyperplasia to hepatocellular Study of Primary Liver Cancer, 3rd ed. Tokyo:
carcinoma. Lancet 336: 1150–1153. Kanehara.
Takayama, T., Makuuchi, M., Hirohashi, S., The liver cancer study group of Japan. 2003.
Sakamoto, M., Yamamoto, J., Shimada, K., General Rules for the Clinical and Pathological
Kosuge, T., Okada, S., Takayasu, K., and Study of Primary Liver Cancer, 2nd English ed.
Yamasaki, S. 1998. Early hepatocellular carci- Tokyo: Kanehara.
noma as an entity with a high rate of surgical Tochio, H., Tomita, S., Kudo, M., Okabe, Y.,
cure. Hepatology 28: 1241–1246. Kashida, H., Hamada, M., Yamamoto, K., Terai,
Takayasu, K., Makuuchi, M., Hirohashi, S., Y., Itani, T., Mimura, J., Hirasa, M., Ibuki, Y.,
Okazaki, N., Muramatsu, Y., Moriyama, N., Komori, H., Orino, A., and Todo, A. 1994.
Takayama, T., and Hasegawa, H. 1989. Imaging Clinical tracing of small hypovascular hepatic
of adenomatous hyperplastic lesions containing nodules associated with chronic liver disease.
and not containing hepatocellular carcinoma in Acta Hepatol. Jpn. 35: 333–346.
the liver. Jpn. J. Gastroenterol. 86: 2404–2412. Tsai, J.F., Jeng, J.E., Ho, M.S., Chang, W.Y.,
Takayasu, K., Moriyama, N., Muramatsu, Y., Hsieh, M.Y., Lin, Z.Y., and Tsai, J.H. 1997.
Makuuchi, M., Hasegawa, H., Okazaki, N., Effect of hepatitis C and B virus infection on risk
260 K. Takayasu
261
262 F. Lodato et al.
Figure 20.2. American association for the study of liver diseases practice guidelines for the algorithm to
be followed for a suspicious nodule detected on US (2005)
vascular pattern or atypical vascular pat- to perform two imaging techniques, and
tern with alfa-fetoprotein > 200 ng/ml is to proceed to liver biopsy only in cases of
sufficient to make a diagnosis of HCC. discordant vascular pattern or atypical vas-
When the nodule is less than 2 cm in cular pattern between the two techniques.
diameter, diagnosis is more difficult and However, no mention is made of the prob-
may require needle liver biopsy. For nod- lems of biopsy, as detailed above, with the
ules smaller than 1 cm, the EASL consensus added consideration of operator error or
recommends serial US examinations every difficulty of being accurate in placing the
3 months until the nodule size exceeds biopsy needle in a lesion < 2 cm.
1 cm, and the same approach is suggested
by AASLD guidelines. When the nodule
is between 1 and 2 cm, EASL consensus SURVEILLANCE FOR
suggests proceeding to needle liver biopsy HCC: RADIOLOGICAL
to confirm the diagnosis, irrespective of the TECHNIQUES
vascular profile at imaging. The American
Association for the Study of Liver Diseases Radiological imaging of cirrhotic patients is
Guidelines have refined this, suggesting not easy and requires particular training and
264 F. Lodato et al.
and high resistive and pulsality indexes. the gold standard to identify hypervascular
Large regenerative nodules and DN do not lesions, the sensitivity of contrast enhanced
have arterial intralesional vessels, so these US was 97% for nodules > 3 cm in diam-
findings support the diagnosis of HCC. eter, 92% for nodules between 2 and 3 cm,
The problem is that in small HCC the and 67% for nodules < 1 cm (Gaiani et al.,
sensitivity of Doppler techniques is low 2004). Another disadvantage is that only
and a typical pulsatile flow arterial aspect one lesion or lesions close together can be
can be shown in less than 50% of nodules examined at one time. Therefore, a cost-
(Lencioni et al., 1996). effective algorithm, of which technique to
The newest microbubble contrast agents use following routine US still needs to be
and the development of specific contrast determined (Fung et al., 2004).
techniques have opened a new era in liver
ultrasonography. The AASLD Guidelines
consider contrast enhanced US as a con- SPIRAL COMPUTED
firmatory technique when a suspicion of TOMOGRAPHY
HCC is made from standard US. Images
are based on non linear acoustic effects The introduction of spiral CT in the early
of bubbles enhancing grey scale, with 1990s has dramatically improved the accu-
high contrast and spatial resolution. This racy of CT in the diagnosis of HCC.
technique improves the accuracy of US Computed Tomography data acquisition
in the diagnosis of focal liver lesions speed is greatly increased so that it is pos-
(Lencioni et al., 2002; Wen et al., 2004). sible to scan the whole liver during a single
Contrast agents are safe and well tolerated. breath hold, allowing a satisfactory evalu-
The advent of second generation contrast ation of different contrast enhancement
agents together with low mechanical index phases. Moreover, 3-D reconstructions are
scanning have led the European Federation available with faster data acquisition. The
for Ultrasound in Medicine and Biology, to standard spiral CT evaluation for HCC
define the guidelines for its use (Albrecht should include unenhanced and contrast
et al., 2004). Contrast-enhanced US is rec- enhanced phase of the arterial phase (25–
ommended to characterize any suspected 30 s after the contrast injection), of the
lesion detected at baseline US in cirrho- portal venous phase (70–80 s after contrast
sis. At contrast enhanced-US examination, injection), and a late phase (180–210 s
HCC typically shows a strong enhancement after contrast injection). As HCC is a
in the arterial phase and a rapid washout vascular tumor that receives almost all
during the portal and late phases in which its blood supply from arterial vessels, it
the nodule remains iso or hypoechoic. is well seen during the arterial phase; the
Regenerative nodules and DN usually do rest of the liver receives the majority of
not show any uptake during the arterial blood supply from the portal vein, and will
phase and their appearance resembles the not enhance until the contrast reaches the
surrounding liver. However, as for non portal system. Fast scans during the dif-
contrast-enhanced US, the sensitivity of ferent phases of contrast flow through the
using ultrasound contrast agents decreases liver help in increasing the differentiation
with smaller lesions. Defining spiral CT as between the tumor and the liver.
266 F. Lodato et al.
a b c
Figure 20.3. MRI of 76 year old man with HBV cirrhosis and 2 HCC. (a) T2 weighted image showing two
high signal lesions. (b) Arterial phase T1 image revealing arterialisation of these lesions. (c) Portal phase T1
image showing washout of the lesions. The signal characteristics of these two lesions are typical of HCCs
268 F. Lodato et al.
example, superparamagnetic iron oxide the nature of these small nodules and,
(SPIO) particles are taken up by Kuppfer most importantly, the relationship between
cells of the reticuloendothelial system these nodules and the development of
(RES-targeted contrast agents) which are HCC is uncertain. It is well accepted that
absent in HCC. They produce a decrease cirrhosis itself represents a preneoplastic
in the intensity of the liver tissue espe- condition, but in carcinogenesis the exact
cially on T2-weighted images as HCC point in time between a lesion that still
nodules show little or no uptake of the has some growth regulation and one that
contrast. Gadolinium and SPIO agents has lost growth regulation is difficult to
could be coupled (Ward and Robinson, define. The only morphological change
2002), but the usefulness of this technique that is considered as intermediate to HCC
is debated (Halavaara et al., 2002). is dysplasia (International Working Party,
Despite substantial progress in MRI still 1995), although foci of dysplasia may
fails in detecting small HCC; the sensitiv- take place within benign regenerative nod-
ity of dynamic MRI ranges from 33% to ules. Moreover, it is not clear whether all
77% in explant studies (Rode et al., 2001; dysplastic nodules will inevitably evolve
Burrel et al., 2003; Krinsky et al., 2002; to HCC. Some reports suggest that high
Teefey et al., 2003) and 78% when using grade dysplastic nodules (HGDN) should
dynamic imaging plus RES-targeted agents be considered as HCC precursors in most
(Bhartia et al., 2003). When considering cases (Borzio et al., 2003). In clinical
only HCC smaller than 1 cm, sensitivity practice a histological diagnosis of HGDN
drops to 4–71%. Nevertheless, a study poses several difficulties in the interpre-
showed that dynamic MRI is superior to tation of this finding, as a conservative
helical CT in the detection of nodules therapeutic approach could affect the out-
between 1 and 2 cm (Burrel et al., 2003). come. More difficulties arise considering
In this study, 50 patient with HCC under- that in the context of a dysplastic nodule,
going liver transplantation were analyzed. foci of HCC can exist but biopsy may
Twenty-nine of them had HCC; the patho- miss the sampling of the malignant por-
logic examination was the gold standard. tion of the lesion. In the setting of cir-
On a per nodule basis, sensitivity of MRI rhosis the picture is also complicated by
was superior to triphasic helical CT (76% vascular lesions that are not malignant
vs. 61%, p = 0.001). Sensitivity for detec- e.g., arteroportal shunts or other perfusion
tion of additional nodules decreased with abnormalities.
size and was superior to CT for nodules 10 The introduction of radiological tech-
to 20 mm (84% vs. 47%, p = 0.016). niques that study the perfusional pattern
of nodules has enormously improved the
accuracy of diagnosis, and the vascular
THE PROBLEM OF SMALL study of lesion has become the main
NODULES parameter used for the radiological diag-
nosis of HCC. Consequently the use of
The improvements in imaging techniques liver biopsy has reduced during the past
have allowed the identification of smaller few years reflecting the EASL recom-
nodules within the cirrhotic liver, but often mendations. This is clearly an advantage
20. Hepatocellular Cancer in Cirrhotic Patients: Radiological Imaging 269
considering that liver biopsy can also be performing biopsy in all lesions > 1 cm and
related to a 0–5% of seeding (Stigliano < 2 cm without a typical vascular pattern.
and Burroughs, 2005) even if associ- However, these recommendations for
ated with percutaneous ablative tech- biopsy open up a problem of the clinical
niques (Stigliano and Burroughs, 2005; applicability to perform biopsy for such
Ebara et al., 2005; Shiina et al., 2005). small nodules, as three factors affect this:
Moreover, when a biopsy is negative for the reliability of obtaining tissue from the
HCC its presence cannot be excluded, right place, the uncertainty of the diagno-
as there is a 10% risk of false negativity sis histologically and the risk of seeding.
especially in the presence of small nodules The consequences of seeding can be dev-
(< 2 cm) (Caturelli et al., 2004). astating. Surgery may no longer be fea-
Bolondi et al. (2005) showed that 78.5% sible because of the spread of neoplastic
of nodules between 1 and 3 cm diameter disease, which has been reported or seed-
with only one positive imaging technique ing may occur after potentially definitive
in terms of hypervascularity suggestive for therapy including liver transplantation. Its
HCC were found to be true HCC at biopsy. occurrence is reported in 0–5.1% of cases
When looking at the rate of hypervascular- following liver biopsy but may also occur
ity demonstrated by two separate radio- following percutaneous treatment for
logical techniques leading to the diagnosis HCC. Percutaneous ethanol injection (PEI)
of HCC, this rate was significantly lower has a risk of 0.13–6.8%. Radiofrequency
in the group of nodules between 1 and Ablation (RFA) has a reported risk of
2 cm than in the group of nodules between 0–3%, that increases to 12.5% in the
2 and 3 cm (44% vs. 84%, p = 0.001). case of superficial nodules (Stigliano and
Amongst the group of nodules between Burroughs, 2005). Prospective studies of
1 and 2 cm with discordant detection of loco-regional therapy that might be con-
hypervascularity by two techniques, a defi- sidered for reducing the risk of seeding
nite diagnosis of HCC by histology was show a seeding rate of 1.96% with PEI
made in 48%. The American Association (Ebara et al., 2005; Shiina et al., 2005) and
for the Study of Liver Diseases Guidelines 2.55% with RFA (Shiina et al., 2005).
recommend performing liver biopsy in In the setting of liver transplantation, the
cases of discordant imaging technique for need of a precise diagnosis can be dictated
nodules > 1 and < 2 cm. Based on the data by the allocation system adopted. In the
by Bolondi et al. (2005) this would be USA and in some European liver trans-
the case for almost half of these nodules plantation centres, the Model for End-
as 44% of small nodules had discordant Stage Liver Disease-HCC scoring system
imaging hypervascularity. The problem is used, in which a single nodule ≥ 2 cm
of hypovascular nodules (20% of cases or multiple nodules increases the priority
in this study) is another important one; on the waiting list (Sharma et al., 2004;
all cases occurred in the group of nodules Hayashi et al., 2004). Here the issue is
between 1 and 2 cm, and 36% were later the need of a precise diagnosis to justify
proven to be HCC. These data support the increased priority, regardless of baseline
approach suggested by the most recent liver function. It is interesting to note that
AASLD guidelines, which recommend explant studies in the USA have shown the
270 F. Lodato et al.
Halavaara, J., Tervahartiala, P., Isoniemi, H., and Levy, I., Greig, P.D., Gallinger, S., Langer, B., and
Hockerstedt, K. 2002. Efficacy of sequential use Sherman, M. 2001. Resection of hepatocellular
of superparamgnetic iron oxide and gadolinium carcinoma without preoperative tumor biopsy.
in liver MR imaging. Acta. Radiol. 43: 180–185. Ann. Surg. 234: 206–209.
Hayashi, P. H., Trotter, J.F., Forman, L., Kugelmas, Lim, J.H., Kim, C.K., Lee, W.J., Park, C.K., Koh,
M., Steinberg, T., Russ, P., Wachs, M., Bak, T., K.C., Paik, S.W., and Joh, J.W. 2000. Detection
Kam, I., and Everson, G.T. 2004. Impact of pre- of hepatocellular carcinomas and dysplastic nod-
transplant diagnosis of hepatocellular carcinoma ules in cirrhotic livers: accuracy of helical CT in
on cadveric liver allocation in the era of MELD. transplant patients. AJR Am. J. Roentgenol. 175:
Liver Transplant. 10: 42–48. 693–698.
International Working Party. 1995. Terminology of Llovet, J.M., Burroughs, A, and Bruix, J. 2003.
nodular hepatocellular lesions. Hepatology 22: Hepatocellular carcinoma. Lancet 363: 1907–1917.
983–989. Murakami, T., Kim, T., Kawata, S., Kanematsu, M.,
Ito, K., Fujita, T., Shimizu, A., Koike, S., Sasaki, K., Federle, M.P., Hori, M., Okada, A., Kumano, S.,
Matsunaga, N., Hibino S., and Yuhara, M. 2004. Sugihara, E., Tomoda, K., and Nakamura, H.
Multiarterial phase dynamic MRI of small early 2003. Evaluation of optimal timing of arterial
enhancing hepatic lesions in cirrhosis or chronic phase imaging for the detection of hypervas-
hepatitis: differentiating between hypervascular cular hepatocellular carcinoma by using triple
hepatocellular carcinomas and pseudolesions. arterial phase imaging with multidetector-row
AJR Am. J. Roentgenol. 183: 699–705. helical computed tomography. Invest. Radiol.
Kim, C.K., Lim, J.H., and Lee, W.J. 2001. Detection 38: 497–503.
of hepatocellular carcinoma and dysplastic nod- Nakayama, A., Imamura, H., Matsuyama, Y., Kitamura,
ules in cirrhotic liver: accuracy of ultrasonogra- H., Miwa, S., Kobayashi, A., Miyagawa, S., and
phy in transplant patients. J. Ultrasound. Med. Kawasaki, S. 2001. Value of lipiodol computed
20: 99–104. tomography and digital subtraction angiography
Krinsky, G.A., Lee, V.S., Theise, N.D., Weinreb, in the era of helical biphasic computed tomogra-
J.C., Morgan, G.R., Diflo, T., John, D., phy as preoperative assessment of hepatocellular
Teperman, L.W., and Goldenberg, A.S. 2002. carcinoma. Ann. Surg. 234: 56–62.
Transplantation for hepatocellular carcinoma Nino-Murcia, M., Olcott, E.W., Jeffrey, R.B.,
and cirrhosis: sensitivity of magnetic resonance Lamm, R.L., Beaulieu, C.F., and Jain, K.A.
imaging. Liver Transplant. 8: 1156–1164. 2000. Focal liver lesions: pattern-based classifi-
Laghi, A., Iannaccone, R., Rossi, P., Carbone, I., cation scheme for enhancement at arterial phase
Ferrari, R., Mangiapane, F., Nofroni, I., and CT. Radiology 215: 746–751.
Passariello, R. 2003. Hepatocellular carcinoma: Oliver, J.H., Baron, L.L., and Carr, B.E. 1997.
detection with triple-phase multidetector row CT imaging of hepatocellular carcinoma:
helical CT in patients with chronic hepatitis. CT-arteriography versus triphasic helical con-
Radiology 226: 543–549. trast CT. Radiology 205: 144 A.
Lencioni, R., Cioni, D., Crocetti, L., Donati, F., Parkin, D.M., Bray, F., Ferlay, J., and Pisani, P. 2001.
Franchini, C., Giusti, S., and Bartolozzi, C. Estimating the world cancer burden: Globocan
2002. Ultrasound imaging of focal liver lesion 2000. Int. J. Cancer 15;94: 153–156.
with a second generation contrast agent. Acad. Peterson, M.S., and Baron, R.L. 2001. Radiologic
Radiol. 9: S 371–374. diagnosis of hepatocellular carcinoma. Clin.
Lencioni, R., Cioni, D., Crocetti, L., Della Pina, C., and Liver Dis. 5: 123–144.
Bartolozzi, C. 2004. Magnetic resonance imag- Rode, A., Bancel, B., Douek, P., Chevallier, M.,
ing of liver tumors. J. Hepatol. 40: 162–171. Vilgrain, V., Picaud, G., Henry, L., Berger, F.,
Lencioni, R., Pinto, F., Armillotta, N., and Bizollon, T., Gaudin, J.L., and Ducerf, C. 2001.
Bartolozzi, C. 1996. Assessment of tumor vas- Small nodule detection in cirrhotic livers: evalu-
cularity in hepatocellular carcinoma: compari- ation with US, spiral CT, and MRI and correla-
son of power-doppler US and color-doppler US. tion with pathologic examination of explanted
Radiology 201: 353–358. liver. J. Comput. Assist. Tomogr. 25: 327–336.
20. Hepatocellular Cancer in Cirrhotic Patients: Radiological Imaging 273
Sangiovanni, A., Del Ninno, E., Fasani, P., De candidates: Prospective comparison of CT, MR
Fazio, C., Ronchi, G., Romeo, R., Morabito, imaging, US and PET. Radiology 226: 533–542.
A., De Franchis, R., and Colombo, M. 2004. Torzilli, G., Minagawa, M., Takayama, T., Inoue, K.,
Increased survival of cirrhotic patients with a Hui, A.M., Kubota, K., Ohtomo, K., and
hepatocellular carcinoma detected during sur- Makuuchi, M. 1999. Accurate preoperative eval-
veillance. Gastroenterology 126: 1005–1014. uation of liver mass lesions without fine-needle
Sharma, P., Balan, V., Hernandez, J.L., Harper, biopsy. Hepatology 30: 889–893.
A.M., Edwards, E.B., Rodriguez-Luna, H., Valls, C., Cos, M., Figueras, J., Andia, E., Ramos, E.,
Byrne, T., Vargas, H.E., Mulligan, D., Rakela, J., Sanchez, A., Serrano, T., and Torras, J. 2004.
and Wiesner, R.H. 2004. Liver transplantation Pretransplantation diagnosis and staging of
for hepatocellular carcinoma: the MELD impact. hepatocellular carcinoma in patients with cir-
Liver Transpl. 10: 36–41. rhosis: value of dual phase helical CT. AJR Am.
Shiina, S., Teratani, T., Obi, S., Sato, S., Tateishi, R., J. Roentgenol. 182: 1011–1017.
Fujishima, T., Ishikawa, T., Koike, Y., Yoshida, H., Ward, J., and Robinson, P.J. 2002. How to detect
Kawabe, T., and Omata, M. 2005. A randomized hepatocellular carcinoma in cirrhosis. Eur
controlled trial of radiofrequency ablation with Radiol. 12: 2258–2272.
ethanol injection for small hepatocellular carci- Wen, Y.L., Kudo, M., Zheng, R.Q., Ding, H.,
noma. Gastroenterology 129: 122–130. Zhou, P., Minami, Y., Chung, H., Kitano,
Stigliano, R., and Burroughs, A.K. 2005. Should M., Kawasaki, T., and Maekawa, K. 2004.
we biopsy each liver mass suspicious for HCC Charaterzation of hepatic tumors: value of
before liver transplantation? – No, please don’t. contrast enhanced coded phase-inversion har-
J. Hepatol. 43: 563–568. monic angio. AJR Am. J. Roentgenol. 182:
Teefey, S.A., Hildeboldt, C.C., Dehdashti, F., 1019–1026.
Siegel, B.A., Peters, M.G., Heiken, J.P., Wiesner, R.H., Freeman, R.B., and Mulligan, D.C.
Brown, J.J., McFarland, E.G., Middleton, W.D., 2004. Liver transplantation for hepatocellular
Balfe, D.M., and Ritter, J.H. 2003. Detection of cancer: the impact of the MELD allocation
primary hepatic malignancyn in liver transplant policy. Gastroenterology 127: S261–267.
B. Treatment
21
Treatment of Hepatocellular Carcinoma
with Thalidomide: Assessment with Power
Doppler Ultrasound
Chiun Hsu, Chiung-Nien Chen, and Ann-Lii Cheng
SUMMARY INTRODUCTION
Hepatocellular carcinoma (HCC) is typically a Hepatocellular carcinoma (HCC) is the
hypervascular tumor and anti-angiogenesis fifth most common cancer in the world,
therapy may be effective for the treatment with an estimated incidence of 1 million
of HCC. Thalidomide has been shown to new cases per year (Parkin, 2001). Most
inhibit angiogenesis induced by various of the patients with HCC are diagnosed at
proangiogenic factors and may produce advanced stage and not eligible for cura-
objective tumor response in various can- tive therapy. However, there is no standard
cers. In HCC, thalidomide may produce systemic therapy that has proven efficacy
complete or partial remission in ~ 4–7% in patients with advanced HCC. HCC
of patients and disease stabilization in patients usually tolerate conventional cyto-
10–20% patients with advanced disease. toxic chemotherapy poorly because of the
We have performed a prospective study underlying cirrhosis and the accompanying
using power Doppler ultrasound to evalu- hypersplenism and peripheral cytopenia
ate the vascular response to thalidomide (Hsu et al., 2004). Therefore, novel thera-
treatment in patients with advanced HCC. peutic agents must be developed to treat
Patients who achieved objective tumor this difficult disease.
response to thalidomide tended to have Molecular targeted therapy, which aims
a higher pretreatment vascularity index, at specific molecular derangements in
defined as the number of colored (vas- cancer cells or their microenvironment,
cular) pixels within a well-demarcated represents breakthrough over conven-
tumor area divided by the number of total tional cytotoxic therapy (Gschwind et
pixels in that area, than patients who did al., 2004). A rapidly expanding list of
not. The pretreatment vascularity index cancers, including non-small cell lung
also correlated with levels of circulating cancer, breast cancer, gastrointestinal stro-
angiogenic factors. The value of power mal tumor, colorectal cancer, and chronic
Doppler ultrasound in future clinical trials myelogeneous leukemia, are now treatable
of anti-angiogenesis therapy of HCC is by agents targeting at specific signaling
discussed. pathways (Krause and Van Etten, 2005).
277
278 C. Hsu et al.
In addition to targeting the cancer cells per decrease the transcription of proangiogenic
se, anti-angiogenic/antivascular therapy factors such as, VEGF, and insulin-like
is another promising approach to treat growth factor (Stephens et al., 2000; Vacca
cancers that are usually refractory to con- et al., 2005). Its potential therapeutic
ventional cytotoxic therapy. Development efficacy for neoplastic disorders has been
of novel molecular targeting agents has reported in patients with refractory multiple
thus become a key endeavor for the global myeloma and human immunodeficiency
pharmacological industry and numerous virus-associated Kaposi’s sarcoma (Singhal
agents are already in the development et al., 1999; Little et al., 2000).
pipeline. Several single-arm clinical trials to test
HCC is typically a hypervascular tumor. the antitumor efficacy of thalidomide for
HCC cells and their surrounding stroma advanced HCC have been published, and
cells have been found to express various their results are summarized in Table 21.1
proangiogenic factors, and these proangio- (Hsu et al., 2003; Wang et al., 2004a, b;
genic factors may be closely associated Lin et al., 2005; Patt et al., 2005; Chen
with high histological grade, portal venous et al., 2005; Schwartz et al., 2005). The
thrombosis, and tumor capsular invasion results are similar among different studies,
(Torimura et al., 1998; Chow et al., 1997, with objective response rate of ~ 4–7% and
1998; Harada et al., 1998). It appears that overall survival of 4–7 months. Dose esca-
a proangiogenic milieu exists to facilitate lation of thalidomide was planned in most
tumor proliferation and invasion in HCC. studies, but the patients usually cannot
It is thus reasonable to hypothesize that tolerate a daily dose of > 300 mg because
anti-angiogenesis therapy may be able to of treatment-related toxicity, including
control this tumor. fatigue, constipation, and neurotoxicity.
The anti-angiogenic activity of thalidomide The short survival and poor tolerance
was first demonstrated in the 1990s. It mostly reflect the advanced disease status
can inhibit in vivo angiogenesis induced and compromised liver function reserves
by vascular endothelial growth factor in this group of patients.
(VEGF) and basic fibroblast growth factor Despite its unsatisfactory overall efficacy
(D’Amato et al., 1994; Kenyon et al., in the treatment of advanced HCC, thali-
1997). Thalidomide has also been shown to domide does produce tumor stabilization
Table 21.1. Clinical trials of thalidomide for patients with advanced HCC.
No. of Daily dose of Objective Disease Overall survival
Author patients thalidomide response rate stabilization rate (months)
Hsu et al. 68 200 mg 6.3% (1 CR, 3 PR) 15.9% 4.3
Wang et al. 99 150–300 mg 7.0% (6 PR) NA NA
Lin et al. 26 200 mg 3.9% (1 PR) 11.5% 4.0
Patt et al. 37 400 mg 5% (1 PR) NA 6.8
Chen et al. 42 200 mg 7% (3 PR) 23.8% 4.0
Schwartz et al. 38 200 mg 5% (1 CR, 1 PR) 18% 5.5
Disease control rate: complete response + partial response + stable disease > 2 months
CR, complete response; PR, partial response
21. Treatment of Hepatocellular Carcinoma with Thalidomide 279
EVALUATION OF VASCULAR
EVALUATION OF TUMOR RESPONSE OF HCC TO
VASCULARITY USING POWER THALIDOMIDE BY POWER
DOPPLER ULTRASOUND DOPPLER ULTRASOUND:
A PROSPECTIVE STUDY
Many methods have been developed to
evaluate tumor vascularity (Choyke et al., We have performed a prospective study of
2002). Among these methods, Doppler thalidomide for the treatment of advanced
ultrasound has the advantage of non- HCC and power Doppler ultrasound was
invasiveness and ease to use for serial used to evaluate the vascular response to
follow-up. Tumor vascularity measured thalidomide treatment (Hsu et al., 2005).
by Doppler ultrasound has been shown to Patients with advanced HCC that were not
provide important prognostic information suitable for curative surgery or other local
for patients with colon, gastric, ovarian, treatment were enrolled to receive thali-
and cervical cancers and has shown good domide treatment. The starting dose was
correlation with vessel density determined 200 mg per day and the dose was escalated
by immunohistochemical staining (Wu by 100-mg increments every 2–3 weeks
et al., 1994; Cheng et al., 1999, 2000, 2002). if no grade 2 or greater treatment-related
The potential of using Doppler ultrasound toxicity developed. Clinical response to
280 C. Hsu et al.
thalidomide treatment was evaluated based stored for later analysis. Each tumor was
on computed tomography (CT) scan find- scanned three times. The stored images
ings and the clinical response. The patients were then retrieved and the examiner
were considered to have response to thali- contoured the tumor margin by using
domide treatment if complete or partial a cursor. Quantification of the vascular
response by WHO criteria (Miller et al., color signals within the demarcated area
1981) or a clinical benefit response was doc- was performed by software (Encomate;
umented. The clinical benefit response was Electronic Business Machine Co., Ltd.,
achieved when all the following three criteria Taipei, Taiwan). The vascularity index (VI)
were met: (1) stable disease on CT imaging, was defined as the number of colored
(2) a more than 50% decrease of alpha- (vascular) pixels within a well-demarcated
fetoprotein level for more than 4 weeks, and tumor area divided by the number of total
(3) improved performance status. pixels in that area. For each tumor the
Sonographic examination was performed mean VI of the three representative tumor
every 1–2 weeks during thalidomide sections, one from each scan, was used for
treatment. The HDI5000 power Doppler statistical analysis.
ultrasound unit (Advanced Technology From April 2000 to March 2003, 144
Laboratories, Bothell, WA) was used. A patients with advanced HCC were screened
2–5 MHz curved array was used for evalu- for enrollment into this study. Tumor vas-
ation of liver tumors and a 5–10 MHz cularity was evaluable by power Doppler
broad-band linear array transducer for ultrasound in only 56 of these patients.
superficial tumors, such as lymph nodes Reasons for nonevaluability included (1)
metastasis. For examination of the liver huge tumor size that could not be encom-
tumors, routine abdominal ultrasound was passed within the field of the sonographic
first performed to identify the index lesion, examination; (2) indiscrete tumor margin
followed by power Doppler evaluation. resulting from diffuse tumor infiltration,
For examination of the superficial tumors, severe cirrhosis, or prior local therapy;
power Doppler ultrasound was directly (3) location of the index tumor that could
used. The power Doppler ultrasound set- not be approached by sonographic exami-
ting was standardized using a medium nation, such as lung or intra-abdominal
wall filter, a color gain of 79%, a pulse lymph nodes. Of the 56 patients with
repetition frequency of 1,000 Hz with evaluable tumor vascularity, 47 (38 men,
moderate-to-long persistence, and a slow 9 women, median age 62.2 years) were
sweep technique to achieve the highest enrolled into this study. The index lesion
sensitivity without apparent background was a liver tumor for 44 patients, a meta-
noise. Focusing depth was set between 1 static neck lymph node for 2 patients,
and 12 cm. Only one representative lesion and a metastatic skin tumor for 1 patient.
was chosen for follow-up power Doppler Patients who completed at least 1 month
examination. of thalidomide treatment were considered
The tumor was scanned carefully in evaluable for response. Forty-four of these
all directions, and the tumor section with patients were evaluable for response to
maximal color signals was captured and thalidomide treatment and all 47 patients
21. Treatment of Hepatocellular Carcinoma with Thalidomide 281
were evaluable for toxicity. Three of the by liver biopsy. He had received chemo-
47 enrolled patients did not complete 1 therapy with doxorubicin before the start
month of treatment because of subjective of thalidomide treatment but progressive
intolerance of side effects, mainly fatigue. disease after chemotherapy was docu-
During thalidomide treatment, the pati- mented. The patient took thalidomide up
ents were followed regularly with physical to 600 mg per day and had stable disease
examination, hematology, serum biochem- for 2 months. Figure 21.1a was from the
istry and alpha-fetoprotein tests. Computed baseline examination and Figure 21.1b
tomography scan was done before the start from examination done at the documenta-
and after 1 month of thalidomide therapy tion of progressive disease to thalidomide
and then every 2 months. Sonographic treatment. An increase in both tumor area
examination was performed before and and vascularity index was noted. Figure
every 1–2 weeks during thalidomide treat- 21.1c and d were from a 72-year-old
ment. Five patients were considered to have woman with hepatitis C-related cirrhosis.
response to thalidomide treatment, includ- Hepatocellular carcinoma was diagnosed
ing one complete and one partial response by compatible imaging findings and high
based on CT scan findings. Another three alpha-fetoprotein levels. She had received
patients fulfilled the criteria of clinical multiple sessions of trans-arterial chemo-
benefit response described above. The total embolization before enrollment into this
response rate was 11.4% (95% C.I.: 1.6– study. The patient received thalidomide
21.2%). Fifteen patients had stable disease 200 mg per day and stable disease was
and 24 patients had progressive disease. achieved for 23 weeks. Figure 21.1c was
The median pretreatment value of VI from the baseline examination and Figure
was 2.73 (range 0–25.36). Pretreatment VI 21.1d from examination done at the docu-
was significantly higher in patients with an mentation of stable disease to thalidomide
objective tumor response to thalidomide treatment. The tumor area remained sta-
(median 7.42 vs. 2.15, p = 0.03). After tha- tionary while a decrease of vascularity
lidomide treatment, four of the five patients index was noted.
with response had a decrease in VI, but 20 In addition to power Doppler ultra-
of the 36 patients without response had sound, we also checked the levels of cir-
also a decrease of VI. The tumors of three culating proangiogenic factors, including
patients became non-evaluable for tumor vascular endothelial growth factor, basic
vascularity during follow-up because of fibroblast growth factor, and placental
progressive tumor infiltration and blurring growth factor. The pretreatment VI was
of the tumor margins. All of them had pro- significantly correlated with pretreatment
gressive disease. levels of basic fibroblast growth factor and
Representative images of power Doppler placental growth factor, suggesting that VI
ultrasound in our patients are shown in may reflect the activity of tumor angiogen-
Figure 21.1. Figure 21.1a and b were esis. However, none of the proangiogenic
from a 51-year-old man who presented factors we measured differed significantly
with multiple liver and lung tumors. between responders and non-responders to
Hepatocellular carcinoma was diagnosed thalidomide treatment.
282 C. Hsu et al.
Figure 21.1. Power Doppler ultrasound findings of HCC patients treated with thalidomide. a and b
were from a 51-year-old man with progressive disease after thalidomide treatment. c and d were from a
72-year-old woman with stable disease for 23 weeks after thalidomide treatment. a and c, before start of
thalidomide treatment; b and d, at documentation of response to thalidomide treatment
is the most extensively studied. The intra- other molecular targeted therapy for HCC
venously injected microbubbles can signi- is ~ 5% (Philip et al., 2005; Schwartz et al.,
ficantly enhance the Doppler signals of 2006; Abou-Alfa et al., 2006). Therefore, it
blood flow and help differential diagnosis is difficult to find a clear-cut threshold VI
of liver tumors (Kim et al., 1998; Vilana value that may predict the response to anti-
et al., 2003; Quaia et al., 2004). The use- angiogenesis therapy. Besides, in our study
fulness of contrast-enhanced ultrasound decrease in VI after thalidomide treatment
in evaluation of tumor blood flow after was also found in patients with progressive
thalidomide treatment has been suggested disease, suggesting that tumors in these
in a pilot study for patients with advanced patients have surpassed the angiogenesis
HCC (Bertolotto et al., 2006). New imag- regulatory mechanisms that may be inhib-
ing techniques, such as pulse inversion ited by thalidomide. The third limitation
harmonic ultrasound and agent detection is the inherent dependence of ultrasound
imaging, may further improve the image on individual examiner’s technique. Repro-
resolution and reduce artifacts (Choi et al., ducibility of these new sonography tech-
2002; Lee et al., 2006). The potential of niques must be validated before they can
these new techniques in the monitoring of be widely used in the clinical settings.
anti-angiogenesis therapy for HCC should Other functional imaging modalities
be further investigated. have also been extensively studied for
Despite these technical advances, there their use in evaluation of anti-angiogenesis
are still limitations in establishing the therapy. Dynamic contrast-enhanced mag-
role of ultrasound for evaluation of anti- netic resonance imaging (DCE-MRI) has
angiogenesis therapy in HCC. The first been widely used to evaluate the vascular
limitation is the patient population. Most of response, including change in vascular
the HCC patients have underlying cirrhosis perfusion and permeability, after antiang-
and had received various local treatments, iogenic therapy (Padhani and Leach, 2005;
such as trans-arterial embolization or per- Liu et al., 2005). Guidelines for its use
cutaneous ethanol injection, before they in early-phase clinical trials have been
started empirical antiangiogenesis therapy. proposed (Leach et al., 2005; Evelhoch
The margins of their tumors are often et al., 2005). DCE-MRI is relatively non-
indiscrete and calculation of VI is there- invasive and can be objectively measured
fore difficult. Besides, patients with large serially. However, detailed definition of
or deep-seated tumors in the liver may not the examination protocols is needed to
be assessable by ultrasound. In our study improve reproducibility. The potential
only ~ 40% of patients with advanced HCC anti-angiogenesis effects of thalidomide
who were otherwise eligible for thalido- in HCC have been evaluated by DCE-
mide therapy were considered evaluable MRI. Patients with stable disease after
by power Doppler ultrasound. The second thalidomide treatment tended to have
limitation is the difficulty in correlating greater reduction of perfusion in their
the vascular response with other clinical tumors, as indicated by a greater decrease
outcomes, such as tumor shrinkage and in peak and maximal signal enhancement,
patient survival. The response rate of curr- than patients with progressive disease (Wang
ently available anti-angiogenesis therapy or et al., 2004b). Other endpoints of vascular
284 C. Hsu et al.
response, such as the transfer constant Cheng, W.F., Lee, C.N., Chu, J.S., Chen, C.A.,
(Ktrans) and the initial area under the gado- Chen, T.M., Shau, W.Y., Hsieh, C.Y., and Hsieh,
F.J. 1999. Vascular index as a novel parameter
linium concentration time curve, should
for the in vivo assessment of angiogenesis in
be evaluated for their validity in future patients with cervical carcinoma. Cancer 85:
clinical trials of anti-angiogenesis therapy 651–657.
for HCC. Choi, B.I., Kim, T.K., Han, J.K., Chung, J.W.,
In conclusion, evaluation of tumor vas- Park, J.H., and Han, M.C. 1996. Power versus
cularity by power Doppler ultrasound conventional color Doppler sonography: com-
parison in the depiction of vasculature in liver
is feasible in a subset of patients with
tumors. Radiology 200: 55–58.
advanced HCC. The usefulness of power Choi, B.I., Kim, A.Y., Lee, J.Y., Kim, K.W.,
Doppler ultrasound and other functional Lee, K.H., Kim, T.K., and Han, J.K. 2002.
imaging modalities should be explored in Hepatocellular carcinoma: contrast enhancement
future clinical trials of anti-angiogenesis with Levovist. J. Ultrasound Med. 21: 77–84.
therapy for HCC. Chow, N.H., Hsu, P.I., Lin, X.Z., Yang, H.B., Chan,
S.H., Cheng, K.S., Huang, S.M., and Su, I.J.
1997. Expression of vascular endothelial growth
factor in normal liver and hepatocellular carci-
REFERENCES
noma: an immunohistochemical study. Human
Abou-Alfa, G.K., Schwartz, L., Ricci, S., Amadori, D., Pathol. 28: 698–703.
Santoro, A., Figer, A., De Greve, J., Douillard, J.Y., Chow, N.H., Cheng, K.S., Lin, P.W., Chan, S.H.,
Lathia, C., Schwartz, B., Taylor, I., Moscovici, M., Su, W.C., Sun, Y.N., and Lin, X.Z. 1998.
and Saltz, L.B. 2006. Phase II study of sorafenib Expression of fibroblast growth factor-1 and
in patients with advanced hepatocellular carcinoma. fibroblast growth factor-2 in normal liver and
J. Clin. Oncol. 24: 4293–4300. hepatocellular carcinoma. Digest. Dis. Sci. 10:
Bertolotto, M., Pozzato, G., Croce, L.S., Nascimben, F., 2261–2266.
Gasparini, C., Cova, M.A., and Tiribelli, C. 2006. Choyke, P.L., Knopp, M.V., and Libutti, S.K. 2002.
Blood flow changes in hepatocellular carcinoma Special techniques for imaging blood flow to
after the administration of thalidomide assessed tumors. Cancer J. 8: 109–118.
by reperfusion kinetics during microbubble infu- D’Amato, R.J., Loughnan, M.S., Flynn, E., and
sion: preliminary results. Invest. Radiol. 41: Folkman, J. 1994. Thalidomide is an inhibitor of
15–21. angiogenesis. Proc. Natl. Acad. Sci. U.S.A. 91:
Chen, C.N., Cheng, Y.M., Liang, J.T., Lee, P.H., 4082–4085.
Hsieh, F.J., Yuan, R.H., Wang, S.M., Chang, Evelhoch, J., Garwood, M., Vigneron, D., Knopp, M.,
M.F., and Chang, K.J. 2000. Color Doppler Sullivan, D., Menkens, A., Clarke, L., and
vascular index can predict distant metastasis and Liu, G. 2005. Expanding the use of magnetic
survival in colon cancer patients. Cancer Res. resonance in the assessment of tumor response
60: 2892–2897. to therapy: workshop report. Cancer Res. 65:
Chen, C.N., Cheng, Y.M., Lin, M.T., Hsieh, F.J., 7041–7044
Lee, P.H., and Chang, K.J. 2002. Association of Gee, M.S., Saunders, H.M., Lee, J.C., Sanzo,
color Doppler vascularity index and microves- J.F., Jenkins, W.T., Evans, S.M., Trinchieri,
sel density with survival in patients with gastric G., Sehgal, C.M., Feldman, M.D., and Lee,
cancer. Ann. Surg. 235: 512–518. W.M. 2001. Doppler ultrasound imaging detects
Chen, L.T., Liu, T.W., Chao, Y., Shiah, H.S., changes in tumor perfusion during antivascu-
Chang, J.Y., Juang, S.H., Chen, S.C., Chuang, lar therapy associated with vascular anatomic
T.R., Chin, Y.H., and Whang-Peng, J. 2005. alterations. Cancer Res. 61: 2974–2982.
α-fetoprotein response predicts survival benefits Goertz, D.E., Yu, J.L., Kerbel, R.S., Burns, P.N.,
of thalidomide in advanced hepatocellular carci- and Foster, F.S. 2002. High-frequency Doppler
noma. Aliment. Pharmacol. Ther. 22: 217–226. ultrasound monitors the effects of antivascular
21. Treatment of Hepatocellular Carcinoma with Thalidomide 285
therapy on tumor blood flow. Cancer Res. 62: Workman, P., and Pharmacodynamic/Pharma-
6371–6375. cokinetic Technologies Advisory Committee,
Gschwind, A., Fischer, O.M., and Ullrich, A. 2004. Drug Development Office, Cancer Research
The discovery of receptor tyrosine kinases: UK. 2005. The assessment of antiangiogenic
targets for cancer therapy. Nat. Rev. Cancer 4: and antivascular therapies in early-stage clini-
361–370. cal trials using magnetic resonance imaging:
Harada, T., Arii, S., Mise, M., Imamura, T., issues and recommendations. Brit. J. Cancer 92:
Higashitsuji, H., Furutani, M., Niwano, M., 1599–1610.
Ishigami, S., Fukumoto, M., Seiki, M., Sato, H., Lee, J.Y., Choi, B.I., Han, J.K., Lee, J.M., and Kim,
and Imamura, M. 1998. Membrane-type matrix S.H. 2006. State-of-the-art ultrasonography of
metalloproteinase-1 (MT1-MMP) gene is over- hepatocellular carcinoma. Eur. J. Radiol. 58:
expressed in highly invasive hepatocellular 177–185
carcinomas. J. Hepatol. 28: 231–239. Lencioni, R., Pinto, F., Armillotta, N., and
Hsu, C., Chen, C.N., Chen, L.T., Wu, C.Y., Yang, Bartolozzi, C. 1996. Assessment of tumor vascu-
P.M., Lai, M.Y., Lee, P.H., and Cheng, A.L. larity in hepatocellular carcinoma: comparison
2003. Low-dose thalidomide treatment for of power Doppler US and color Doppler US.
advanced hepatocellular carcinoma. Oncology Radiology 201: 353–358.
65: 242–249. Lin, A.Y., Brophy, N., Fisher, G.A., So, S., Biggs, C.,
Hsu, C., Cheng, J.C., and Cheng, A.L. 2004. Yock, T.I., and Levitt, L. 2005. Phase II
Recent advances in non-surgical treatment for study of thalidomide in patients with unre-
advanced hepatocellular carcinoma. J. Formos. sectable hepatocellular carcinoma. Cancer 103:
Med. Assoc. 103: 483–495. 119–125.
Hsu, C., Chen, C.N., Chen, L.T., Wu, C.Y., Hsieh, Little, R.F., Wyvill, K.M., Pluda, J.M., Welles, L.,
F.J., and Cheng, A.L. 2005. Effect of thalidomide Marshall, V., Figg, W.D., Newcomb, F.M.,
in hepatocellular carcinoma: assessment with Tosato, G., Feigal, E., Steinberg, S.M.,
power Doppler US and analysis of circulating Whitby, D., Goedert, J.J., and Yachoan, R. 2000.
angiogenic factors. Radiology 235: 509–516. Activity of thalidomide in AIDS-related Kaposi’s
Kenyon, B.M., Browne, F., and D’Amato, R. 1997. sarcoma. J. Clin. Oncol. 18: 2593–2602.
Effects of thalidomide and related metabolites Liu, G., Rugo, H.S., Wilding, G., McShane, T.M.,
in a mouse corneal model of neovascularization. Evelhoch, J.L., Ng, C., Jackson, E., Kelcz, F.,
Exp. Eye Res. 64: 971–978. Yeh, B.M., Lee, F.T. Jr., Charnsangavej, C.,
Kim, A.Y., Choi, B.I., Kim, T.K., Han, J.K., Park, J.W., Ashton, E.A., Steinfeldt, H.M.,
Yun, E.J., Lee, K.Y., and Han, M.C. 1998. Pithavala, Y.K., Reich, S.D., and Herbst, R.S.
Hepatocellular carcinoma: power Doppler 2005. Dynamic contrast-enhanced magnetic
US with a contrast agent: preliminary results. resonance imaging as a pharmacodynamic measure
Radiology 209: 135–140. of response after acute dosing of AG-013736,
Korn, E.L., Arbuck, S.G., Pluda, J.M., Simon, R., an oral angiogenesis inhibitor, in patients with
Kaplan, R.S., and Christian, M.C. 2001. advanced solid tumors: results from a phase I
Clinical trial designs for cytostatic agents: are study. J. Clin. Oncol. 23: 5464–5473.
new approaches needed? J. Clin. Oncol. 19: Miller, A.B., Hoogstraten, B., Staquet, M., and
265–272. Winkler, A. 1981. Reporting results of cancer
Krause, D.S., and Van Etten, R.A. 2005. Tyrosine treatment. Cancer 47: 207–214.
kinases as targets for cancer therapy. N. Engl. J. Padhani, A.R., and Leach, M.O. 2005. Antivascular
Med. 353: 172–187. cancer treatments: functional assessments by
Leach, M.O., Brindle, K.M., Evelhoch, J.L., dynamic contrast-enhanced magnetic resonance
Griffiths, J.R., Horsman, M.R., Jackson, A., imaging. Abd. Imaging 30: 324–341
Jayson G.C., Judson, I.R., Knopp, M.V., Maxwell, Parkin, D.M. 2001. Global cancer statistics in the
R.J., McIntyre, D., Padhani, A.R., Price, P., year 2000. Lancet Oncol. 2: 533–543.
Rathbone, R., Rustin, G.J., Tofts, P.S., Tozer, Patt, Y.Z., Hassan, M.M., Lozano, R.D., Nooka,
G.M., Vennart, W., Waterton, J.C., Williams, S.R., A.K., Schnirer, I.I., Zeldis, J.B., Abbruzzese, J.L.,
286 C. Hsu et al.
and Brown, T.D. 2005. Thalidomide in the treat- Stephens, T.D., Bunde, C.J., and Fillmore, B.J.
ment of patients with hepatocellular carcinoma: 2000. Mechanism of action in thalidomide tera-
a phase II trial. Cancer 103: 749–755. togenesis. Biochem. Pharmacol. 59: 1489–1499.
Philip, P.A., Mahoney, M.R., Allmer, C., Thomas, J., Torimura, T., Sata, M., Ueno, T., Kin, M., Tsuji, R.,
Pitot, H.C., Kim, G., Donehower, R.C., Fitch, T., Suzaku, K., Hashimoto, O., Sugawara, H., and
Picus, J., and Erlichman, C. 2005. Phase II Tanikawa, K. 1998. Increased expression of
study of erlotinib (OSI-774) in patients with vascular endothelial growth factor is associated
advanced hepatocellular cancer. J. Clin. Oncol. with tumor progression in hepatocellular carci-
23: 6657–6663. noma. Human Pathol. 29: 986–991.
Quaia, E., Calliada, F., Bertolotto, M., Rossi, S., Vacca, A., Scavelli, C., Montefusco, V., Di Pietro, G.,
Garioni, L., Rosa, L., and Pozzi-Mucelli, R. 2004. Neri, A., Mattioli, M., Bicciato, S., Nico, B.,
Characterization of focal liver lesions with Ribatti, D., Dammacco, F., and Corradini, P. 2005.
contrast-specific US modes and a sulfur Thalidomide downregulates angiogenic genes in
hexafluoride-filled microbubble contrast agent: bone marrow endothelial cells of patients with
diagnostic performance and confidence. Radiology active multiple myeloma. J. Clin. Oncol. 23:
232:420–430. 5334–5346.
Rubin, J.M., Bude, R.O., Carson, P.L., Bree, Vilana, R., Llovet, J.M., Bianchi, L., Sanchez, M.,
R.L., and Adler, R.S. 1994. Power Doppler Pages, M., Sala. M., Gilabert, R., Nicolau, C.,
US: a potentially useful alternative to mean- Garcia, A., Ayuso, C., Bruix, J., and Bru, C.
frequency-based color Doppler US. Radiology 2003. Contrast-enhanced power Doppler sonog-
190: 853–856. raphy and helical computed tomography for
Schwartz, J.D., Sung, M., Schwartz, M., Lehrer, D., assessment of vascularity of small hepatocellular
Mandeli, J., Liebes, L., Goldenberg, A., and carcinomas before and after percutaneous abla-
Volm, M. 2005. Thalidomide in advanced hepa- tion. J. Clin. Ultrasound 31: 119–128.
tocellular carcinoma with optional low-dose Wang, J., Chen, L.T., Tsang, Y.M., Liu, T.W., and
interferon-α2a upon progression. Oncologist 10: Shih, T.T. 2004a. Dynamic contrast-enhanced
718–727. MRI analysis of perfusion changes in advanced
Schwartz, J.D., Schwartz, M., Lehrer, D., Cohen, E., hepatocellular carcinoma treated with an antian-
Sung, M., Kinkhabwala, M., Siegel, A., giogenic agent: a preliminary study. Am. J.
Holloway, S., Ocean, A., and Wadler, S. 2006. Roentgenol. 183: 713–719.
Bevacizumab in unresectable hepatocellular Wang, T.E., Kao, C.R., Lin, S.C., Chang, W.H.,
carcinoma (HCC) for patients without metastasis Chu, C.H., Lin, J., and Hsieh, R.K. 2004b.
and without invasion of the portal vein. Proc. Salvage therapy for hepatocellular carcinoma
Am. Soc. Clin. Oncol. 24: abstr 4144. with thalidomide. World J. Gastroenterol. 10:
Singhal, S., Mehta, J., Desikan, R., Ayers, D., 649–653.
Roberson, P., Eddlemon, P., Munshi, N., Anaissie, Wu, C.C., Lee, C.N., Chen, T.M., Shyu, M.K.,
E., Wilson, C., Dhodapkar, M., Zeddis, J., Hsieh, C.Y., Chen, H.Y., and Hsieh, F.J. 1994.
and Barlogie, B. 1999. Antitumor activity of Incremental angiogenesis assessed by color
thalidomide in refractory multiple myeloma. N. Doppler ultrasound in the tumorigenesis of ovarian
Engl. J. Med. 341: 1565–1571 neoplasm. Cancer 73: 1251–1256.
22
Perfusion Scintigraphy with Integrated
Single Photon Emission Computed
Tomography/Computed Tomography
in the Management of Transarterial
Treatment of Hepatic Malignancies
Timm Denecke, Bert Hildebrandt, and Enrique Lopez-Hänninen
287
288 T. Denecke et al.
most experiences with this approach has advantage of systemic drug concentrations
been gained with the treatment of “liver provided by the intraarterial application of
only”-metastases of colorectal can- 5-FU (Lorenz and Muller, 2000; Sadahiro
cer (Cohen and Kemeny, 2003; MAGC, et al., 2004).
1996). In addition, limited data are avail- Unlike hepatic arterial FUDR, which
able for other liver neoplasms such as is usually delivered through surgically
hepatocellular (Tanaka et al., 2005) or implanted infusion pumps, regional appli-
cholangiocellular carcinoma (Cantore cations of 5-FU demand the use of external
et al., 2005). Recently, the introduction of pumps, due to the maximal drug concen-
minimal invasively implanted port systems tration of 25 mg/ml. Therefore, hepatic
led to a reappraisal of this approach (Ricke arterial 5-FU can only be delivered by
et al., 2004). intermittent percutaneous access, or
For colorectal cancer patients with unre- through port systems. There is extensive
sectable liver metastases or after hepatic experience with the use of port catheters in
metastasectomy, more than a dozen of ran- a number of specialized treatment centers
domized trials have compared HAIC using worldwide, but this technique is associ-
fluoropyrimidines floxuridine (FUDR) or ated with a considerable proportion of
5-flourouracil (5-FU) with intravenous primary failure which has been reported to
drug application (Cohen and Kemeny, approximate one third of patients sched-
2003; Nelson and Freels, 2004). Summing uled for 5-FU-based HAIC in recent mul-
up those results, HAIC proved to improve ticenter trials (Lorenz and Muller, 2000;
response rates in patients with unresect- Kerr et al., 2003). Once HAIC has been
able disease and time to hepatic progres- initiated, port catheters are associated with
sion for both indications, but results were higher complication rates than surgically
counterbalanced by the lack of a survival implanted pumps, whereas secondary fail-
benefit in most of the studies. ure rates have been reported to be similar
The reasons preventing a clear-cut sur- (Heinrich et al., 2003). Presently, it is
vival benefit for HAIC have been dis- accepted that the low performance level of
cussed in detail (Cohen and Kemeny, surgically implanted port catheters repre-
2003; MAGC, 1996; Nelson and Freels, sents one major reason why the theoretical
2004). Particularly in the earlier trials, advantage of 5-FU-based HAIC could
small patient numbers and inappropriate not be transferred to convincing clinical
control groups may have led to mislead- improvements.
ing interpretations. In addition, exces- In this context, interventionally
sively high rates of extrahepatic failure implanted port catheter systems (IIPCS)
were observed in studies on intraarte- have evolved a promising alternative to
rial FUDR. Adjacent trials completed in surgically implanted devices. The use of
North America suggested that systemic IIPCS enables initiation of HAIC with-
control rates of intraarterial FUDR can be out laparotomy, and available data sug-
improved by combinations with intrave- gest favorable complication and failure
nous 5-FU (Kemeny et al., 1999). In addi- rates with this approach. However, studies
tion, European and Asian investigators published so far mainly focused on either
aimed to optimize results by taking the technical or clinical aspects of treatment
22. Perfusion Scintigraphy with Integrated Single Photon Emission 289
(Ricke et al., 2004). Further trials are nec- of these patients resection was enabled by
essary to more clearly define the value of therapy related shrinkage of the tumors.
this promising novel technique. Concerning colorectal cancer metas-
tases, the most relevant (randomized phase
Transarterial Radioembolization III) trial so far included 71 patients who
received intraarterially either FUDR alone
As another innovative modality, the arte-
or FUDR in combination with Y90 resin
rial administration of Y90 loaded glass
microspheres. There was a significant dif-
and resin microspheres to the liver for
ference observed concerning the mean
treatment of hepatic malignancies was
time to disease progression in both groups,
currently approved by the North American
which were favorable for those patients
Food and Drug Administration (FDA) for
who were treated with radioembolization
specific indications, the stand alone or pre-
(9.7 vs. 15.9 months, p = 0.001) (Gray
surgical or pretransplant radiation therapy
et al., 2001). Similar survival benefits
of irresectable hepatocellular carcinoma
were demonstrated in other trials for RE
(glass device, 2000) and the treatment of
plus chemotherapy vs. chemotherapy alone
irresectable colorectal metastases in com-
with a mean survival time of 29.4 vs. 12.8
bination with intraarterial FUDR (resin
months (Van Hazel et al., 2004). Studies
device, 2002). Currently, the widest experi-
employing the glass microsphere device
ence comprises colorectal liver metastases
also achieved encouraging results with sta-
and hepatocellular carcinoma, and there
ble disease or partial response in more than
is evidence that both devices are suitable
50% of cases (Andrews et al., 1994).
for the treatment of hepatocellular carci-
noma and colorectal metastases (Murthy
et al., 2005). However, radioembolization TECHNIQUE OF
of other tumor entities in the liver (e.g.,
breast cancer, gall bladder cancer, thyroid
TRANSARTERIAL
cancer, non-small cell lung cancer) has TREATMENT
also shown encouraging results concern-
Hepatic Arterial Infusion Chemotherapy
ing local control and may prove beneficial
in the future. While intraarterial application of FUDR is
In hepatocellular carcinoma, a phase II exceptionally performed using surgically
trial has shown improved mean survival implanted infusion pumps, 5-FU can be
in those patients receiving a higher dose administered intermittently via a percutane-
of Y90 glass microspheres (> 104 Gy, 635 ous access. This can be done over a subcu-
days; < 104 Gy, 323 days) (Dancey et al., taneous port reservoir linked to a catheter
2000). Data of another trial, summarizing which is inserted into the arterial system.
80 patients with hepatocellular carcinoma Surgical intraarterial port systems are usually
treated with different stages of disease implanted within the bounds of a laparotomy
and doses of 47–270 Gy showed similar with hepatic resection or resection of the
survival times as reported for transarterial primary. A subcutaneous port reservoir is
chemoembolization. The response rate in placed ventrally at the upper abdomen with
a trial including 71 patients was 27%, in 4 the catheter inserted retrogradely into the
290 T. Denecke et al.
Radioembolization
dissected gastroduodenal artery (GDA).
The perfusion territory is controlled ini- Radioembolization (RE) is a one-time
tially by injection of e.g., methylene blue. procedure and is usually not thought to be
The surgical placement of intraarterial port repeated. The therapeutic agent consists
systems is definite. In case of system dys- of microspheres, 20–40 nm in diameter,
function, e.g., catheter or artery occlusion, which contain a radionuclide emitting
a revision is difficult. Thus, dysfunction of beta-radiation (e.g., Yttrium90) (Gray
surgical ports usually leads to a replace- et al., 2001). The microspheres are deliv-
ment of HAIC by systemic treatment ered into the hepatic artery and embolize
(Heinrich et al., 2003). the microvessels and capillaries of tissues.
Employing novel catheterization tech- Again, the dominance of arterial supply
niques, minimally invasive interventional of hepatic malignancies in contrast to the
implantations of intraarterial port systems normal liver is used for a semiselectivity
became possible (Ricke et al., 2004). For of drug delivery. The typical vascular
this alternative approach, port catheters are architecture of liver metastases leads to a
usually positioned through a femoral artery predominant seeding of the microspheres
access under DSA guidance to position the into the outer regions of the tumors, which
catheter tip in the hepatic artery. The subcu- are hypervascularized and thus hyperoxy-
taneous port reservoir is placed just inferior genated, making radiation in these areas
to the groin. A subclavicular approach is also more effective (Campbell et al., 2001).
practicable. In principle, minimal invasively The transarterial access is similar to that
implanted port systems appear to be superior described for interventional radiological
to surgically implanted port catheters with port placement. Alternatively, the injection
respect to primary and secondary failure and can be performed over an intraarterial port
the possibility to easily repair or replace the system supplying the liver. For the procedure
systems in case of dysfunction (Ricke et al., of drug application, radiation protection
2004; Sadahiro et al., 2004). An additional is mandatory. Therefore, specific devices
advantage is the minimal risk of the proce- for application were developed in order
dure as compared with the mortality rate of to avoid direct and prolonged contact of
surgical port placement (Kerr et al., 2003). the radiologist to the therapeutic agent. To
However, the correct position of intraar- preserve a hepatic reserve, a fractionized
terial catheters regarding hepatic and application selectively to one and sub-
extrahepatic distribution of therapeuti- sequently to the other liver lobe is under
cally applied drugs, unlike in surgically evaluation.
implanted catheters, cannot be confirmed Avoidance of extrahepatic implantation of
by intra-operative application of e.g., microspheres, potentially causing gastritis,
methylene blue, and dislocations of the ulcers, pancreatitis and other gastrointestinal
catheter after initial placement and dur- morbidities, is even more decisive than in
ing the therapy course have to be consid- HAIC (Murthy et al., 2005). Therefore,
ered (Ricke et al., 2004). Therefore, valid a sufficient and careful planning of the
imaging procedures are mandatory to pre- procedure is mandatory. The interven-
serve the advantages of minimal-invasive tion planning includes an arterial selec-
catheter implantation in clinical practice. tive catheterization with application of
22. Perfusion Scintigraphy with Integrated Single Photon Emission 291
low dose X-ray tube enabling both X-ray tracer retrogradely into arteries originating
based attenuation correction for enhanced proximally from the tip of the port line,
image quality of SPECT and anatomic suggesting a catheter dislocation. Just after
mapping on inherently fused SPECT-CT sequential imaging, static planar acquisi-
images with sufficient spatial resolution tion is started. Anterior, right oblique and
to delineate the liver silhouette (Bocher right lateral planar image acquisitions of
et al., 2000). Other comparable devices, the upper abdomen (each up to 800,000
which became available more recently, counts) as well as anterior projections of
consist of a dual head gamma camera and the lower abdomen to rule out leaking
an integrated diagnostic multirow detector and of the thorax to assess shunting to the
CT scanner. lung are obtained. In patients with two
A recommendable scintigraphy proto- port systems, the activity can be divided
col for a low dose SPECT-CT includes by two and is injected into the port sys-
planar scintigraphy covering pelvis, abdo- tems, each injection followed by dynamic
men, and thorax, as well as tomographic and static planar imaging as described
imaging of the relevant organs in the above. Then (∼ 20 min after tracer injec-
upper abdomen with image fusion for ana- tion) a SPECT-CT of the upper abdomen
tomical mapping. The patient is prepared is performed using a dual-head gamma
with a thyroid blockage (e.g., perchlorate, camera (LEHR collimators; FOV, 540 ×
potassium iodide). This is important to 400 mm; matrix, 128 × 128; 120 frames;
avoid active Tc-99m uptake of the gastric 3° steps; 15 s/frame) with an integrated
mucosa to enable the identification of rotating X-ray tube (fixed tube current,
accidental embolization of Tc-99m-MAA 2.5 mA; 140 kV; slice thickness, 10 mm;
into the stomach wall. After testing the matrix, 128 × 128; rotation time, 13.5 s;
port system for patency with a bolus of step and shoot technique) for attenuation
isotonic saline (0.9%), 200 MBq Tc-99m- correction (“Hawkeye”Millenium VG, GE
MAA in 2 ml saline (the syringe must Medical Systems). With this device, total
be shook before injection to optimize acquisition time is ∼ 27 min including the
the distribution) is injected into the port initial transmission scan using the low-
capsule followed by another saline flush dose X-ray tube (9 min for 40 cm scan
(∼ 5–10 ml) under dynamic anterior pla- length) and the subsequent SPECT result-
nar image acquisition of the abdomen (1 ing in a delay of SPECT acquisition from
image/5 s over 1 min) with a low energy tracer injection of ∼30 min. After iterative
high resolution (LEHR) collimator to rule reconstruction (OSEM algorithm; number
out tracer extravasation. The injection has of iterations, 2; number of subsets, 10;
to be performed carefully and slowly. prefiltering, Butterworth (frequency, 0.25;
This is important because of two major order, 10); postfiltering, Butterworth (fre-
problems: (1) the viscosity of MAA needs quency, 0.5; order, 10) and attenuation
a relatively high injection pressure (espe- correction (attenuation map derived from
cially when thin supraselective catheters the low dose CT), the reconstructed data
are used as port lines), which may cause are visualized in sagital, coronal, and axial
rupture of the port system, and (2) a high slices, with the SPECT, the CT and the
injection rate can cause a reflux of the fusion images side by side. Inherent image
22. Perfusion Scintigraphy with Integrated Single Photon Emission 293
alike, are to be mentioned. The cystic artery implantation, the incidence of endoscopi-
arises usually from the right hepatic artery cally confirmed mucosal lesions is 36%,
(71%). This implies the risk of therapy compared to 3% in cases with sufficient
related cholecystitis. If the gall bladder has coiling. Thus, it is important to exclude
not been resected previously, intraarterial insufficient coiling with recanalization of
drug delivery should be performed distally the vessel (4% of cases), subsequent open-
from the origin of the cystic artery. The ing collateral vessels of the right gastric
right gastric artery is a minor contributor artery before application of therapeutic
to the gastric blood supply. Variation of agents proximally to its origin. The GDA
the right gastric artery is common. It may and its branches may be responsible for
arise from the proper hepatic artery (51%), duodenal ulceration or pancreatitis dur-
the left (23%) or the right (3%) hepatic ing regional therapy, when coiled insuffi-
artery, the origin of the GDA (3%) or the ciently or subsequently opening collaterals
common hepatic artery (9%). The iden- were initially not visible during coiling
tification of this vessel is imperative for and MAA test injection (Liu et al., 2005).
regional treatment, as inflammation with If extrahepatic perfusion of non-target ves-
gastric necrosis, mucosal ulceration, and sels is detected before therapy application
perforation are potential complications (Figure 22.1), a reintervention or a change
of inadvertent delivery of RE similarly of the therapeutic management becomes
as in intraarterial infusion chemotherapy. necessary (Denecke et al., 2005).
Here, insufficient coiling of the right gas- Tc-99m-MAA deposition to non-target
tric artery during minimally invasive port vessels can easily be seen when there
Figure 22.1. Port perfusion SPECT-CT of a 64-year-old male who recieved coil embolization of the
right gastric artery and the gastroduodenal artery before transarterial radioembolisation with Y90 micro-
spheres. The red target demonstrates tracer accumulation in the gastric wall. In a second angiography
session, a collateral vessel supplying the gastric wall was identified and occluded
22. Perfusion Scintigraphy with Integrated Single Photon Emission 295
is accumulation in the spleen, indicating are change of the liver silhouette due to
reflux or dislocation of port lines into the surgical treatment (e.g., hemihepatectomy,
celiac trunk. Accumulation in the gastric enucleation) or tumor and treatment related
wall in relevant amounts can be depicted dystrophy. To distinguish sufficient supply
on planar images, if a focal or longitudinal with heterogeneous accumulation (that
tracer retention occurs clearly separated are not affecting therapy efficacy) from
from the liver silhouette. Discrete accumu- true hypoperfusion or perfusion defects
lations in the right gastric wall close to the on scintigraphic images, it is necessary
liver, however, can be obscured by the inten- to have anatomic images for correlation.
sive uptake of the adjacent hepatic paren- Ideally, image fusion helps to compare
chyma. These subtle findings can usually the scintigraphically visualized perfusion
be depicted on axial and coronal SPECT territory and the liver outlines, which is
images. This, as well as accumulation in important in cases with atypical liver con-
the gall bladder or the GDA bed, is best figuration. Additionally, heterogeneous
diagnosed on SPECT-CT fusion images MAA accumulations on SPECT images
(Figure 22.1). The clear outline of the liver can be correlated to the size, shape, and
on the anatomic images help to discriminate distribution of hepatic tumor lesions on
extrahepatic foci from hepatic uptake. This CT or MRI to clearly identify unsup-
especially holds true in the cases with an plied liver segments, especially those with
untypical configuration of the liver or a het- tumor involvement.
erogeneous intrahepatic perfusion pattern. A lack of supply can be caused by sev-
For an effective regional chemotherapy eral problems. A (subtotal) port line occlu-
of liver metastases, the intrahepatic tran- sion is diagnosed easily by unsuccessful
sarterial supply with chemotherapy agents saline injection and requires recanaliza-
should address the entire liver or at least tion. System leakage and disconnection is
reach a sufficient level in those segments discussed above. Occlusion of the arterial
with tumor burden. Heterogeneous per- vessels supplied by the port catheter can
fusion detected on Tc-99m-MAA SPECT also occur. This can be due to intima irri-
can be due to several reasons, such as tation by the catheter or the chemotherapy
hyper- or hypoperfusion of tumors or agent, resulting in dissection and/or throm-
vascular alterations. Heterogeneous int- bosis. Tracer reflux to proximal non-target
rahepatic perfusion patterns that do not vessels is seen in the case of a total occlu-
represent malsupply of tumor baring seg- sion of the hepatic bed. This is sometimes
ments and do not require an alteration of accompanied by a prominent focal tracer
the therapeutic management, can poten- retention in the region of the catheter tip,
tially be misinterpreted on scintigraphic which, however, is not a pathognomonic
images because of the lack of anatomical sign for arterial occlusion. A partial lack
background information. These are for of hepatic supply can be observed in the
example, necrotic areas within hepatic case of occlusion of the right or left liver
tumors and areas previously treated with artery only. Just as perfusion of non-
chemoembolization or interstitial tumor target vessels, these are important findings
ablation (e.g., radiofrequency ablation). and require recanalization efforts before
Other potential causes of misinterpretation resumption of therapy infusion.
296 T. Denecke et al.
Another reason for hypoperfusion or is possible by placing the catheter tip into
perfusion defects within the liver are aber- the proper hepatic artery. In patients with
rant hepatic arteries. There exist numer- aberrant hepatic arteries, however, a tactic
ous variants of the anatomy of hepatic is needed, which allows for perfusion of the
arteries. The relevant variants for minimal entire liver without changing the location
invasive implantation and scintigraphic of agent release. There are different strate-
control of intraarterial liver ports are listed gies in use: (1) catheter implantation into
in the following. An aberrant or accessory the major feeder and neglect of the minor
right hepatic artery occurs in 10–31% hepatic arteries (Figure 22.2); (2) catheter
of cases, with 96% of these originating implantation into the dominant vessel and
from the superior mesenteric artery (Liu proximal coil embolization of the minor
et al., 2005). The caliber of accessory hepatic artery; (3) Catheter implantation
right hepatic arteries varies as well as the into the splenic artery with the therapeutic
supplied volume of liver. An accessory or agent entering the liver over the portal
replaced left hepatic artery is present in vein; or (4) double port implantation into
12–21% of cases. It usually arises from the hepatic arteries of separate origin with a
left gastric artery and thus communicates port reservoir in each groin are not to be
with the gastric and esophageal bed. Other favored as the supply of hepatic malignan-
relevant variants are double hepatic arter- cies is predominantly arterial not portal,
ies with separate origin from the aorta or and a double port system poses a doubled
an accessory middle hepatic artery (Liu risk of dysfunction or complication (Ricke
et al., 2005). et al., 2004).
Regarding the application of therapeutic To ignore a minor accessory vessel results
agents for HAIC as a continuous proce- in a parallel blood flow. Macroaggregated
dure or for RE as a one-time procedure, albumine (MAA) as well as therapeutic
specific considerations have to be made agents injected in one hepatic artery reaches
for both intervention and scintigraphic only this territory, while the neighboring
control. For a one-time procedure, it is territory is fed by the ignored vessel. This
possible to divide the amount of therapeutic is represented by a signal defect on per-
agents into two or more portions and to fusion SPECT (Denecke et al., 2005). To
infuse them over an angiographic cath- interpret this defect correctly, the outline
eter selectively into the supplying arteries of the liver (visualized on SPECT-CT)
of the liver or single liver lobes. This is as well as the arterial anatomy and the
advantageous not only because this tech- desired location of the port line are needed
nique is independent from anatomic vari- (Figure 22.2).
ants with separate hepatic arteries but also To avoid hypoperfusion in tumor bearing
for fractionation of the therapy into two segments which are supplied by accessory
single lobe treatment sessions in order to vessels, a different strategy is needed. It is
preserve a hepatic reserve. In contrast, con- proposed that proximal coil embolization
tinuing infusions of chemotherapy require of one liver artery induces intrahepatic
a single fix location of the port line, flood- collateralization from the remaining feeder
ing the arterial bed of the entire liver. In by lowering the filling pressure in the
patients with regular arterial anatomy, this depending arterial bed. Coil embolization
22. Perfusion Scintigraphy with Integrated Single Photon Emission 297
a b
Figure 22.2. Fifty-one-year-old male with bilobar liver metastases from colorectal cancer: CT-angiography
(Figure 22.2a, volume rendering reconstruction) shows the port line (5) inside the abdominal aorta and the
celiac trunc with its tip (4) in the common hepatic artery just proximal from the origin of the coil embol-
ized gastroduodenal (3) artery. The port system supplies the right hepatic artery (2), while the replaced
left hepatic (1) artery arises separately from the left gastric artery. Consecutively, the port perfusion
SPECT-CT (Figure 22.2c) shows sufficient supply of the right liver lobe and a hypoperfused area in the
left lateral segments. No accumulation of tracer is seen outside the liver, stating the sufficient emboliza-
tion of the gastroduodenal artery. As a consequence, proximal coiling of the left hepatic artery should be
performed to induce intrahepatic collateralization from the right lobe to improve the therapeutic access to
the metastases in the left lateral segments as seen on the CT scan (Figure 22.2b)
studies have shown a relevant impact of correct placement and homogenous and
port perfusion scintigraphy with SPECT exclusive supply of the liver (Lehner et al.,
and SPECT-CT on the therapeutic strategy 1987; Lubin et al., 1987; Morimoto et al.,
in 33% (8/24) of patients. In two of these 1999; Seki et al., 2003; Pelosi et al., 1999;
patients, the relevant finding leading to anSuzuki et al., 1991; Denecke et al., 2005).
alteration of the therapeutic management In transarterial RE with Y90-microspheres,
was only depicted on SPECT-CT fusion intraaterial hepatic perfusion scintigraphy
images. The consequences were repre- is mandatory to evaluate the shunt volume
sented by repositioning of the port lines, to the lung and, as in HAIC, to exclude
recanalization therapy of port lines and/ extrahepatic intraabdominal accumulation.
or vessels or coil embolization of non-target
For perfusion scintigraphy of intraarte-
vessels, not visualized by initial DSA. The rial port systems promising results have
opportunity of repositioning is, compared been reported (Lubin et al., 1987; Pelosi
with the formally common surgical tech- et al., 1999). These evaluations have
nique of port placement, a characteristic focused specifically on the exclusion
advantage of angiographically implanted of extrahepatic perfusion to avoid addi-
ports. This even further strengthens the tional morbidity of regional chemotherapy.
value of port perfusion scintigraphy, since With the employment of SPECT it is
the consequence of inadequate perfusion possible to get additional information
territories is not to stop HAIC and start about the intrahepatic distribution of
systemic chemotherapy instead, but to tracer injected over the intra-arterial port
optimize drug delivery. (Suzuki et al., 1991).
In RE patients, the therapeutic impact Many patients receiving HAIC or RE
of perfusion scintigraphy is 100% as the have unusual anatomic configurations of
modality is mandatory as a planning tool the liver (e.g., owing to surgical or locally
in every single patient. This, however, is to
ablative therapy of hepatic metastases,
assess the hepatopulmonary shunt volume excessive tumor burden, dysmorphic liver).
than to visualize the hepatic perfusion pat-Hence, an estimation of hepatic supply
tern. With SPECT-CT as a highly accurate with chemotherapy agents over intraarterial
tool, obviously superior to SPECT alone port systems is difficult, if planar images
(Denecke et al., 2005), for identifica- or SPECT is used only, particularly in
tion of extrahepatic tracer accumulations patients with inhomogeneous hepatic tracer
and intrahepatic malperfusions, it is ide- accumulation. In these cases, the reviewer
ally suited to improve safety and efficacy extrapolates the imaging findings to a nor-
of liver directed administration of Y90- mal liver silhouette without morphologic
microspheres. correlation. However, this interpretation
may be unreliable. In this context, CT with
intraport injection of contrast agent has
DISCUSSION been reported to be superior compared to
scintigraphy including SPECT (Morimoto
Several imaging modalities have been et al., 1999). In assessing the intrahepatic
employed for control of intra-arterial port perfusion pattern, slow infusion MRA is
systems for HAIC with respect to patency, considered even superior to infusion CT
22. Perfusion Scintigraphy with Integrated Single Photon Emission 299
on the tissues hardly be stopped. For safety Cantore, M., Mambrini, A., Fiorentini, G., Rabbi, C.,
reasons, the best possible modality for Zamagni, D., Caudana, R., Pennucci, C.,
Sanguinetti, F., Lombardi, M., and Nicoli, N.
exclusion of inadvertent gastrointestinal
2005. Phase II study of hepatic intraarterial
implantation of diagnostic and therapeutic epirubicin and cisplatin, with systemic 5-fluor-
agents should be used, which is currently ouracil in patients with unresectable biliary tract
perfusion scintigraphy and Bremsstrahlung tumors. Cancer 103: 1402–1407.
scintigraphy with SPECT-CT. Cohen, A.D., and Kemeny, N.E. 2003. An update
In conclusion, SPECT-CT improves on hepatic arterial infusion chemotherapy for
colorectal cancer. Oncologist 8: 553–566.
identification and localization of areas
Dancey, J.E., Shepherd, F.A., Paul, K., Sniderman,
with reduced chemotherapy agent supply K.W., Houle, S., Gabrys, J., Hendler, A.L.,
and eventual extrahepatic perfusion in and Goin, J.E. 2000. Treatment of nonresect-
patients receiving HAIC. This has a rele- able hepatocellular carcinoma with intra-
vant impact on the therapeutic strategy and hepatic 90Y-microspheres. J. Nucl. Med. 41:
enables treatment optimization. In patients 1673–1681.
Denecke, T., Hildebrandt, B., Lehmkuhl, L., Peters, N.,
scheduled for RE with Y90-microspheres,
Nicolaou, A., Pech, M., Riess, H., Ricke, J.,
SPECT-CT is a very valuable tool to detect Felix, R., and Amthauer, H. 2005. Fusion imag-
or exclude inadvertent deposition of tracer ing using a hybrid SPECT-CT camera improves
into gastrointestinal non-target vessels, port perfusion scintigraphy for control of hepatic
and thus helps to increase the safety of this arterial infusion of chemotherapy in colorectal
procedure. cancer patients. Eur. J. Nucl. Med. Mol. Imaging
32(9): 1003–1010.
Gray, B., Van Hazel, G., Hope, M., Burton, M.,
REFERENCES
Moroz, P., Anderson, J., and Gebski, V. 2001.
Amthauer, H., Denecke, T., Rohlfing, T., Ruf, J., Randomised trial of SIR-Spheres plus chemo-
Bohmig, M., Gutberlet, M., Plockinger, U., therapy vs. chemotherapy alone for treating
Felix, R., and Lemke, A.J. 2005. Value of image patients with liver metastases from primary large
fusion using single photon emission computed bowel cancer. Ann. Oncol. 12: 1711–1720.
tomography with integrated low dose computed Heinrich, S., Petrowsky, H., Schwinnen, I., Staib-
tomography in comparison with a retrospective Sebler, E., Gog, C., El-Ganainy, A., Gutt, C.,
voxel-based method in neuroendocrine tumours. Muller, H.H., and Lorenz, M. 2003. Technical
Eur. Radiol. 15: 1456–1462. complications of continuous intra-arterial chem-
Andrews, J.C., Walker, S.C., Ackermann, R.J., otherapy with 5-fluorodeoxyuridine and 5-fluor-
Cotton, L.A., Ensminger, W.D., and Shapiro, B. ouracil for colorectal liver metastases. Surgery
1994. Hepatic radioembolization with yttrium-90 133: 40–48.
containing glass microspheres: preliminary Kemeny, N., Huang, Y., Cohen, A.M., Shi, W.,
results and clinical follow-up. J. Nucl. Med. 35: Conti, J.A., Brennan, M.F., Bertino, J.R.,
1637–1644. Turnbull, A.D., Sullivan, D., Stockman, J.,
Bocher, M., Balan, A., Krausz, Y., Shrem, Y., Blumgart, L.H., and Fong, Y. 1999. Hepatic
Lonn, A., Wilk, M., and Chisin, R. 2000. arterial infusion of chemotherapy after resection
Gamma camera-mounted anatomical X-ray of hepatic metastases from colorectal cancer. N.
tomography: technology, system characteris- Engl. J. Med. 341: 2039–2048.
tics and first images. Eur. J. Nucl. Med. 27: Kerr, D.J., McArdle, C.S., Ledermann, J., Taylor, I.,
619–627. Sherlock, D.J., Schlag, P.M., Buckels, J., Mayer,
Campbell, A.M., Bailey, I.H., and Burton, M.A. D., Cain, D., and Stephens, R.J. 2003. Medical
2001. Tumour dosimetry in human liver fol- Research Council’s colorectal cancer study group;
lowing hepatic yttrium-90 microsphere therapy. European Organisation for Research and Treatment
Phys. Med. Biol. 46: 487–498. of Cancer colorectal cancer study group.
22. Perfusion Scintigraphy with Integrated Single Photon Emission 301
Intrahepatic arterial versus intravenous fluor- 1999. Hepatic arterial infusion chemotherapy
ouracil and folinic acid for colorectal cancer for unresectable confined liver metastases: pre-
liver metastases: a multicentre randomised trial. diction of systemic toxicity with the applica-
Lancet 361: 368–373. tion of a scintigraphic and pharmacokinetic
Lehner, K., Reiser, M., Gebhardt, U., Heuck, A., approach. Cancer Chemother. Pharmacol. 44:
and Schaff, J. 1987. DSA – control of implanted 505–510.
devices for arterial hepatic perfusion. Cardiovasc. Ricke, J., Hildebrandt, B., Miersch, A., Nicolaou, A.,
Intervent. Radiol. 10: 71–74. Warschewske, G., Teichgraber, U., Lopez
Liu, D.M., Salem, R., Bui, J.T., Courtney, A., Hanninen, E., Riess, H., and Felix, R. 2004.
Barakat, O., Sergie, Z., Atassi, B., Barrett, K., Hepatic arterial port systems for treatment of
Gowland, P., Oman, B., Lewandowski, R.J., liver metastases: factors affecting patency and
Gates, V.L., Thurston, K.G., and Wong, C.Y. adverse events. J. Vasc. Interv. Radiol. 15:
2005. Angiographic considerations in patients 825–833.
undergoing liver-directed therapy. J. Vasc. Interv. Sadahiro, S., Suzuki, T., Ishikawa, K., Yasuda, S.,
Radiol. 16: 911–935. Tajima, T., Makuuchi, H., Saitoh, T., and
Lorenz, M., and Muller, H.H. 2000. Randomized, Murayama, C. 2004. Prophylactic hepatic arte-
multicenter trial of fluorouracil plus leucov- rial infusion chemotherapy for the prevention
orin administered either via hepatic arterial or of liver metastasis in patients with colon carci-
intravenous infusion versus fluorodeoxyurid- noma: a randomized control trial. Cancer 100:
ine administered via hepatic arterial infusion 590–597.
in patients with nonresectable liver metastases Seki, H., Ozaki, T., Takaki, S., Ooi, H., Oda, J.,
from colorectal carcinoma. J. Clin. Oncol. 18: and Shiina, M. 2003. Using slow-infusion MR
243–254. arteriography and an implantable port system
Lubin, E., Cyjon, A., and Neuman, M. 1987. to assess drug distribution at hepatic arterial
Technetium-99m macroaggregates for the study infusion chemotherapy. AJR Am. J. Roentgenol.
of the distribution of arterial infusion chemother- 180: 681–686.
apy in the liver. Clin. Nucl. Med. 12: 385–388. Suzuki, Y., Kobayashi, S., and Yasuda, S. 1991.
Meta-Analysis Group in Cancer (MAGC). 1996. Application of single-photon emission com-
Reappraisal of hepatic arterial infusion in the puted tomography (SPECT) with 99mTc-MAA
treatment of nonresectable liver metastases from in evaluation of perfusion patterns during hepatic
colorectal cancer. J. Natl. Cancer Inst. 88: infusion chemotherapy. Ann. Nucl. Med. 5:
252–258. 123–126.
Morimoto, M., Satake, M., Sekiguchi, R., Haruno, M., Tanaka, S., Shimada, M., Shirabe, K., Maehara, S.,
and Moriyama, N. 1999. Harimoto, N., Tsujita, E., Sugimachi, K.,
Optimal injection protocol for CT evaluation during and Maehara, Y. 2005. A novel intrahepatic
hepatic arterial infusion chemotherapy. Invest. arterial chemotherapy after radical resec-
Radiol. 34: 744–750. tion for advanced hepatocellular carcinoma.
Murthy, R., Nunez, R., Szklaruk, J., Erwin, W., Hepatogastroenterology 52: 862–865.
Madoff, D.C., Gupta, S., Ahrar, K., Wallace, Van Hazel, G., Blackwell, A., Anderson, J., Price, D.,
M.J., Cohen, A., Coldwell, D.M., Kennedy, A.S., Moroz, P., Bower, G., Cardaci, G., and Gray, B.
and Hicks, M.E. 2005. Yttrium-90 microsphere 2004 Randomised phase 2 trial of SIR-Spheres
therapy for hepatic malignancy: devices, indica- plus fluorouracil/leucovorin chemotherapy ver-
tions, technical considerations, and potential sus fluorouracil/leucovorin chemotherapy alone
complications. Radiographics 25(1): S41–S55. in advanced colorectal cancer. J. Surg. Oncol.
Nelson, R.L., and Freels, S. 2004. A systematic 88: 78–85.
review of hepatic artery chemotherapy after Ziessman, H.A., Gyves, J.W., Juni, J.E., Wahl,
hepatic resection of colorectal cancer metastatic F.L., Thrall, J.H., Ensminger, W.D., Goldstein,
to the liver. Dis. Colon Rectum. 47: 739–745. H.A., and Dubiansky, V. 1985. Atlas of hepatic
Pelosi, E., Bar, F., Battista, S., Bello, M., Bucchi, arterial perfusion scintigraphy. Clin. Nucl. Med.
M.C., Alabiso, O., Molino, G., and Bisi, G. 10: 675–681.
23
Postoperative Interferon Alpha
Treatment of Patients with Hepatocellular
Carcinoma: Expression of p48 Using
Tissue Microarray
Hui-Chuan Sun
303
304 H.-C. Sun
On the other hand, IFN-α therapy pro- surgical team at the Liver Cancer Institute
duced some side effects in patients who (Fudan University, Shanghai, China) were
did not respond therapeutically to IFN-α. enrolled to test the effect of postoperative
Therefore, a predictive marker of response IFN-α therapy on the prevention of recur-
to IFN-α is needed to avoid unnecessary rence. All patients were randomized into
treatment costs and side effects. either the IFN-α therapy group (5 MU three
A wealth of information is available times a week for 18 months or until recur-
on the molecular processes underlying rence was diagnosed) or the control group
IFN-α-induced signaling. Binding of (no antitumor treatment, until recurrence
IFN-α to the type I receptor brings the was diagnosed). The details of this trial
receptor-associated tyrosine kinases Tyk2 have been reported elsewhere (Sun et al.,
and Jak1 in close proximity, which allows 2006).
the kinases to phosphorylate, leading to In this study, 80 samples from the IFN-α
their activation and the generation of recep- group (group 1) and 75 samples from the
tor docking sites for the latent cytoplasmic control group (group 2) were retrieved from
protein STAT (Wu et al., 2004). Tyrosine- the Department of Pathology, Zhongshan
phosphorylated STAT1 and STAT2 also Hospital. All samples were from patients
associate as heterodimers to form IFN- who followed the treatment protocol.
α-stimulated gene factor 3 (ISGF3α) and
translocate to the nucleus, where they asso- Tissue Microarray and
ciate with the DNA-binding protein ISGF3γ Immunohistochemistry
(P48) to form ISGF3, which then binds After careful screening of hematoxylin-
to the IFN-stimulated response element eosin-stained slides for optimal tumor con-
to induce transcription of several IFN- tent, we constructed tissue microarray slides
inducible genes (Ghislain et al., 2001). Our in collaboration with Shanghai Biochip
previous study showed that transduction Co. Briefly, two 0.8-mm-diameter punch
of P48 in resistant HCC cells can restore cores were taken from each formalin-fixed,
sensitivity to IFN-α, a finding that further paraffin-embedded HCC sample. Samples
confirms the importance of Jak/STAT sig- were obtained from nonnecrotic tumor foci.
naling in the antiproliferative effects of Immunohistochemistry for P48 was per-
IFN-α in HCC cells (Wu et al., 2005). We formed using the avidin-biotin complex
believe that P48 expression in tumor cells (Vector Laboratories, Burlingame, CA),
is related to the effect of IFN-α. Therefore, including heat-induced antigen retrieval
we hypothesized that P48 expression would procedures. Incubation with polyclonal anti-
predict outcome in the patients receiving bodies against P48 (P48:C20, 1:100; Santa
IFN-α therapy who were enrolled in the Cruz Biotechnology, Santa Cruz, CA) was
randomized controlled trial presented here. carried out at 4°C for 18 h (Clifford et al.,
2000). Negative controls were treated identi-
cally but with the primary antibody omitted.
MATERIALS AND METHODS
In this randomized controlled trial, 232 Scoring of P48 Immunohistochemistry
patients who underwent curative resection Three pathologists who were blinded to
of HCC (pathologically proven) by the same patient data independently evaluated
23. Postoperative Interferon Alpha Treatment of Patients with Hepatocellular Carcinoma 305
a b c
Figure 23.1. Immunohistochemistry of P48. a: Overview of tissue dot with positive staining of P48 in
hepatocellular carcinoma (avidin-biotin complex [ABC], 40×). b: Overview of tissue dot with negative
staining of 48 in hepatocellular carcinoma (40×). c: Positive staining of P48 in cytoplasm in hepatocellular
carcinoma (ABC, 400×)
306 H.-C. Sun
P48 was stained mainly in the cytoplasm group 2 (hazard ratio, 3.9; 95% CI, 1.7–9.2;
of HCC cells but occasionally in the P = 0.001). The mean overall survival was
nucleus of HCC cells and in the bile duct 57.3 months (95% CI, 53.0–61.5 months)
(Figure 23.1C). There was no statistically in group 1 and 41.5 months (95% CI,
significant difference in clinicopatholog- 34.0–49.0 months) in group 2, but the differ-
ical data, except gender, between patients ence was not statistically significant (hazard
with positive and those with negative P48 ratio, 1.5; 95% CI, 0.8–2.9; P = 0.246).
expression. There was no statistically
significant difference from patients who
received IFN-α (group 1) and those who Prognostic Factors for Disease-Free
did not (group 2) between patients with Survival and Overall Survival
positive and those with negative P48 in Group 1
expression. Univariate analysis of prognostic factors for
DFS revealed that of all evaluated factors,
only liver cirrhosis (P = 0.000) and P48
Survival
expression (P = 0.011) had a significant prog-
In group 1 (IFN-α therapy group), DFS nostic influence. After multivariate analysis,
was significantly better in patients with only liver cirrhosis (present vs. absent: risk
positive P48 expression compared with ratio [RR] 4.9; 95% CI, 1.4–16.9; P = 0.012)
those with negative P48 expression (haz- and pattern of P48 staining (negative vs. pos-
ard ratio, 1.8; 95% CI, 1.0–3.2; P = 0.036). itive: RR, 2.0; 95% CI, 1.1–3.8; P = 0.028)
The cumulative 1-, 3-, and 5-year DFS remained significant prognostic factors for
rates were 84%, 55%, and 36% among DFS . Receiver operating characteristic anal-
patients with positive P48 expression and ysis showed that these two prognostic factors
53%, 34%, and 29% among those with have good predictive accuracy with regard
negative P48 expression, respectively. In to 1-year DFS (area under the curve [AUC]
the control group, however, there was no = 0.649; 95% CI, 0.5–0.8; P < 0.01) and
significant difference in DFS between acceptable accuracy in terms of 2-year DFS
patients with positive and those with nega- (AUC = 0.604; 95% CI, 0.5–0.7; P = 0.05).
tive staining (hazard ratio, 1.4; 95% CI, Univariate analysis of prognostic fac-
0.8–2.6; P = 0.208). tors for overall survival revealed that of
In group 1, overall survival was signifi- all evaluated factors, liver cirrhosis (P =
cantly better in patients with positive P48 0.032), HBeAg positivity (P = 0.009),
expression than in those with negative P48 and P48 expression (P = 0.014) had a sig-
expression (hazard ratio, 2.7; 95% CI, nificant prognostic influence. After multi-
1.2–6.2; P = 0.014). Again, in group 2, variate analysis, only Edmondson’s stage
there was no significant difference in over- (high grade [III/IV] vs. low grade [I/II]:
all survival between patients with positive RR, 5.6; 95% CI, 2.0–15.7; P = 0.001),
and those with negative staining (hazard and pattern of P48 staining (negative vs.
ratio, 1.2; 95% CI, 0.6–2.3; P = 0.545). positive: RR, 3.8; 95% CI, 1.4–10.3; P =
When we analyzed data from patients 0.008) were significant prognostic factors
with positive P48 expression, we found that for overall survival. ROC analysis showed
overall survival in group 1 was better than in that these two prognostic factors have a
23. Postoperative Interferon Alpha Treatment of Patients with Hepatocellular Carcinoma 307
impressive predictive accuracy for overall also show that patients with poorly dif-
survival in the first 4 years (AUCs for 1-, ferentiated HCC and/or liver cirrhosis may
2-, 3-, and 4-year survival were 0.883 [P < have a poorer outcome. We noticed that the
0.01], 0.737 [P < 0.01], 0.646 [P = 0.01], predictive accuracy of P48 expression in
and 0.634 [P = 0.03], respectively). terms of DFS and overall survival dimin-
ished over time. The reason is that IFN-α
was administered for 18 months to patients
DISCUSSION in the treatment group, and P48 expression
was used to predict the effect of IFN-α
A gene expression profile and IFN-α/type therapy, so the predictive accuracy of P48
2 interferon receptor (IFNAR2) have been expression was good in the first several
reported to predict response of HCC to the years and diminished after IFN-α therapy
combination of IFN-α and 5-fluorouracil was stopped. As we found in our clinical
(Kurokawa et al., 2004; Ota et al., 2005). trial (Sun et al., 2006), 18 months of IFN-α
In melanoma and transitional cell carci- therapy delayed tumor recurrence but did
noma, a decrease in P48 expression was not reduce the incidence of recurrence over
associated with poor response to IFN-α 5 years; a number of patients developed
(Matin et al., 2001). Other molecular tumor recurrence after IFN-α therapy was
markers, such as SOCS-1 and SOCS-3, stopped. However, overall survival within
acted as negative feedback regulators of the 5-year period was improved by IFN-α
Jak/STAT and determined sensitivity to therapy, because of delayed recurrence in
IFN (Takayama et al., 2000). However, the treatment group.
until now there have been no reports on a Tissue microarray analysis was first
predictive marker for outcome after post- introduced by Kononen et al. in 1998. It
operative IFN-α therapy in patients with is a high-throughput approach for histo-
HBV-related HCC. logic examination of tissue for the pur-
In the present tissue microarray analy- pose of determining protein expression.
sis, we found no significant differences However, high-throughput platforms are
in DFS and in overall survival between compromised by a loss of completeness of
control subjects (patients not treated with data. Therefore, several important issues
postoperative IFN-α) with positive and need to be considered in conducting tissue
those with negative P48 expression; how- microarrays.
ever, significant differences were found First, adequate sample size is important.
between treated patients with positive Tumor heterogeneity is of major concern in
and those with negative P48 expression. interpreting results from tissue microarrays
Furthermore, overall survival in patients (Rubin et al., 2002). In well-established
with positive P48 expression who were methods of immunohistochemistry, about
treated with IFN-α was better than over- 0.05% of tissue in a routine section rep-
all survival in patients with positive P48 resents 1 cc of tumor tissue (Camp et al.,
expression who were not treated with IFN- 2000), whereas a tissue microarray section
α. These data clearly demonstrate that P48 represents a greatly reduced amount of tis-
expression is associated with outcome after sue. Obtaining more tissue microarray cores
postoperative IFN-α therapy. Our results should yield a more accurate result with
308 H.-C. Sun
less variability, but this is not always pos- Tissue microarray analysis is the best tool
sible, especially when the archived block for evaluating the clinical significance
has been repeatedly used or when a fine- of a specific gene expression in a larger
needle biopsy sample is used. Furthermore, number of specimens with different sub-
obtaining too many cores reduces the value types of tumor.
of tissue microarray analysis. Rubin et It is also notable that approximately
al. (2002) analyzed the results of a tis- 16% of patients with positive P48 expres-
sue microarray test for Ki67 expression sion received a diagnosis of recurrence
in prostate cancer and found that three within 1 year. One explanation for this
0.8-mm-diameter cores are necessary and may be that expression of P48 is not suf-
enough to obtain accurate results, but other ficient for antiproliferation in Jak/STAT
researchers have reported that one or two signaling. A previous study found that lack
cores are enough (Camp et al., 2000; Lugli of STAT1 expression also led to resistance
et al., 2004). Currently, there is no hard- to growth inhibitory effects of IFN-α in
and-fast guideline for determining optimal a liver cancer cell line (Murphy et al.,
sample size (Hewitt, 2006). In this study, 2001). In addition to antiproliferation,
two 0.8-mm-diameter cores were taken IFN-α has other antitumor mechanisms,
from each specimen. such as immunomodulation and antian-
Second, nonrepresentative samples should giogenesis (Jonasch and Haluska, 2001).
be avoided. The location of sampling sites Our previous study showed that the antian-
may play an important role. In this study, the giogenesis mechanism involved another
location of each core was decided in col- signal transduction pathway, despite P48
laboration with an experienced pathologist. deficiency in the HCC cell line (Wu et
The principle is to take cores from the non- al., 2005). Meanwhile, the natural limita-
necrotic portion of the sample and to collect tions of tissue microarray analysis, such as
as many intact tumor cells as possible. tumor heterogeneity, nonuniform tissue-
The third challenge is to establish a fixation processes, or even interpretation
scoring system for evaluation of P48 of immunoactivity, may also compromise
expression. In the present study, three predictive accuracy.
pathologists reviewed every slide and As pointed out by Kononen et al. (1998),
noted a mean of 348, 338, and 361 positive the tissue microarray test can be used with
tumor cells per patient. Standard devia- a large number of samples to profile more
tions were 355, 348 and 359, and standard general trends related to the tissue type
deviation did not increase as the corre- being studied; the test is not geared toward
sponding mean increased. determining the exact protein expression
An implication of this study is that tissue of an individual case. Tissue microarrays
microarray analysis could be used to deter- are designed to analyze a large number
mine the significance of gene expression of samples from archived paraffin blocks
in a larger sample size. Although we found but are not suitable for routine clinical
P48 expression to be associated with sen- analysis. For the routine clinical access of
sitivity of liver cancer cells to IFN-α, we P48 expression as a screening tool, a pro-
are not sure about the significance of P48 spective study with a standard section and
expression in patients who receive IFN-α. staining protocol is still needed.
23. Postoperative Interferon Alpha Treatment of Patients with Hepatocellular Carcinoma 309
In conclusion, the results of this study put molecular profiling of tumor specimens. Nat.
indicate that P48 assessed immunohisto- Med. 4: 844–847.
Kubo, S., Nishiguchi, S., Hirohashi, K., Tanaka, H.,
chemically from surgical samples may be
Shuto, T., Yamazaki, O., Shiomi, S., Tamori, A.,
a useful marker for predicting patient out- Oka, H., Igawa, S., Kuroki, T., and Kinoshita, H.
come after postoperative IFN-α therapy. 2001. Effects of long-term postoperative inter-
Postoperative IFN-α therapy should be feron-alpha therapy on intrahepatic recurrence
administered on the basis of tumor P48 after resection of hepatitis C virus-related hepa-
expression status, the presence of cirrhosis, tocellular carcinoma. A randomized, controlled
trial. Ann. Intern. Med. 134: 963–967.
and Edmondson’s grade, to provide maxi-
Kurokawa, Y., Matoba, R., Nagano, H., Sakon, M.,
mal benefit and avoid IFN-α toxicities. Takemasa, I., Nakamori, S., Dono, K., Umeshita, K.,
Ueno, N., Ishii, S., Kato, K., and Monden, M. 2004.
REFERENCES Molecular prediction of response to 5-fluorouracil
and interferon-alpha combination chemotherapy in
Camp, R.L., Charette, L.A., and Rimm, D.L. 2000. advanced hepatocellular carcinoma. Clin. Cancer
Validation of tissue microarray technology in Res. 10: 6029–6038.
breast carcinoma. Lab. Invest. 80: 1943–1949. Lin, S.M., Lin, C.J., Hsu, C.W., Tai, DI., Sheen,
Clifford, J.L., Menter, D.G., Yang, X., Walch, E., I.S., Lin, D.Y., and Liaw, Y.F. 2004. Prospective
Zou, C., Clayman, G.L., Schaefer, T.S., El-Naggar, randomized controlled study of interferon-alpha
A.K., Lotan, R., and Lippman, S.M. 2000. in preventing hepatocellular carcinoma recur-
Expression of protein mediators of type I inter- rence after medical ablation therapy for primary
feron signaling in human squamous cell carci- tumors. Cancer 100: 376–382.
noma of the skin. Cancer Epidemiol. Biomarkers Lin, S.M., Yu, M.L., Lee, C.M., Chien, R.N., Sheen,
Prev. 9: 993–997. I.S., Chu, C.M., and Liaw, Y.F. 2007. Interferon
Ghislain, J.J., Wong, T., Nguyen, M., and Fish, therapy in HBeAg positive chronic hepatitis
E.N. 2001. The interferon-inducible Stat2:Stat1 reduces progression to cirrhosis and hepatocel-
heterodimer preferentially binds in vitro to a con- lular carcinoma. J. Hepatol. 46: 45–52.
sensus element found in the promoters of a sub- Lugli, A., Tornillo, L., Mirlacher, M., Bundi, M.,
set of interferon-stimulated genes. J. Interferon Sauter, G., and Terracciano, L.M. 2004.
Cytokine Res. 21: 379–388. Hepatocyte paraffin 1 expression in human
Hewitt, S.M. 2006. The application of tissue normal and neoplastic tissues: tissue microarray
microarrays in the validation of microarray analysis on 3,940 tissue samples. Am. J. Clin.
results. Meth. Enzymol. 410: 400–415. Pathol. 122: 721–727.
Ikeda, K., Arase, Y., Saitoh, S., Kobayashi, M., Matin, S.F., Rackley, R.R., Sadhukhan, P.C., Kim,
Suzuki, Y., Suzuki, F., Tsubota, A., Chayama, K., M.S., Novick, A.C., and Bandyopadhyay, S.K.
Murashima, N., and Kumada, H. 2000. Interferon 2001. Impaired alpha-interferon signaling in
beta prevents recurrence of hepatocellular car- transitional cell carcinoma: lack of p48 expres-
cinoma after complete resection or ablation of sion in 5637 cells. Cancer Res. 61: 2261–2266.
the primary tumor – a prospective randomized Murphy, D., Detjen, K.M., Welzel, M., Wiedenmann, B.,
study of hepatitis C virus-related liver cancer. and Rosewicz, S. 2001. Interferon-alpha delays
Hepatology 32: 228–232. S-phase progression in human hepatocellular
Jonasch, E., and Haluska, F.G. 2001. Interferon carcinoma cells via inhibition of specific cyclin-
in oncological practice: review of interferon dependent kinases. Hepatology 33: 346–356.
biology, clinical applications, and toxicities. Ota, H., Nagano, H., Sakon, M., Eguchi, H, Kondo, M.,
Oncologist 6: 34–55. Yamamoto, T., Nakamura, M., Damdinsuren, B.,
Kononen, J., Bubendorf, L., Kallioniemi, A., Wada, H., Marubashi, S., Miyamoto, A., Dono, K.,
Bärlund, M., Schraml, P., Leighton, S., Torhorst, J., Umeshita, K., Nakamori, S., Wakasa, K., and
Mihatsch, M.J., Sauter, G., and Kallioniemi, Monden, M. 2005. Treatment of hepatocellular
O.P. 1998. Tissue microarrays for high-through- carcinoma with major portal vein thrombosis by
310 H.-C. Sun
combined therapy with subcutaneous interferon- Takayama, T., Sekine, T., Makuuchi, M., Yamasaki,
alpha and intra-arterial 5-fluorouracil; role of S., Kosuge, T., Yamamoto, J., Shimada, K.,
type 1 interferon receptor expression. Br. J. Sakamoto, M., Hirohashi, S., Ohashi, Y., and
Cancer 93: 557–564. Kakizoe, T. 2000. Adoptive immunotherapy to
Parkin, D.M., Bray, F., Ferlay, J., and Pisani, P. lower postsurgical recurrence rates of hepatocel-
2005. Global cancer statistics. 2002. CA Cancer lular carcinoma: a randomised trial. Lancet 356:
J. Clin. 55: 74–108. 802–807.
Poon, R.T. 2007. Optimal initial treatment for Wang, L., Tang, Z.Y., Qin, L.X., Wu, X.F., Sun,
early hepatocellular carcinoma in patients H.C., Xue, Q., and Ye, S.L. 2000. High-dose and
with preserved liver function: transplantation or long-term therapy with interferon-alfa inhibits
resection? Ann. Surg. Oncol. 14: 541–547. tumor growth and recurrence in nude mice bear-
Regimbeau, J.M., Abdalla, E.K., Vauthey, J.N., ing human hepatocellular carcinoma xenografts
Lauwers, G.Y., Durand, F., Nagorney, D.M., Ikai, I., with high metastatic potential. Hepatology 32:
Yamaoka, Y., and Belghiti, J. 2004. Risk factors 43–48.
for early death due to recurrence after liver resec- Wang, L., Wu, W.Z., Sun, H.C., Wu, X.F., Qin,
tion for hepatocellular carcinoma: results of a L.X., Liu, Y.K., Liu, K.D., and Tang, Z.Y. 2003.
multicenter study. J. Surg. Oncol. 85: 36–41. Mechanism of interferon alpha on inhibition of
Rubin, M.A., Dunn, R., Strawderman, M., and metastasis and angiogenesis of hepatocellular
Pienta, K.J. 2002. Tissue microarray sampling carcinoma after curative resection in nude mice.
strategy for prostate cancer biomarker analysis. J. Gastrointest. Surg. 7: 587–594.
Am. J. Surg. Pathol. 26: 312–319. Wu, W.Z., Sun, H.C., Gao, Y.Q., Li, Y., Wang, L.,
Sun, H.C., and Tang, Z.Y. 2003. Preventive treat- Zhou, K., Liu, K.D., Iliakis, G., and Tang, Z.Y.
ments for recurrence after curative resection of 2004. Reduction in p48-ISGFgamma levels con-
hepatocellular carcinoma – a literature review of fers resistance to interferon-alpha2a in MHCC97
randomized control trials. World J. Gastroenterol. cells. Oncology 67: 428–440.
9: 635–640. Wu, W.Z., Sun, H.C., Shen, Y.F., Chen, J., Wang, L.,
Sun, H.C., Tang, Z.Y., Ma, Z.C., Qin, L.X., Wang, L., Tang, Z.Y., Iliakis, G., and Liu, K.D. 2005.
Ye, Q.H., Fan, J., Wu, Z.Q., and Zhou, X.D. Interferon alpha 2a down-regulates VEGF
2005. The prognostic factor for outcome follow- expression through PI3 kinase and MAP kinase
ing second resection for intrahepatic recurrence signaling pathways. J. Cancer Res. Clin. Oncol.
of hepatocellular carcinoma with a hepatitis B 131: 169–178.
virus infection background. J. Cancer Res. Clin. Yan, P., and Yan, L.N. 2003. Staging of hepatocel-
Oncol. 131: 284–288. lular carcinoma. Hepatobiliary Pancreat. Dis.
Sun, H.C., Tang, Z.Y., Wang, L., Qin, L.X., Ma, Int. 2: 491–495.
Z.C., Ye, Q.H., Zhang, B.H., Qian, Y.B., Wu, Yu, M.L., Lin, S.M., Chuang, W.L., Dai, C.Y.,
Z.Q., Fan, J., Zhou, X.D., Zhou, J., Qiu, S.J., and Wang, J.H., Lu, S.N., Sheen, I.S., Chang, W.Y.,
Shen, Y.F. 2006. Postoperative interferon alpha Lee, C.M., and Liaw, Y.F. 2006. A sustained
treatment postponed recurrence and improved virological response to interferon or interferon/
overall survival in patients after curative resec- ribavirin reduces hepatocellular carcinoma and
tion of HBV-related hepatocellular carcinoma: improves survival in chronic hepatitis C: a
a randomized clinical trial. J. Cancer Res. Clin. nationwide, multicentre study in Taiwan. Antivir.
Oncol. 132: 458–465. Ther. 11: 985–994.
C. Prognosis
24
Hepatocellular Carcinoma:
Overexpression of Homeoprotein
Six1 as a Marker for Predicting Survival
Kevin Tak-Pan Ng and Kwan Man
313
314 K. Tak-Pan Ng and K. Man
according to the manufacturer’s instruc- TBS/T for 10 min each and incubated
tions (Invitrogen, Carlsbad, CA). Each with secondary antibody at room tempera-
complementary DNA was synthesized ture for 1 h. Protein signal was detected by
from 1 μg of RNA extract using the High ECL Plus system (Amersham Biosciences,
capacity cDNA Kit (Applied Biosystems, Piscataway, NJ). Antibodies Six1 and
Foster City, CA) under the condition of β-Actin were purchased from Santa Cruz
25°C for 5 min followed by 37°C for 2 h. Biotechnology (Santa Cruz, CA).
For clinical samples, PCR reaction for
Six1 was performed using the Taq PCR Statistical Analysis
Kit (Promega, Madison, WI) under the
following PCR cycles: 95°C for 5 min, Statistical analysis was carried out using
40 cycles at 95°C for 1 min, 57°C for SPSS 11.5 for Windows (SPSS Inc.,
1 min and 72°C for 1 min. For internal Chicago, IL). The association of Six1
control 18S, 30 PCR cycles were used. protein expression and clinicopathological
For cell line samples, 30-cycle multiplex parameters including sex, age, pTNM
PCR reaction was performed by combin- staging, venous infiltration, encapsulation,
ing Six1 and 18S primers. PCR products tumor size, alpha-fetoprotein level, and
were visualized by 2% agarose gel electro- hepatitis B surface antigen was analyzed
phoresis stained with ethidium bromide. by Chi-square test. The prognostic value
Primer sets used were as follows: for Six1, of Six1 protein for predicting the overall
sense 5′ AAG GAG AAG TCG AGG survival of HCC patients after hepatic
GGT GT-3′, antisense 5′-TGC TTG TTG resection was calculated by Kaplan-Meier
GAG GAG GAG TT-3′; for 18S ribosomal analysis with the log-rank test. Cox pro-
RNA, sense 5′-CTC TTA GCT GAG TGT portional hazard regression model was
CCC GC-3′, antisense 5′-CTG ATC GTC performed with univariable and multi-
TTC GAA CCT CC-3′. variable analyses to test factors that were
significantly associated with the overall
Western Blot survival of the HCC patients. Logistic
regression analysis was also performed
Proteins from clinical specimens and HCC
to compare those factors for predicting
cell lines were prepared by using Urea
1- and 5-year overall survival of the HCC
buffer (8M Urea, pH 8.0). The amount
patients. A P value of < 0.05 was consid-
of protein lysates used for Western blot
ered to be statistically significant.
analysis was 100 and 50 mg for clini-
cal HCC samples and HCC cell lines,
respectively. Protein extracts were sepa-
rated by 12% SDS-PAGE and transferred RESULTS
to PDMF membrane (Millipore, Billerica,
Six1 Expression in HCC Cell Lines
MA) according to the standard proto-
col. After blocking with 5% nonfat milk Three nonmetastatic HCC cell lines
at room temperature for 1 h, Six1 anti- (Hep3B, Huh7 and PLC) and two meta-
body with 1:1,000 dilution was hybridized static HCC cell lines (MHCC97L and
with the membrane at 4°C overnight. The MHCC97H) were checked for Six1 gene
membrane was washed three times with expression in terms of mRNA and protein
316 K. Tak-Pan Ng and K. Man
levels (Figure 24.1A). Using multiplex Hep3B, MHCC97L, and MHCC97H when
RT-PCR analysis, Six1 mRNA was compared with the mRNA level in Huh7
detected with a relatively higher level in and PLC. The metastatic HCC cell line
18S
18S
18S
Figure 24.1. (a) mRNA and protein expression analysis of Six1 gene among nonmetastatic (Huh7, 3B and
PLC) and metastatic (MHCC97L and MHCC97H) HCC cells. (b) mRNA and protein expression pattern of
Six1 gene in HCC patients. T, tumor tissue; NT, paired nontumor tissue. (c) mRNA and protein expression
analysis of Six1 genes in normal liver tissue. 1–8, liver tissues from different normal donors; C, positive
control. Ribosomal RNA 18S and β-actin were internal control for RT-PCR and Western blot
24. Hepatocellular Carcinoma: Overexpression of Homeoprotein Six1 317
Table 24.1. Summary of Six1 expression in liver tissues from HCC patients and
normal donors.
HCC patients
Six1 expression Tumor Nontumor Normal
RNA Positive 61 (84.7%) 6 (8.3%) 2 (10%)
Negative 11 (15.3%) 66 (91.7%) 18 (90%)
Protein Positive 61 (59.2%) 0 (0%) 0 (0%)
Negative 42 (40.8%) 103 (100%) 20 (100%)
318 K. Tak-Pan Ng and K. Man
Table 24.2. Correlation of Six1 protein expression and clinicopathological features of HCC patients.
Six1 protein expression (n)
Clinicopathological features Number (n) Negative Positive P
Sex
Male 87 35 52
Female 16 7 9 0.792
Age
≤55 years 59 25 34
>55 years 44 17 27 0.703
pTNM stage
Early stage (I–II) 40 24 16
Advanced stage (III–IV) 63 18 45 0.002*
Venous infiltration
Absent 44 25 19
Present 59 17 42 0.004*
Cirrhosisa
Absent 33 13 20
Present 69 29 40 0.800
Encapsulationa
Absent 48 22 26
Present 25 11 14 0.881
Tumor size
<5 cm 26 13 13
≥5 cm 77 29 48 0.268
AFP level
≤20 ng/ml 41 20 21
>20 ng/ml 62 22 40 0.179
Hepatitis B surface antigena
Negative 16 8 8
Positive 86 34 52 0.520
P
TNM = pathologic tumor-node-metastasis;
AFP = alpha-fetoprotein.
a
Total number less than 103 due to missing data.
*
Significant difference.
and multivariate Cox proportional hazard the 1- and 5-year overall survival showed
regression analyses of these factors that Six1 protein was the most influential
corresponding to the overall survival of factor for predicting the 1-year overall sur-
HCC patients were performed. Univariate vival (OR = 5.405, 95% CI 1.427–20.474,
Cox proportional hazard regression analy- p = 0.013) while pTNM staging was the
sis showed that Six1 protein (HR = 1.956, most important factor for predicting the
95% CI 1.011–3.784, p = 0.046) and other 5-year overall survival (OR = 12.152, 95%
factors were significantly associated with CI 1.652–89.378, p = 0.004, Table 24.4).
the overall survival of HCC patients after
hepatic resection (Table 24.3). However,
multivariate analysis indicated that pTNM DISCUSSION
staging was the only independent factor
for predicting the overall survival of HCC Homeobox transcription factor Six1 located
patients (HR = 7.698, 95% CI 1.891–31.33, at 14q23 of the chromosome is involved in
p = 0.004, Table 24.3). Logistic regression the early development of many organs
analysis of these factors associated with including the brain, ear, eye, muscle and
Table 24.3. Cox proportional hazard regression analysis of Six1 protein expression and clinico-
pathological parameters in relation to the overall survival of HCC patients.
Univariate analysis Multivariate analysis
HR (95% CI) P HR (95% CI) P
Six1 protein
Positive vs. negative 1.956 (1.011–3.784) 0.046 1.29 (0.624–2.503) NS
pTNM stage
Advanced vs. early 4.952 (2.077–11.808) 0.000 7.698 (1.891–31.33) 0.004
Venous infiltration
Presence vs. absence 3.302 (1.572–6.934) 0.002 0.493 (0.136–1.780) NS
AFP level
> 20 ng/ml vs. ≤ 20 ng/ml 3.062 (1032–4.118) 0.04 1.735 (0.783–3.844) NS
HR = hazard ratio; CI = confidence interval; NS = not significant
Table 24.4. Logistic Regression Analysis of Six1 Protein and Clinicopathological Parameters on
Predicting the 1- and 5-year Overall Survival of HCC Patients.
1-year survival 5-year survival
OR (95% CI) P OR (95% CI) P
Six1 protein
Positive vs. negative 5.405 (1.427–20.474) 0.013 1.044 (0.401–2.719) NS
pTNM stage
Advanced vs. early 2.914 (0.245–34.63) NS 12.152 (1.652–89.378) 0.004
Venous infiltration
Presence vs. absence 1.557 (0.155–15.636) NS 0.478 (0.072–3.19) NS
AFP level
>20 ng/ml vs. <20 ng/ml 1.754 (0.554–5.552) NS 2.239 (0.846–5.929) NS
OR = odd ratio; CI = confidence interval; NS = not significant.
320 K. Tak-Pan Ng and K. Man
Hepatic resection is one of the major HCC patients may still provide a valuable
curative options for HCC patients. Different index for effective management of HCC
centers have reported achieving different patients after hepatectomy especially during
survival rates (Llovet et al., 2005). Identifi- early year.
cation of novel risk markers after hepa- Up to now, the understanding of the
tectomy is beneficial for the manage- mechanism of Six1 in pathogenesis of can-
ment of HCC patients and improvement of cers is still limited. Overexpression of Six1
their survival. Our results showed that in breast cancer cells can promote the can-
overexpression of Six1 protein in HCC cer cells to escape from the G2 cell cycle
patients was significantly associated with checkpoint after X-ray irradiation (Ford
poor overall survival after hepatectomy et al., 1998). The cell cycle regulatory
(p = 0.0423, Figure 24.2), suggesting that activity of Six1 in breast cancer is regu-
this protein may be a novel marker related lated by casein kinase II which inactivates
to poor overall survival of HCC patients Six1 through phosphorylation (Ford et al.,
after hepatectomy. 2000). Cyclin A1 is a downstream effector
Although multivariate Cox proportional for Six1 in breast cancer where over-
hazard regression analysis showed that expression of Six1 promotes cyclin A1
Six1 protein was not an independent factor expression and subsequently increases
for predicting the overall survival, logistic cell proliferation and progression (Coletta
regression analysis indicated that Six1 pro- et al., 2004). Gene amplification of
tein was the most influential factor asso- Six1 is a probable mechanism which
ciated with poor 1-year overall survival contributes to tumorigenesis in breast
(p = 0.013, Table 24.4) rather than 5-year cancer (Reichenberger et al., 2005).
overall survival (Table 24.4), proposing Overexpression of Six1 in RMS cells
that Six1 may be regarded as a predictor can boost their pulmonary metastasis
for short-term overall survival of HCC potential, whereas downregulation of
patients after hepatectomy. Hepatocellular Six1 suppresses their metastatic abil-
carcinoma patients with a higher rate of ity (Yu et al., 2004). Yu et al. (2006)
early-year death after hepatectomy may be also demonstrated that Six1 potentially
caused by higher malignancy of HCC that activates several oncogenes including
pathologically is, in part, determined to be cyclin D1, c-Myc and Ezrin. Moreover, the
at the advanced pTNM stage or the presence ability of Six1 to promote metastasis of
of venous infiltration. Overexpression of RMS requires the function of Ezrin. The
Six1 protein in HCC patients was deter- above research in breast cancer and RMS
mined to be significantly associated with provides important clues to clarify the
these malignant factors indicating the pos- functional roles of Six1 in HCC because
sible role of Six1 on promoting HCC different cancers may have different regu-
malignancy. Even though pTNM staging latory mechanisms. In fact, Cyclin A is
was determined to be the most important overexpressed in HCC and its overexpres-
factor in predicting long-term overall sur- sion is associated with poor survival of
vival of HCC patients after hepatectomy HCC patients (Chao et al., 1998; Ohashi
(Tables 24.3 and 24.4), assessment of Six1 et al., 2001). Casein kinase II shows
protein level of resected tumor tissues of an important involvement in transforming
322 K. Tak-Pan Ng and K. Man
growth factor-beta1-induced HCC (Gavin Fields, A.C., Cotsonis, G., Sexton, D., Santoianni, R.,
et al., 2003). The relationship between Six1 and Cohen, C. 2004. Survivin expression in
hepatocellular carcinoma: correlation with pro-
and Cyclins as well as casein kinase II in
liferation, prognostic parameters, and outcome.
regulating the tumorigenesis and metastasis Mod. Pathol. 17: 1378–1385.
of HCC thus needs further clarification. Ford, H.L., Kabingu, E.N., Bump, E.A., Mutter,
The specificity of Six1 protein expression G.L., and Pardee, A.B. 1998. Abrogation of the
in HCC metastatic cells may also provide a G2 cell cycle checkpoint associated with over-
good bridge to study the functions of Six1 expression of HSIX1: a possible mechanism
of breast carcinogenesis. Proc. Natl. Acad. Sci.
involved in metastasis of HCC through
USA 95: 12608–12613.
suppression strategy such as using antisense Ford, H.L., Landesman-Bollag, E., Dacwag, C.S.,
or RNA interference means. Stukenberg, P.T., Pardee, A.B., and Seldin, D.C.
In conclusion, our data indicated that 2000. Cell cycle-regulated phosphorylation of
Six1 protein was specifically and fre- the human SIX1 homeodomain protein. J. Biol.
quently expressed in HCC tumor tissues. Chem. 275: 22245–22254.
Gavin, L.G., Romieu-Mourez, R., Panta, G.R., Sun,
Overexpression of Six1 protein in HCC
J., Factor, V.M., Thorgeirsson, S.S., Sonenshein,
patients was significantly associated with G.E., and Arsura, M. 2003. Inhibition of CK2
advanced pTNM stage and venous infiltra- activity by TGF-beta1 promotes IkappaB-alpha
tion. In addition, Six1 protein could be a protein stabilization and apoptosis of immortal-
novel marker for predicting the short-term ized hepatocytes. Hepatology 38: 1540–1551.
overall survival of HCC patients after Ikeguchi, M., Ueda, T., Sakatani, T., Hirooka, Y.,
and Kaibara, N. 2002. Expression of survivin
hepatic resection. Further functional stud-
messenger RNA correlates with poor progno-
ies are worthwhile to explore the precise sis in patients with hepatocellular carcinoma.
mechanism and eventually develop potential Diagn. Mol. Pathol. 11: 33–40.
therapies targeting Six1 in liver cancer Jemal, A., Siegel, R., Ward, E., Murray, T., Xu, J.,
recurrence and metastases. and Thun, M.J. 2007. Cancer statistics 2007. CA.
Cancer J. Clin. 57: 43–66.
Laclef, C., Hamard, G., Demignon, J., Souil, E.,
REFERENCES Houbron, C., and Maire, P. 2003. Altered myo-
Befeler, A.S., and Di Bisceglie, A.M. 2002. genesis in Six1-deficient mice. Development
Hepatocellular carcinoma: diagnosis and treat- 130: 2239–2252.
ment. Gastroenterology 122: 1609–1619. Lee, T.K., Man, K., Poon, R.T., Lo, C.M., Yuen,
Bosch, F.X., Ribes, J., and Borras, J. 1999. A.P., Ng, I.O., Ng, K.T., Leonard, W., and Fan,
Epidemiology of primary liver cancer. Semin. S.T. 2006. Signal transducers and activators of
Liver Dis. 19: 271–285. transcription 5b activation enhances hepatocellu-
Chao, Y., Shih, Y.L., Chiu, J.H., Chau, G.Y., Lui, lar carcinoma aggressiveness through induction
W.Y., Yang, W.K., Lee, S.D., and Huang, T.S. of epithelial-mesenchymal transition. Cancer
1998. Overexpression of cyclin A but not Skp 2 Res. 66: 9948–9956.
correlates with the tumor relapse of human hepa- Li, C.M., Guo, M., Borczuk, A., Powell, C.A.,
tocellular carcinoma. Cancer Res. 58: 985–990. Wei, M., Thaker, H.M., Friedman, R., Klein, U.,
Coletta, R.D., Christensen, K., Reichenberger, Tycko, B. 2002. Gene expression in Wilms’
K.J., Lamb, J., Micomonaco, D., Huang, L., tumor mimics the earliest committed stage in the
Wolf, D.M., Muller-Tidow, C., Golub, T.R., metanephric mesenchymal-epithelial transition.
Kawakami, K., and Ford, H.L. 2004. The Six1 Am. J. Pathol. 160: 2181–2190.
homeoprotein stimulates tumorigenesis by reac- Li, Y., Tang, Z.Y., Ye, S.L., Liu, Y.K., Chen, J., Xue,
tivation of cyclin A1. Proc. Natl. Acad. Sci. USA Q., Chen, J., Gao, D.M., and Bao, W.H. 2001.
101: 6478–6483. Establishment of cell clones with different
24. Hepatocellular Carcinoma: Overexpression of Homeoprotein Six1 323
potential from the hepatocellular carcinoma and postoperative outcome. Liver Transpl. 10:
cell line MHCC97. World J. Gastroenterol. 7: S39–S45.
630–636. Reichenberger, K.J., Coletta, R.D., Schulte, A.P.,
Llovet, J.M., Schwartz, M., and Mazzaferro, V. 2005. Varella-Garcia, M., and Ford, H.L. 2005. Gene
Resection and liver transplantation for hepatocel- amplification is a mechanism of Six 1over-
lular carcinoma. Semin. Liver Dis. 25: 181–200. expression in breast cancer. Cancer Res. 65:
Lo, C.M., and Fan, S.T. 2004. Liver transplanta- 2668–2675.
tion for hepatocellular carcinoma. Br. J. Surg. Relaix, F., and Buckingham, M. 1999. From insect
91: 131–133. eye to vertebrate muscle: redeployment of a
Muller, C. 2006. Hepatocellular carcinoma – rising regulatory network. Genes Dev. 13: 3171–3178.
incidence, changing therapeutic strategies. Wien. Ueki, T., Fujimoto, J., Suzuki, T., Yamamoto, H.,
Med. Wochenshr. 156: 404–409. and Okamoto, E. 1997. Expression of hepatocyte
Ohashi, R., Gao, C., Miyazaki, M., Hamazaki, K., growth factor and its receptor c-met proto-onco-
Tsuji, T., Inoue, Y., Uenura, T., Hirai, R., gene in hepatocellular carcinoma. Hepatology
Shimizu, N., and Namba, M. 2001. Enhanced 25: 619–623.
expression of cyclin E and cyclin A in human Xu, P.X., Zheng, W., Huang, L., Maire, P., Laclef, C.,
hepatocellular carcinomas. Anticancer Res. 21: and Silvius, D. 2003. Six1 is required for
657–662. the early organogenesis of mammalian kidney.
Oliver, G., Wehr, R., Jenkins, N.A., Copeland, Development 130: 3085–3094.
N.G., Cheyette, B.N., Hartenstein, V., Zipursky, Yu, Y., Davicioni, E., Triche, T.J., and Merlino, G.
S.L., and Gruss, P. 1995. Homeobox genes and 2006. The homeoprotein six1 transcriptionally
connective tissue patterning. Development 121: activates multiple protumorigenic genes but
693–705. requires ezrin to promote metastasis. Cancer
Ozaki, H., Nakamura, K., Funahashi, J., Ikeda, Res. 66: 1982–1989.
K., Yamada, G., Tokano, H., Okamura, H.O., Yu, Y.L., Khan, J., Khanna, C., Helman, L.,
Kitamura, K., Muto, S., Kotaki, H., Sudo, K., Meltzer, P.S., and Merlino, G. 2004. Expression
Horai, R., Iwakura, Y., and Kawakami, K. 2004. profiling identifies the cytoskeletal organ-
Six1 controls patterning of the mouse otic vesi- izer ezrin and the developmental homeopro-
cle. Development 131: 551–562. tein Six-1 as key regulators. Nat. Med. 10:
Parkin, D.M., Bray, F., Ferlay, J., and Pisani, P. 2001. 175–181.
Estimating the world cancer burden: Globocan Zheng, W., Huang, L., Wei, Z.B., Silvius, D.,
2000. Int. J. Cancer 94: 153–156. Tang, B., and Xu, P.X. 2003. The role of Six1
Poon, R.T., and Fan, S.T. 2004. Hepatectomy in mammalian auditory system development.
for hepatocellular carcinoma: patient selection Development 130: 3989–4000.
25
Hepatocellular Carcinoma:
KiSS-1 Overexpression
as a Prognostic Factor
Katharina Schmid, Isabella Mosberger, and Fritz Wrba
325
326 K. Schmid et al.
levels of KiSS-1 mRNA than normal pan- (2003) reported that overexpression of
creas tissue. Ikeguchi et al. (2004) observed KiSS-1 correlated with tumor progression
that the loss of KiSS-1 gene expression in in HCC. Also, Martin et al. (2005) found
esophageal squamous carcinoma was a sig- that the expression of KiSS-1 increased in
nificant predictor of lymph node metastasis. patients with aggressive breast cancer.
According to Dhar et al. (2004), down- In order to investigate a possible correla-
regulation of KiSS-1 expression was the tion between KiSS-1 expression in HCC
strongest independent prognostic factor for and patient prognosis, we performed an
gastric carcinoma in multivariate analysis. immunohistochemical study: expression
Kostadima et al. (2007) found that the levels of the KiSS-1 gene product and its
KiSS-1 gene was silent in the vast majority cognate receptor AXOR12 were analysed
of resected node-positive breast carcino- in HCC specimens on the protein level,
mas, strongly supporting the antimetastatic and results were correlated with clinical
role of the gene. Conversely, Ikeguchi et al. follow-up (Figure 25.1).
a b
c d
Figure 25.1. Cases with high immunoreactivity of (a) KiSS-1 and (b) AXOR12 in HCC (both magnifi-
cation x200); (c) shows KiSS-1 (magnification x100) and (d, right side) AXOR12 immounoreactivity in
cirrhotic liver tissue, and (d, left side) negative immunostaining of tumor cells (magnification x200)
25. Hepatocellular Carcinoma: KiSS-1 Overexpression as a Prognostic Factor 327
Lee, J.H., Miele, M.E., Hicks, D.J., Phillips, K.K., Takatsu, Y., Masuda, Y., Ishibashi, Y., Watanabe, T.,
Trent, J.M., Weissman, B.E., and Welch, D.R. Asada, M., Yamada, T., Suenaga, M., Kitada, C.,
1996. KiSS-1, a novel human malignant Usuki, S., Kurokawa, T., Onda, H., Nishimura, O.,
melanoma metastasis-suppressor gene. J. Natl. and Fujino, M. 2001. Metastasis suppressor
Cancer Inst. 88:1731–1737. gene KiSS-1 encodes peptide ligand of
Lee, J.H., and Welch, D.R. 1997. Suppression a G-proteincoupled receptor. Nature 411:
of metastasis in human breast carcinoma 613–617.
MDA-MB-435 cells after transfection with the Pokorny, H., Gnant, M., Rasoul-Rockenschaub, S.,
metastasis suppressor gene, KiSS-1. Cancer Gollackner, B., Steiner, B., Steger, G., Steininger,
Res. 57:2384–2387. R., and Muehlbacher, F. 2005. Does additional
Martin, T.A., Watkins, G., and Jiang, W.G. 2005. doxorubicin chemotherapy improve outcome in
KiSS-1 expression in human breast cancer. Clin. patients with hepatocellular carcinoma treated
Exp. Metastasis 22:503–511. by liver transplantation? Am. J. Transplant
Masui, T., Doi, R., Mori, T., Toyoda, E., Koizumi, M., 5:788–794.
Kami, K., Ito, D., Peiper, S.C., Broach, J.R., Sanchez-Carbayo, M., Capodieci, P., and Cordon-
Oishi, S., Niida, A., Fujii, N., and Imamura, M. Cardo, C. 2003. Tumor suppressor role of KiSS-1
2004. Metastin and its variant forms suppress in bladder cancer. Loss of KiSS-1 expression is
migration of pancreatic cancer cells. Biochem. associated with bladder cancer progression and
Biophys. Res. Commun. 315:85–92. clinical outcome. Am. J. Pathol. 162:609–617.
Muir, A.I., Chamberlain, L., Elshourbagy, N.A., Schmid, K., Wang, X., Sieghart, W., Peck-
Michalovich, D., Moore, D.J., Calamari, A., Radosavljevic, M., Bodingbauer, M., Rasoul-
Szekeres, P.G., Sarau, H.M., Chambers, J.K., Rockenschaub, S., and Wrba, F. 2007. KiSS-1
Murdock, P., Steplewski, K., Shabon, U., Miller, J.E., overexpression as an independent prognostic
Middleton, S.E., Darker, J.G., Larminie, C.G., marker in hepatocellular carcinoma: an immu-
Wilson, S., Bergsma, D.J., Emson, P., Faull, nohistochemical study. Virchows Arch. 450:
R., Philpott, K.L., and Harrison, D.C. 2001. 143–149.
AXOR12, a novel human G protein-coupled Shirasaki, F., Takata, M., Hatta, N., and Takehara,
receptor, activated by the peptide KiSS-1. K. 2001. Loss of expression of the metastasis
J. Biol. Chem. 276:28969–28975. suppressor gene KiSS-1 during melanoma pro-
Ohtaki, T., Shintani, Y., Honda, S., Matsumoto, H., gression and its association with LOH of chromo-
Hori, A., Kanehashi, K., Terao, Y., Kumano, S., some 6q16.3–q23. Cancer Res. 61:7422–7425.
26
Hepatocellular Carcinoma: Prognosis
Using Hepatoma-Derived Growth Factor
Immunohistochemistry
Hideji Nakamura, Kenya Yoshida, and Yasuhiko Tomita
333
334 H. Nakamura et al.
the nucleus rather than in the cytoplasm HDGF is highly expressed in oval cells,
(Everett et al., 2001). The gene-specific that are progenitor cells with bipotential
region of HDGF, at least the bipartite NLS activity for differentiating into hepatocytes
sequence and both the N-and C-terminal and bile duct cells. These findings sug-
neighboring portions, is essential for the gest that HDGF plays important roles in
mitogenic activity (Kishima et al., 2002a). the proliferation of immature hepatocytes
Thus, HDGF exerts its proliferating activity and hepatic progenitor cells including oval
via two different pathways: (1) via a putative cells, showing significant involvement of
plasma membrane-located HDGF receptor HDGF in liver development.
for which signaling depends on the hath In the fetus, HDGF was also abundantly
region, resulting in MAP kinase activation, expressed in other organs. Hepatoma-
and (2) via targeting to the nucleus by NLS. derived growth factor is highly expressed
in fetal rat aorta, conotruncus, and heart
Developmentally Regulated Expression of (Everett et al., 2000, 2001), and is widely
Hepatoma-Derived Growth Factor distributed in the renal anlage at the early
Cancers develop by the accumulation of stages of renal development and disap-
dysregulated gene expression from multistep peared from the adult kidney (Oliver and
genetic mutations of oncogenes and/or Al-Awqati, 1998). Hepatoma-derived
suppressor genes. These oncogenic proteins growth factor is highly expressed in the
and tumor suppressor proteins generally endothelial cells of non-muscularized
include growth factors, their receptors, forming blood vessels of the fetal lung
intracellular signal transduction molecules, (Everett et al., 2004), and its expression is
and transcriptional regulatory factors. enhanced dominantly in the bronchial and
Genes that are reduced significantly during alveolar epithelial cells including type II
organ development are frequently absent pneumocytes by bleomycin-instillation in
or little expressed in the adult tissues, mice (Mori et al., 2004). This growth fac-
and these developmentally regulated genes tor is also highly expressed in hind gut epi-
may be reactivated in human cancers. In thelia as well as atrial myocytes in the fetal
the fetus, HDGF is abundantly expressed stage. Furthermore, HDGF is expressed in
in various organs including liver. the nucleus of the colonic epithelial cells,
Hepatoma-derived growth factor is dominantly in the bottom of the intestinal
highly expressed in fetal liver of the mid- crypts by immunohistochemical analysis
gestation stage, and is markedly decreased (Nakamura et al., 2002). Thus, HDGF
near birth. The expression of this factor was is one of the developmentally regulated
significantly decreased with differentiation genes which are abundantly expressed in
in fetal hepatocytes induced by oncostatin cancer cells.
M treatment in in vitro primary culture dif-
ferentiation system (Enomoto et al., 2002). Role in Hepatocarcinogenesis
Adenoviral introduction of HDGF antisense The Fatty Liver Shionogi (FLS) mouse is
cDNA into the fetal hepatocytes suppresses an inbred strain that develops spontaneous
their proliferation, and the inhibitory effects fatty liver without obesity, and results in
of the HDGF antisense virus were recov- liver tumor development to ~ 45% in 52
ered by exogenous HDGF. Furthermore, weeks and 90% at 72 weeks after birth
26. Hepatocellular Carcinoma: Prognosis Using Hepatoma-Derived Growth Factor 335
in male mice; these tumors have been was dramatically increased in human colorec-
histologically diagnosed as hepatocellular tal cancers, especially in tumors proficient
adenoma and carcinoma. In the liver of in DNA mismatch repair (Lepourcelet et
FLS mice, HDGF expression has already al., 2005). Conversely, down regulation of
increased at an early stage before the HDGF by use of HDGF-siRNA had mini-
tumors developed microscopically in the mal effect on anchorage-dependent growth
liver (Yoshida et al., 2003). Hepatoma- but significantly reduced anchorage-
derived growth factor may be induced and independent growth of non-small cell lung
secreted or released from hepatocytes and cancer (NSCLC) cells in soft agar (Zhang
macrophages, enhancing the cell cycle et al., 2006). Thus, these findings suggest
progression of hepatocytes, inducing their that HDGF is an oncogenic protein, and
transformation, and promoting the pro- plays a role in cancer development.
liferation of HCC cells in an intracrine,
autocrine and/or paracrine manner. Role in Cancer Progression and
Furthermore, HDGF expression was Angiogenesis
strongly detected in an oval cell line, Hepatoma-derived growth factor protein
Oc 15–5, by immunohistochemistry, which is abundantly expressed in various human
is established from the liver of Long- cancers. In liver, HDGF expression is
Evans-Cinnamon rats. Hepatocellular car- higher in HCC than in the adjacent tissues
cinoma is considered to be developed (Yoshida et al., 2003, 2006). The HDGF-
from oval cells induced in the regen- over-expressing hepatoma cell line HepG2
eration process after liver injury. By dif- proliferates more rapidly and produces
ferential subtractive chain reaction from larger tumors, showing more rapid growth
strong anchorage-independent growth to in nude mice than neomycin-resistant cells
its negative HCC cells, HDGF is cloned in vivo. Recombinant HDGF stimulates
as one of the genes related to anchorage- the growth of cancer cells, while antisense
independency (Huang et al., 2004). These HDGF oligonucleotides or polyclonal anti-
findings strongly suggest that HDGF poten- HDGF antibody suppress their growth
tially participates in hepatocarcinogenesis (Kishima et al., 2002b; Nakamura et al.,
and in the early stage of HCC. 2002). Exogenously supplied HDGF pro-
Hepatoma-derived growth factor-over- motes the proliferation of bronchial squa-
expressing NIH3T3 cells generate sarco- mous cell carcinoma cell line, A549 cells
matous tumors in nude mice. In our study, (Mori et al., 2004). Knock-down expres-
HDGF-over-expressing NIH3T3 cells did sion of HDGF by small interfering RNA
not show significant anchorage-independ- (siRNA) in NSCLC cells significantly
ent growth in soft agar assay; however, shows slower growth, less colony for-
HDGF-over-expressing NIH3T3 cells mation in soft agar, and lesser in vitro
developed more small colonies in soft agar invasion activity (Zhang et al., 2006).
than parent or neomycin-resistant cells By proteomic differential display analysis
(Okuda et al., 2003). Hepatoma-derived and mass spectrometry, HDGF is down-
growth factor was more highly expressed regulated in regressive cancer cells as
in colon cancer cells than non-transformed compared with that in inflammatory cell-
intestinal epithelial cells, and its expression promoting progressive cells of the murine
336 H. Nakamura et al.
fibrosarcoma cell line (Hayashi et al., and through the induction of VEGF in the
2005). The higher expression of HDGF nucleus (Okuda et al., 2003).
shows more malignant potentials, suggest-
ing that HDGF is a candidate factor for IMMUNOHISTOCHEMICAL
cancer progression.
Hepatocellular carcinoma is a hyper-
AND ANALYTICAL
vascular tumor. Hepatoma-derived growth METHODS
factor is intrinsically related to angiogen-
Materials
esis and vasculogenesis and its protein
expression is induced in the regenerating 1. Rabbit polyclonal antibody raised
process of vascular vessels in wound repair against C-terminal amino acids (amino
(Everett et al., 2000). Additionally, HDGF acids 231–240) of the human HDGF.
is a candidate endothelial growth factor 2. Dulbecco’s phosphate-buffered saline
for renal glomerulus formation (Oliver and without magnesium and calcium – Tween
Al-Awqati, 1998). 20 (PBS-T): 200 mg potassium chloride,
Tumors developed from HDGF-over- 200 mg monobasic potassium phosphate,
expressing NIH3T3 cells inoculated 8 g sodium chloride, and 2.9 g dibasic
in nude mice are macroscopically red- sodium phosphate·12 H2O, bring volume
colored and histologically rich in vascu- to 11 ml with deionized distilled water, pH
lature, and immunohistochemical analysis 7.4. Add 0.1 ml Tween 20 and mix well.
by anti-CD31 antibody shows prominent 3. Fixative: 10% formalin in distilled water.
new vessel formation (Okuda et al., 2003). 4. 10 mM citric acid buffer (pH 6.0): 294.10 g
Hepatoma-derived growth factor protein trisodium citrate·2 H2O, bring volume to
stimulates the proliferation and tubule for- 11 ml with deionized distilled water to
mation of human umbilical vein endothe- prepare 1 M trisodium citrate. 210.14 g
lial cells and the migration of human citric acid·H2O, bring volume to 11 ml
pulmonary microvascular endothelial cells. with deionized distilled water to prepare
Using chick chorioallantoic membrane 1 M citric acid. Add 5 ml of 1 M trisodium
(CAM) assay, recombinant HDGF stimu- citrate and 0.8 ml of 1 M citric acid to
lates blood vessel formation, and stimu- 494.2 ml of deionized distilled water
lates cellular reorganization within the to prepare 10 mM citric acid, pH 6.0.
CAM from a loose network into a more 5. Methanol containing 0.2% hydrogen
compact, linear alignment reminiscent peroxide: 9 ml of 31% hydrogen perox-
of tube formation (Everett et al., 2004). ide in 150 ml methanol.
The more potent growth stimulating activ- 6. Blocking serum (goat serum) diluted at
ity of HDGF in vivo than in vitro must 1/50 in PBS-T.
be brought on by both the direct cell 7. Primary antibody (rabbit polyclonal
growth activity and its angiogenic activity. anti-HDGF antibody) diluted at 1/5,000
Furthermore, HDGF induces vascular in PBS-T.
endothelial growth factor (VEGF) gene 8. Secondary antibody (biotinylated goat
expression, and is a potent angiogenic fac- polyclonal anti-rabbit antibody) diluted
tor that functions by its own direct stimu- at 1/100 in PBS-T.
lation of the proliferation of endothelial 9. Avidin–biotin-complex solution: Add
cells and vascular smooth muscle cells 20 μl of VECTASTAIN Elite reagent
26. Hepatocellular Carcinoma: Prognosis Using Hepatoma-Derived Growth Factor 337
A and B to 1 ml PBS-T. Mix well and 13. Incubate the slides with secondary
let it ,stand for more than 30 min. antibody diluted in PBS-T for 30 min
10. 3,3 -Diaminobenzidine, tetrahydro- in a humidity chamber at 37°C.
chloride (DAB) solution: 75 mg of 14. Rinse the slide three times in PBS-T
DAB powder in 150 ml PBS-T. Stir for 5 min each.
well. Add 1 ml of 31% hydrogen per- 15. Incubate the slides with Avidin–
oxide to 150 ml of DAB solution. biotin-complex solution for 45 min in
a humidity chamber at 37°C.
16. Rinse the slide three times in PBS-T
Method
for 5 min.
1. Fix samples in 10% formalin, dehy- 17. Prepare DAB solution.
drated, and embedded in paraffin. 18. Immerse the slide in DAB solution for
2. Cut the paraffin block at 4 μm thick- 1 min.
ness and mount sections on poly-L- 19. Wash the slide in distilled water five
lysine-coated glass slides. times.
3. Deparaffinize the slide in xylene three 20. Immerse the slide in methyl green for
times for 15 min each, and in 100% 10 min.
ethanol three times for 5 min each. 21. Wash the slide in distilled water five
4. Immerse the slide in PBS-T. times.
5. For antigen retrieval, heat the slide 22. Dehydrate with 100% ethanol three
in 10 mM citrate buffer (pH 6.0) for times for 5 min each, and xylene three
15 min using microwave at 600 W. times for 5 min each, then mount with
6. Rinse the slide three times in PBS-T mounting medium.
for 5 min.
7. Immerse the slide in methanol con-
taining 0.2% hydrogen peroxide for
EVALUATION OF HEPATOMA-
20 min to block the endogenous per- DERIVED GROWTH
oxide activity at room temperature. FACTOR EXPRESSION
8. Rinse the slide three times in PBS-T IN HEPATOCELLULAR
for 10 min. CARCINOMA
9. Incubate the slides with 2% goat serum
in PBS-T for 30 min in a humidity Observe the slide under light microscope.
chamber at room temperature to block Bile ducts in the noncancerous lesions
the nonspecific binding of the primary uniformly show positive HDGF stain-
antibody. ing both for the nucleus and cytoplasm
10. Lightly tap off excess serum from each (Yoshida et al., 2006). Cancer cells showing
slide, and apply anti-HDGF primary staining similar to or stronger than the
antibody diluted in PBS-T. Incubate noncancerous bile ducts can be regarded
the slides overnight in a humidity as positive. Carefully examine the positivity
chamber at 4°C. ratio of all the cancer cells in the slide
11. Rinse the slide three times in PBS-T for both the nucleus and the cytoplasm.
for 5 min each. The HDGF labeling index (LI) can be
12. Prepare Avidin–biotin-complex solu- determined as follows: samples with more
tion. than 90% of cancer cells showing positive
338 H. Nakamura et al.
staining both for nucleus and cytoplasm nucleus and cytoplasm of HCC cells
were regarded as HDGF LI Level 2, and than in the adjacent normal hepatocytes
others as Level 1 (Figure 26.1). (Figure 26.2) (Yoshida et al., 2003, 2006).
Hepatocellular carcinomas with HDGF
LI Level 2 were regarded as having a
PROGNOSTIC SIGNIFICANCE positive HDGF index, and the remaining
OF HEPATOMA-DERIVED HCC with HDGF LI Level 1 as a nega-
GROWTH FACTOR IN tive HDGF index. Rate of positive HDGF
index was 27% in all HCC, and 29.9%
HEPATOCELLULAR among HCC with Edmondson’s differ-
CARCINOMA entiation grade I, II, and III. The HDGF
expression level by immunohistochemis-
Immunohistochemical analysis using try was significantly correlated to the dif-
specific anti-C terminus of HDGF anti- ferentiation of HCC, but it was not shown
body revealed that HDGF was more in other types of cancers. Hepatoma-
strongly and frequently expressed in the derived growth factor expression was
as a prognostic factor for disease-free and nostic marker for cancer patients. It is
overall survival of patients with HCC. expected in the future that any tool regu-
Besides HCC, HDGF is a prognsostic lating HDGF expression or HDGF signal
factor of pancreatic ductal cancer, lung pathways may be a useful candidate for
cancer, gastric cancer, and esophageal the suppression of carcinogenesis and can-
cancer. In patients with HCC or early cer progression.
stage NSCLC who underwent curative sur-
gery, the disease-free and disease-specific REFERENCES
survival and overall survival were lower
Abouzied, M.M., El-Tahir, H.M., Prenner, L.,
in those with higher HDGF expression Haberlein, H., Gieselmann, V., and Franken, S.
indexes than in those with lower HDGF 2005. Hepatoma-derived growth factor.
indexes (Ren et al., 2004; Iwasaki et al., Significance of amino acid residues 81–100 in
2005). The high expression of HDGF cell surface interaction and proliferative activity.
showed distant metastasis and shortened J. Biol. Chem. 280: 10945–10954.
survival time in patients with NSCLC. In Dietz, F., Franken, S., Yoshida, K., Nakamura, H.,
Kappler, J., and Gieselmann, V. 2002. The family of
gastric cancer, the patients with abundant hepatoma-derived growth factor proteins:
expression of HDGF showed significantly characterization of a new member HRP-4 and
higher rates of infiltrative tumor growth classification of its subfamilies. Biochem. J.
and vascular and lymphatic invasion, as 366: 491–500.
compared to those with lower expression Enomoto, H., Yoshida, K., Kishima, Y., Kinoshita, T.,
(Yamamoto et al., 2006). Pancreatic ductal Yamamoto, M., Everett, A.D., Miyajima, A., and
Nakamura, H. 2002. Hepatoma-derived growth
cancers with higher HDGF expression in factor is highly expressed in developing liver
the nucleus showed a poorer 5-year sur- and promotes fetal hepatocyte proliferation.
vival rate than those with lower expression Hepatology 36: 1519–1527.
(Uyama et al., 2006). These clinical find- Everett, A.D., Lobe, D.R., Matsumura, M.E.,
ings and in vitro data of pleiotropic HDGF Nakamura, H., and McNamara, C.A. 2000.
activities described above suggest that the Hepatoma-derived growth factor stimulates smooth
muscle cell growth and is expressed in vascular
higher expression of HDGF shows more development. J. Clin. Invest. 105: 567–575.
malignant potential including invasive and Everett, A.D., Stoops, T., and McNamara, C.A.
metastatic activity. 2001. Nuclear targeting is required for hepato-
In conclusion, HDGF is a unique nuclear ma-derived growth factor-stimulated mitogen-
targeting growth factor, which is expressed esis in vascular smooth muscle cells. J. Biol.
abundantly in cancer cells, promotes their Chem. 276: 37564–37568.
Everett, A.D., Narron, J.V., Stoops, T., Nakamura, H.,
malignant potential, and generates tumors and Tucker, A. 2004. Hepatoma-derived growth
and promotes their growth in vivo via its factor is a pulmonary endothelial cell-expressed
mitogenic activity and angiogenic activity angiogenic factor. Am. J. Physiol., Lung Cell.
deriving from both its own direct ang- Mol. Physiol. 286: L1194–1201.
iogenic activity and the induction of VEGF. Hayashi, E., Kuramatsu, Y., Okada, F., Fujimoto, M.,
Multivariate analysis of the relationship Zhang, X., Kobayashi, M., Iizuka, N., Ueyama,
Y., and Nakamura, K. 2005. Proteomic profil-
of HDGF expression and recurrence-free ing for cancer progression: differential display
and overall survival in patients with HCC analysis for the expression of intracellular pro-
confirmed that HDGF has the potential to teins between regressive and progressive cancer
become a significantly efficacious prog- cell lines. Proteomics 5: 1024–1032.
26. Hepatocellular Carcinoma: Prognosis Using Hepatoma-Derived Growth Factor 341
Hu, T.H., Huang, C.C., Liu, L.F., Lin, S.Y., epithelial growth. Am. J. Respir. Cell Mol. Biol.
Chang, H.W., Changchien, C.S., Lee, C.M., 30: 459–469.
Chuang, J.H., and Tai, M.H. 2003. Expression Nakamura, H., Kambe, H., Egawa, T., Kimura,
of hepatoma-derived growth factor in hepatocel- Y., Ito, H., Hayashi, E., Yamamoto, H., Sato, J.,
lular carcinoma. Cancer 98: 1444–1456. and Kishimoto, S. 1989. Partial purification and
Huang, J.S., Chao, C.C., Su, T.L., Yeh, S.H., Chen, characterization of hepatoma-derived growth
D.S., Chen, C.T., Chen, P.J., and Jou, Y.S. 2004. factor. Clin. Chim. Acta 183: 273–284.
Diverse cellular transformation capability of Nakamura, H., Izumoto, Y., Kambe, H., Kuroda,
overexpressed genes in human hepatocellular T., Mori, T., Kawamura, K., Yamamoto, H.,
carcinoma. Biochem. Biophys. Res. Commun. and Kishimoto, T. 1994. Molecular cloning
315: 950–958. of complementary DNA for a novel human
Iwasaki, T., Nakagawa, K., Nakamura, H., hepatoma-derived growth factor. J. Biol. Chem.
Takada, T., Matsui, K., and Kawahara, K. 2005. 269: 25143–25149.
Hepatoma-derived growth factor as a prognostic Nakamura, H., Yoshida, K., Ikegame, K., Kishima,
marker in completely resected non-small-cell Y., Uyama, H., and Enomoto, H. 2002.
lung cancer. Oncol. Rep. 13: 1075–1080. Antibodies against hepatoma-derived growth
Izumoto, Y., Kuroda, T., Harada, H., Kishimoto, T., and factor and mucosal repair in ulcerative colitis.
Nakamura, H. 1997. Hepatoma-derived growth J. Gastroenterol. 37 (Suppl. 14): 8–14.
factor belongs to a gene family in mice showing Nakamura, H., and Hada, T. 2004. Hepatoma-
significant homology in the amino terminus. derived growth factor in ontogeny and tumori-
Biochem. Biophys. Res. Commun. 238: 26–32. genesis. Recent Res. Devel. Biophys. Biochem.
Kishima, Y., Yamamoto, H., Izumoto, Y., Yoshida, K., 4: 17–27.
Enomoto, H., Yamamoto, M., Kuroda, T., Ito, H., Okuda, Y., Nakamura, H., Yoshida, K., Enomoto, H.,
Yoshizaki, K., and Nakamura, H. 2002a. Uyama, H., Hirotani, T., Funamoto, M., Ito,
Hepatoma-derived growth factor stimulates cell H., Everett, A.D., Hada, T., and Kawase, I.
growth after translocation to the nucleus by 2003. Hepatoma-derived growth factor induces
nuclear localization signals. J. Biol. Chem. 277: tumorigenesis in vivo through both direct
10315–10322. angiogenic activity and induction of vascular
Kishima, Y., Yoshida, K., Enomoto, H., endothelial growth factor. Cancer Sci. 94:
Yamamoto, M., Kuroda, T., Okuda, Y., Uyama, 1034–1041.
H., and Nakamura, H. 2002b. Antisense oli- Oliver, J.A., and Al-Awqati, Q. 1998. An endothe-
gonucleotides of hepatoma-derived growth lial growth factor involved in rat renal develop-
factor (HDGF) suppress the proliferation of ment. J. Clin. Invest. 102: 1208–1219.
hepatoma cells. Hepatogastroenterology 49: Ren, H., Tang, X., Lee, J.J., Feng, L., Everett, A.D.,
1639–1644. Hong, W.K., Khuri, F.R., and Mao, L. 2004.
Lepourcelet, M., Tou, L., Cai L., Sawada, J., Expression of hepatoma-derived growth factor
Lazar, A.J.F., Glickman, J.N., Williamson, is a strong prognostic predictor for patients with
J.A., Everett, A.D., Redston, M., Fox, E.A., early-stage non-small cell lung cancer. J. Clin.
Nakatani, Y., and Shivdasani, R.A. 2005. Oncol. 22: 3230–3237.
Insights into development mechanisms and can- Uyama, H., Tomita, Y., Nakamura, H., Nakamori, S.,
cers in the mammalian intesitine derived from Zhang, B., Hoshida, Y., Enomoto, H., Okuda, Y.,
serial analysis of gene expression and study of Sakon, M., Aozasa, K., Kawase, I., Hayashi, N.,
the hepatoma-derived growth factor (HDGF). and Monden, M. 2006. Hepatoma-derived
Development 132: 415–427. growth factor is a novel prognostic factor for
Mori, M., Morishita, H., Nakamura, H., Matsuoka, H., patients with pancreatic cancer. Clin. Cancer
Yoshida, K., Kishima, Y., Zhou, Z., Kida, H., Res. 12: 6043–6048.
Funakoshi, T., Goya, S., Yoshida, M., Kumagai, Yamamoto, S., Yasuhiko, T., Hoshida, Y., Takiguchi,
T., Tachibana, I., Yamamoto, Y., Kawase, I., and S., Fujiwara, Y., Yasuda, T., Doki, Y., Yoshida, K.,
Hayashi, S. 2004. Hepatoma-derived growth fac- Aozasa, K., Nakamura, H., and Monden, M.
tor is involved in lung remodeling by stimulating 2006. Expression of hepatoma-derived growth
342 H. Nakamura et al.
factor is correlated with lymph node metastasis and Aozasa, S., Nagano, H., Sakon, M., Kawase, I.,
prognosis of gastric carcinoma. Clin. Cancer Monden, M., and Nakamura, H. 2006. Hepatoma-
Res. 12: 117–122. derived growth factor is a novel prognostic fac-
Yoshida, K., Nakamura, H., Okuda, Y., Enomoto, tor for hepatocellular carcinoma. Ann. Surg.
H., Kishima, Y., Uyama, H., Ito, H., Hirasawa, Oncol. 13: 159–167.
T., Inagaki, S., and Kawase, I. 2003. Expression Zhang, J., Ren, H., Yuan, P., Lang, W., Zhang, L.,
of hepatoma-derived growth factor in hepato- and Mao, L. 2006. Down-regulation of
carcinogenesis. J. Gastroenterol. Hepatol. 18: hepatoma-derived growth factor inhibits
1293–1301. anchorage-independent growth and invasion
Yoshida, K., Tomita, T., Okuda, Y., Yamamoto, S., of non-small cell lung cancer cells. Cancer
Enomoto, H., Uyama, H., Ito, H., Hoshida, Y., Res. 66: 18–23.
27
Hepatitis C Virus-Related Human
Hepatocellular Carcinoma: Predictive
Markers Using Proteomic Analysis
(Methodology)
Yasuhiro Kuramitsu and Kazuyuki Nakamura
343
344 Y. Kuramitsu and K. Nakamura
HSC71, HSP70.1, HSP60, glutamine syn- three protein spots of Mr 42,000 and pI
thetase, triosephosphate isomerase, ATP 6.4–6.8 on the 2-DE gel.
synthetase beta chain and alpha-enolase (7) Triosephosphate isomerase
(Takashima et al., 2003, 2005; Kuramitsu Triosephosphate isomerase is an enzyme of
et al., 2006). The functions of these protein glycolytic pathway, and catalyzes the reac-
spots are described below. tion between D-glyceraldehyde-3-phosphate
(1) GRP75 and dihydroxyacetone phosphate.
Glucose-regulated 75 kDa protein (GRP75) (8) Alpha-enolase
is located in the mitochondria, and inhibits Alpha-enolase is an isoenzyme of eno-
cell division and controls cell proliferation lase, a key protein that catalyzes the
and cellular aging. conversion of 2-phosphoglycerate to
(2) GRP78 phosphoenolpyruvate in the glycolytic
Glucose-regulated 78 kDa protein (GRP78) pathway. Three enolase isoenzymes
is a molecular chaperone. This molecule is have been identified. The alpha form is
involved in the unfolded protein response, present in most embryonic tissues, the
and plays a role in facilitating the assembly beta form is expressed in muscle tissues,
of multimeric protein complexes inside the and the gamma form is found only in
endoplasmic reticulum. neuronal tissues. Alpha enolase has been
(3) HSC70 detected not only in the cytoplasm but also
Heat shock cognate 71 kDa protein at the membrane surface. Although the
(HSC70) is a constitutively expressed mechanism of surface expression and the
chaperone that is also inducible by heat orientation on the membrane are not well
shock stress. It binds to nascent polypep- understood, it is known that the C-terminal
tides to facilitate correct folding and also amino acid, lysine, is exposed at the cell
participates in the formation and repair surface and is involved in binding to
of higher order protein structures such as plasminogen, which is then activated and
that of ATPase. converted to plasmin. Plasmin, stabilized
(4) HSP70.1 at the cell surface, induces fibrinolysis.
Heat shock 70 kDa protein 1 (HSP70.1) (9) Adenosine triphosphate (ATP) syn-
stabilizes preexistent proteins against thase beta chain
aggregation and mediates the folding of Adenosine triphosphate synthase is an
newly translated polypeptides within the enzyme that synthesizes ATP from adeno-
cytosol as well as within in cooperation sine diphosphate (ADP) and phosphate.
with other chaperones. Beta chain is a subunit of this enzyme.
(5) HSP60
Although heat shock 60 kDa protein
(HSP60) is one of the molecular chaper- Proteins With Decreased Expression
ones, this molecule does not belong to the By means of the analysis with 2-DE, 11
HSP70 family. spots with expression intensity decreased
(6) Glutamine synthetase less than half in cancerous tissues than
Glutamine synthetase is an enzyme to noncancerous tissues were detected, and
synthesize glutamine from glutamate and identified as aldolase, arginase 1, enoyl-CoA
NH3. This protein was identified from the hydratase, ketohexokinase, smoothelin,
348 Y. Kuramitsu and K. Nakamura
compared to that of noncancerous tis- Kim, W., Lim, S.O., Kim, J-S., Ryu, Y.H., Byeon,
sues. Each of the spots was matched to an J.Y., Kim, H.J., Kim, Y.I., Heo, J.S., Park,
equivalent spot on staining gels. To iden- Y.M., and Jung, G. 2003. Comparison of pro-
tify these four immunoreactive proteins, teome between hepatitis B virus- and hepatitis C
the spots were digested, and were identi- virus-associated hepatocellular carcinoma. Clin.
Cancer Res. 9: 5493–5500.
fied by MS as HSP70.1, glyceraldehyde
Kuramitsu,Y.,Harada,T.,Takashima,M.,Yokoyama,Y.,
3-phosphate dehydrogenase (GAPDH), Hidaka, I., Iizuka, N., Toda, T., Fujimoto, M.,
peroxiredoxin, and Mn-SOD (Takashima Zhang, X., Sakaida, I., Okita, K., Oka, M. and
et al., 2006). Figure 27.3 shows a scheme Nakamura, K. 2006. Increased expression, and
of the detection system for autoantibodies phosphorylation of liver glutamine synthetase
from HCC-patients sera. in well-differentiated hepatocellular carcinoma
tissues of patients infected with hepatitis C virus.
Electrophoresis 27: 1651–1658.
REFERENCES
Li, C., Tan, Y-X., Zhou, H. Ding, S.J., Li, S.J.,
El-Serag, H. B., and Mason, A. C. 1999. Rising Ma, D.J., Man, X.B., Hong, Y., Zhang, L.,
incidence of hepatocellular carcinoma in the Li, L., Xia, Q.C., Wu, J.R,, Wang, H.Y., and
United States. N. Eng. J. Med. 340: 745–750. Zeng, R. 2005. Proteomic analysis of hepa-
Ha, G.H., Lee, S.U., Kang, D.G., Ha, N.Y., Kim, titis B virus-associated hepatocellular carci-
S.H., Kim, J., Bae, J.M., Kim, J.W., and Lee, noma: Identification of potential tumor markers.
C.W. 2002. Proteome analysis of human stom- Proteomics 5: 1125–1139.
ach tissue: separation of soluble proteins by Okuda, K. 1992. Hepatocellular carcinoma: recent
two-dimensional polyacrylamide gel electro- progress. Hepatology 15: 948–963.
phoresis and identification by mass spectrom- Okuda, K. 2000. Hepatocellular carcinoma. J.
etry. Electrophoresis 23: 2513–2524. Hepatol. 32: 225–237.
350 Y. Kuramitsu and K. Nakamura
Paradis, V., Degos, F., Dargere, D., Pham, N., shock protein 70 family members as biomarkers
Belghiti, J., Degott, C., Janeau, J.L., Bezeaud, for hepatitis C virus-related hepatocellular carci-
A., Delforge, D., Cubizolles, M., Laurendeau, noma. Proteomics 3: 2487–2493.
I., and Bedossa, P. 2005. Identification of a new Takashima, T., Kuramitsu, Y., Yokoyama, Y.,
marker of hepatocellular carcinoma by serum Iizuka, N., Fujimoto, M., Nishisaka, T., Sakaida,
profiling of patients with chronic liver diseases. I., Okita, K., Oka, M., and Nakamura, K. 2005.
Hepatology 41: 40–47. Overexpression of alpha enolase in hepatitis C
Seow, T. K., Liang, R. C. Y., Leow, C. K., and virus-related hepatocellular carcinoma: asso-
Chung, M. C. M., 2001. Hepatocellular carci- ciation with tumor progression as determined by
noma: from bedside to proteomics. Proteomics proteomic analysis. Proteomics 5: 1686–1692.
1: 1249–1263. Takashima, T., Kuramitsu, Y., Yokoyama, Y.,
Simpson, R.J., Connolly, L.M., Eddes, J.S., Iizuka, N., Fujimoto, M., Sakaida, I., Okita, K.,
Pereira, J.J., Moritz, R.L., and Reid, G.E. 2000. Oka, M., and Nakamura, K. 2006. Proteomic
Proteomic analysis of the human colon carci- analysis of autoantibodies in patients with hepa-
noma cell line (LIM 1215): development of a tocellular carcinoma. Proteomics 6: 3894–3900.
membrane protein database. Electrophoresis 21: Wasinger, V.C., Cordwell, S.J., Cerpa-Poljak,
1707–1732. A., Yan, J.X., Gooley, A.A., Wilkins, M.R.,
Steel, L.F., Mattu, T.S., Mehta, A., Hebestreit, Duncan, M.W., Harris, R., Williams, K.L., and
H., Dwek, R., Evans, A. A., London, W.T., and Humphery-Smith, I. 1995. Progress with gene-
Block, T. 2001. A proteomic approach for the product mapping of the Mollicutes: Mycoplasma
discovery of early detection markers of hepato- genitalium. Electrophoresis 16: 1090–1094.
cellular carcinoma. Dis. Markers 17: 179–189. Yano, K., Yatsuhashi, H., and Yano, M., 2003.
Steel, L.F., Shumpert, D., Trotter, M., Seeholzer, Hepatitis C. Nippon Rinsho 61: 241–244.
S.H., Evans, A.A., London, W.T., Dwek, R., and Yatsuhashi, H., 2004. General remarks on chronic
Block, T.M. 2003. A strategy for the compara- hepatitis C. Nippon Rinsho 62: 400–404.
tive analysis of serum proteomes for the discov- Yokoyama, Y., Kuramitsu, Y., Takashima, T., Iizuka,
ery of biomarkers for hapatocellular carcinoma. N., Toda, T., Terai, S., Sakaida, I., Oka, M.,
Proteomics 3: 601–609. Nakamura, K., and Okita, K. 2004. Proteomic
Takashima, T., Kuramitsu, Y., Yokoyama, Y., Iizuka, profiling of proteins decreased in hepatocellular
N., Toda, T., Saiakida, I., Okita, K., Oka, M., and carcinoma from patients infected with hepatitis
Nakamura, K. 2003. Proteomic profiling of heat C virus. Proteomics 4: 2111–2116.
A. Diagnosis
28
Liver Metastases from Colorectal
Cancer: Ultrasound Imaging
Søren R. Rafaelsen
355
356 S.R. Rafaelsen
such as computed tomography (CT), mag- lobes provides a better signal-to-noise ratio
netic resonance imaging (MRI) and positron and reduces artifacts. Destructive sound-
emission tomography (PET), ultrasound wave effects from the body wall are also
has continued to play an important role. reduced. Choudhry et al. (2000) found that
Ultrasound has no ionizing radiation and in patients with a body mass index of 30 or
can be performed at low cost compared more tissue harmonic imaging was clearly
to the former modalities. With the advent better for lesion visibility and confidence
of intraoperative- and larparoscopic ultra- of diagnosis. The authors recommend rou-
sound, tissue harmonic imaging, color-, tine use of harmonic imaging for abdominal
power Doppler, contrast enhancement, pulse ultrasound studies in all adult patients.
inversion imaging, microflow imaging, 3D, However, for patients with diffuse fatty
and 4D ultrasound still have an important infiltration in the liver, penetration of the
role in diagnosing liver metastases as well low frequency ultrasound waves is often
as response evaluation, ultrasound guided better on conventional ultrasound images
biopsy and RFA. than when using tissue harmonic imaging.
Ultrasound guided biopsy is often used
to obtain cytological and histological veri-
ULTRASOUND SCANNING fication of liver metastases from colorectal
TECHNIQUE cancer prior to chemotherapy. Because of
the risk of needle-track tumor seeding,
Hepatic ultrasound is usually performed liver surgeons advise against percutane-
with the patient in the supine position. ous biopsy. A recent study has described
Turning the patient with the right side a negative effect on patient long-term
up may also be helpful in some cases. survival after liver resection. The 4-year
Images are obtained in sagital, transverse, survival was 32% in the group of patients
and oblique planes. The transducer is who had a preoperative liver biopsy versus
positioned in the right subcostal margin 47% in those without (Jones et al., 2005).
and by angling or moving the transducer In addition to biopsy assistance, ultrasound
it is possible to obtain images in various is used as needle guidance of percutaneous,
planes. Deep inspiration is often helpful intraoperative and larparoscopic RFA.
by displacing the liver caudally inferior Realtime virtual sonography enables
to the curvature. Intercostal scanning is the realtime demonstration of multiplanar
also used as a visualization technique. reconstructed CT imaging corresponding
Harmonic waves are generated from non- to the realtime ultrasound image side-by-
linear distortion of an acoustic signal as an side on the screen. The system uses an
ultrasound wave insonates tissues in the electromagnetic position sensor mounted
body. These harmonic ultrasound waves on the transducer. This can be helpful
are numeral multiples of a fundamental in ultrasound-guided RFA and biopsies.
transmitted frequency. Potential advan- The sonographical visualization of focal
tages of harmonic imaging include better lesions in a steatotic liver can be difficult.
axial resolution due to higher frequencies The association between hepatic steatosis
and improved lateral resolution due to nar- and oral 5-fluorouracil (5-FU) agents
rower beams. Decreasing noise from side is well described (Miyake et al., 2005).
28. Liver Metastases from Colorectal Cancer 357
exhibited specific tumoricidal activity and At present the measurements of the size
may damage the blood vessels within the of the metastases should incorporate the
tumor (Smyth et al., 2003). However, halo in the total diameter. It is essential
it might be too early to conclude that a to include the halo in resection of liver
natural tumor necrotic factor could be an metastases. In addition, the halo must be
explanation for a correlation between the encompassed in the ablation field in inter-
registered few central Doppler signals and stitial therapy studies. The neovascularity
a low incidence of local recurrence. in solid tumors can hinder thermal ablation
A large-scale follow-up study is required as well as large vessels surrounding the
to answer the important question of a pos- lesion. The hypervascularity present in the
sible relationship between flow pattern and margin of the metastasis can act as a heat
growth and response to chemotherapy. sink that limits the heating of tissue used in
In the future, power Doppler examina- radiofrequency ablation. In the flow study
tion of neovascularity of hepatic metastases local recurrence was observed in 66% of
could be a possible adjunct in the clinical the patients with a power Doppler signal in
evaluation of antiangioneogenic therapy the center but in none of the patients who
(Du et al., 2005). Contrast-enhanced ultra- had no such power Doppler signal. The
sound has currently been described as a tool clinical conclusion of this could be that
for assessing differences in the microcircu- attention should be paid to the possibility
lation of animal tumors in studies of antian- of local recurrence where hepatic resection
giogenic agents (Pollard et al., 2002). or tissue ablation is considered.
Bevacizumab (Avastin), the first approved It is concluded that the peripheral halo
therapy designed to inhibit tumor angio- in hepatic metastases indicates tissue with
genesis, has significant clinical benefits in increased vascularity. A central power
the management of colorectal cancer (Diaz- Doppler signal within the liver metas-
Rubio and Schmoll, 2005). tases implies the co-existence of more
With the advent of three-dimensional advanced disease. Further survival studies
power Doppler ultrasound it is now pos- are needed to assess the prognostic value
sible to visualize all detectable vessels of of this new finding and its usefulness as a
the metastasis in one 3-D image and thus tool for monitoring the possible effect of
reduce operator dependability at the same anti-angiogenic therapy.
time using contrast media (Forsberg et al., An in vitro study showed that ultra-
2002b). Furthermore, new contrast specific sound contrast microbubbles targeted
methods and contrast agents are now avail- via linkage with the tripeptide arginine-
able. Phase inversion imaging and flash- arginine-leucine could specifically adhere
echo imaging with intravenously infused to tumor angiogenic endothelium (Weller
microbubble contrast have shown promis- et al., 2005). Endothelial cells of angiogenic
ing results in measuring vascular flow and tumor vasculature are characterized by
tissue perfusion (Lucidarme et al., 2004). altered expression of molecular markers on
The new perfluro gasses replacing air and their surface. Targeted microbubbles may
multipulse scanning technique are very thus offer a noninvasive contrast-enhanced
sensitive in detecting the microbubble sig- ultrasound imaging technique for the func-
nals (Cosgrove and Blomley, 2004). tional imaging of tumor neovascularization
28. Liver Metastases from Colorectal Cancer 363
a standard ultrasound device. The remote Brannigan, M., Burns, P.N., and Wilson, S.R.
ultrasound scanning system agreed in at 2004. Blood flow patterns in focal liver lesions
at microbubble-enhanced US. Radiographics
least 80% of the 52 cases with the diagnosis
24:921–935.
obtained from conventional scanning Bruneton, J.N., Raffaelli, C., Balu-Maestro, C.,
(Courreges, 2005). Padovan I, B., Chevallier, P., and Mourou, M.Y.
Computer-aided detection may have the 1996. Sonographic diagnosis of solid liver nodules
potential to increase the accuracy of diag- in cancer patients. Eur. Radiol. 6:439–442.
nosis of focal liver lesions in ultrasound Choudhry, S., Gorman, B., Charboneau, J.W.,
Tradup, D.J., Beck, R.J., Kofler, J.M., and Groth,
images, although the literature is very
D.S. 2000. Comparison of tissue harmonic imag-
scarce (Yoshida, 2003). ing with conventional US in abdominal disease.
The development of new contrast agents Radiographics 20:1127–1135.
and ultrasound software still continues, Cosgrove, D., and Blomley, M. 2004. Liver tumors.
i.e., micro flow imaging. In the future, Abdom. Imaging 29:446–454
ultrasound might be adopted in various Courreges, F., Vieyres, P., Istepanian, R.S., Arbeille, P.,
and Bru C. 2005. Clinical trials and evaluation
drug delivery and other therapeutic appli-
of a mobile, robotic tele-ultrasound system. J.
cations. Microspheres and nanoparticles of Telemed. Telecare 11 Suppl 1:46–49.
different sizes with or without drugs, stem Diaz-Rubio, E., and Schmoll, H.J. 2005. The future
cells, and/or radioisotopes, can be used development of bevacizumab in colorectal can-
for drug delivery or diagnostic purposes. cer. Oncology 69 Suppl 3:34–45.
Ultrasound will then facilitate the deli- Du, W.H., Yang, W.X., Xiong, X.Q., Wang, X.,
Zhou, Y., and Wang H. 2005. Contrast-enhanced
very of chemotherapeutic drugs into liver
ultrasonographic imaging diagnosis on assess-
metastases and promote gene therapy to ment of vascularity in liver metastatic lesions.
targeted tissues under real-time imaging World J. Gastroentol. 11:3610–3613.
(Mitragotri, 2005). Ultrasound will pro- Easson, A.M., Barron, P.T., Cripps, C., Hill, G.,
vide information that will benefit patient Guindi, M., and Michaud, C. 1996. Calcification
management now and in the future. in colorectal hepatic metastases correlates with
longer survival. J. Surg. Oncol. 63:221–225.
Eberhardt, S.C., Choi, P.H., Bach, A.M., Funt,
REFERENCES
S.A., Felderman, H.E., and Hann, L.E. 2003.
Adam, R., Avisar, E., Ariche, A., Giachetti, S., Utility of sonography for small hepatic lesions
Azoulay, D., Castaing, D., Kunstlinger, F., found on computed tomography in patients with
Levi, F., and Bismuth, F. 2001. Five-year survival cancer. J. Ultrasound Med. 22:335–343.
following hepatic resection after neoadjuvant Faivre, J., Manfredi, S., and Bouvier, A.M. 2003.
therapy for nonresectable colorectal cancer. Ann. Epidemiology of colorectal cancer liver metas-
Surg. Oncol. 8:347–353. tases. Bull. Acad. Nath. Med. 187:815–823.
Bang, N., Nielsen, M.B., Rasmussen, A.N., Figueredo, A., Rumble, R.B., Maroun, J., Earle,
Osterhammel, P.A., and Pedersen, J.F. 2001. C.C., Cummings, B., McLeod, R., Zuraw, L.,
Hepatic vein transit time of an ultrasound con- and Zwaal, C. Gastrointestinal Cancer Disease
trast agent: simplified procedure using pulse Site Group of Cancer Care Ontario’s Program
inversion imaging. Br. J. Radiol. 74:752–755. in Evidence-based Care. 2003. Follow-up of
Bernatik, T., Becker, D., Neureiter, D., Hansler, patients with curatively resected colorectal can-
J.H., Frieser, M., Schaber, S., Wein, A., Hahn, cer: a practice guideline. Bio. Med. Central.
E.G., and Strobel, D. 2004. Hepatic transit Cancer 6;3:26
time of an echo enhancer: an indicator of meta- Forsberg, F., Piccoli, C.W., Liu, J.B., Rawool, N.M.,
static spread to the liver. Eur. J. Gastroenterol. Merton, D.A., Mitchell, D.G., and Goldberg, B.B.
Hepatol. 16:313–317. 2002a. Hepatic tumor detection: MR imaging
28. Liver Metastases from Colorectal Cancer 365
and conventional US versus pulse-inversion Guyot, F., Faivre, J., Manfredi, S., Meny, B.,
harmonic US of NC100100 during its reticu- Bonithon-Kopp, C., and Bouvier, A,M. 2005.
loendothelial system-specific phase. Radiology Time trends in the treatment and survival of
222:824–829. recurrences from colorectal cancer. Ann. Oncol.
Forsberg, F., Rawool, N.M., Merton, D.A., Liu, J.B., 16:756–761.
and Goldberg, B.B. 2002b. Contrast enhanced Hosten, N., Puls, R., Lemke, A.J., Steger, W.,
vascular three-dimensional ultrasound imaging. Zendel, W., Zwicker, C., and Felix, R. 1999.
Ultrasonics 40:117–122. Contrast-enhanced power Doppler sonography:
Furudoi, A., Tanaka, S., Haruma, K., Kitadai, Y., improved detection of characteristic flow pat-
Yoshihara, M., Chayama, K., and Shimamoto, F. terns in focal liver lesions. J. Clin. Ultrasound
2002. Clinical significance of vascular endo- 27:107–115.
thelial growth factor C and angiogenesis at the Jenkins, L.T., Millikan, K.W., Bines, S.D., Staren,
deepest invasive site of advanced colorectal E.D., and Doolas, A. 1997. Hepatic resec-
carcinoma. Oncology 62:157–166. tions for metastatic colorectal cancer. Am. Surg.
Gillams, A.R. 2005. The use of radiofrequency in 63:605–610.
cancer. Br. J. Cancer 92:1825–1829. Jensen, J.A., Holm, O., Jensen, L.J., Bendsen, H.,
de Goede, E., Filez, L., Janssens, J., and Van Nikolov, S.I., Tomov, B.G., Munk, P., Hansen, M.,
Cutsem, E. 1998. Follow-up of colon cancer: Salomonsen, K., Hansen, J., Gormsen, K.,
detection of liver metastases: benefit and perio- Pedersen, H.M., and Gammelmark, K.L. 2005.
dicity. Acta Gastroenterol. Belg. 61:8–10. Ultrasound research scanner for real-time synthetic
Goldberg, B.B., Merton, D.A., Forsberg, F., Liu, aperture data acquisition. IEEE. Trans. Ultrason.
J.B., and Rawool, N. 1996. Color Amplitude Ferroelectr. Freq. Control 52:881–891.
Imaging: Preliminary results using vascular Jones, O.M., Rees, M., John, T.G., Bygrave, S., and
sonographic contrast agents. J. Ultrasound Med. Plant, G. 2005. Biopsy of resectable colorectal
15:127–134. liver metastases causes tumour dissemination
Gordon, M.S., and Cunningham, D. Managing and adversely affects survival after liver resec-
patients treated with bevacizumab combination tion. Br. J. Surg. 92:1165–1168.
therapy. 2005. Oncology 69 Suppl 3:25–33. Jorgensen, T., and Rafaelsen, S. 1992. Gallstones
Graham, K.C., Wirtzfeld, L.A., MacKenzie, L.T., and colorectal cancer – there is a relationship,
Postenka, C.O., Groom, A.C., MacDonald, I.C., but it is hardly due to cholecystectomy. Dis.
Fenster, A., Lacefield, J.C., and Chambers, A.F. Colon Rectum 35:24–28.
2005. Three-dimensional high-frequency ultra- Khatri, V.P., Petrelli, N.J., and Belghiti, J. 2005.
sound imaging for longitudinal evaluation of Extending the frontiers of surgical therapy for
liver metastases in preclinical models. Cancer hepatic colorectal metastases: is there a limit?
Res. 65:5231–5237. J. Clin. Oncol. 23:8490–8499.
Green, B., Bree, R.L., Goldstein, H.M., and Stanley, Kruskal, J.B., Thomas, P., Nasser, I., Cay, O., and
C. 1977. Gray scale ultrasound evaluation of Kane, RA. 2000. Hepatic colon cancer in mice:
hepatic neoplasms: patterns and correlations. Dynamic in vivo correlation with hypoechoic
Radiology 124:203–208. rims visible at US. Radiology 215:852–857.
Gruenberger, T., Zhao, J., King, J., Chung, T, Clingan Lassau, N., Koscielny, S., Opolon, P., De Baere, T.,
P.R., and Morris, D.L. 2002. Echogenicity of Peronneau, P., Leclere, J., and Roche, A. 2001.
liver metastases from colorectal carcinoma is Evaluation of contrast-enhanced color Doppler
an independent prognostic factor in patients ultrasound for the quantification of angiogenesis
treated with regional chemotherapy. Cancer in vivo. Invest. Radiol. 36:50–55.
94:1753–1759. Leen, E., Angerson, W.J., Warren, H.W., O’Gorman, P.,
Gunven, P., Makuuchi, M., Takayasu, K., Moule, B., Carter, E.C., and Mc Ardel, C.S.
Moriyama, N., Yamasaki, S., and Hasegawa, H. 1994. Increased sensitivity of color Doppler
1985. Preoperative imaging of liver metastases. flow imaging of colorectal hepatic metastases
Comparison of angiography, CT scan, and ultra- using galactose microparticles: a preliminary
sonography. Ann. Surg. 202:573–579. report. Br. J. Surg. 8:252–254.
366 S.R. Rafaelsen
Leen, E., Goldberg, J.A., Robertson, J., Angerson, with low MI real time contrast enhanced sonogra-
W.J., Sutherland, G.R., Cooke, T.G., and phy and SonoVue. Ultraschall Med. 26:277–284.
McArdle, C.S. 1993. Early detection of occult Pollard, R.E., Sadlowski, A.R., Bloch, S.H., Murray,
colorectal hepatic metastases using duplex colour L., Wisner, E.R., Griffey, S., and Ferrara, K.W.
Doppler sonography. Br. J. Surg. 80:1249–1251. 2002. Contrast-assisted destruction-replenishment
Lucidarme, O., Nguyen, T., Kono, Y., Corbeil, J., ultrasound for the assessment of tumor microvas-
Choi, SH., Varner, J., and Mattrey, R.F. 2004. culature in a rat model. Technol. Cancer Treat.
Angiogenesis model for ultrasound contrast 1:459–470.
research. Acad. Radiol. 11:4–12. Rafaelsen, S.R., Kronborg, O., Larsen, C., and
Lundstedt, C., Ekberg, H., Hederstrom, E., Fenger, C. 1995. Intraoperative ultrasonography
Stridbeck, H., Torfason, B., and Transberg, K.G. in detection of hepatic metastases from colorec-
1987. Radiologic diagnosis of liver metastases tal cancer. Dis. Colon Rectum 38:355–360.
in colo-rectal carcinoma. Prospective evaluation Rafaelsen, S.R. and Solvig, J. 2004. Ultrasound
of the accuracy of angiography, ultrasonography, imaging of flow patterns in liver metastases
computed tomography and computed tomo- from colorectal cancer. Scand. J. Gastroenterol.
graphic angiography. Acta Radiol. 28:431–438. 39:761–765.
Marchal, G.J., Plyser, K., Tshibwabwa-Tumba, Roumen, R.M., Scheltinga, M.R., Slooter, G.D.,
E.A., Verbeken, E.K., Oyen, R.H., Baert, A.L., and van der Linden, A.W. 2005. Doppler per-
and Lauweryns, J.M. 1985. Anechoic halo in fusion index fails to predict the presence of
solid liver tumors. Radiology 156:479–483. occult hepatic colorectal metastases. Eur. J. Surg.
Maruyama, H., Matsutani, S., Saisho, H., Mine, Y., Oncol. 31:521–527.
Kamiyama, N., Hirata, T., and Sasamata, M. Smyth, M.J., Takeda, K., Haykawa, Y., Peschon,
2005. Real-time blood-pool images of contrast J.J., van den Brink, M.R., and Yagita, H. 2003.
enhanced ultrasound with Definity in the detec- Nature´s TRAIL – on a path to cancer immuno-
tion of tumour nodules in the liver. Br. J. Radiol. therapy. Immunity 18:1–6.
78:512–518. Tanaka, S., Ioka, T., Oshikawa, O., Hamada, Y.,
Milsom, J.W., Jerby, B.L., Kessler, H., Hale, and Yoshioka, F. 2001. Dynamic sonography of
J.C., Herts, B.R., and O’Malley, C.M. 2000. hepatic tumors. Am. J. Radiol. 177:799–805.
Prospective, blinded comparison of laparoscopic Taylor, R.H., Gilbert, J.M., Evans, M.G., Lane,
ultrasonography vs. contrast-enhanced com- H.G., and Cassell, P.G. 1984. Prospective serial
puterized tomography for liver assessment in liver ultrasound scanning in resectable colorec-
patients undergoing colorectal carcinoma surgery. tal cancer treated with adjuvant razoxane. Clin.
Dis. Colon Rectum 43:44–49. Exp. Metastasis 2:321–331.
Mitragotri, S. 2005. Healing sound: the use of Torzilli, G., Del Fabbro, D., Palmisano, A., Donadon,
ultrasound in drug delivery and other therapeutic M., Bianchi, P., Roncalli, M., Balzarini, L.,
applications. Nat. Rev. Drug Discov. 4:255–260. and Montorsi, M. 2005. Contrast-enhanced
Miyake, K., Hayakawa, K., Nishino, M., intraoperative ultrasonography during hepatec-
Morimoto, T., and Mukaihara, S. 2005. Effects tomies for colorectal cancer liver metastases.
of oral 5-fluorouracil drugs on hepatic fat con- J. Gastrointest. Surg. 9:1148–1154.
tent in patients with colon cancer. Acad. Radiol. Weller, G.E., Wong, M.K., Modzelewski, R.A.,
12:722–727. Lu, E., Klibanov, A.L., Wagner, W.R., and
Nies, C., Leppek, R., Sitter, H., Klotter, H.J., Riera, J., Villanueva, F.S. 2005. Ultrasonic imaging
Klose, K.J., Schwerk, W.B., and Rothmund, M. of tumor angiogenesis using contrast micro-
1996. Prospective evaluation of different diag- bubbles targeted via the tumor-binding pep-
nostic techniques for the detection of liver metas- tide arginine-arginine-leucine. Cancer Res.
tases at the time of primary resection of colorectal 65:533–539.
carcinoma. Eur. J. Surg. 162:811–816. Wernecke, K., Henke, L., Vassaloo, P., von
Oldenburg, A., Hohmann, J., Foert, E., Skrok, J., Bassewitz, D.B., Diederich, S., Peters, P.E., and
Hoffmann, C.W., Frericks, B., Wolf, K.J., and Edel, G. 1992. Pathologic explanation for hyp-
Albrecht, T. 2005. Detection of hepatic metastases oechoic halo seen on sonograms of malignant
28. Liver Metastases from Colorectal Cancer 367
liver tumors: An in vitro correlative study. Am. packet-based texture analysis for differentiation
J. Radiol. 159:1011–1116. between benign and malignant liver tumours
Yoo, S.K., Kim, D.K., Jung, S.M., Kim, E.K., in ultrasound images. Phys. Med. Biol. 48:
Lim, J.S., and Kim, J.H. 2004. Performance of a 3735–753.
web-based, realtime, tele-ultrasound consultation Yoshida, T., Matsue, H., Suzuki, M., Okazaki, N.,
system over high-speed commercial telecommu- Yoshino, M., Moriya, Y., and Hojo, K. 1989.
nication lines. J. Telemed. Telecare 10:175–179. Preoperative ultrasonography screening for liver
Yoshida, H., Casalino, D.D., Keserci, B., Coskun, metastases of patients with colorectal cancer.
A., Ozturk, O., and Savranlar, A. 2003. Wavelet- Jpn. J. Clin. Oncol. 19:112–115.
29
Preclinical Liver Metastases:
Three-Dimensional High-Frequency
Ultrasound Imaging
Kevin C. Graham, Lauren A. Wirtzfeld, James C. Lacefield, and Ann F. Chambers
369
370 K.C. Graham et al.
The variability in the timing of metastatic needs of each research project when select-
development, while still present, is no ing an imaging modality. Criteria specific
longer a factor in a longitudinal experiment to each research project include the type of
because the researcher can study metastatic metastasis model being used, the anatomical
progression within each individual mouse location at which metastases are expected,
and gather data or initiate a treatment the desired resolution, the availability of
based on the unique stage of progression equipment and funding, the necessity and
in each mouse. By decreasing the impact availability of contrast agents, and access
of experimental variability, smaller groups to personnel skilled in the acquisition and
of animals are required to achieve a statis- interpretation of the images.
tically significant result, thereby reducing For researchers who have access to MRI
both the cost and time required to perform facilities, it offers high-resolution imag-
an experiment. ing, inherent soft tissue contrast, and the
The importance of small animal imaging availability of a variety of contrast agents
in metastasis research is reflected by the for vascular imaging (Evelhoch et al.,
number of technologies dedicated to this task. 2000; Gillies et al., 2000). These features
Many of the technologies commonly used have led to the selection of MRI for pre-
clinically, such as magnetic resonance clinical studies involving the quantifica-
imaging (MRI), x-ray computed tomogra- tion of tumor burden in liver metastasis
phy (CT), and positron emission tomogra- models (Cai et al., 2005; Wu et al., 2003)
phy (PET), have been redesigned to give and in the measurement of tumor blood
spatial resolutions that are adequate for volume and vascular permeability through
small animal imaging (Weissleder, 2002). dynamic contrast enhanced (DCE) MRI
Other imaging technologies, because of (Leach, 2001). Current studies are extend-
their inherent limitations in imaging depth ing the use of MRI to study metastasis to
or their requirement for the genetic mani- the brain (Heyn et al., 2006). Widespread
pulation of the cell line being studied, are use of MRI to noninvasive study of cancer
not widely used clinically, but have been models is largely hampered by long image
developed for and utilized by small animal acquisition and processing times, and the
researchers. These technologies include high costs associated with equipment pur-
many of the optical techniques based on chase and operation (Weissleder, 2002).
fluorescent or bioluminescent proteins. X-ray CT offers high-resolution imag-
The general attributes that make an ing and inherent contrast between bone
imaging modality attractive for metastasis and soft tissue, but poor contrast within
research include high spatial resolution to the various soft tissues themselves. Due
allow metastatic progression to be tracked to these characteristics, CT has been pre-
from early stages, fast image acquisition dominately selected for studies of bone
to increase animal throughput, affordable metastasis (Paulus et al., 2000). The devel-
capital and operating costs, and inherent opment of novel iodinated contrast agents
contrast between the metastases and the which provide liver parenchyma – liver
normal parenchyma. No single imaging metastases contrast has introduced the
modality can fulfill all of these criteria, so possibility of using CT to study liver
the researcher must evaluate the particular metastasis models (Weber et al., 2004).
372 K.C. Graham et al.
This development may make CT useful morbidity and mortality in cancer patients.
for endpoint studies of tumor burden, but Preclinical models of liver metastasis are
longitudinal imaging may be limited due invaluable for our understanding of the
to repeated ionizing radiation exposure metastatic process and in the development
during imaging. of novel therapeutics to target this process.
Positron emission tomography is We have developed an experimental
extremely sensitive, yet suffers from poor metastasis model, injecting cultured tumor
resolution. The strength of PET imag- cells into the mesenteric vein, to produce
ing lies in the use of the specific radi- metastases in the liver of mice (Morris
oisotope-labeled tracers that can serve et al., 1993). This method can be utilized
as reporters for specific molecular proc- with a variety of cell lines, of both mouse
esses (Gambhir, 2002). As opposed to and human origin, from a variety of tumor
the anatomical imaging of MRI and CT, types. We opted to pursue an experimental
PET is able to image biological processes metastasis model to be able to control the
such as glucose metabolism or cellular number and timing of cells delivered to
proliferation by linking the radioisotope the liver. Since the timing of delivery was
to the appropriate tracer. By combining known, we were able to study the earliest
PET and CT, imaging information can be steps in the metastatic process. This led us
gleaned on both the anatomical location to identify a number of key biological and
and the biological activity of tumor cells molecular processes that govern metastatic
(Gambhir, 2002). efficiency in a secondary site (Chambers et
Similar to PET imaging, fluorescent al., 1995, 2002). The resolution required
and bioluminescent imaging have limited to study the early stages of metastasis was
spatial resolution, but are extremely ver- on the single cell level, which necessitated
satile in that the reporter gene or protein the use of high resolution in vivo videomi-
can be manipulated to provide a signal croscopy (IVVM). This is an inherently
in response to a variety of biological invasive procedure that is terminal for the
processes. Fluorescent and bioluminescent animal being studied.
imaging technologies are widely available Expanding on these studies, we are work-
and relatively inexpensive to purchase and ing to identify key biological and molecu-
operate. For these reasons, optical imag- lar processes that govern tumor growth at
ing has found a wide range of applications later stages of metastasis, with the ultimate
in the study of cancer models from the goal of using these discoveries to help
quantification of tumor burden to monitor- identify effective therapeutic strategies.
ing the delivery of cancer gene therapies With the focus on following the develop-
(Choy et al., 2003; Lyons, 2005). ment of micrometastases into tumors of a
A large portion of the metastasis research clinically relevant size, our studies require
in our laboratory has focused on the biology an imaging modality capable of longitu-
of the metastatic process with the liver as dinal imaging at sub-millimeter resolu-
our model organ (Chambers et al., 1995, tions. The imaging modality that best fit
2002). Liver metastases are common from the biological requirements of these studies
primary tumors in the colon and breast, was high-frequency ultrasound. High-
and are a frequent contributor to both frequency ultrasound offers cost effective,
29. Preclinical Liver Metastases: Three-Dimensional High-Frequency Ultrasound Imaging 373
high-resolution imaging with rapid image the 20–60 MHz range. Increasing the fre-
acquisition and processing, and allows for quency restricts imaging penetration to
the non-invasive longitudinal study of liver 5–15 mm, which is sufficient for many
metastasis models without the requirement preclinical applications in mouse models.
for contrast agents. Imaging penetration does limit clinical
use of high-frequency ultrasound to very
specialized applications, such as ocular
METHOD: HIGH-FREQUENCY or intravascular imaging, where the ultra-
ULTRASOUND sound transducer can be placed in close
proximity to the tissue of interest (Foster
Ultrasound images are produced by send- et al., 2000, 2002).
ing short duration acoustic pulses from The exact specifications of a high-fre-
the ultrasound transducer into the tissue of quency ultrasound system will depend on
interest. When the sound wave encounters the design and vary slightly between sys-
an interface between two tissues hav- tems. The system that we have utilized for
ing different acoustic properties, some of preclinical cancer studies is the Vevo 660
the energy is reflected back towards the produced by VisualSonics Inc. The design
transducer. The time it takes for the echo and performance of this scanner has been
to return to the transducer allows for the previously reviewed (Foster et al., 2002).
calculation of the distance to this interface. In brief, the Vevo 660 transducer operates
The resulting image displays the location at a 40 MHz center frequency with a 6 mm
and magnitude of sound reflection at these focal distance. The spatial resolution at the
interfaces. focus is 40 × 80 × 80 μm3.
In ultrasound imaging there is an Similar to the clinical situation, the
inherent tradeoff between the resolution acquisition of high quality preclinical
and depth of penetration. As the ultra- images with high-frequency ultrasound
sound frequency increases, the resolution depends on appropriate preparation of the
increases, while the imaging penetration subject to be imaged and the technical
decreases. For this reason, clinical ultra- experience of the sonographer. In order
sound is often performed in the range of to image the mouse liver, the abdomen
3–12 MHz, allowing for a resolution of must first be depilated with commercial
0.5–2.0 mm and an imaging depth in the hair removal cream. The mouse must be
range of 5–10 cm. For preclinical imag- lightly anesthetized to keep the animal
ing, mouse organs have linear dimen- immobile during imaging, a task most
sions that are approximately tenfold easily accomplished with an inhaled anes-
smaller than humans (Foster et al., 2002). thetic such as 1.5% isoflurane in oxygen.
To resolve the same tissue characteristics Unlike clinical imaging, which uses a
seen in the clinical setting, the preclinical very thin layer of ultrasound contact gel
ultrasound system must achieve a resolu- to allow transmission of sound waves into
tion that is approximately tenfold greater the body, when imaging a small animal a
than the clinical systems. This increase 0.5 cm thick layer of ultrasound contact
in resolution is achieved by increasing gel is used. This allows the transducer
the frequency of the sound waves into to slide easily over the curvature of the
374 K.C. Graham et al.
abdomen without losing contact with the shape and thus would be expected to give
gel. Ultrasound is strongly reflected by a more accurate measurement of tumor
bone and air-filled structures, which hin- volume. Indeed, previous work with a
ders imaging of any tissues located below clinical ultrasound system has shown that
bony structures or the lungs. In the evalu- three-dimensional imaging is more accu-
ation of the mouse liver, we have found rate than two-dimensional imaging for
that the liver is partially obscured by the phantoms of regular and irregular shapes
ribcage, but significant portions of the left (Tong et al., 1998; Xu et al., 2003).
lateral, left medial, and right medial liver To acquire three-dimensional ultrasound
lobes are accessible for imaging (Graham images with the Vevo 660, the transducer
et al., 2005). The liver lobes move in rela- is fixed to a stepper motor and acquires
tion to the bony anatomy and in relation two-dimensional images at 30 μm inter-
to each other over time. This movement vals as it moves in the out-of-plane dimen-
necessitates landmarks internal to each sion. The two-dimensional images are
liver lobe to be noted during longitudinal then interpolated and a three-dimensional
imaging. For liver metastases, landmarks image is reconstructed on-line (Fenster
such as the specific liver lobe in which et al., 2001). A typical 3-D image of
the metastasis develops, tumor shape, and the mouse liver containing a metastasis
proximity to readily identifiable structures of interest would take ∼20 s to acquire
such as major blood vessels and liver lobe and reconstruct. This rapid processing of
edges may be used to reliably identify an the imaging data is advantageous to the
individual metastasis over multiple imag- researcher as it allows the acquired image
ing sessions. to be evaluated and approved before mov-
ing on to the next subject. Such rapid
analysis is not possible with many other
THREE-DIMENSIONAL imaging modalities.
IMAGING AND VOLUME We have compared the calculated volume
CALCULATION of preclinical liver metastases after two-
dimensional measurement and assumption
A calculation of the volume of a region of an ellipsoid volume with the volume
of interest from two-dimensional ultra- calculated after segmentation of three-
sound images involves an assumption of dimensional images. We have shown that,
a defined three-dimensional shape. For for murine B16F1 melanoma liver metas-
instance, when calculating the volume of tases, the two-dimensional method on
a tumor in preclinical models, the tumor is average yields a smaller value than the
often assumed to have an ellipsoid volume three-dimensional method with a mean
that can be calculated after measurement percent difference of 8.8 ± 23.5% (Graham
of the length and diameter of the tumor. et al., 2005). The large standard deviation
The accuracy of this method is dependent indicates that the two-dimensional method
on how closely the tumor of interest can yield large overestimations or under-
matches the assumed shape. The ability to estimations of tumor volume in compari-
acquire three-dimensional images elimi- son to the three-dimensional method. To
nates the need to assume a predetermined determine which method is more accurate,
29. Preclinical Liver Metastases: Three-Dimensional High-Frequency Ultrasound Imaging 375
the true tumor volume of each metastasis variation (COV), the standard deviation
would have to be known. There is currently divided by the mean volume measure-
no method of measuring tumor volume that ment, for each set of volume measure-
is accepted as the “gold standard” to which ments. For B16F1 liver metastases with
other methods can be compared. Tumor volumes less than 2 mm3, the COV was
dimensions seen through non-invasive 6.8 ± 2.6% (Hastie et al., 2004). For
imaging may be compared to serial histo- HT-29 liver metastases, with volumes
logical sections to gain a sense of whether less than 2 mm3, the COV was 13 ± 5.7%
the two methods yield similar results, but (Hastie et al., 2004). The larger measure-
values obtained through histology cannot ment variability in the HT-29 model, in
be considered the true values because the comparison to the B16F1 model, may
extent of tissue distortion during histologi- reflect the more complex geometries of
cal processing is unknown. Without such the HT-29 metastases and the superior
a “gold standard”, the accuracy of any delineation of tumor boundaries in the
volume measurement cannot be known, but B16F1 model (Hastie et al., 2004).
it is reasonable to assume that the three- Since a measured change in tumor vol-
dimensional method is more accurate in ume is only significant if the change is
measuring tumor volume. The improved greater than the measurement variation,
accuracy of the three-dimensional method knowledge of measurement variability and
reflects the ability to measure complex tumor growth rate can be used to ensure
shapes, and the three-dimensional segmen- that imaging is only performed when
tation is done offline, providing more time significant changes in tumor volume are
for the operator to evaluate the images. In expected. In the case of B16F1 metastases,
the two-dimensional method, the operator with an average doubling time of 1.2 days
must quickly decide which imaging plane (Graham et al., 2005) and a measure-
contains the maximum tumor diameter, a ment variability of 6.8 ± 2.6%, signifi-
decision that may change between observers cant changes in tumor volume could be
or even for the same observer over multiple detected with daily imaging sessions. The
imaging sessions. HT-29 metastases show variable growth
For a three-dimensional imaging modal- rates, with doubling times between 3.7 and
ity to be useful for longitudinal studies 4.8 days (Graham et al., 2005) and a meas-
of tumor progression, the tumor volume urement variation of 13 ± 5.7%. In this
measurement must be accurate and con- case, daily imaging may be unnecessary,
sistent. The variability in repeated volume with more time necessary to ensure sig-
measurements of the same tumor will nificant changes in tumor volume between
dictate the minimum change in volume imaging sessions.
that can be measured. We have measured In studies of tumor growth or treatment
the intraobserver variability in repeated response, it is advantageous to use an
volume measurements of murine B16F1 imaging modality and volume measure-
melanoma and human HT-29 colon car- ment method with the smallest possible
cinoma cells growing as liver metastases. measurement variability so that small
Intraobserver measurement variability is changes in volume can be distinguished.
frequently reported by the coefficient of The reported coefficients of variation for
376 K.C. Graham et al.
tumor volume measurement with high- on the ultrasound images. The minimum
frequency ultrasound, between 6.8% and metastasis size that has been detected
13% depending on the tumor model, are by ultrasound imaging, and confirmed
similar to the clinical results obtained for through histological methods is (maxi-
the measurement of liver tumor volumes mum diameter → volume) 0.22 mm →
with MRI. It has been reported that quan- 0.01 mm3, 0.47 mm → 0.03 mm3, 0.66 mm
tification of tumor volumes from MRI by → 0.08 mm3 and 0.78 mm → 0.17 mm3
manual segmentation had an intraobserver for B16F1, HT-29, MDA-MB-435/HAL
COV of 9.3%, while volume measurement and PAP2 tumor, respectively (Graham
by point counting and stereological cor- et al., 2005). For the B16F1 model a
rection had a COV of 8.4% (Mazonakis large number of mice have been imaged
et al., 2004). Studies of measurement vari- throughout the early stages of metastasis
ation among preclinical imaging modali- development, thus the reported detection
ties are notably absent from the literature, threshold may be at or near the true detec-
which is an issue that needs to be addressed tion limit. For the other three models the
to ensure the proper use of these emerging absolute detection limit may be less than
technologies. the reported values, because only a limited
number of animals were imaged over vary-
ing time frames.
The experimental detection limits of
HIGH-FREQUENCY high-frequency ultrasound compare favo-
ULTRASOUND IMAGING rably to many of the other non-invasive
OF PRECLINICAL LIVER imaging modalities. It has been reported
METASTASES that contrast-enhanced micro-CT and
contrast-enhanced micro-MRI may detect
Using experimental metastasis models, we liver metastases as small as 0.3 mm in
have shown that high-frequency ultrasound diameter (Kobayashi et al., 2001; Weber
can reliably detect murine liver metastases et al., 2004). Non-invasive optical imaging
(Graham et al., 2005). We have shown that with green fluorescent protein may detect
ultrasound detection of liver metastases extremely small tumors, but imaging sen-
shows excellent agreement with post-mor- sitivity is dependent on tissue depth. For
tem analysis, including gross pathology liver metastasis models, we identified the
and histological sections (Figure 29.1). majority of the liver metastases at the
Ultrasound detection of liver metastases 2–3 mm depth range. It has been reported
was demonstrated with four cell lines of that for green fluorescent protein (GFP)-
different primary tumor origin, includ- labeled tumors at a depth of 2.2 mm, the
ing B16F1 (murine melanoma), HT-29 minimum tumor size detected is 1.86 mm
(human colon carcinoma), MDA-MB-435/ (Yang et al., 2000).
HAL (human breast carcinoma), and PAP2 It is important to note that unlike micro-
(murine oncogene-transformed fibrob- CT, micro-MRI or optical methods, the
lasts). With all four models, liver metas- imaging protocol to identify liver tumors
tases were hypoechoic, appearing darker by high-frequency ultrasound does not
than the surrounding liver parenchyma, involve the use of genetically encoded or
29. Preclinical Liver Metastases: Three-Dimensional High-Frequency Ultrasound Imaging 377
Figure 29.1. Identification of B16F1 (a) and HT-29 (b) liver metastases by non-invasive ultrasound
imaging. Two-dimensional ultrasound images show excellent agreement with corresponding histologi-
cal sections. The maximum diameter of the metastasis shown in the ultrasound image is 1.56 mm (a)
and 0.47 mm (b). Bar on the ultrasound images, 1.00 mm. Bar on histological sections, 1.00 mm (a) and
0.10 mm (b). (Reprinted with permission from Graham et al., 2005.)
exogenously given contrast agents. The contrast agent may also limit longitudi-
inherent tumor to liver parenchyma con- nal imaging. In the case of genetically
trast seen in ultrasound imaging offers encoded marker proteins, such as GFP, the
significant advantages in terms of animal necessity of introducing a foreign reporter
throughput and in the number of animal gene into each cell line to be studied may
models able to be used. For ultrasound limit the number of animal models that
imaging, each animal can be anesthetized, can be utilized. Additionally, the expres-
prepared, imaged, the images reconstructed sion of a foreign reporter protein can lead
and the content reviewed, and the animal to increased immunogenicity of the tumor
recovered in a time window of 15 min cell line being studied (Steinbauer et al.,
per animal. The use of a contrast agent, 2003).
usually delivered intravenously, may be The high-throughput, detection sensitivity
technically difficult and time consum- and the inherent tumor to liver parenchyma
ing, which decreases animal throughput. contrast makes high-frequency ultrasound
The clearance kinetics and toxicity of a uniquely suited for studying the longitudinal
378 K.C. Graham et al.
also detect and track tumor development (Goertz et al., 2003), as this technique is
from a much earlier stage than caliper less sensitive to image artifacts than color
measurement (Cheung et al., 2005). The Doppler (Lencioni et al., 1996).
continued development of tumors past Many cancer therapies induce apoptosis,
a size of 1 mm is thought to rely on the an active response to cellular injury that
formation of a neovasculature through the results in characteristic changes to the cell
process of angiogenesis. Thus, the ability nucleus and membrane, eventually leading
to detect and track small tumors may be to cell death. The ability to non-invasively
useful for studying biological processes, quantify apoptosis would be useful in pre-
such as angiogenesis, that occur early in clinical studies to assess the efficacy of
tumor development. novel cancer therapeutics. To date, quan-
Beyond the study of tumor growth, tification of apoptosis is largely limited to
angiogenesis research would benefit from biochemical analysis on harvested tissues,
non-invasive methods of quantifying vas- although single photon emission computed
cular parameters. For determination of tomography with radiolabeled annexin V
red blood cell velocity, Doppler ultra- has shown promise for non-invasive in vivo
sound may be used. Doppler ultrasound is analysis in the clinical setting (Blankenberg
based on the principle that when incident et al., 1998; Narula et al., 2001). For preclin-
ultrasound waves hit a moving object the ical research, the low resolution of SPECT
frequency of the return signal is shifted in imaging may be limiting, so alternative
proportion to the velocity of the object. In methods of quantifying apoptosis may be
color Doppler ultrasound, the frequency required. In a series of in vitro experiments,
shift in the return signal from moving red utilizing cell pellets of human acute myeloid
blood cells is displayed by mapping the leukemia cells, it has been demonstrated
frequency shifts to a blue-red color map that the structural changes associated with
with positive frequency shifts being red apoptosis may be detectable by analyzing
and negative frequency shifts being blue the radiofrequency backscatter signal in
in the resulting image. In an alternative high-frequency ultrasound (Kolios et al.,
method of processing Doppler frequency 2002; Tunis et al., 2005). It is not yet known
shift data, termed Power Doppler, the if the in vitro results will be translatable to
magnitude of frequency shift, and hence the more complex in vivo model, where the
red blood cell velocity, is ignored. In this presence of tissue structures other than the
method, all frequency shifts contribute tumor cells may affect the method of analysis
equally to the display data, meaning the (Kolios et al., 2002).
displayed value is related to the number
of red blood cells moving through a given
area. The utility of high-frequency power FUTURE DEVELOPMENTS –
Doppler ultrasound in the non-invasive ULTRASOUND CONTRAST
evaluation of anti-vascular therapies has AGENTS
been demonstrated in an intradermal
melanoma model (Goertz et al., 2002). It The use of high-frequency ultrasound in
is likely that Power Doppler will be fur- the identification of primary and metastatic
ther developed for high-frequency systems tumors, and in the longitudinal evaluation
29. Preclinical Liver Metastases: Three-Dimensional High-Frequency Ultrasound Imaging 381
of tumor progression, is well documented between 250 and 1,000 nm, have also been
(Graham et al., 2005; Wirtzfeld et al., investigated (Couture et al., 2006; Lanza
2005). These studies utilized the inherent et al., 1997).
ultrasound contrast, present between tumor Sensitivity of microbubble detection
and normal tissue, to delineate the tumors. may be further enhanced based on the
As outlined in this review, the longitudinal non-linear reflection of ultrasound by
evaluation of tumor growth allows for the microbubbles. This characteristic allows
study of a variety of processes, including the echoes produced from microbubbles
tumor dormancy and treatment response, to be differentiated from those of normal
and allows the researcher greater con- tissue using appropriate signal processing
trol when using models that exhibit large techniques. On clinical ultrasound systems
variability in the timing of tumor develop- this is often accomplished by transmitting
ment. Although the ability to study tumor at one frequency and receiving signals
growth based on inherent ultrasound con- at integer multiples of the transmit fre-
trast is clearly a significant development quency, referred to as harmonic imaging.
in its own right, the current development The strength of a nonlinear echo at a given
and application of novel contrast agents is frequency depends on the transmitted fre-
poised to expand the utility of preclinical quency and on the size of the microbubble.
ultrasound imaging in the fields of vas- For preclinical high-frequency ultrasound,
cular imaging, molecular imaging, and with a 40 MHz transmitting frequency, a
targeted drug delivery. microbubble will produce a much stronger
Inherent ultrasound contrast can be nonlinear echo at one-half of the trans-
enhanced by adding an agent to the imag- mit frequency, the subharmonic, than at
ing field of view that has an acoustic the higher harmonic frequencies. For this
impedance that differs from the tissue reason, subharmonic imaging may be
of interest. Intravenously injected micro- more sensitive than harmonic imaging
bubbles are commonly used to achieve for detecting microbubble contrast agents
this objective. Microbubbles are typically with preclinical imaging systems (Goertz
composed of a gaseous perfluorocarbon et al., 2005).
core surrounded by an outer phospholipid While Doppler ultrasound can provide
or albumin shell. The large difference in valuable information about tumor blood
acoustic impedance between the perfluor- flow, the tissue motion artifacts created
ocarbon gas and the surrounding tissue by the movement of internal organs cur-
leads to greatly enhanced ultrasound signal rently limit the utility of Doppler sys-
reflection and the ability to detect small tems. This limitation may be overcome
numbers of microbubbles in the ultrasound by the development of array transducers
image. Microbubbles are commonly pro- for preclinical systems or through the use
duced with diameters between 1 and 5 μm of intravenously injected microbubbles.
to allow them to pass freely through the Microbubbles enhance the signal from
circulatory system without being arrested blood in B-mode images, and hence are
by size restriction. Recently, the prop- less susceptible to artifacts from tissue
erties of much smaller contrast agents, motion. In addition, continuous imaging
liquid filled nanoparticles with diameters after the contrast agent is administered
382 K.C. Graham et al.
may also be applied to targeted gene Anwer, K., Kao, G., Proctor, B., Anscombe, I.,
delivery. For gene therapy, the applica- Florack, V., Earls, R., Wilson, E., McCreery,
T., Unger, E., Rolland, A., and Sullivan, S.M.
tion of ultrasound is known to increase
2000. Ultrasound enhancement of cationic
the transfection efficiency of traditional lipid-mediated gene transfer to primary tumors
cationic lipid based transfection complexes following systemic administration. Gene Ther.
(Anwer et al., 2000), although performing 7:1833–1839.
ultrasound in the presence of microbubbles, Blankenberg, F.G., Katsikis, P.D., Tait, J.F., Davis,
causing cavitation and disruption of cellular R.E., Naumovski, L., Ohtsuki, K., Kopiwoda,
S., Abrams, M.J., Darkes, M., Robbins, R.C.,
membranes may be more effective than
Maecker, H.T., and Strauss, H.W. 1998. In vivo
ultrasound alone (Lawrie et al., 2000). In detection and imaging of phosphatidylserine
both of these procedures the application of expression during programmed cell death. Proc.
ultrasound can be localized to a target site Natl. Acad. Sci. U S A 95:6349–6354.
to enhance therapeutic effect, while mini- Cai, S.R., Garbow, J.R., Culverhouse, R., Church,
mizing systemic toxicities. R.D., Zhang, W., Shannon, W.D., and McLeod,
H.L. 2005. A mouse model for developing treat-
ment for secondary liver tumors. Int. J. Oncol.
27:113–120.
CONCLUSIONS Chambers, A.F., MacDonald, I.C., Schmidt, E.E.,
Koop, S., Morris, V.L., Khokha, R., and Groom,
High-frequency ultrasound offers cost A.C. 1995. Steps in tumor metastasis: new con-
effective, non-invasive, high-resolution cepts from intravital videomicroscopy. Cancer
Metastasis Rev. 14:279–301.
imaging with rapid image acquisition and
Chambers, A.F., Groom, A.C., and MacDonald,
processing. These attributes make high- I.C. 2002. Dissemination and growth of can-
frequency ultrasound an attractive option cer cells in metastatic sites. Nat. Rev. Cancer
for the preclinical study of primary and 2:563–572.
metastatic cancer models. The longitudi- Cheung, A.M., Brown, A.S., Hastie, L.A., Cucevic,
nal study of primary tumor development V., Roy, M., Lacefield, J.C., Fenster, A., and
Foster, F.S. 2005. Three-dimensional ultrasound
and metastatic progression may provide
biomicroscopy for xenograft growth analysis.
further insight into dynamic processes Ultrasound Med. Biol. 31:865–870.
such as tumor dormancy and angiogen- Choy, G., Choyke, P., and Libutti, S.K. 2003.
esis. With the development of ultrasound Current advances in molecular imaging: non-
contrast agents for diagnostic and thera- invasive in vivo bioluminescent and fluores-
peutic purposes, high-frequency ultra- cent optical imaging in cancer research. Mol.
Imaging 2:303–312.
sound will be an invaluable tool in the
Couture, O., Bevan, P.D., Cherin, E., Cheung, K.,
preclinical characterization and evalua- Burns, P.N., and Foster, F.S. 2006. Investigating
tion of these agents, eventually leading to perfluorohexane particles with high-frequency
the clinical application of these emerging ultrasound. Ultrasound Med. Biol. 32:73–82.
technologies. Dayton, P.A., and Ferrara, K.W. 2002. Targeted
imaging using ultrasound. J. Magn. Reson.
Imaging 16:362–377.
REFERENCES Demicheli, R., Terenziani, M., and Bonadonna,
Albrecht, T., Hohmann, J., Oldenburg, A., Skrok, G. 1998. Estimate of tumor growth time for
J., and Wolf, K.J. 2004. Detection and charac- breast cancer local recurrences: rapid growth
terisation of liver metastases. Eur. Radiol. 14 after wake-up? Breast Cancer Res. Treat. 51:
Suppl. 8:P25–P33. 133–137.
384 K.C. Graham et al.
Ellegala, D.B., Leong-Poi, H., Carpenter, J.E., Fenster, A., Lacefield, J.C., and Chambers, A.F.
Klibanov, A.L., Kaul, S., Shaffrey, M.E., Sklenar, 2005. Three-dimensional high-frequency ultra-
J., and Lindner, J.R. 2003. Imaging tumor ang- sound imaging for longitudinal evaluation of
iogenesis with contrast ultrasound and micro- liver metastases in preclinical models. Cancer
bubbles targeted to alpha(v)beta3. Circulation Res. 65:5231–5237.
108:336–341. Hastie, L.A., Graham, K.C., Groom, A.C.,
Evelhoch, J.L., Gillies, R.J., Karczmar, G.S., MacDonald, I.C., Chambers, A.F., Fenster, A.,
Koutcher, J.A., Maxwell, R.J., Nalcioglu, O., and Lacefield, J.C. 2004. Variability of three-
Raghunand, N., Ronen, S.M., Ross, B.D., and dimensional high-frequency ultrasound measure-
Swartz, H.M. 2000. Applications of magnetic ments of small tumor volumes. IEEE Ultrason.
resonance in model systems: cancer therapeu- Sympos. Proc. 3:2185–2188.
tics. Neoplasia 2:152–165. Heyn, C., Ronald, J.A., Mackenzie, L.T.,
Fenster, A., Downey, D.B., and Cardinal, H.N. MacDonald, I.C., Chambers, A.F., Rutt, B.K.,
2001. Three-dimensional ultrasound imaging. and Foster, P.J. 2006. In vivo magnetic resonance
Phys. Med. Biol. 46:R67–R99. imaging of single cells in mouse brain with opti-
Foster, F.S., Pavlin, C.J., Harasiewicz, K.A., cal validation. Magn. Reson. Med. 55:23–29.
Christopher, D.A., and Turnbull, D.H. 2000. Holmgren, L., O’Reilly, M.S., and Folkman,
Advances in ultrasound biomicroscopy. J. 1995. Dormancy of micrometastases: bal-
Ultrasound Med. Biol. 26:1–27. anced proliferation and apoptosis in the pres-
Foster, F.S., Zhang, M.Y., Zhou, Y.Q., Liu, G., Mehi, ence of angiogenesis suppression. Nat. Med. 1:
J., Cherin, E., Harasiewicz, K.A., Starkoski, 149–153.
B.G., Zan, L., Knapik, D.A., and Adamson, S.L. Karrison, T.G., Ferguson, D.J., and Meier, P. 1999.
2002. A new ultrasound instrument for in vivo Dormancy of mammary carcinoma after mastec-
microimaging of mice. Ultrasound Med. Biol. tomy. J. Natl. Cancer Inst. 91:80–85.
28:1165–1172. Khanna, C., and Hunter, K. 2005. Modeling metas-
Gambhir, S.S. 2002. Molecular imaging of cancer tasis in vivo. Carcinogenesis 26:513–523.
with positron emission tomography. Nat. Rev. Kobayashi, H., Saga, T., Kawamoto, S., Sato, N.,
Cancer 2: 683–693. Hiraga, A., Ishimori, T., Konishi, J., Togashi,
Gillies, R.J., Bhujwalla, Z.M., Evelhoch, J., K., and Brechbiel, M.W. 2001. Dynamic micro-
Garwood, M., Neeman, M., Robinson, S.P., magnetic resonance imaging of liver micrometas-
Sotak, C.H., and Van Der Sanden, B. 2000. tasis in mice with a novel liver macromolecular
Applications of magnetic resonance in model sys- magnetic resonance contrast agent DAB-Am64-
tems: tumor biology and physiology. Neoplasia (1B4M-Gd)(64). Cancer Res. 61:4966–4970.
2:139–151. Kolios, M.C., Czarnota, G.J., Lee, M., Hunt,
Goertz, D.E., Yu, J.L., Kerbel, R.S., Burns, P.N., J.W., and Sherar, M.D. 2002. Ultrasonic spec-
and Foster, F.S. 2002. High-frequency Doppler tral parameter characterization of apoptosis.
ultrasound monitors the effects of antivascu- Ultrasound Med. Biol. 28:589–597.
lar therapy on tumor blood flow. Cancer Res. Krix, M., Kiessling, F., Vosseler, S., Farhan, N.,
62:6371–6375. Mueller, M.M., Bohlen, P., Fusenig, N.E.,
Goertz, D.E., Yu, J.L., Kerbel, R.S., Burns, P.N., and Delorme, S. 2003. Sensitive noninvasive
and Foster, F.S. 2003. High-frequency 3-D monitoring of tumor perfusion during antiang-
color-flow imaging of the microcirculation. iogenic therapy by intermittent bolus-contrast
Ultrasound Med. Biol. 29:39–51. power Doppler sonography. Cancer Res. 63:
Goertz, D.E., Cherin, E., Needles, A., Karshafian, 8264–8270.
R., Brown, A.S., Burns, P.N., and Foster, F.S. Lanza, G.M., Wallace, K.D., Fischer, S.E., Christy,
2005. High frequency nonlinear B-scan imag- D.H., Scott, M.J., Trousil, R.L., Cacheris, W.P.,
ing of microbubble contrast agents. IEEE Trans. Miller, J.G., Gaffney, P.J., and Wickline, S.A.
Ultrason. Ferroelectr. Freq. Control 52:65–79. 1997. High-frequency ultrasonic detection
Graham, K.C., Wirtzfeld, L.A., MacKenzie, L.T., of thrombi with a targeted contrast system.
Postenka, C.O., Groom, A.C., MacDonald, I.C., Ultrasound Med. Biol. 23:863–870.
29. Preclinical Liver Metastases: Three-Dimensional High-Frequency Ultrasound Imaging 385
Lawrie, A., Brisken, A.F., Francis, S.E., Cumberland, created by targeted microbubble destruction with
D.C., Crossman, D.C., and Newman, C.M. 2000. ultrasound. Circulation 98:1264–1267.
Microbubble-enhanced ultrasound for vascular Steinbauer, M., Guba, M., Cernaianu, G., Kohl,
gene delivery. Gene Ther. 7:2023–2027. G., Cetto, M., Kunz-Schughart, L.A., Geissler,
Leach, M.O. 2001. Application of magnetic reso- E.K., Falk, W., and Jauch, K.W. 2003. GFP-
nance imaging to angiogenesis in breast cancer. transfected tumor cells are useful in examining
Breast Cancer Res. 3:22–27. early metastasis in vivo, but immune reaction
Lencioni, R., Pinto, F., Armillotta, N., and precludes long-term tumor development studies
Bartolozzi, C. 1996. Assessment of tumor vascu- in immunocompetent mice. Clin. Exp. Metastasis
larity in hepatocellular carcinoma: comparison 20:135–141.
of power Doppler US and color Doppler US. Tong, S., Cardinal, H.N., McLoughlin, R.F.,
Radiology 201:353–358. Downey, D.B., and Fenster, A. 1998. Intra- and
Lindner, J.R. 2004. Microbubbles in medical imag- inter-observer variability and reliability of pros-
ing: current applications and future directions. tate volume measurement via two-dimensional
Nat. Rev. Drug Discov. 3:527–532. and three-dimensional ultrasound imaging.
Lyons, S.K. 2005. Advances in imaging mouse Ultrasound Med. Biol. 24:673–681.
tumour models in vivo. J. Pathol. 205:194–205. Tunis, A.S., Czarnota, G.J., Giles, A., Sherar, M.D.,
Mazonakis, M., Damilakis, J., Mantatzis, M., Hunt, J.W., and Kolios, M.C. 2005. Monitoring
Prassopoulos, P., Maris, T., Varveris, H., and structural changes in cells with high-frequency
Gourtsoyiannis, N. 2004. Stereology versus ultrasound signal statistics. Ultrasound Med.
planimetry to estimate the volume of malignant Biol. 31:1041–1049.
liver lesions on MR imaging. Magn. Reson. Turnbull, D.H., Ramsay, J.A., Shivji, G.S.,
Imaging 22:1011–1016. Bloomfield, T.S., From, L., Sauder, D.N., and
Miller, D.L., and Quddus, J. 2000. Diagnostic Foster, F.S. 1996. Ultrasound backscatter micro-
ultrasound activation of contrast agent gas bod- scope analysis of mouse melanoma progression.
ies induces capillary rupture in mice. Proc. Natl. Ultrasound Med. Biol. 22:845–853.
Acad. Sci. U S A 97:10179–10184. Udagawa, T., Fernandez, A., Achilles, E.G.,
Morris, V.L., MacDonald, I.C., Koop, S., Schmidt, Folkman, J., and D’Amato, R.J. 2002. Persistence
E.E., Chambers, A.F., and Groom, A.C. 1993. of microscopic human cancers in mice: altera-
Early interactions of cancer cells with the tions in the angiogenic balance accompanies loss
microvasculature in mouse liver and mus- of tumor dormancy. FASEB J. 16:1361–1370.
cle during hematogenous metastasis: videomi- Unger, E.C., Matsunaga, T.O., McCreery, T.,
croscopic analysis. Clin. Exp. Metastasis 11: Schumann, P., Sweitzer, R., and Quigley, R.
377–390. 2002. Therapeutic applications of microbubbles.
Narula, J., Acio, E.R., Narula, N., Samuels, Eur. J. Radiol. 42:160–168.
L.E., Fyfe, B., Wood, D., Fitzpatrick, J.M., Weber, S.M., Peterson, K.A., Durkee, B., Qi,
Raghunath, P.N., Tomaszewski, J.E., Kelly, C., C., Longino, M., Warner, T., Lee, F.T., Jr.,
Steinmetz, N., Green, A., Tait, J.F., Leppo, and Weichert, J.P. 2004. Imaging of murine
J., Blankenberg, F.G., Jain, D., and Strauss, liver tumor using microCT with a hepatocyte-
H.W. 2001. Annexin-V imaging for noninvasive selective contrast agent: accuracy is dependent
detection of cardiac allograft rejection. Nat. on adequate contrast enhancement. J. Surg. Res.
Med. 7:1347–1352. 119:41–45.
Paulus, M.J., Gleason, S.S., Kennel, S.J., Hunsicker, Weissleder, R. 2002. Scaling down imaging:
P.R., and Johnson, D.K. 2000. High resolution molecular mapping of cancer in mice. Nat. Rev.
X-ray computed tomography: an emerging tool Cancer 2:11–18.
for small animal cancer research. Neoplasia Welch, D.R. 1997. Technical considerations for
2:62–70. studying cancer metastasis in vivo. Clin. Exp.
Price, R.J., Skyba, D.M., Kaul, S., and Skalak, T.C. Metastasis 15:272–306.
1998. Delivery of colloidal particles and red Weller, G.E., Wong, M.K., Modzelewski, R.A.,
blood cells to tissue through microvessel ruptures Lu, E., Klibanov, A.L., Wagner, W.R., and
386 K.C. Graham et al.
Villanueva, F.S. 2005. Ultrasonic imaging of efficacy of the cytosine analogue 1-[2-C-cyano-
tumor angiogenesis using contrast microbubbles 2-deoxy-beta-D-arabino-pentofuranosyl]-N(4)-
targeted via the tumor-binding peptide arginine- palmitoyl cytosine (CS-682) in a liver-metastasis
arginine-leucine. Cancer Res. 65:533–539. athymic nude mouse model. Cancer Res.
Wirtzfeld, L.A., Wu, G., Bygrave, M., Yamasaki, 63:2477–2482.
Y., Sakai, H., Moussa, M., Izawa, J.I., Downey, Xu, H.X., Yin, X.Y., Lu, M.D., Liu, G.J., and Xu,
D.B., Greenberg, N.M., Fenster, A., Xuan, Z.F. 2003. Estimation of liver tumor volume
J.W., and Lacefield, J.C. 2005. A new three- using a three-dimensional ultrasound volumetric
dimensional ultrasound microimaging technol- system. Ultrasound Med. Biol. 29:839–846.
ogy for preclinical studies using a transgenic Yang, M., Baranov, E., Jiang, P., Sun, F.X., Li, X.M.,
prostate cancer mouse model. Cancer Res. 65: Li, L., Hasegawa, S., Bouvet, M., Al-Tuwaijri,
6337–6345. M., Chishima, T., Shimada, H., Moossa, A.R.,
Wu, M., Mazurchuk, R., Chaudhary, N.D., Spernyak, Penman, S., and Hoffman, R.M. 2000. Whole-
J., Veith, J., Pera, P., Greco, W., Hoffman, R.M., body optical imaging of green fluorescent pro-
Kobayashi, T., and Bernacki, R.J. 2003. High- tein-expressing tumors and metastases. Proc.
resolution magnetic resonance imaging of the Natl. Acad. Sci. U S A 97:1206–1211.
30
Colorectal Liver Metastases:
18
F-Fluorodeoxyglucose-Positron
Emission Tomography
Stéphanie Truant, Damien Huglo, and François-René Pruvot
387
388 S. Truant et al.
beginning of the use of FDG-PET, only host cell and tumor glucose metabolism
few studies fulfilled the inclusion cri- may alter appearance of tumor on FDG-
teria and the guidelines established by PET. Patients with diabetes mellitus may
the authors, reflecting the limitations of have an increased accumulation of FDG in
published reports (Huebner et al., 2000; skeletal muscle related to insulin adminis-
Kinkel et al., 2002). Over time, however, tration and a subsequent decreased uptake
studies on FDG-PET have focused on in malignant tissue, affecting both sensi-
more definite settings and have included tivity and specificity of FDG-PET imaging
more systematically histological verifica- in this patient population. In overweight
tions of FDG-PET findings, resulting in a patients, the significant amount of soft tis-
better assessment of FDG-PET accuracy sue can cause excessive scattering, and it
and impact. In this chapter, the yield of is recommended to increase the injected
FDG-PET in colorectal liver metastases activity and/or the acquisition time for
will be preferentially analyzed through both emission and transmission scans
these latter studies. to improve the image quality. Systemic
therapy such as chemotherapy may also
deteriorate FDG-PET accuracy by signifi-
PRINCIPLE OF FDG-PET cantly decreasing the activity of the key
AND GENERAL PITFALLS glycolytic enzyme hexokinase (Akhurst
et al., 2005). It is, therefore, recommended
FDG-PET is based on the increased uptake to delay FDG-PET scanning for 2–3 weeks
of glucose and, therefore, of FDG in after achieving chemotherapy. Mucinous
cancer cells. Many factors are involved adenocarcinoma and thus mucinous-
in FDG accumulation within a tumor or differentiated metastases use 18F-FDG at
metastases, such as the number of viable the same rate as normal surrounding tis-
tumor cells within a lesion, the amount of sues owing to the relative abundance of
necrotic tissue as well as regional blood mucin and the subsequent relative hypocel-
flow, and tumor hypoxia has proven to lularity of such tumors. The sensitivity of
increase FDG uptake. Besides, FDG is FDG-PET for detection of mucinous ade-
not a specific tracer of malignant cells. nocarcinoma is significantly lower than
The uptake of this tracer may occur in the nonmucinous ones (Whiteford et al.,
any condition associated with increased 2000). Finally, the risk of FDG-PET mis-
tissue metabolism, such as inflammatory interpretation may be aggravated by the
cells – as those observed in nonneoplas- lack of anatomic information. It has been
tic stroma tissues within tumor tissue shown that the recently-developed hybrid
– infection or healing. There also exist PET/CT machines, by better localizing
many sources of physiologic increased FDG uptakes, were able to reduce the rate
activity at FDG-PET, such as uptake in of inaccurate readings of FDG-PET alone.
brain, muscle, thyroid, digestive or urinary The knowledge of these pitfalls empha-
tract. This lack of FDG specificity can sizes the importance of correlating FDG
lead to its false-positive uptake in cancer uptake with clinical findings and results of
patients. Moreover, conditions that alter conventional imaging studies.
30. Colorectal Liver Metastases 389
higher accuracy for FDG-PET compared metastatic lesions were identified by the
with CT or other conventional diagnostic former. Accordingly, another meta-analy-
methods in detecting liver metastases in sis came to the conclusion that FDG-PET
recurrent colorectal cancer (Vitola et al., had a significantly higher sensitivity for
1996a, b). In a second meta-analysis, the detection of colorectal liver metastases on
mean sensitivity of FDG-PET (90%) was a per-patient basis compared with that of
significantly higher than that of ultrasound other modalities such as CT or MRI, but
(55%), CT (72%), and MRI (76%) in the showed comparable sensitivity on a per-
detection of liver metastases from color- lesion basis (Bipat et al., 2005).
ectal, gastric, and esophageal cancer, but This poor sensitivity of the liver is partly
any conclusion was difficult to draw due a result of the relatively high hexokinase
to comparison of 15-year range of publi- activity and thereby high FDG uptake
cation for CT with 3-year range for FDG- of normal hepatocytes. The liver usually
PET (Kinkel et al., 2002). The sensitivity displays a faint and homogeneous FDG
of FDG-PET was comparable to that of the uptake that is worsened by respiratory
most sensitive test for detection of hepatic motion leading to FDG smearing. Large
metastases, i.e., CT portography (90% vs. metastases are generally readily identified,
97%), with a much higher specificity – appearing as discrete foci of increased
expected by the high rate of false-positive activity in the liver or rings of increased
findings of CT portography (100% vs. 9%) activity in the case of necrotic centers.
(Vitola et al., 1996a, b). The adjunctive In contrast, small lesions < 10 mm can be
of intraoperative sonography also limited missed by FDG-PET due to its limited res-
impact on treatment selection in patients olution. The sensitivity of FDG-PET in the
screened preoperatively by FDG-PET. detection of liver metastases was confirmed
Over time however, with improve- to be related to lesion size: only 14–25%
ment in study designs, the superiority of hepatic lesions < 1 cm were detected
of FDG-PET over three phase CT or by FDG-PET in contrast with 80–100%
contrast-enhanced MRI in the detection of lesions > 1 cm, resulting in an overall
of colorectal liver metastases was not rate of undetected lesions of 30% (Fong
confirmed (Truant et al., 2005). This et al., 1999). Preoperative chemotherapy
was particularly true in studies using has proven to significantly alter tumor FDG
a systematic histological verification of uptake, resulting in rates of undetected
PET/CT discordances and an analysis lesions as high as 37% and possible missed
by lesion that appeared to be more lesions > 3 cm (Akhurst et al., 2005).
accurate than an analysis by patient. In FDG-PET is also unable to detect muci-
lesion-by-lesion analyses of resected liver nous-differentiated metastases owing to
specimens, FDG-PET demonstrated only the relative hypocellularity of such tumors.
70–79% of liver metastases, namely not Another major drawback of FDG-PET
better than conventional diagnostic meth- regarding liver metastases detection is the
ods (Truant et al., 2005). Rohren et al. lack of anatomic details, such as anatomi-
(2002), in comparing FDG-PET with sur- cal landmarks and relationship to hepatic
gical inspection and intraoperative ultra- vasculature. This method is often unhelp-
sound, found that only 79% of hepatic ful in evaluating anatomical resectability of
30. Colorectal Liver Metastases 391
that FDG-PET is a valuable tool for the of the abundance of the lymphocyte cells
detection of extrahepatic metastases. in the region, which are very glucose avid.
Benign causes of FDG accumulation also
Abdominal Cavity exist, such as fistulas, abscesses, diver-
In overall abdomen, FDG-PET has shown ticulitis, inflammatory bowel disease, and
high rates of sensitivity (63–100%), specifi- adenoma. As a rule, diffuse colonic FDG
city (98–100%), and accuracy (93–100%) uptake pattern, regardless of degree, is
(Flamen et al., 1999; Truant et al., 2005). predictive of normal findings at colonos-
FDG-PET was more sensitive than CT copy, whereas a segmental high uptake
(sensitivity 25–57%), while specificity may imply an inflammatory condition.
rates were comparable (CT specificity of In contrast, focal intense activity in the
91–99%). These results were conditioned colon is predictive of pathologic find-
by the prior knowledge of the specific ings and should be further evaluated with
pitfalls and artifacts encountered in the colonoscopy to exclude a neoplastic proc-
abdominal cavity owing to the many sources ess. Moderate focal FDG uptake can also
of physiologic and benign increased activ- be seen at stomas sites or in laparotomy
ity at FDG-PET, such as uptake in urinary scars. Such increased FDG uptake associ-
and gastrointestinal tract. ated with healing may persist for several
months. Because FDG is excreted in the
Physiologic and Benign Increased Activity of urine, intense FDG activity is encountered
FDG in the Abdominal Cavity
in the intrarenal collecting systems, ure-
Increased activity can be seen in most ters, and bladder. Other benign uptakes
segments of the digestive tract. The are possible, such as increased activity of
mechanism underlying this phenomenon FDG in the uterus during menstruation,
is unclear, partly due to smooth muscle inflammatory changes within the ovary
activity associated with peristalsis, meta- during ovulation, hyperventilation-induced
bolically active mucosa, and bacterial uptake in the diaphragm, or aortic FDG
uptake. Gastric activity can be seen and accumulation due to atherosclerosis. All of
has a characteristic J-shape. Uptake along these sites have potential for tumor recur-
the esophagus is also common, especially rence, and in the case of abdominal 18F-
in the distal portion and at the gastro- FDG uptake, correlation with clinical and
esophageal junction; the esophagus is best CT findings, as well as consideration for
identified on frontal and sagittal views. A biopsy, are mandatory. In case of inflam-
physiologic uptake can also occur in the matory or infectious processes, repeated
small bowel and is usually identified by FDG-PET, in two to four months or after
its pattern that has often a linear “string of a course of appropriate therapy (e.g., anti-
beads” appearance. However, more limited biotics), will often show an absence of or
bowel uptake, which forms discrete foci substantial decrease in uptake intensity at
of increased activity, can be impossible the site. The use of laparoscopy to assess
to distinguish from tumor deposits. The intra-abdominal burden before laparotomy
colonic activity is also variable. Uptake should also be recommended but is not
in the caecum and right colon is usually always possible, given existing adhesions
higher than in the rest of the colon because and the extent of previous surgery.
30. Colorectal Liver Metastases 393
Figure 30.1. A 68-year-old man with prior colectomy for carcinoma presented with two liver metastases.
Preoperative FDG-PET (b, c) showed not only the two already-known liver metastases but also a small
left pelvic uptake, suggesting local recurrence. In contrast, conventional workup by pelvic CT (a) and
transrectal endoscopic ultrasonography was negative. The patient underwent an exploratory laparotomy
that provided a histological confirmation of local recurrence
with the distorted pelvic anatomy. In 62 were 98%, 96%, 90%, 97%, and 93% for
patients, sensitivity, specificity, positive PET/CT and 82%, 65%, 73%, 75%, and
predictive value, negative predictive value 74% for PET, respectively (Even-Sapir
and accuracy for differentiating malignant et al., 2004). In that study, the most common
from benign 18F-FDG uptake in the pelvis cause for false-positive interpretation of
30. Colorectal Liver Metastases 395
Figure 30.2. A 63-year-old man, 6 months after colectomy for advanced adenocarcinoma of the colon
(pT3N2M1) was referred for FDG-PET (b, c, d) for preoperative staging of synchronous liver metastases
after receiving neoadjuvant chemotherapy. In contrast to abdominal CT (a), FDG-PET showed carcinosis
that was confirmed by surgical biopsies and contraindicated hepatectomy
30. Colorectal Liver Metastases 397
Figure 30.3. A 46-year-old man underwent FDG-PET as preoperative staging of colorectal liver metas-
tases. FDG-PET showed multiple liver metastases in the right lobe in association with a right supracla-
vicular uptake that was related to a metastatic lymph node at sonography
by upstaging disease from resectable to while FDG-PET did not confer any advan-
unresectable, reducing futile morbidity or tage over conventional diagnostic methods
seldom performance of more extensive sur- in the assessment of hepatic metastases.
gery by disproving a positive CT finding. Consequently, FDG-PET improved
It may lead to commencement of palliative resectability rates. In two series (Strasberg
chemotherapy, resulting in improved out- et al., 2001; Truant et al., 2005), 95% of
come or much better palliation. Integration patients who underwent laparotomy actu-
of FDG-PET in the presurgical evaluation ally underwent hepatic resection, similar
of patients with liver metastases from to that achieved by the use of staging
colorectal cancer has also been shown to laparoscopy to detect unresectable tumors
substantially reduce overall costs. Due to before laparotomy. In comparison, 20% of
the lack of randomized controlled clinical patients who underwent exploration on the
trial to assess FDG-PET impact, only stud- basis of conventional imaging were found
ies that focused on preoperative staging of to have unresectable disease at the time
liver metastases have been considered for of surgery. On the other hand, in these
this part of the analysis. series, FDG-PET also falsely upstaged and
understaged up to 7% and 8% of patients,
respectively (Topal et al., 2001; Truant
Additional Disease and Change of
et al., 2005) and showed very poor yield
Management
in patients who were on chemotherapy
When FDG-PET was added to a complete at the time of FDG-PET (Akhurst et al.,
conventional work-up in preoperative 2005). In the preoperative staging of liver
planning for patients with hepatic metas- metastases judged resectable on the basis
tases, additional sites of disease were of conventional imaging findings, FDG-
identified in 9.5–25% of patients (Truant PET resulted in a change of management
et al., 2005). FDG-PET contribution came in 10–25% (Ruers et al., 2002). Selzner
essentially from the discovery of extra- et al. (2004) studied whether the novel
hepatic sites, mainly in the abdomen, PET/CT imaging modality impacted the
30. Colorectal Liver Metastases 399
treatment of patients with metastatic color- before surgical exploration because these
ectal cancer of the liver. As for FDG-PET patients have a low rate of detection of
alone, the main impact of PET/CT came extrahepatic disease and even a non-
from the discovery of extrahepatic disease. negligible rate of false positive findings
New findings in the PET/CT resulted in a (Schussler-Fiorenza et al., 2004; Truant
change in the therapeutic strategy in 21% et al., 2005). As FDG-PET is costly and
of the patients. Therefore, the impact of not widely available, FDG-PET should be
FDG-PET in the particular setting of pre- restricted to high risk patients in centers
sumed resectable liver metastases seems with low availability of FDG-PET in order
to be less than the mean impact of 29% to delay a surgery.
reported by Huebner et al. (2000) in their
meta-analysis on FDG-PET contribution
in overall recurrent colorectal cancer. Survival Impact
The major part of the additional diag- The addition of FDG-PET imaging to pre-
nostic value of FDG-PET resided in surgical evaluation may improve survival
the subgroup of patients at high risk for by appropriately eliminating ineffective
unresectable disease, as predicted by the surgery in patients with inoperable dis-
Memorial Sloan Kettering Cancer Centre ease. Results are, however, highly contro-
clinical risk score based on five criteria: versial regarding the survival benefit of
(1) node positive primary, (2) disease-free FDG-PET. It has been reported that the
interval from primary to liver metastases use of preoperative FDG-PET lessened
< 12 months, (3) largest hepatic tumour the recurrence rate in patients undergo-
> 5 cm, (4) number of hepatic tumours > 1 ing hepatic resection for colorectal liver
(5) CEA level >200 ng/ml (Fong et al., metastases by detection of disease not
1999). Among patients with a low clinical found on conventional imaging. This had a
risk score, 0–3% of patients had additional significant impact on the survival of these
disease detected by FDG-PET, compared patients. In 100 patients having hepatic
with 14–17% for patients with a high clini- resection between March 1995 and June
cal risk score (Schussler-Fiorenza et al., 2002, the 5-year disease-free and actu-
2004; Truant et al., 2005). Thus, patients arial overall survival in patients screened
with large (> 5 cm), multiple and synchro- with FDG-PET was 34.8% and 58%,
nous metastases were more likely to have respectively, which was higher than previ-
extrahepatic disease detected by FDG-PET. ously reported (Fernandez et al., 2004).
In 40 patients being considered for hepatic However, this series included 63% patients
resection but with a high clinical risk score, with a single tumor while in other series,
FDG-PET altered or influenced the clinical resections of solitary liver metastases rep-
management in 16 patients (40%) (Fong et resented the minority of cases and the
al., 1999). This FDG-PET impact is higher resectioning of ten or more tumors was not
than that reported in the aforementioned uncommon. Another study did not confirm
series without consideration of the clini- this survival benefit: 3-year overall and
cal risk score. Some authors now consider disease-free survival rates after exclud-
that patients at low risk of metastatic ing the patients who did not undergo liver
disease should not undergo FDG-PET resection were comparable with those for
400 S. Truant et al.
all patients (51% vs. 47% and 27% vs. However, rising CEA often occurs in asso-
24%, respectively) (Topal et al., 2001). ciation with negative or equivocal results
Most patients selected for surgery by of conventional imaging studies, leading
FDG-PET indeed still had recurrence of to the diagnosis of tumor recurrences at
disease because of the presence of dis- a non-resectable stage. The advantage
ease that was not detected by FDG-PET. of FDG-PET over anatomic imaging to
In the study by Fong et al. (1999), after screen the whole-body in one examination
a relatively short median follow-up (7 session was assumed to be of particular
months), 40% patients had a recurrence value in this setting. Due to the lack of
after liver resection, whereas other series specific study, FDG-PET contribution in
have reported recurrences as early as 3 the postoperative follow-up of resected
months after resection (Strasberg et al., liver metastases can only be extrapolated
2001). Although a fraction of these find- from studies on patients with suspected
ings are likely due to disease that is below recurrent colorectal cancer on the basis of
the level of detection by FDG-PET, at least rising CEA.
some of these early recurrences are due to In asymptomatic patients with a rising
true–false negative findings on FDG-PET. CEA and nondiagnostic radiologic workup,
On the other hand, patients in whom FDG- the sensitivity of FDG-PET for detection of
PET provided additional information had a recurrence has been reported as 79–100%,
dismal prognosis with a 3-year overall sur- specificity as 50–83% and accuracy as
vival of 0% that could be explained by the 74–93% on a patient-by-patient analysis
extent of their disease (Topal et al., 2001). (Flanagan et al., 1998). FDG-PET was
This may be the true survival impact of more sensitive than both CT and serum
FDG-PET: identifying the small number CEA for detection of disease recurrence,
of patients that have no chance for cure except for the liver, and the specificity of
despite reassuring conventional imaging. FDG-PET was equivalent to that of CEA.
FDG-PET showed high positive predictive
value (89–97%) and negative predictive
Postoperative Follow-Up After Resection
value (71–100%) (Flanagan et al., 1998).
of Colorectal Liver Metastases and
FDG-PET was also more accurate than
Rising Carcinoembryonic Antigen
conventional diagnostic methods in pre-
(CEA)
dicting the resectability or non-resectabil-
After hepatic resection, recurrences occur ity of the recurrence, though it did show
in up to 60% of patients, mostly within 2 a low sensitivity on a lesion-by-lesion
years of surgery and with an equal distri- analysis (57%) in 28 patients who under-
bution between hepatic and extrahepatic went second-look laparotomy (Libutti
sites. It is assumed that recurrences diag- et al., 2001). By contrast, in patients with
nosed by systematic follow-up of asymp- normal CEA levels and symptoms, FDG-
tomatic patients are more often resectable PET showed a high rate of false-positive
with a curative intent. Serial determina- results and was inaccurate in predicting
tions of the plasma CEA concentration is recurrence. In the postoperative follow-up
the most frequently used method for the of liver metastases, specific pitfalls have
detection of such asymptomatic recurrences. to be known such as FDG uptake in the
30. Colorectal Liver Metastases 401
scar of previous partial hepatectomy in apy outcome earlier than 2 months after
the case of important inflammatory reac- onset to therapy because resolution of
tion, e.g., postoperative biloma. Studies therapy-induced anatomic changes lags
are now needed to address the impact of behind tumor cell mortality. As metabolic
early diagnosis by FDG-PET on survival changes induced by chemotherapy precede
and its cost-effectiveness in the postopera- the morphological changes, 18F-FDG-PET
tive follow-up of resected liver metastases was assumed to predict response to therapy
before large-scale introduction in clinical at an earlier stage than conventional diag-
practice can be advocated. nostic methods. Evaluation of response
to therapy has involved careful compari-
son of pretreatment and post-treatment
FDG-PET for Monitoring the Response FDG uptake, as assessed by qualitative
to Systemic or Regional Therapy of visual interpretation, standardized uptake
Colorectal Liver Metastases value (SUV) or extensive kinetic analysis.
Hepatic metastases can be treated with Studies performed in patients with hepatic
a curative or palliative intent with sys- metastases from colorectal cancer demon-
temic therapy (chemotherapy, new tar- strated a relatively rapid decline (within
geted therapies) or regional therapy to 72 h) in the SUV in responding tumors after
the liver, including local ablative therapy a single infusion of 5-fluorouracil and foli-
(radiofrequency ablation, cryoablation), nic acid, whereas nonresponding tumors
chemotherapy administered through the showed an increase, no change, or only a
hepatic artery using infusion pomps, selec- small decline in 18F-FDG uptake. Using
tive chemoembolization or radiolabeled SUV in tumors 4–5 weeks after the start
90
Y-microspheres. of treatment with 5-fluorouracil (Findlay
et al., 1996) or after FOLFOX therapy
Anticancer Systemic Therapy (Dimitrakopoulou-Strauss et al., 2003) fur-
ther predicted the nonresponse with a high
Systemic Chemotherapy
accuracy. Quantitative, dynamic FDG-PET
Most patients with colorectal cancer meta- also identified patients not responsive to
static to the liver will receive chemotherapy, treatment with the FOLFOX regimen,
using mostly 5-fluorouracil in combina- although the aggressiveness of the method
tion chemotherapy regimen including (multiple arterial samplings), its sensitiv-
oxaliplatin (FOLFOX) and/or irinotecan ity to the noise and the long time of the
(FOLFIRI). When liver metastases are dynamic acquisition (1 h on a limited field)
non-resectable, chemotherapy alone may are limiting for its use in carcinologic
supply a significant survival benefit pro- domain (Dimitrakopoulou-Strauss et al.,
viding that patient is a responder. Accurate 2003).
prediction of treatment failure may allow FDG-PET negativity following exten-
the oncologist to apply an alternative treat- sive chemotherapy does not, however,
ment regimen without subjecting the patient necessarily mean complete tumor destruc-
to the toxicity of the full treatment scheme. tion because injured but alive tumor cells
Monitoring with radiologic imaging meth- may not accumulate FDG to the same
ods like CT and MRI cannot assess ther- extent as untreated cells. Conversely,
402 S. Truant et al.
a marked increase in 18F-FDG metabolism monitoring the response during the course
has been described in lesions responding of therapy, further clinical trials are needed
after initiation of chemotherapy, because to assess the beneficial effect of early
of major inflammatory response and tran- FDG-PET scanning on the treatment of
sient increase in stromal reaction, and patients and the ultimate outcome. The
may result in overestimation of the frac- method of quantifying tumor 18F-FDG
tion of viable cells. In a tumor mouse uptake, through visual or semiquantitative
model treated by chemotherapy, the FDG methods such as SUV or kinetic analysis,
uptake was stabilized by day 8–10 despite and remain challenging, as well as the
a further reduction in viable tumor cell exact timing of FDG-PET.
fraction (Spaepen et al., 2003). In 18
patients with colorectal liver metastases,
New Targeted Therapies (Monoclonal
the reduction in FDG uptake 1–2 weeks
Antibodies)
after onset of chemotherapy was not sig-
nificantly different between the respond- Systemic chemotherapy is now used in
ing and non-responding groups and was association with new targeted therapies,
false-positive for response in 20% of the such as the monoclonal antibodies bev-
patients (Findlay et al., 1996). Optimal acizumab and cetuximab that specifi-
timing of FDG-PET is thus crucial and cally block the receptors of the vascular
imaging 1–2 weeks after completion of endothelial growth factor (VEGF) and
chemotherapy is recommended to avoid of the epidermal growth factor (EGF),
transient fluctuations in tumor FDG respectively. The VEGF stimulates the
uptake confounding response assessment. proliferation and migration of vascularly
These data suggest that a late 18F-FDG derived endothelial cells and is overex-
examination, namely at 4–5 weeks after pressed in colon cancer. Several studies
the start of treatment, may be better for have suggested that FDG-PET could allow
therapy outcome. FDG-PET may further follow-up of patients treated by an angio-
be of prognostic value in patients under- genesis inhibitor. Endothelial cells have
going chemotherapy. The first FDG-PET indeed proven to contribute to 18F-FDG
study prior to onset of chemotherapy was uptake, with a positive correlation between
predictive for the therapy outcome, with SUV and microvessel density. Recently,
pretreated tumors displaying a low FDG FDG-PET showed an increased metabolic
uptake more likely to have a worse prog- activity in colorectal liver metastases fol-
nosis after chemotherapy than those with lowing removal of the primary tumor, as
a high SUV (Dimitrakopoulou-Strauss et reflected by a rise in FDG uptake. As this
al., 2003). An early response to treatment phenomenon coincided with a decrease in
with 5-fluorouracil or FOLFOX regimen, serum anti-angiogenic factor levels, the
as shown on FDG-PET, was also well cor- authors came to the conclusion that the
related with the ultimate response seen in primary tumor suppressed angiogenesis in
clinical, radiographic or pathologic find- its distant metastases and that its removal
ings a few weeks or months later. resulted in an increase in metabolic activity
Despite interesting findings of sev- in liver metastases, emphasizing the inter-
eral studies investigating FDG-PET for est of angiogenesis inhibitors in colorectal
30. Colorectal Liver Metastases 403
cancer (Peeters et al., 2005). In the only of recurrence with follow-up FDG-PET,
series of seven patients treated with beva- whereas MRI and CT showed only 43%
cizumab and irinotecan for colorectal liver and 50% of these recurrences, respectively
metastases, FDG-PET findings correlated (Anderson et al., 2003). In lesions with
better than those of CT with pathology a positive FDG-PET after treatment, the
(Goshen et al., 2006). Further studies are positive predictive value for the detection
needed to confirm these results, as well of local recurrence was as high as 80%;
as studies on FDG-PET in the evaluation the negative predictive value was 100%
of treatment of colorectal liver metastases (Langenhoff et al., 2002). The authors
with cetuximab. advised to administer a second ablative
treatment on any lesions that remain posi-
Regional Therapies tive. The dual-modality PET/CT, that was
comparable to FDG-PET alone in terms of
Local Ablative Therapy (Cryosurgery sensitivity (65%) in a series of 11 patients
Ablation and Radiofrequency Ablation)
(Veit et al., 2006), might be of great value
Local ablative techniques are used as for guidance of reinterventions. However,
adjuncts to chemotherapy or hepatic resec- detection of residual tumor remains chal-
tion in a curative or palliative intent when lenging, as increased glucose metabolism
complete resection of tumor in the liver due to tissue regeneration surrounding
cannot be achieved. These techniques can the ablative necrosis or abscess formation
be performed laparoscopically or during in the treated area may superimpose on
open surgery, or for radiofrequency abla- small areas of residual tumor, resulting
tion alone by a percutaneous procedure in false-negative FDG-PET results. Some
guided by CT or sonography. In absence authors recommended follow-up FDG-
of evolving extrahepatic lesions, as con- PET directly after radiofrequency abla-
firmed by pretherapy FDG-PET, these tion or at least within 24 h to avoid these
local ablative therapies have been associ- artefacts (Veit et al., 2006). Further studies
ated with improved median survival rates. are needed to confirm these results and to
However, the frequency of incomplete determine the best timing of FDG-PET.
treatment, and thus local recurrence, is still
relatively high due to limitations in ade-
quately monitoring the destruction process Hepatic Arterial Chemotherapy
at the time of treatment. Moreover, after Patients with unresectable liver metastases
local ablative therapy, local recurrences can be proposed hepatic arterial infusional
in the liver are not easily identified with chemotherapy through an hepatic artery
CT, except at advanced stage of relapse, pump placed during a short laparotomy.
because of difficulties in differentiating This treatment provides at least equivalent
viable tumor tissue from adjacent necrosis or even increased response rate for color-
or therapy-induced hyperperfusion. ectal liver metastases when compared to
The time of detection of recurrence by intravenous chemotherapy. It is, however
FDG-PET has proven to be earlier than contraindicated in the presence of extrahe-
the detection by CT alone. Comparative patic disease due to limited systemic effi-
studies reported a fully accurate prediction ciency. Because patients with unresectable
404 S. Truant et al.
liver metastases have a high risk of occult benefit of resected patients needs to be
extrahepatic disease, FDG-PET may spare confirmed. In the postoperative follow-up
patients an unnecessary laparotomy in this of resected liver metastases, FDG-PET
particular setting. may be useful in locating recurrences in
asymptomatic patients with a rising CEA
Other Regional Therapies and nondiagnostic radiologic workup,
although a specific study is still needed in
Some authors have studied the effects of
this setting, as well as an analysis of the
regional chemoembolization therapy using
impact of early diagnosis by FDG-PET
FDG uptake as a criterion and found that
on survival. Results of preliminary studies
decreased FDG uptake correlated with
have also suggested the relevance of
response, while presence of residual uptake
FDG-PET in monitoring the response to
was used to guide further regional therapy
systemic or regional therapy of colorectal
(Vitola et al., 1996a, b). A few studies have
liver metastases. The improvement of ther-
also suggested the effectiveness of FDG-
apeutic management and the cost savings
PET in monitoring the therapeutic response
associated with the use of FDG-PET justify
of hepatic metastases to 90Y-glass micro-
its generalization in clinical practice.
spheres, that are introduced via the hepatic
artery that supplies the majority of blood
flow to the metastases and exerts a local REFERENCES
radiotherapeutic effect (Wong et al., 2004). Akhurst, T., Kates, T. J., Mazumdar, M., Yeung,
H., Riedel, E. R., Burt, B. M., Blumgart, L.,
Jarnagin, W., Larson, S. M., and Fong, Y. 2005.
CONCLUSION Recent chemotherapy reduces the sensitivity
of [18F]fluorodeoxyglucose positron emission
In the preoperative staging of presumed tomography in the detection of colorectal metas-
resectable colorectal liver metastases, the tases. J. Clin. Oncol. 23: 8713–8716.
Anderson, G. S., Brinkmann, F., Soulen, M. C.,
best results of FDG-PET are obtained in the Alavi, A., and Zhuang, H. 2003. FDG positron
detection of extra-hepatic disease, mainly emission tomography in the surveillance of
in the abdomen, while high-quality CT may hepatic tumors treated with radiofrequency abla-
be as accurate as FDG-PET in depicting tion. Clin. Nucl. Med. 28: 192–197.
liver or lung metastases. FDG-PET may be Bipat, S., van Leeuwen, M. S., Comans, E. F.,
especially valuable in patients with high Pijl, M. E., Bossuyt, P. M., Zwinderman, A. H.,
and Stoker, J. 2005. Colorectal liver metastases:
clinical risk scores or synchronous metas- CT, MR imaging, and PET for diagnosis meta-
tases, which represent a greater proportion analysis. Radiology 237: 123–131.
of patients referred for surgery. Conversely, Dimitrakopoulou-Strauss, A., Strauss, L. G., and
in those instances in which FDG-PET Rudi, J. 2003. PET-FDG as predictor of therapy
findings result in the patient being denied response in patients with colorectal carcinoma.
potentially curative surgery, further inves- Q. J. Nucl. Med. 47: 8–13.
Even-Sapir, E., Parag, Y., Lerman, H., Gutman, M.,
tigations are needed, particularly in low Levine, C., Rabau, M., Figer, A., and Metser, U.
risk patients. The main impact involves 2004. Detection of recurrence in patients with
patients that are contraindicated before rectal cancer: PET/CT after abdominoperineal or
undergoing futile surgery, whereas survival anterior resection. Radiology 232: 815–822.
30. Colorectal Liver Metastases 405
Fernandez, F. G., Drebin, J. A., Linehan, D. C., MR imaging, PET): a meta-analysis. Radiology
Dehdashti, F., Siegel, B. A., and Strasberg, S. 224: 748–756.
M. 2004. Five-year survival after resection of Lai, D. T., Fulham, M., Stephen, M. S., Chu, K. M.,
hepatic metastases from colorectal cancer in Solomon, M., Thompson, J. F., Sheldon, D. M.,
patients screened by positron emission tomogra- and Storey, D. W. 1996. The role of whole-body
phy with F-18 fluorodeoxyglucose (FDG-PET). positron emission tomography with [18F]fluoro-
Ann. Surg. 240: 438–447; discussion 447–450. deoxyglucose in identifying operable colorectal
Findlay, M., Young, H., Cunningham, D., Iveson, cancer metastases to the liver. Arch. Surg. 131:
A., Cronin, B., Hickish, T., Pratt, B., Husband, 703–707.
J., Flower, M., and Ott, R. 1996. Noninvasive Langenhoff, B. S., Oyen, W. J., Jager, G. J., Strijk,
monitoring of tumor metabolism using fluorode- S. P., Wobbes, T., Corstens, F. H., and Ruers,
oxyglucose and positron emission tomography T. J. 2002. Efficacy of fluorine-18-deoxyglu-
in colorectal cancer liver metastases: correlation cose positron emission tomography in detecting
with tumor response to fluorouracil. J. Clin. tumor recurrence after local ablative therapy for
Oncol. 14: 700–708. liver metastases: a prospective study. J. Clin.
Flamen, P., Stroobants, S., Van Cutsem, E., Dupont, Oncol. 20: 4453–4458.
P., Bormans, G., De Vadder, N., Penninckx, Libutti, S. K., Alexander, H. R., Jr., Choyke, P.,
F., Van Hoe, L., and Mortelmans, L. 1999. Bartlett, D. L., Bacharach, S. L., Whatley,
Additional value of whole-body positron emis- M., Jousse, F., Eckelman, W. C., Kranda, K.,
sion tomography with fluorine-18–2-fluoro-2- Neumann, R. D., and Carrasquillo, J. A. 2001.
deoxy-D-glucose in recurrent colorectal cancer. A prospective study of 2-[18F] fluoro-2-de-
J. Clin. Oncol. 17: 894–901. oxy-D-glucose/positron emission tomography
Flanagan, F. L., Dehdashti, F., Ogunbiyi, O. A., scan, 99mTc-labeled arcitumomab (CEA-scan),
Kodner, I. J., and Siegel, B. A. 1998. Utility of and blind second-look laparotomy for detecting
FDG-PET for investigating unexplained plasma colon cancer recurrence in patients with increas-
CEA elevation in patients with colorectal cancer. ing carcinoembryonic antigen levels. Ann. Surg.
Ann. Surg. 227: 319–323. Oncol. 8: 779–786.
Fong, Y., Saldinger, P. F., Akhurst, T., Macapinlac, Moore, H. G., Akhurst, T., Larson, S. M., Minsky,
H., Yeung, H., Finn, R. D., Cohen, A., Kemeny, B. D., Mazumdar, M., and Guillem, J. G. 2003.
N., Blumgart, L. H., and Larson, S. M. 1999. A case-controlled study of 18-fluorodeoxyglu-
Utility of 18F-FDG positron emission tom- cose positron emission tomography in the detec-
ography scanning on selection of patients for tion of pelvic recurrence in previously irradiated
resection of hepatic colorectal metastases. Am. rectal cancer patients. J. Am. Coll. Surg. 197:
J. Surg. 178: 282–287. 22–28.
Goshen, E., Davidson, T., Zwas, S. T., and Aderka, Peeters, C. F., de Geus, L. F., Westphal, J. R., de
D. 2006. PET/CT in the Evaluation of Response Waal, R. M., Ruiter, D. J., Wobbes, T., Oyen,
to Treatment of Liver Metastases from Colorectal W. J., and Ruers, T. J. 2005. Decrease in circu-
Cancer with Bevacizumab and Irinotecan. lating anti-angiogenic factors (angiostatin and
Technol. Cancer. Res. Treat. 5: 37–44. endostatin) after surgical removal of primary
Huebner, R. H., Park, K. C., Shepherd, J. E., colorectal carcinoma coincides with increased
Schwimmer, J., Czernin, J., Phelps, M. E., and metabolic activity of liver metastases. Surgery
Gambhir, S. S. 2000. A meta-analysis of the 137: 246–249.
literature for whole-body FDG PET detection Rohren, E. M., Paulson, E. K., Hagge, R., Wong, T.
of recurrent colorectal cancer. J. Nucl. Med. 41: Z., Killius, J., Clavien, P. A., and Nelson, R. C.
1177–1189. 2002. The role of F-18 FDG positron emission
Kinkel, K., Lu, Y., Both, M., Warren, R. S., and tomography in preoperative assessment of the
Thoeni, R. F. 2002. Detection of hepatic metas- liver in patients being considered for curative
tases from cancers of the gastrointestinal tract resection of hepatic metastases from colorectal
by using noninvasive imaging methods (US, CT, cancer. Clin. Nucl. Med. 27: 550–555.
406 S. Truant et al.
Ruers, T. J., Langenhoff, B. S., Neeleman, N., Jager, curable colorectal liver metastases. Eur. J. Surg.
G. J., Strijk, S., Wobbes, T., Corstens, F. H., and Oncol. 27: 175–179.
Oyen, W. J. 2002. Value of positron emission Truant, S., Huglo, D., Hebbar, M., Ernst, O.,
tomography with [F-18]fluorodeoxyglucose in Steinling, M., and Pruvot, F. R. 2005. Prospective
patients with colorectal liver metastases: a pro- evaluation of the impact of [18F]fluoro-2-deoxy-
spective study. J. Clin. Oncol. 20: 388–395. D-glucose positron emission tomography of
Schussler-Fiorenza, C. M., Mahvi, D. M., resectable colorectal liver metastases. Br. J.
Niederhuber, J., Rikkers, L. F., and Weber, S. M. Surg. 92: 362–369.
2004. Clinical risk score correlates with yield Veit, P., Antoch, G., Stergar, H., Bockisch, A.,
of PET scan in patients with colorectal hepatic Forsting, M., and Kuehl, H. 2006. Detection of
metastases. J. Gastrointest. Surg. 8: 150–157; residual tumor after radiofrequency ablation of
discussion 157–158. liver metastasis with dual-modality PET/CT:
Selzner, M., Hany, T. F., Wildbrett, P., McCormack, initial results. Eur. Radiol. 16: 80–87.
L., Kadry, Z., and Clavien, P. A. 2004. Does the Vitola, J. V., Delbeke, D., Meranze, S. G., Mazer,
novel PET/CT imaging modality impact on the M. J., and Pinson, C. W. 1996a. Positron emis-
treatment of patients with metastatic colorectal sion tomography with F-18-fluorodeoxyglucose
cancer of the liver? Ann. Surg. 240: 1027–1034; to evaluate the results of hepatic chemoemboli-
discussion 1035–1036. zation. Cancer 78: 2216–2222.
Spaepen, K., Stroobants, S., Dupont, P., Bormans, Vitola, J. V., Delbeke, D., Sandler, M. P., Campbell,
G., Balzarini, J., Verhoef, G., Mortelmans, L., M. G., Powers, T. A., Wright, J. K., Chapman,
Vandenberghe, P., and De Wolf-Peeters, C. 2003. W. C., and Pinson, C. W. 1996b. Positron emis-
[(18)F]FDG PET monitoring of tumour response sion tomography to stage suspected metastatic
to chemotherapy: does [(18)F]FDG uptake cor- colorectal carcinoma to the liver. Am. J. Surg.
relate with the viable tumour cell fraction? Eur. 171: 21–26.
J. Nucl. Med. Mol. Imaging 30: 682–688. Whiteford, M. H., Whiteford, H. M., Yee, L. F.,
Strasberg, S. M., Dehdashti, F., Siegel, B. A., Ogunbiyi, O. A., Dehdashti, F., Siegel, B. A.,
Drebin, J. A., and Linehan, D. 2001. Survival of Birnbaum, E. H., Fleshman, J. W., Kodner, I. J.,
patients evaluated by FDG-PET before hepatic and Read, T. E. 2000. Usefulness of FDG-PET
resection for metastatic colorectal carcinoma: scan in the assessment of suspected metastatic
a prospective database study. Ann. Surg. 233: or recurrent adenocarcinoma of the colon and
293–299. rectum. Dis. Colon Rectum 43: 759–767; discus-
Tanaka, T., Kawai, Y., Kanai, M., Taki, Y., sion 767–770.
Nakamoto, Y., and Takabayashi, A. 2002. Wong, C. Y., Salem, R., Qing, F., Wong, K. T.,
Usefulness of FDG-positron emission tomogra- Barker, D., Gates, V., Lewandowski, R., Hill,
phy in diagnosing peritoneal recurrence of color- E. A., Dworkin, H. J., and Nagle, C. 2004.
ectal cancer. Am. J. Surg. 184: 433–436. Metabolic response after intraarterial 90Y-glass
Topal, B., Flamen, P., Aerts, R., D’Hoore, A., microsphere treatment for colorectal liver metas-
Filez, L., Van Cutsem, E., Mortelmans, L., tases: comparison of quantitative and visual
and Penninckx, F. 2001. Clinical value of analyses by 18F-FDG PET. J. Nucl. Med. 45:
whole-body emission tomography in potentially 1892–1897.
A. Diagnosis
31
Biliary Cystic Tumors:
Clinicopathological Features
Yasuni Nakanuma, Hiroko Ikeda, Yasunori Sato, Kenichi Harada,
Koichi Nakamura, and Yoh Zen
411
412 Y. Nakanuma et al.
a b c
Figure 31.1. Radiological and pathological features of peribiliary cysts. (a) An image of computed
tomography (CT) reveals cystic lesions along the hilar bile ducts. (b) A macroscopic image shows small
cysts in periductal connective tissue. (c) Microscopically, cysts are lined by a single-layered columnar
epithelium (H&E staining, ×40)
As for the underlying disease, peri- which arise from the dilatation of VMC
biliary cysts were originally found in or biliary microhamartoma (Kida et al.,
cases of advanced liver disease such as 1992; Qian et al., 2003). Parenchymal
liver cirrhosis and other portal hyper- cysts are predominant in some cases,
tensive diseases, and in patients with while peribiliary cysts are predominant
other hepatobiliary diseases (Nakanuma, in others. Peribiliary cysts are reportedly
2004) (see below). Peribiliary cysts are identified by computed tomography (CT)
clinically suggested by computed tomog- in 73% of adult polycystic liver disease.
raphy (CT) in ~ 10% of cirrhotic livers That is, in adult polycystic liver disease,
and more in autopsied cirrhotic livers intrahepatic peribiliary glands show fre-
by histopathological studies. In these quent and variable cystic dilatation. Such
diseases, peribiliary cysts may develop peribiliary cysts may comprise a propor-
secondarily, possibly related to disturbed tion of numerous cysts of adult polycystic
circulatory changes. Peribiliary cysts in liver. The genetic events responsible for the
cases of these chronic liver diseases are development of liver cysts in the paren-
not infrequently associated with inflam- chyma may also be a factor in the develop-
matory and edematous changes (Terada ment of peribiliary cysts from peribiliary
and Nakanuma, 1990a). glands. Peribiliary cysts also develop to a
Peribiliary cysts are frequently found lesser degree and frequency in livers with
in adult polycystic liver disease with solitary nonparasitic cysts. In this context,
autosomal-dominant polycystic kidney peribiliary cysts could be a component of
disease or alone, in addition to the cysts the hepatobiliary fibropolycystic disease
in the parenchyma (parenchymal cysts) family. In contrast, peribiliary glands of
414 Y. Nakanuma et al.
the extrahepatic bile ducts show no cystic defects in cholangiography, though their
dilatation in most cases of liver cirrhosis etiopathogenesis, particularly their relation
or adult polycystic liver disease. to the peribiliary cysts in the hepatic hilus
Clinically, peribiliary cysts do not usu- and the portal tracts described above,
ally show any symptom. However, other remains speculative.
hepatobiliary diseases such as ascending
cholangitis or obstructive jaundice may
Hepatic Foregut Cyst
develop directly or indirectly due to com-
pression by multiple enlarged peribiliary Ciliated hepatic foregut cysts (CHFCs)
are rare cystic lesions, usually located
cysts on the adjoining bile ducts. Follow-up
radiological studies have disclosed that in the subcapsular region (Portman and
Nakanuma, 2006; Terada et al., 1990)
peribiliary cysts gradually increase in size
and number during the course of liver cir- immediately beneath the hepatic capsule,
rhosis, and the increase can be used as a mostly in Segment IV. They are solitary
marker of progression of liver cirrhosis. and unilocular, usually < 3 cm in diameter,
Peribiliary cysts can be demonstrated byand filled with viscous mucinous fluid
CT, ultrasonography (US), and magnetic surrounded by fibrous tissue. Histologically,
resonance imaging (MRI) in the vicin- the cyst wall consists of four layers: pseu-
ity of the hepatic hilum (Figure 31.1A). dostratified ciliated columnar epithelia with
Radiologically, these peribiliary cysts aremucous cells, subepithelial connective tis-
seen as tiny cysts anywhere in the large sue, smooth-muscle bundles immunoreac-
portal tracts and/or hepatic hilum. Drip- tive to actin and desmin, and an outermost
infusion cholangiography CT clearly dem- fibrous capsule. The lining epithelial cells
onstrates the exact location of the cysts. contain neutral, carboxylated, and sulfated
Magnetic resonance (MR) cholangiography mucus, and are also positive for keratin and
epithelial membrane antigen, immunohis-
easily clarifies the distribution of tiny cysts
tochemically. Cilia are immunoreactive to
just along the large intrahepatic bile ducts,
actin and tubulin. Ultrastructurally, definite
hepatic bile ducts or both. While the adjoin-
ing bile ducts are occasionally compressed cilia arranged in a 9 + 2 pattern as well as
and locally deformed by the cysts, there ismucous cells are observed in these cells.
no communication between the cysts and These morphologic features are similar to
those of normal bronchi or bronchogenic
the bile ducts. The differential diagnosis is
cysts, suggesting that CHFC arises from
based on the diffuse or local dilation of int-
rahepatic bile ducts, VMC, and cholangitis.the embryonic foregut and differentiates
Ultrasonography shows round or tubular toward bronchial structures in the liver,
anechoic lesions around the large portal analogous to a bronchogenic cyst of the
mediastinum (Terada et al., 1990).
tracts mimicking dilatation of the bile ducts.
Clinically, CHFCs are symptomless and
Peribiliary Cysts Around the Extrahepatic incidentally detected radiologically or at
Bile Ducts
autopsy. CHFCs are a cystic lesion. With
Peribiliary glands around the extrahepatic US, CHFCs appear anechoic or hypoechoic.
bile ducts also occasionally show grossly These cystic lesions include high-density
visible cystic changes and exhibit filling contents, due to thick mucus (Okuda,
31. Biliary Cystic Tumors: Clinicopathological Features 415
2001). Therefore, unenhanced CT demon- tally in 1–6% of reported autopsy series. Von
strates the lesion to be a high-density or Meyenburg complex (VMC) is a develop-
iso-density lesion relative to the surround- mental malformation that is thought to rep-
ing liver parenchyma. Magnetic resonance resent persistence of the embryonic ductal
imaging (MRI) demonstrated the lesion to plate. The histological appearance is quite
be hyperintense on T1-weighted image. characteristic. Individual lesions are usually
These cystic lesions do not show a contrast < 5 mm in diameter, but sometimes up to
enhancement. When a subcapsular nodular 10 mm. Von Meyenburg complexes (VMC)
lesion shows the above-mentioned find- consist of a variable number of irregular bile
ings, a diagnosis of CHFC should be con- ductular structures embedded in a fibrocol-
sidered (Kadoya et al., 1990). This type of lagenous stroma and are located along the
cyst may be mistaken for a hypovascular portal tracts. They are typically adjacent to
solid tumor. a portal area and consist of a fibrous stroma
that contains irregularly shaped duct-like
Variants: Recently, a histologic variant of
structures lined by a low cuboidal epithe-
CHFC without a smooth muscle layer or
lium. The ducts are usually somewhat dilated
with inconspicuous muscle elements has
and often U-shaped or branching. They may
also been reported. It was < 2.0 cm in diam-
contain proteinous or bile-stained secretions.
eter, and was located subcapusularly in the
In some lesions, one of the ducts may dilate
medial segment (S4) of the liver. It had an
sufficiently to form a macroscopic cyst.
epithelial lining of ciliated pseudostratified
Multiple VMCs are considered part of the
cells with occasional goblet cells, and lacked
spectrum of hepatic fibropolycystic disease.
a smooth muscle layer. The lining epithe-
It is generally believed that both solitary bile
lium contained cells positive for immunohis-
duct cyst and cysts of polycystic disease
tochemical staining of surfactant apoprotein
arise from the dilatation of VMC.
A, suggesting an embryonic foregut origin
Clinically, VMCs are symptomless
of the cysts and differentiation toward bron-
and occasionally misdiagnosed as malig-
chiolar structures (Sato et al., 2006).
nant liver disease including metastasis.
Biliary Hamartoma Imaging findings of VMC are nonspecific.
Ultrasonography (US), CT, and MRI show
The following two types are representative multiple periportal small nodular or dot-like
of biliary hamartoma. lesions. Von Meyenburg complex (VMC)
shows multiple, or numerous tiny hypoat-
Von Meyenburg Complexes
tenuating lesons, which are small hypoe-
Von Meyenburg complexes (VMCs), also choic structures occasionally associated
known as biliary microhamartoma or mul- with comet-tail echoes on US, and hyperin-
tiple bile duct hamartoma, are benign liver tesne lesions on T2-weighted MRI (Okuda,
malformations (Okuda, 2001). These com- 2001).
plexes are usually multiple, though their
Multicystic Biliary Hamartoma
number varies on the individual basis. They
are present in ~ 90% of livers with adult This lesion is a solitary and localized
polycystic liver disease, in ~ 20% of livers cystic lesion ~ 2–5 cm in diameter, which
with solitary nonparasitic cysts, and inciden- when located around the hepatic capsule
416 Y. Nakanuma et al.
of the hepatic lobe, characteristically pro- lesions). These lesions might be related to
trudes from the liver (Figure 31.2a). Some developmental abnormalities of the biliary
cases not associated with inflammatory tract or embryonal foregut.
changes have a honeycomb appearance. Ultrasonography (US) and CT reveal rel-
Histologically, these nodular lesions consist atively well-circumscribed nodules in the
of ductal structures with ductal epithelium medial segment of the liver, some of which
resembling biliary epithelium and periduc- are protruding from the liver. This nodule
tal glands resembling peribiliary glands, is enhanced on CT by contrast medium
and fibrous connective tissues containing and sustained until the delayed phase.
blood vessels (Zen et al., 2006c). Biliary Computed tomography (CT) reveales a
epithelial cells and glandular epithelial relatively well-circumscribed nodule just
cells express biliary-type cytokeratins such beneath the hepatic capsule. Angiographies
as cytokeratins 7 and 19. Smooth muscle revealed that this lesion is supplied by the
bundles focally surround ductal structures hepatic artery. This nodule shows a multi-
in some cases (Figure 31.2B). Bile-like locular cystic lesion containing many small
materials are observed within some ducts. cystic spaces on radiologic images.
Two cases are associated with xanthogran-
ulomatous inflammation around bile-like
materials. This cystic lesion shares several BILIARY CYSTIC NEOPLASM
pathologic characteristics of CHFCs, such
as their location (around the falciform Biliary cystic neoplasms are heterogene-
ligament) and periductal smooth muscle ous and can be divided into several cat-
bundles, but does not fulfill the diagnostic egories. Hepatobiliary cystadenoma and
criteria (no ciliated cells and multilocular adenocarcinoma are a prototype of neo-
a b
Figure 31.2. Pathological features of multicystic biliary hamartoma. (a) A nodular lesion is protruding
from the liver. This lesion shows a honeycomb appearance consisting of many ductal structures and sur-
rounding connective tissue. (b) Microscopically, this cystic lesion is covered by a biliary type epithelium.
Periductal glands are observed in fibrous connective tissue (H&E staining, ×100)
31. Biliary Cystic Tumors: Clinicopathological Features 417
a b c
Comparison with Mucinous Cystic Neoplasm cystic disease is also often mistaken for
of the Pancreas hepatobiliary cystadenocarcinoma.
Pancreatic MCN to which mucinous cys-
tadenoma and cystadenocarcinoma of the Intraductal Papillary Neoplasm
pancreas belong in addition to MCN of of the Bile Duct (IPNB)
the retroperitoneum are cystic epithelial Clinicopathological Features
neoplasms occurring almost exclusively in
women. They are composed of a colum- Biliary papillary tumors differing in malig-
nar, mucin-producing epithelium, sup- nant potential and biological behavior from
ported by ovarian-type cellular stroma. ordinary tubular cholangiocarcinomas are
Tumors may be classified as adenoma, known to develop in the intrahepatic and
borderline (low grade malignant), and extrahepatic bile ducts (Adbores-Saavedra
non-invasive or invasive adenocarcinoma. et al., 2000), though their clinicopatho-
Pancreatic MCN shows no communica- logic features remain to be clarified in
tion with the pancreatic ductal system. detail. As described below, these biliary
Interestingly, hepatobiliary cystadenoma/ papillary tumors share clinicopathologi-
cystadenocarcinoma also presents these cal features of pancreatic IPMN. We pro-
clinicopathological features of pancreatic posed referring to these papillary biliary
MCN. In this context, hepatobiliary cysta- tumors collectively as intraductal papil-
denoma/adenocarcinoma are a counterpart lary neoplasm of the bile duct (IPNB) as a
of pancreatic MCN and could be called biliary counterpart of pancreatic IPMN
hepatobiliary mucinous cystic neoplasm (Chen et al., 2001; Zen et al., 2006b).
(hepatobiliary MCN) (Zen et al., 2006a). Both IPNBs and pancreatic IPMN
As for stromal components of these present with dilatation or multilocular
MCNs, they might be derived from the cystic changes of the affected ducts and
ovarian primodium. The ovarian-like the pancreatic ducts, respectively.
stroma in mucinous cystadenomas can
(1) Classification into three subgroups
originate from gonadal epithelium during
based on clinicopathological features
early embryonic development. In human
embryos, preceding their ‘descent’, the Our recent studies showed that IPNB is
gonads are situated directly under the characterized by three key pathologic
diaphragm, dorsal to the liver, the tail of lesions in various components: (1) papillary
the pancreas and the spleen, but separated proliferation of neoplastic biliary epithelial
from these organs by the peritoneal cavity. cells (dysplastic or adenocarcinomatous)
In contrast to the peritoneal epithelium with delicate fibrovascular stalks within the
elsewhere, the cells covering the gonads bile ducts, (2) mucin hypersecretion (mac-
show an activated morphology. The ovar- roscopic mucin is recognizable in some
ian stroma in mucinous cystadenomas of cases), and (3) variable dilatation or mul-
the liver and pancreas suggests a com- tilocular cystic changes of the affected bile
mon origin in epithelial cells that cover ducts (Figure 31.4B). Some IPNB cases
the embryonic gonads in early fetal life. have been referred to as biliary papilloma
Cholangiocarcinoma arising in liver with and papillary cholangiocarcinoma or the
420 Y. Nakanuma et al.
a b c
Figure 31.4. Radiological and pathological features of intraductal papillary mucinous neoplasm of
the bile duct (IPNB). A multilocular cystic lesion with a mural nodule is observed in the liver. (b)
Macroscopically, the multilocular cystic space is occupied by papillary structures. (c) Tumor cells show
prominent papillary proliferation with distinct fibrovascular cores (H&E staining, ×40)
the molecular and genetic pathogenesis as MMP-2, -7 and MT1-MMP in IPNB are
and progression processes of IPNB. As low or weakly positive in comparison with
for microsatellite instability, high-level conventional cholangiocarcinoma (CC) and
instability was present in 12 % of IPNB a balance of expression of MMP-2/tissue
cases, and low-level instability was present inhibitor of matrix metalloproteinase-2
in 35 % of IPNB cases (Abraham et al., (TIMP-2) is also significantly decreased in
2002). No IPNB cases showed methyla- conventional CC compared to IPNB. The
tion of hMLH1. These results indicate participation of MMP/TIMP in the devel-
that microsatellite instability is a relatively opment and progression of IPNB remains
frequent event in IPNB but is not associated to be clarified.
with hMLH1 promoter hypermethylation. As for K-ras gene, activating muta-
Interestingly, the pattern of allelic shifts tions in codon 12 of the K-ras oncogene
was frequently not identical between the are present in 30 % of IPNB cases and
intraductal and invasive cholangiocarci- identical K-ras mutations are present in
noma components of IPNB, suggesting that both the intraductal and invasive com-
there is significant genetic heterogeneity ponents, suggesting that these mutations
within these neoplasms. arise early in the tumorigenesis of IPNB
E-cadherin is one subtype of trans- (Abraham et al., 2003). This frequency
membrane glycoprotein, whose cytoplas- of K-ras gene mutations is lower than
mic domain is bound to α- and β-catenin that previously reported for pancreatic
(Ashida et al., 1998), and β-catenin acts IPMNs. The expression of cyclin D1,
as a key structural molecule in E-cadherin- p21WAF1/CIP, pRb, p53, and MDM2 is
mediated cell-to-cell adhesion. Membranous gradually increased with the progression
β-catenin and E-cadherin expression is well of IPNB. p53 expression is found in 40 %
preserved in non-neoplastic epithelium. of dysplastic IPNB and the expression
Expression levels of both molecules are rate reaches a plateau in borderline malig-
significantly decreased in parallel in inva- nant or non-invasive IPNB and invasive
sive IPNB as compared with non-invasive IPNB. Cyclin D1 and c-myc, a target of
counterparts, respectively, suggesting that Wnt signaling, are frequently present in
a loss or reduction of the membranous IPNB, and interestingly, nuclear β-catenin
expression of β-catenin and E-cadherin is staining, reflecting activation of Wnt sign-
involved in the progression of IPNB. A aling, is observed in 25 % of non-invasive
reduction of β-catenin in the cell membrane IPNB and also in invasive IPNB without
is presumed to disrupt cell to cell adhesive a significant difference. This suggests
stability, and therefore, such a reduction that nuclear staining of β-catenin, prob-
may be followed by invasion and then ably reflecting stimulation of the Wnt
metastasis of carcinoma cells. Matrix met- signaling pathway, is probably involved
alloproteinase (MMP) and tissue inhibitor at an early stage in the progression of
of metalloproteinases (TIMP)-2 also play IPNB. However, no β-catenin gene muta-
important roles in the degradation of the tions have been demonstrated in IPNB.
basement membrane during tumor inva- Interestingly, p53 expression, K-ras muta-
sion. It is well known that the expression or tion, and the nuclear accumulation of
activities of matrix metalloproteinases such β-catenin are greater in conventional CC
424 Y. Nakanuma et al.
than IPNB, suggesting that the carcino- cytoplasm. Our results of Dpc4 expression
genesis is different in these two types of are not in line with the results of previous
biliary neoplasm. studies. This discrepancy might be depend-
Inactivation of p16INK4A, a cancer ent on the cases examined. Our study con-
suppressor gene, is reportedly involved sisted of Asian cases, some of which were
in the carcinogenesis in several organs. associated with hepatolithiasis.
Our recent studies showed that while
p16INK4A is frequently expressed in IPNB Corresponds to IPMN of Main
non-neoplastic, stone-containing bile Pancreatic Duct Type
ducts, its expression is markedly reduced The disease entity of pancreatic IPMN is
in low grade IPNB to invasive IPNB in well established, and it arises in the main
hepatolithiasis (Ishikawa et al., 2004b). pancreatic duct (main-duct type) and also
Methylation-specific PCR revealed that in its major branches (branch-duct type).
54.6 % of 44 IPNB foci harbored p16INK4a The histologic grade and invasiveness, and
promoter hypermethylation, and such foci phenotypes of papillary epithelium compo-
are significantly correlated with decreased nent, and the degree of mucin secretion and
expression of p16INK4, suggesting that mucin phenotypes are different depending
p16INK4a inactivation mainly due to its on these two types of pancreatic IPMN. For
promoter hypermethylation, is a frequent example, while pancreatic IPMN is divided
and early event of IPNB. DeHaan et al. into benign, borderline, and malignant non-
(2007) recently reported that a homozygous invasive or invasive lesions. Benign lesions
9p21 deletion is frequently detected in are common in the branch-duct IPMN and
cholangiocarcinoma with complete loss of borderline and malignant lesions are com-
p16INK4a. And this may be the case in mon in the main-duct type.
IPNB. This inactivation may be partially Interestingly, the gross and microscopic
responsible for the step-wise increase in the characteristics and corresponding radio-
cell proliferative activities of biliary epithe- logic features of IPNB are similar to those
lial cells during the progression of INPB, of pancreatic IPMN, particularly those
followed by further genetic and epigenetic arising in the main pancreatic duct. The
alterations of other cell cycle regulators affected ducts of IPNB and pancreatic
shown in IPNB. DPC4, the gene encod- IPMN are cystically or cylindrically dilated
ing Smad4/Dpc4, is a tumor-suppressor and some appear to be multilocular. They
gene located on chromosome 18q. Dpc4 are characterized by the prominent intra-
protein is almost universally expressed in ductal papillary proliferation of biliary
pancreatic IPMN and pancreatic cancer lining cells. Phenotypically, IPNB and pan-
derived from pancreatic IPMN. Abraham creatic IPMN show one of three epithelial
et al. (2003) revealed that Dpc4 expression phenotypes; pancreaticobiliary, intestinal
was intact in all cases of IPNB examined and gastric types. In addition, an oncocytic
by immunohistochemistry. However, in our variant belonging to the pancreatobiliary
studies, Dpc4 expression was lost in 21.4 % phenotype is also reported in IPNB and
of invasive IPNB cases, while Dpc4 expres- pancreatic IPMN. Like pancreatic IPMN,
sion was diffusely observed in all cases of IPNB could progress to tubular adenocar-
non-neoplastic biliary epithelium in their cinoma or mucinous carcinoma. These
31. Biliary Cystic Tumors: Clinicopathological Features 425
remains to be clarified whether all hepato- nosis of and distinguishing between these
biliary “cystadenoma and cystadenocarci- cystic lesions are clinically important in
noma” with luminal communication to the the choice of treatment, and genetic and
bile ducts and those without ovarian-like molecular approaches are needed to clarify
stroma are IPNB, and also whether uncate- the etiology and pathogenesis of these
gorized biliary neoplasm(s) may constitute cystic diseases.
these unusual cystic tumors.
Other Types of Neoplastic Cystic REFERENCES
Diseases Abraham, S.C., Lee, J.H., Hruban, R.H., Argani, P.,
Furth, E.E., and Wu, T.T. 2003. Molecular and
There may be other types or categories of immunohistochemical analysis of intraductal
biliary cystic tumors which are different papillary neoplasms of the biliary tract. Hum.
from hepatobiliary MCN or IPNB with Pathol. 34: 902–910.
cystic dilatation or multilocular changes Abraham, S.C., Lee, J.H., Boitnott, J.K., Argani,
of the affected bile ducts. For example, P., Furth, E.E., and Wu, T.T. 2002. Microsatellite
instability in intraductal papillary neoplasms of
cholangiocarcinoma is known to develop the biliary tract. Mod. Pathol. 15: 1309–1317.
in biliary diseases showing cystic dilata- Adbores-Saavedra, J., Scoazec, J.C., Wittekind,
tion such as congenital choledochal cyst C., Sripa, B., Menck, H.R., Soehendra, N., and
and Caroli’s disease, and also in hepatic Sriram, P.V.J. 2000. Carcinoma of the gallbladder
cystic diseases such as adult polycystic and extrahepatic bile ducts. In: Hamilton, S.R.,
liver disease or solitary liver cysts, and Aaltonen, L.A., eds. World Health Organization
Classification of Tumours. Pathology and
rarely in the peribiliary glands or cysts Genetics of Tumours of the Digestive System.
(Nakanuma et al., 2000; Terada and Lyon: IARC Press, 207–214.
Nakanuma, 1990b). Cholangiocarcinoma Ashida, K., Terada, T., Kitamura,Y., and Kaibara, N.
arising in these diseases may present cystic 1998. Expression of E-cadherin, alpha-catenin,
changes within or adjacent to the tumor beta-catenin, and CD44 (standard and variant iso-
tissue as suggested by Fujii et al. (2005). forms) in human cholangiocarcinoma: an immu-
nohistochemical study. Hepatology 27:974–82.
In conclusion, biliary cystic tumors are Chen, T.C., Nakanuma, Y., Zen, Y., Chen, M.F.,
largely divided into non-neoplastic, largely Jan, Y.Y., Yeh, T.S., Chiu, C.T., Kuo, T.T.,
hamartomatous, and neoplastic etiologies. Kamiya, J., Oda, K., Hamaguchi, M., Ohno, Y.,
The former includes peribiliary cysts and Hsieh, L.L., and Nimura, Y. 2001. Intraductal
multicystic biliary hamartoma. The lat- papillary neoplasia of the liver associated with
ter includes hepatobiliary cystadenoma/ hepatolithiasis. Hepatology 34: 651–658.
DeHaan, R.D., Kipp, B.R., Smyrk, T.C., Abraham,
adenocarcinoma and IPNB, particularly S.C., Roberts, L.R., and Halling, K.C. 2007. An
those with cystic dilatation or multilocular assessment of chromosomal alterations detected
cystic changes of the affected bile ducts by fluorescence in situ hybridization and p16
(cystic variant of IPNB). Intraductal pap- expression in sporadic and primary scleros-
illary neoplasm of the bile duct (IPNB) ing cholangitis-associated cholangiocarcinomas.
may be a biliary equivalent of pancreatic Hum. Pathol. 38: 491–499.
Devaney, K., Goodman, Z.D., and Ishak, K.G.
IPMN, and hepatobiliary cystadenoma/cys- 1994. Hepatobiliary cystadenoma and cystade-
tadenocarcinoma share clinicopathological nocarcinoma. A light microscopic and immuno-
features of pancreatic MCT, and therefore histochemical study of 70 patients. Am. J. Surg.
can be called biliary MCT. An exact diag- Pathol. 18: 1078–1091.
31. Biliary Cystic Tumors: Clinicopathological Features 427
Fujii, T., Harada, K., Katayanagi, K., Kurumaya, H., normal and pathologic biliary tract in humans,
and Nakanuma, Y. 2005. Intrahepatic cholangi- including blood supply. Microsc. Res. Tech. 38:
ocarcinoma with multicystic, mucinous appearance 552–570.
and oncocytic change. Pathol. Int. 55: 206–209. Nakanuma, Y., Leong, A.S.Y., Sripa, B.,
Hamazaki, K., Gochi, A., Shimamura, H., Kaihara, Ponchon, T., Vatanasapt, V., and Ishak, K.G.
A., Maruo, Y., Doi, Y., Orita, K., Lygidakis, N.J. 2000. Intrahepatic cholangiocarcinoma. In:
1996. Serum levels of circulating intercellular Hamilton, S.R., Aaltonen, L.A., eds. World
adhesion molecule 1 in hepatocellular carci- Health Organization Classification of Tumours.
noma. Hepatogastroenterology 43: 229–234. Pathology and Genetics of Tumours of the
Ishikawa, A., Sasaki, M., Ohira, S., Ohta, T., Oda, Digestive System. Lyon: IARC Press, 173–180.
K., Nimura, Y., Chen, M.F., Jan, Y.Y., Yeh, T.S., Nakanuma, Y. 2004. Peribiliary cysts have at least
and Nakanuma, Y. 2004a. Aberrant expression two different pathogeneses. J. Gastroenterol. 39:
of CDX2 is closely related to the intestinal 407–408.
metaplasia and MUC2 expression in intraductal Okuda, K. 2001. Okuda, K., Mitchell, D.G., Itai,
papillary neoplasm of the liver in hepatolithiasis. Y., Ariyama, J., eds. Hepatobiliary Diseases.
Lab. Invest. 84: 629–638. Oxford: Blackwell Science.
Ishikawa, A., Sasaki, M., Sato, Y., Ohira, S., Chen, Portman, B.C., and Nakanuma, Y. 2006. Diseases
M.F., Huang, S.F., Oda, K., Nimura, Y., and of bile ducts. Burt, A.D., Portman, B.C. Ferrell,
Nakanuma, Y. 2004b. Frequent p16ink4a inac- L.D., eds. MacSween’s Pathology of the Liver.
tivation is an early and frequent event of intra- Churchill Livingstone 5th ed., pp. 517–581.
ductal papillary neoplasm of the liver arising in Qian, Q., Li, A., King, B.F., Kamath, P.S., Lager,
hepatolithiasis. Hum. Pathol. 35: 1505–1514. D.J., Huston, J., 3rd, Shub, C., Davila, S.,
Kadoya, M., Matsui, O., Nakanuma, Y., Yoshikawa, Somlo, S., and Torres, V.E. 2003. Clinical profile
J., Arai, K., Takashima, T., Amano, M., and of autosomal dominant polycystic liver disease.
Kimura, M. 1990. Ciliated hepatic foregut cyst: Hepatology 37: 164–71.
radiologic features. Radiology 175: 475–477. Sato, Y., Kitagawa, S., Zen, Y., Minato, H., and
Kida, T., Nakanuma, Y., and Terada, T. 1992. Nakanuma, Y. 2006. Ciliated hepatic cyst with-
Cystic dilatation of peribiliary glands in livers out smooth muscle layer: a variant of ciliated
with adult polycystic disease and livers with hepatic foregut cyst? Pathol. Int. 56: 340–344.
solitary nonparasitic cysts: an autopsy study. Sato, M., Watanabe, Y., Tokui, K., Kohtani, T.,
Hepatology 16: 334–340. Nakata, Y., Chen, Y., and Kawachi, K. 2003.
Lee, S.S., Kim, M.H., Lee, S.K., Jang, S.J., Song, Hepatobiliary cystadenocarcinoma connected to
M.H., Kim, K.P., Kim, H.J., Seo, D.W., Song, the hepatic duct: a case report and review of the lit-
D.E., Yu, E., Lee, S.G., and Miin, Y.I. 2004. erature. Hepatogastroenterology 50: 1621–1624.
Clinicopathologic review of 58 patients with bil- Sudo, Y., Harada, K., Tsuneyama, K., Katayanagi,
iary papillomatosis. Cancer 100: 783–793. K., Zen, Y., and Nakanuma, Y. 2001. Oncocytic
Longnecker, D.S., Hruban, R.H., Adler, G., and biliary cystadenocarcinoma is a form of intra-
Kloppel, G. 2000. Intraductal papillary-mucinous ductal oncocytic papillary neoplasm of the liver.
neoplasms of the pancreas. In: Hamilton, S.R., Mod. Pathol. 14: 1304–1309.
Aaltonen, L.A., eds. World Health Organization Terada, T., and Nakanuma, Y. 1990a. Pathological
Classification of Tumours. Pathology and observations of intrahepatic peribiliary glands
Genetics of Tumours of the Digestive System. in 1,000 consecutive autopsy livers. III. Survey
Lyon: IARC Press, 237–240. of necroinflammation and cystic dilatation.
Nakanuma, Y., Kurumaya, H., and Ohta, G. Hepatology 12: 1229–1233.
1984. Multiple cysts in the hepatic hilum and Terada, T., Nakanuma, Y., and Ohta, G. 1987.
their pathogenesis. A suggestion of periductal Glandular elements around the intrahepatic bile
gland origin. Virchows. Arch. A. Pathol. Anat. ducts in man; their morphology and distribution
Histopathol. 404: 341–350. in normal livers. Liver 7: 1–8.
Nakanuma, Y., Hoso, M., Sanzen, T., and Sasaki, Terada, T., Nakanuma, Y., Kono, N., Ueda, K.,
M. 1997. Microstructure and development of the Kadoya, M., and Matsui, O. 1990. Ciliated
428 Y. Nakanuma et al.
hepatic foregut cyst. A mucus histochemical, cystic tumors with bile duct communication: a
immunohistochemical, and ultrastructural study cystic variant of intraductal papillary neoplasm
in three cases in comparison with normal bronchi of the bile duct. Mod. Pathol. 19: 1243–1254.
and intrahepatic bile ducts. Am. J. Surg. Pathol. Zen, Y., Fujii, T., Itatsu, K., Nakamura, K., Minato,
14: 356–363. H., Kasashima, S., Kurumaya, H., Katayanagi,
Terada, T., and Nakanuma, Y. 1990b. Pathological K., Kawashima, A., Masuda, S., Niwa, H.,
observations of intrahepatic peribiliary glands in Mitsui, T., Asada, Y., Miura, S., Ohta, T., and
1,000 consecutive autopsy livers. II. A possible Nakanuma, Y. 2006b. Biliary papillary tumors
source of cholangiocarcinoma. Hepatology 12: share pathological features with intraductal
92–97. papillary mucinous neoplasm of the pancreas.
Wheeler, D.A., and Edmondson, H.A. 1985. Hepatology 44: 1333–1343.
Cystadenoma with mesenchymal stroma (CMS) Zen, Y., Terahata, S., Miyayama, S., Mitsui, T.,
in the liver and bile ducts. A clinicopathologic Takehara, A., Miura, S., Nobata, K., Kitao, A.,
study of 17 cases, 4 with malignant change. Kakuda, K., Kiyohara, K., and Nakanuma, Y.
Cancer 56: 1434–1445. 2006c. Multicystic biliary hamartoma: a hitherto
Zamboni, G., Longnecker, D.S., Kloppel, G., Adler, undescribed lesion. Hum. Pathol. 37: 339–344.
G., and Hruban, R.H. 2000. Mucinous cystic neo- Zen, Y., Sasaki, M., Fujii, T., Chen, T.C., Chen,
plasm of the pancreas. In: Hamilton SR, Aaltonen M.F., Yeh, T.S., Jan, Y.Y., Huang, S.F., Nimura,
LA, eds. World Health Organization Classification Y., and Nakanuma, Y. 2006d. Different expres-
of Tumours. Pathology and Genetics of Tumours sion patterns of mucin core proteins and
of the Digestive System. Lyon: IARC Press, pp. cytokeratins during intrahepatic cholangiocar-
234–236. cinogenesis from biliary intraepithelial neo-
Zen, Y., Fujii, T., Itatsu, K., Nakamura, K., Konishi, plasia and intraductal papillary neoplasm of
F., Masuda, S., Mitsui, T., Asada, Y., Miura, S., the bile duct – an immunohistochemical study
Miyayama, S., Uehara, T., Katsuyama, T., Ohta, of 110 cases of hepatolithiasis. J. Hepatol. 44:
T., Minato, H., and Nakanuma, Y. 2006a. Biliary 350–358.
32
Cholangiocarcinoma: Intraductal
Sonography
Kazuo Inui, Hironao Miyoshi, and Junji Yoshino
INTRODUCTION
METHODS
The recent development of magnetic reso-
Instruments
nance cholangiopancreatography (MRCP)
and multiple-detector-row computed tom- Several types of ultrasonic probes, a 2.5 mm
ography (MD-CT) facilitate diagnosis of diameter miniature ultrasonic probe (UM-
the obstruction site in patients with cholan- 2R or UM-3R, Olympus Medical Systems,
giocarcinoma. However, the extension of Tokyo, Japan) and a 2.1 mm diameter ultra-
cholangiocarcinoma is still difficult to diag- sonic probe (UM-4R, Olympus Medical
nose precisely. The correct diagnosis of Systems) have been developed. The UM-2R
cancer extension, especially invasion to ultrasonic probe incorporates a radial scan-
the portal vein and pancreas, is necessary ning system with a frequency of 12 MHz, and
to decide the operative procedures for a the UM-3R and UM-4R with 20 MHz. These
complete resection. Intraductal ultrasonog- are connected to an endoscopic ultrasonic
raphy (IDUS) has been reported by several observation unit (EU-M30 or EU-M2000,
authors, (Kuroiwa et al., 1998; Fujita et al., Olympus Medical Systems) and incorpo-
1998; and Tamada et al., 2001a), as a rated by a probe-driving unit (MAJ 335 or
reliable method for the evaluation of extra- MAJ 2000, Olympus Medical Systems). The
hepatic cholangiocarcinoma and aiding in probes are easily passed through the 2.8 mm
its accurate staging. We, Kanemaki et al. diameter biopsy channel of a duodenoscope
(1997) and Inui et al. (1998), reported on the (JF-200, Olympus Medical Systems) or the
usefulness of three-dimensional intraductal 3.6 mm diameter biopsy channel of a new
ultrasonography (3D-IDUS) as a compre- electroduodenoscope (JF-V260, Olympus
hensive image display, and the instruments Medical Systems).
and processor systems of 3D-IDUS have We use the 3D-IDUS system developed
been developed. In this chapter, we introduce by Olympus Medical Systems (Figure
the methodology of IDUS and 3D-IDUS for 32.1); a 3D ultrasonic probe, UM-3D-2R
the diagnosis of cholangiocarcinoma. (12 MHz) or UM-3D-3R (20 MHz), with
429
430 K. Inui et al.
Insertion Methods
Transpapillary Approach
The patient is placed on the fluoroscopy
table in the prone position. Premedication
consists of 0.5 mg atropine sulfate, 15 mg
pentazocine, 2 mg scopolamine butylbro-
Figure 32.1. Photograph of a 3D-IDUS system, mide intramuscularly, and 10 mg diazepam
which includes a 3D ultrasonic probe and a probe intravenously. After insertion of the elec-
driving unit, the MAJ 2000. The ultrasonic probe troduodenoscope into the second portion
(arrow), a UM-DG20–35R, is a new type of ultra- of the duodenum, endoscopic retrograde
sonic probe with a ropeway system cholangiopancreatography (ERCP) is
performed. After ERCP, we insert the
a diameter of 3.4 mm; and a probe driving ultrasonic probe into the stricture of the
unit, the MAJ 355 beginning in 1995 and bile duct through the biopsy channel of
the MAJ 2000 since 2002. Recently, we the electroduodenoscope (Figure 32.2).
usually use a new type of ultrasonic probe Ultrasonography is performed, and the
with a ropeway system, UM-DG20-35R area observed by the probe is confirmed
(Figure 32.1). These probes are connected by fluoroscopy.
to an endoscopic ultrasonic observation
Percutaneous Transhepatic Approach
unit (EU-M30 or EU-M2000). The 3D
ultrasonic probe consists of an external In patients with obstructive jaundice, per-
tube as an outer sheath and the probe itself cutaneous transhepatic biliary drainage
with a diameter of 2.4 mm and a 12 or (PTBD) is performed immediately after
20 MHz radial scan transducer at its tip. admission to our hospital. We dilate the
32. Cholangiocarcinoma: Intraductal Sonography 431
Indications
We have treated patients with several
types of pancreatobiliary disease since
1989. The diseases include extrahepatic
cholangiocarcinoma, gallbladder carci-
noma, carcinoma of the duodenal Papilla
Vater, pancreatic carcinoma, intraductal
papillary mucinous neoplasia of the pan-
creas, benign biliary strictures, chronic
pancreatitis, autoimmune pancreatitis
and other rare conditions. Four patients
Figure 32.2. An ultrasonic probe is inserted into with cholangiocarcinoma had 3D-IDUS
the bile duct through the biopsy channel of a duo- after treatment by Nd-YAG laser irradia-
denoscope along with a guide-wire tion. In only four patients, the ultrasonic
probes could not be inserted into the bile
duct and the stricture. In about one third
of patients, we used the percutaneous
sinus tract of the PTBD to 16 F and per- approach.
form percutaneous transhepatic cholangi- Intraductal ultrasonography and
oscopy (PTCS), IDUS, and/or 3D-IDUS 1 3D-IDUS may be indicated for the diag-
week after PTBD. nosis of remnant stones after endoscopic
The patient is placed on the fluoroscopy lithotripsy of common bile duct stones.
table in the supine position. Premedication They also may be indicated for the
consists of 0.5 mg atropine sulfate, 15 mg diagnosis of biliary strictures and the
pentazocine intramuscularly, and 10 mg precise extension of cancer in patients
diazepam intravenously. We observe the with cholangiocarcinoma and intraduc-
biliary stricture with PTCS and take biop- tal papillary mucinous neoplasia of the
sies from the stricture. After PTCS, a pancreas.
432 K. Inui et al.
Figure 32.4. 3D-IDUS of cholangiocarcinoma. The DPR images, including the radial and linear images,
easily reveal the relationship between the tumor (T) and the portal vein (PV)
an increased success rate using a trans- of IDUS in assessing tumor invasion to the
papillary approach was reported by Fujita pancreas was reported to be 88.9–100 %
et al. (2000). Using the ropeway system (Inui et al., 2002). When the high-echoic
with a guide-wire, scanning of the bile layer between the tumor and a vessel dis-
duct was achieved in 98.2–99.2 % of the appears, it is diagnosed as positive vessel
patients, regardless of the site and length invasion. The accuracy of IDUS in assess-
of the biliary strictures. Once the guide- ing tumor invasion to the right hepatic
wire is passed through the biliary stenosis, artery and portal vein was reported to be
no further technique is required to intro- 92.3–100 % (Inui et al., 2002). IDUS is
duce the ultrasonic probe. useful for the staging of cholangiocarci-
The clinical indications for IDUS of noma, especially invasion to the portal
the biliary tract include the diagnosis of vein and pancreas. Menzel and Dosmschke
choledocholithiasis (Palazzo, 1997), the (1999) reported on the usefulness of IDUS
differential diagnosis of biliary strictures for diagnosis of cholangiocarcinoma, and
(Menzel et al., 2000) and the local stag- showed that tumor extent was diagnosed
ing of cholangiocarcinoma. Wada et al. correctly in 76.8 % of cases when using
(2000) reported that the diagnosis of bil- IDUS and in 53.6 % cases in EUS.
iary metallic stent obstruction and evalu- Tamada et al. (2001b) reported that an
ation for microwave coagulation therapy accurate distinction of T1 and T2 tumor
were new clinical applications of IDUS. was difficult. In these cases, even if the
Menzel et al. (2000) proposed that tumor was limited to the inside low-echoic
malignancy should be suspected for hyp- layer, it could be a T2 tumor (the tumor
oechoic masses with irregular margins and invaded the peri-muscular connective tis-
inhomogeneous echo-poor areas invading sue) or a T1 tumor (the tumor was con-
the surrounding tissue. Penetration was fined to the fibromuscular layer).
defined as a continuation of the main The diagnosis of the longitudinal spread
hypo-echoic tumor mass into adjacent of a tumor is the most important issue
structures. On the other hand, homogene- for a curable resection, but this is diffi-
ous echo patterns and smoothly defined cult for IDUS. When the irregular thick-
margins were most likely not malignant. ness of the bile duct can be observed
They reported that the sensitivity, specifi- consecutively and away from the main
city, and accuracy of IDUS were 91.1 %, lesion, the longitudinal spread of tumor
80 %, and 89.1 %, respectively. Tamada can be diagnosed. The overall accuracy
et al. (1998) reported that if the inter- of IDUS for intraductal tumor spread was
ruption of the bile duct wall structure or 84.6%, and the result was not satisfactory.
a lesion with a diameter of greater than However, Tamada et al. (2001b) reported
8.0 mm were diagnosed as malignant, the that transpapillary IDUS prior to biliary
accuracy of IDUS was 76%. drainage is a tactic that can minimize the
When a tumor protrudes into the pan- influence of the biliary drainage catheter.
creatic parenchyma, or if the outer echoic Nevertheless, their results were the same
layer of the intra-pancreatic bile duct is as our prior study (Inui et al., 2002) that
interrupted by a tumor, invasion to the used a percutaneous approach, with an
pancreas can be diagnosed. The accuracy accuracy of 85 %.
436 K. Inui et al.
the evaluation of patients with biliary strictures Tamada, K., Nagai, H., Yasuda, Y., Tomiyama, T.,
and no abdominal mass on computed tomogra- Ohashi. A., Wada, S., Kanai, N., Satoh, Y., Ido,
phy. Endoscopy 37:715–721. K., Sugano, K. 2001b. Transpapillary intraductal
Tamada, K., Ueno, N., Tomiyama, T., Oohashi, ultrasonography prior to biliary drainage in the
A., Wada, S., Nishizono, T., Tano, S., assessment of longitudinal spread of extrahe-
Aizawa, T., Ido, K., and Kimura, K. 1998. patic bile duct carcinoma. Gastrointest. Endosc.
Characterization of biliary structures using 53:300–307.
intraductal Ultrasonography: comparison Wada, S., Tamada, K., Tomiyama, T., Ohashi, A.,
with percutaneous cholangioscopic biopsy. Utsunomiya, K., Higashisawa, T., Satoh, Y.,
Gastrointest. Endosc. 47:341–349. Miyata, T., Sugano, K. 2000. Endoscopic micro-
Tamada, K., Inui, K., Menzel, J. 2001a. Intraductal wave coagulation therapy for bile duct cancer with
ultrasonography of the bile duct system. intraductal ultrasonographic monitoring: brief case
Endoscopy 33:878–885. report. Am. J. Gastroenterol. 95:1104–1105.
B. Prognosis
33
Extrahepatic Bile Duct Carcinoma:
Role of the p53 Protein Family
Alexander I. Zaika and Seung-Mo Hong
Epidemiology Etiology
Cholangiocarcinoma is a malignant neo- Several different risk factors may give rise
plasm arising from the epithelium lining to the injury and proliferation of the bile
441
442 A.I. Zaika and S.-M. Hong
duct epithelial cells. These factors are asso- and undifferentiated carcinomas (Hong
ciated with the development of EBD carci- et al., 2005). By tumor growth feature,
noma, which included primary sclerosing EBD carcinoma can be classified as poly-
cholangitis associated with inflammatory poid, nodular, and infiltrating types. One
bowel disease, hepatolithiasis, infection of the often seen histologic characteristics
with liver flukes (Clonorchis sinensis and of EBD is marked desmoplastic stro-
Opisthorchis viverrini), familial adeno- mal reaction, which is extensive stromal
matous polyposis coli, von Meyenberg fibrosis surrounding tumor cells caused
complexes, anomalous pancreaticobiliary by cancer cell infiltration. A few malig-
junction, and choledochal cysts (Choi nant tumors from other organs can show
et al., 2004). desmoplastic stromal reaction. Other than
desmoplastic stromal reaction, micro-
Diagnosis scopic findings of EBD carcinoma are
similar to adenocarcinomas from other
The patients’ symptoms, which include
organs. Usually tumor cells form tubular
abdominal pain, jaundice, and weight
glands. Based on the degree of glandular
loss, are non specific. Therefore, the
formation (histologic grade), EBD carci-
evaluation of patients in regard to the
noma can be divided into well, moderate,
extent of EBD carcinoma is determined
or poorly differentiated adenocarcinomas.
by several radiologic examinations, such
For example, when more than 75 % of all
as ultrasonography, helical-CT, direct
tumor cells make glandular structures, it
cholangiography, and MRCP (magnetic
can be classified as a well-differentiated
resonance cholangiopancreatography)
carcinoma. On the other hand, when less
(Seyama and Makuuchi, 2007). In some
than 25 % of cancer cells form tubular
situations, it is difficult to make a definite
structures, EBD carcinoma can be called
diagnosis with these imaging tools. Bile
poorly differentiated carcinoma. Most
cytology or brushing biopsies are the
EBD carcinomas are classified as well to
tools of choice in this case (Seyama and
moderately differentiated adenocarcinomas
Makuuchi, 2007).
(Albores-Saavedra et al., 2000). Nuclei of
tumor cells show significant difference
Pathology
in size (pleomorphism), are stained more
EBD carcinoma usually appears as an ill- darkly by hematoxylin and eosin staining
demarcated, firm, whitish tan mass with (hyperchromasia), and demonstrate atypi-
induration or ulceration of the bile duct cal, asymmetric mitotic features. In about
wall. Adenocarcinoma is the most com- 10 % of EBD carcinoma cases, dysplasia is
mon histologic type, and consists of about observed in the peritumoral noncancerous
80 % of EBD carcinomas. Other histo- epithelial cells (Albores-Saavedra et al.,
logic variants that compose the remain- 2000). At the time of diagnosis, many EBD
ing 20 % include: papillary carcinomas, carcinoma cases show tumor cell invasion
intestinal-type adenocarcinomas, adenos- into lymphatic system, blood vessels, nerve
quamous carcinomas, mucinous carcino- fibers, lymph nodes, and other adjacent
mas, clear cell carcinomas, signet ring organs such as pancreas, duodenum, and
cell carcinomas, small cell carcinomas, liver.
33. Extrahepatic Bile Duct Carcinoma: Role of the p53 Protein Family 443
Figure 33.1. Architectures of human p53, p63 and p73 genes and proteins. TP53, TP63 and TP73 genes
encode a transactivation domain (TAD), DNA-binding domain (DBD), oligomerization domain (OD),
and have two promoters (P1 and P2). The P1 promoter drives expression of transactivation-competent
full-length proteins (TA isoforms). Conversely, the P2 promoter produces TAD-deficient proteins (ΔN
isoforms) with dominant-negative functions. p73 and p63 proteins have also additional domain, termed
SAM (sterile alpha motif) that is not found in p53. Different isoforms are designated with Greek letters
a b
Figure 33.2. Immunohistochemical staining for p63 and p73 proteins in extrahepatic bile duct carcinoma.
(a). Moderately differentiated EBD adenocarcinoma shows nuclear staining of p63 (X200) (b). Strong
staining for p73 in papillary carcinoma of the bile duct (X200)
33. Extrahepatic Bile Duct Carcinoma: Role of the p53 Protein Family 447
patients and interpreted based on isoform- mice mutant for p63 and p73: evidence for
specific analysis. broader tumor suppressor functions for the p53
family. Cancer Cell 7: 363–373.
Green, F.L., Page, D.L., Fleming, I.D., Fritz, A.G.,
REFERENCES Balch, C.M., Haller, D.G., and Morrow, M.
2002. AJCC Cancer Staging Manual, Springer,
Ahrendt, S.A., Rashid, A., Chow, J.T., Eisenberger, New York.
C.F., Pitt, H.A., and Sidransky, D. 2000. p53 Hong, S.M., Cho, H., Moskaluk, C.A., Yu, E., and
overexpression and K-ras gene mutations in Zaika, A.I. 2007. p63 and p73 expression in ext-
primary sclerosing cholangitis-associated biliary rahepatic bile duct carcinoma and their clinical
tract cancer. J. HepatoBiliary-Pancreat. Surg. 7: significance. J. Mol. Histol. 38: 167–175.
426–431. Hong, S.M., Kim, M.J., Pi, D.Y., Jo, D., Cho, H.J.,
Albores-Saavedra, J., Henson, D.E., and Yu, E., and Ro, J.Y. 2005. Analysis of extrahe-
Klimstra, D.S. 2000. Tumors of the Gall blad- patic bile duct carcinomas according to the New
der, Extrahepatic Bile Ducts and Ampulla of American Joint Committee on Cancer staging
Vater. Atlas of Tumor Pathology, Armed Forces system focused on tumor classification problems
Institute of Pathology, Washington, DC. in 222 patients. Cancer 104: 802–810.
Argani, P., Shaukat, A., Kaushal, M., Wilentz, Khan, S.A., Thomas, H.C., Toledano, M.B., Cox,
R.E., Su, G.H., Sohn, T.A., Yeo, C.J., Cameron, I.J., and Taylor-Robinson, S.D. 2005. p53
J.L., Kern, S.E., and Hruban, R.H. 2001. Mutations in human cholangiocarcinoma: a
Differing rates of loss of DPC4 expression review. Liver Int. 25: 704–716.
and of p53 overexpression among carcinomas Lee, Y.C., Song, S.Y., Chung, J.B., Kang, J.K.,
of the proximal and distal bile ducts. Cancer and Park, I.S. 1996. p53 protein expression in
91: 1332–1341. extrahepatic bile duct cancer. Yonsei Med. J. 37:
Choi, B.I., Han, J.K., Hong, S.T., and Lee, K.H. 112–117.
2004. Clonorchiasis and cholangiocarcinoma: Leong, C.O., Vidnovic, N., DeYoung, M.P., Sgroi,
etiologic relationship and imaging diagnosis. D., and Ellisen, L.W. 2007. The p63/p73 net-
Clin. Microbiol. Rev. 17: 540–552. work mediates chemosensitivity to cisplatin in
Concin, N., Becker, K., Slade, N., Erster, S., a biologically defined subset of primary breast
Mueller-Holzner, E., Ulmer, H., Daxenbichler, cancers. J. Clin. Invest. 117: 1370–1380.
G., Zeimet, A., Zeillinger, R., Marth, C., and Malhi, H., and Gores, G.J. 2006.
Moll, U.M. 2004. Transdominant DeltaTAp73 Cholangiocarcinoma: modern advances in
isoforms are frequently up-regulated in ovarian understanding a deadly old disease. J. Hepatol.
cancer. Evidence for their role as epigenetic p53 45: 856–867.
inhibitors in vivo. Cancer Res. 64: 2449–2460. Momoi, H., Okabe, H., Kamikawa, T., Satoh,
Diamantis, I., Karamitopoulou, E., Perentes, S., Ikai, I., Yamamoto, M., Nakagawara, A.,
E., and Zimmermann, A. 1995. p53 protein Shimahara, Y., Yamaoka, Y., and Fukumoto,
immunoreactivity in extrahepatic bile duct M. 2001. Comprehensive allelotyping of human
and gallbladder cancer: correlation with tumor intrahepatic cholangiocarcinoma. Clin. Cancer
grade and survival, Hepatology 22: 774–779, Res. 7: 2648–2655.
Baltimore, MD. Munro, A.J., Lain, S., and Lane, D.P. 2005.
Farazi, P.A., Zeisberg, M., Glickman, J., Zhang, Y., P53 abnormalities and outcomes in colorectal
Kalluri, R., and DePinho, R.A. 2006. Chronic cancer: a systematic review. Br. J. Cancer 92:
bile duct injury associated with fibrotic matrix 434–444.
microenvironment provokes cholangiocarci- Murata, T., Nagasaka, T., Kamiya, J., Nimura,
noma in p53-deficient mice. Cancer Res. 66: Y., Wakai, K., Yoshida, K., and Nakashima, N.
6622–6627. 2002. P53 labeling index in cholangioscopic
Flores, E.R., Sengupta, S., Miller, J.B., Newman, biopsies is useful for determining spread of
J.J., Bronson, R., Crowley, D., Yang, A., McKeon, bile duct carcinomas. Gastrointest. Endos. 56:
F., and Jacks, T. 2005. Tumor predisposition in 688–695.
33. Extrahepatic Bile Duct Carcinoma: Role of the p53 Protein Family 449
Nagakawa, T., Kayahara, M., Ikeda, S., Futakawa, Immunohistochemical detection of proliferating
S., Kakita, A., Kawarada, H., Matsuno, M., cell nuclear antigen and p53 expression in car-
Takada, T., Takasaki, K., Tanimura, H., Tashiro, cinoma of the extrahepatic bile duct. Oncology
S., and Yamaoka, Y. 2002. Biliary tract cancer 54: 407–413.
treatment: results from the Biliary Tract Cancer Tan, G., Yilmaz, A., De Young, B.R., Behling,
Statistics Registry in Japan. J. HepatoBiliary- C., Lehman, A., and Frankel, W.L. 2004.
Pancreat. Surg. 9: 569–575. Immunohistochemical analysis of biliary tract
Nathan, H., Pawlik, T.M., Wolfgang, C.L., Choti, lesions. Appl. Immunohistochem. Mol. Morphol.
M.A., Cameron, J.L., and Schulick, R.D. 2007. 12: 193–197.
Trends in survival after surgery for cholangi- Tannapfel, A., Engeland, K., Weinans, L., Katalinic,
ocarcinoma: a 30-year population-based SEER A., Hauss, J., Mossner, J., and Wittekind, C.
database analysis. J. Gastrointest. Surg. 11: 1999. Expression of p73, a novel protein related
1488–1496; discussion: 1496–1487. to the p53 tumour suppressor p53, and apoptosis
Nomoto, K., Tsuneyama, K., Cheng, C., Takahashi, in cholangiocellular carcinoma of the liver. Br. J.
H., Hori, R., Murai, Y., and Takano, Y. 2006. Cancer 80: 1069–1074.
Intrahepatic cholangiocarcinoma arising in cir- Vilgelm, A., El-Rifai, W., and Zaika, A. 2008.
rhotic liver frequently expressed p63-positive Therapeutic prospects for p73 and p63: Rising
basal/stem-cell phenotype. Pathol. Res. Pract. from the shadow of p53. Drug Resist. Updat.
202: 71–76. 11: 152–163.
Seyama, Y., and Makuuchi, M. 2007. Current Yang, B., House, M.G., Guo, M., Herman, J.G., and
surgical treatment for bile duct cancer. World J. Clark, D.P. 2005. Promoter methylation profiles
Gastroenterol. 13: 1505–1515. of tumor suppressor genes in intrahepatic and
Suto, T., Sugai, T., Nakamura, S., Uesugi, N., extrahepatic cholangiocarcinoma. Mod. Pathol.
Sasaki, R., Kanno, S., and Saito, K. 1997. 18: 412–420.
34
Extrahepatic Bile Duct Carcinoma:
Mucin 4, a Poor Prognostic Factor
Michiyo Higashi, Shugo Tamada, Kohji Nagata, Masamichi Goto,
and Suguru Yonezawa
451
452 M. Higashi et al.
MUC5B and MUC6), and soluble mucin central large tandem repeat domain con-
(MUC7) (Hollingsworth and Swanson, taining 16-amino acid repetitive units,
2004). regions rich in potential N-glycosylation
sites, two cystein-rich domains, a putative
GDPH proteolytic cleavage site, three epi-
MUC4 MUCIN dermal growth factor (EGF)-like domains,
a hydrophobic transmembrane domain,
MUC4, a transmembrane mucin, has and a short cytoplasmic tail (Singh
gained significant biological concern in et al., 2007). Approximately two thirds of
the past few years. Like MUC1, MUC4 the MUC4 protein sequence consists of
is a membrane mucin; however, MUC4 identical 16-amino-acid tandem repeats
functions through different mechanisms (Moniaux et al., 2004).
than those of MUC1. As a signaling mol- MUC4/Sialomucin complex (SMC) is
ecule, MUC1 acts as a docking protein, a rat homologue of human mucin gene
whereas MUC4 acts as a receptor ligand MUC4 and its transmembrane subunit acts
(Ramsauer et al., 2003). The difference as an intramembrane ligand for the recep-
of the expression patterns suggests the tor tyrosine kinase ErbB2 to induce the
possibility of different mechanisms in phosphorylation of Tyr-1248 of the ErbB2
the expression of MUC1 and MUC4, (DiGiovanna and Stern, 1995; Jepson
both of which are membrane mucins pos- et al., 2002; Ramsauer et al., 2003). The
sessing cell signaling functions. MUC4 heterodimeric glycoprotein complex con-
extends at least 1.12–2.12 μm above the sists of ASGP-1 (ascites sialoglycopro-
cell membrane, far above all other mem- tein-1), an O-glycosylated mucin subunit,
brane-associated proteins, such as adhe- and ASGP-2, an N-glycosylated trans-
sion molecules (Moniaux et al., 1999). membrane subunit (Singh et al., 2007).
MUC4 with its rigid and extended struc- MUC4/SMC leads to the expression of
ture is considered a modulator for cell–cell the cell-cycle inhibitor p27 (Ramsauer
and cell-extracellular matrix interactions et al., 2003). On the other hand, MUC4/
(Komatsu et al., 1999). SMC acts with neuregulin synergistically
The gene encoding the MUC4 mucin to enhance phosphorylation of both ErbB2
was first identified in 1991 from a tracheo- and ErbB3, resulting in the downregula-
bronchial cDNA library (Porchet et al., tion of p27 and activation of protein kinase
1991). However, due to its large size and B/Akt. It is proposed that complex forma-
complex structure, its complete genomic tion of MUC4/SMC and ErbB2 has an effect
organization could not be established until on epithelial cell behavior, as a switch in
2000 (Moniaux et al., 1999; Moniaux epithelial differentiation and proliferation
et al., 2000). MUC4 gene is located on (Jepson et al., 2002). Furthermore, Singh
chromosome 3 in the q29 region that et al. (2004) proposed that MUC4 partici-
codes a family of 25 alternatively spliced pates in tumor growth and metastasis by
transcripts. The deduced full-length amino directly altering the tumor cell properties,
acid sequence of MUC4 apoprotein shows and/or via modulating ErbB2 expression.
the presence of a leader peptide, serine and Komatsu et al. (1999) showed that SMC
threonin rich non-tandem repeat region, disrupts integrin-mediated cell adhesion
34. Extrahepatic Bile Duct Carcinoma: Mucin 4, a Poor Prognostic Factor 453
a a
b b
Figure 34.1. Expression of MUC4 in EHBDC. Figure 34.2. Expression of MUC4 in duodenal
(a). MUC4, clone 8G7 (b) MUC4, clone 1G8. The mucosa. a. MUC4, clone 8G7 The clone 8G7 is
clone 8G7 antibody is expressed in the cytoplasm not expressed in the duodenal mucosa. b. MUC4,
and/or membrane of the carcinoma cells, and is clone 1G8 The clone 1G8 stains the luminal sur-
negative for normal peribilliary glands (arrow- face of the duodenal epithelia
heads). The clone 1G8 is weakly expressed in the
membrane of the carcinoma cells, and also stains amino acids (Hollingsworth and Swanson,
the endothelial cells of the small blood vessels. This is 2004). In normal individuals, there is a
negative for normal peribilliary glands (arrowheads)
variation in the number of repeats per allele
(from 25 to > 125 repeats), implying that
MUC1 has a relatively simple architec- the length of the repetitive array may be
ture. The MUC1 gene codes a membrane- critical for the normal function of MUC1.
associated protein with distinct domains; The large extracellular central domain is
an amino-terminal domain consisting of a heavily O-glycosylated. Glycosylation
20-amino acid signal peptide and degener- varies among different tissues, and tumor-
ate tandem repeats; a large central domain associated MUC1 has been reported to be
(two-thirds of the core protein sequence) either less glycosylated or more glyco-
made up of variable numbers of a 20-amino sylated than the forms expressed by normal
acid repeat; a carboxyl terminus made of tissue (Higashi et al., 1999; Horinouchi
degenerate tandem repeats; a hydrophobic et al., 2003). Another important domain of
membrane-spanning domain of 31 amino MUC1 is the cytoplasmic domain, which is
acids, and a cytoplasmic domain of 69 highly conserved across mammalian species
34. Extrahepatic Bile Duct Carcinoma: Mucin 4, a Poor Prognostic Factor 455
b c
d e
Figure 34.3. Expression of MUC1 in EHBDC. a. Hematoxylin and eosin, b. MUC1/CORE, c. MUC1/
DF3, d. MUC1/MY1E12, e. MUC1/HMFG-2. MUC1/DF3 (c) and MUC1/MY1E12 (d) are more widely
expressed in the carcinoma cells than in the MUC1/CORE (b) and MUC1/HMFG-2 (e). MUC1/HMFG-2
is partially glycosylated MUC1 (Burchell et al., 1983)
bile duct and pancreatic duct epithelia are the pancreaticoduodenectomy. A suitable
especially prone to undergo degenera- opening procedure is necessary because
tion (autolytic change), thus much atten- the anatomy of pancreas head is com-
tion should be paid to the specimens of plex. These guidelines are very useful not
34. Extrahepatic Bile Duct Carcinoma: Mucin 4, a Poor Prognostic Factor 457
only for histological diagnosis, but also We should avoid opening both common
for immunohistochemical and genomic bile duct and main pancreatic duct, because
examination. the shape of the resected specimens may
Here, we describe the proper handling be distorted and the orientation will be
procedure of pancreaticoduodenectomy lost. Only common bile duct thus should
specimens. First of all, the orientation of be opened. It is better to stop opening the
duodenum is recognized. There are two bile duct before the papilla of Vater for the
points to distinguish the proximal and dis- preservation of its structure of the papilla
tal sides of the duodenum. One of the dis- of Vater. Photographs with a scale bar are
tinguishing features is at the proximal free taken, and fixation is completed.
end, which is shorter than that at the distal The orientation and the shape should be
free end. The other is that the stomach adequately kept when chemical fixation is
(pyloric region), is sometimes attached at performed. A corkboard, and pins or nee-
the proximal margin. Subsequently, bile dles are useful tools, because the mucosa
duct should be identified. When the gall- of the duodenum and bile duct is kept in
bladder is attached to the resected speci- the correct position when the mucosa is
mens, it is relatively easy to identify the immobilized with these tools on the cork-
bile duct. We follow the gallbladder, and board. The corkboard may be replaced
the cystic duct to common bile duct ante- by a board made of Styrofoam or rubber.
gradely. When the gallbladder has been Injection of fixative solution into the pan-
removed, we have to find the suture thread creas parenchyma before attaching it on
of bile duct ligation. Usually, the distal the board is also recommended in order
side of the bile duct is dilated; and these to avoid autolysis. An 18 gauge is suitable
procedures are easy. But in some cases it is for this purpose. Injection of too much
difficult to find the bile duct. In such cases, fixative should be avoided especially for
we follow the bile duct from the papilla of adipose tissue because the adipose tissue
Vater retrogradely after opening the duo- shows edematous change.
denum. We have to identify the anterior The cases of simple extrahepatic bile
and posterior sides before opening the duct excision are relatively simply handled
duodenum. The superior mesenteric vessel when compared to those of pancreati-
sometimes indents the uncinate process to coduodenectomy. After the identification
form a groove called the pancreatic notch. of the anatomical orientation, the bile
The opening of the duodenum should be duct is opened in a longitudinal direction.
performed after the correct orientation of Injection fixation is useful if the dissection is
the above sites has been noted. It is better not done. Pinning is necessary in all cases.
to open the duodenal wall near the pancreas All specimens are fixed in 10 % buffered
at the posterior side than opening it at the formalin or 10 % formalin. Sampling for
opposite side (lateral side), because with histological examination should be done
the former method it is easy to observe the within 48 h after fixation. Prolonged fixta-
papilla of Vater. The common bile duct is tion may produce formalin pigments and
also opened at the posterior side after open- loss of the antigenecity. Further fixation is
ing the duodenum. A long metal probe of expected by cutting out thin blocks from
small diameter can be used as a guide wire. the large and thick specimens. Sections are
458 M. Higashi et al.
embedded in paraffin, and cut into 4 μm PBS. The sections then receive Elite ABC
thick serial sections for usual hematoxylin(Vector Laboratories, Burlingame, CA)
and eosin staining, and further immunohis- for 30 min. After washing with PBS three
tochemistry. Silan-coated glass slides are times, the sections are finally reacted
suitable for immunohistochemistry. with diaminobenzidine substrate (0.2 mg/
ml) for 10 min for visualization, rinsed
with tap water, counterstained with hema-
IMMUNOHISTOCHEMISTRY toxylin and mounted. Reaction products
are not present when nonimmune serum
Immunohistochemical staining is per- or PBS is used instead of the primary
formed by the immunoperoxidase method antibodies.
using the avidin–biotinylated horserad-
ish peroxidase complex (ABC complex),
REFERENCES
(Vector Laboratories, Burlingame, CA).
The sections are deparaffinized with Burchell, J., Durbin, H., and Taylor-Papadimitriou,
J. 1983. Complexity of expression of antigenic
xylene for 30 min followed by 100 %
determinants recognized by monoclonal anti-
ethanol three times. After hydration in bodies HMFG-1 and HMFG-2 in normal and
decreasing concentrations of ethanol in malignant human mammary epithelial cells. J.
water (95 %, 70 %, 50 %, 30 % and 20 %, Immunol. 131: 508–513.
5 min for each), the sections are washed Chen, A. J., Meng, Q. H., Long, B., and Yang,
in 0.01 mol/l phosphate buffered saline, G. L. 2002. Relationship between p27 expres-
sion and prognosis of colorectal carcinoma. Ai
pH 7.4 (PBS). Endogenous peroxidase
Zheng. 21: 1075–1077.
is blocked by incubating the sections in DiGiovanna, M. P., and Stern, D. F. 1995. Activation
0.3 % hydrogen peroxidase in absolute state-specific monoclonal antibody detects tyro-
methanol at room temperature for 30 min. sine phosphorylated p185neu/erbB-2 in a subset
For the epitope retrieval of MUC4, the of human breast tumors overexpressing this
slides are immersed in a water bath pre- receptor. Cancer Res. 55: 1946–1955.
Higashi, M., Yonezawa, S., Ho, J. J., Tanaka, S.,
treatment at 80°C for 20 min in 0.01 M
Irimura, T., Kim, Y. S., and Sato, E. 1999.
citrate buffer (pH6.0). After heating, the Expression of MUC1 and MUC2 mucin antigens
samples are allowed to cool for 15–20 min in intrahepatic bile duct tumors: its relation-
at room temperature. ship with a new morphological classification of
The sections are washed twice with PBS cholangiocarcinoma. Hepatology 30: 1347–1355.
after cooling. Then, 2 % horse serum in Hollingsworth, M. A., and Swanson, B. J. 2004.
Mucins in cancer: protection and control of the
PBS is applied for 30 min at room tem-
cell surface. Nat. Rev. Cancer 4: 45–60.
perature to prevent nonspecific staining. Horinouchi, M., Nagata, K., Nakamura A., Goto
The sections are washed twice with PBS, M., Takao S., Sakamoto M., Fukushima N., Miwa
and then incubated with diluted anti- A., Irimura T., Imai K., Sato E. and Yonezawa
MUC4 monoclonal antibody (clone 8G7 S. 2003. Expression of different glycoforms
1:3000, clone 1G8 1:100) in PBS with 1 % of membrane mucin (MUC1) and secretory
mucin (MUC2, MUC5AC and MUC6) in pan-
bovine serum albumin for 16 h at 4°C. The
creatic neoplasms. Acta Histochem. Cytochem.
sections are washed three times with PBS 36: 443–453.
and incubated with the biotinylated anti- Jepson, S., Komatsu, M., Haq, B., Arango, M.
mouse IgG, and washed three times with E., Huang, D., Carraway, C. A., and Carraway,
34. Extrahepatic Bile Duct Carcinoma: Mucin 4, a Poor Prognostic Factor 459
K. L. 2002. Muc4/sialomucin complex, the Ramsauer, V. P., Carraway, C. A., Salas, P. J.,
intra-membrane ErbB2 ligand, induces spe- and Carraway, K. L. 2003. Muc4/sialomucin
cific phosphorylation of ErbB2 and enhances complex, the intramembrane ErbB2 ligand,
expression of p27(kip), but does not activate translocates ErbB2 to the apical surface in
mitogen-activated kinase or protein kinaseB/Akt polarized epithelial cells. J. Biol. Chem. 278:
pathways. Oncogene 21: 7524–7532. 30142–30147.
Komatsu, M., Yee, L., and Carraway, K. L. 1999. Sgambato, A., Migaldi, M., Leocata, P., Ventura,
Overexpression of sialomucin complex, a rat L., Criscuolo, M., Di Giacomo, C., Capelli, G.,
homologue of MUC4, inhibits tumor killing by Cittadini, A., and De Gaetani, C. 2000. Loss
lymphokine-activated killer cells. Cancer Res. of p27Kip1 expression is a strong independent
59: 2229–2236. prognostic factor of reduced survival in N0 gas-
Moniaux, N., Escande, F., Batra, S. K., Porchet, tric carcinomas. Cancer 89: 2247–2257.
N., Laine, A., and Aubert, J. P. 2000. Alternative Singh, A. P., Chaturvedi, P., and Batra, S. K. 2007.
splicing generates a family of putative secreted Emerging roles of MUC4 in cancer: a novel
and membrane-associated MUC4 mucins. Eur. J. target for diagnosis and therapy. Cancer Res. 67:
Biochem. 267: 4536–4544. 433–436.
Moniaux, N., Nollet, S., Porchet, N., Degand, P., Singh, A. P., Moniaux, N., Chauhan, S. C., Meza,
Laine, A., and Aubert, J. P. 1999. Complete J. L., and Batra, S. K. 2004. Inhibition of MUC4
sequence of the human mucin MUC4: a putative expression suppresses pancreatic tumor cell growth
cell membrane-associated mucin. Biochem. J. and metastasis. Cancer Res. 64: 622–630.
338: 325–333. Taguchi, K., Aishima, S., Asayama, Y., Kajiyama,
Moniaux, N., Varshney, G. C., Chauhan, S. K., Kinukawa, N., Shimada, M., Sugimachi, K.,
C., Copin, M. C., Jain, M., Wittel, U. A., and Tsuneyoshi, M. 2001. The role of p27kip1
Andrianifahanana, M., Aubert, J. P., and Batra, protein expression on the biological behavior of
S. K. 2004. Generation and characterization intrahepatic cholangiocarcinoma. Hepatology
of anti-MUC4 monoclonal antibodies reactive 33: 1118–1123.
with normal and cancer cells in humans. J. Tamada, S., Goto, M., Nomoto, M., Nagata, K.,
Histochem. Cytochem. 52: 253–261. Shimizu, T., Tanaka, S., Sakoda, K., Imai, K.,
Porchet, N., Nguyen, V. C., Dufosse, J., Audie, J. and Yonezawa, S. 2002. Expression of MUC1
P., Guyonnet-Duperat, V., Gross, M. S., Denis, and MUC2 mucins in extrahepatic bile duct car-
C., Degand, P., Bernheim, A., and Aubert, J. cinomas: its relationship with tumor progression
P. 1991. Molecular cloning and chromosomal and prognosis. Pathol. Int. 52: 713–723.
localization of a novel human tracheo-bronchial Tamada, S., Shibahara, H., Higashi, M., Goto, M.,
mucin cDNA containing tandemly repeated Batra, S. K., Imai, K., and Yonezawa, S. 2006.
sequences of 48 base pairs. Biochem. Biophys. MUC4 is a novel prognostic factor of extrahe-
Res. Commun. 175: 414–422. patic bile duct carcinoma. Clin. Cancer Res. 12:
Price, M. R., Rye, P. D., Petrakou, E., Murray, 4257–4264.
A., Brady, K., Imai, S., Haga, S., Kiyozuka, Y., Utsunomiya, T., Yonezawa, S., Sakamoto, H.,
Schol, D., Meulenbroek, M. F., Snijdewint, F. Kitamura, H., Hokita, S., Aiko, T., Tanaka, S.,
G., von Mensdorff-Pouilly, S., Verstraeten, R. A., Irimura, T., Kim, Y. S., and Sato, E. 1998.
Kenemans, P., Brockzjil, A., Nilsson, K., Nilsson, Expression of MUC1 and MUC2 mucins in gas-
O., Reddish, M., Suresh, M. R., Koganty, R. R., tric carcinomas: its relationship with the prognosis
Fortier, S., Baronic, L., Berg, A., Longenecker, of the patients. Clin. Cancer Res. 4: 2605–2614.
M. B., Hilgers, J. et al. 1998. Summary report Yamamoto, M., Bhavanandan, V. P., Nakamori, S.,
on the ISOBM TD-4 Workshop: Analysis of 56 and Irimura, T. 1996. A novel monoclonal anti-
monoclonal antibodies against the MUC1 mucin. body specific for sialylated MUC1 mucin. Jpn.
Tumour Biol. 19(Suppl 1): 1–20. J. Cancer Res. 87: 488–496.
C. Treatment
35
Hilar Cholangiocarcinoma: Photodynamic
Therapy and Stenting
Marcus Wiedmann, Joachim Mössner, and Helmut Witzigmann
463
464 M. Wiedmann et al.
this increase may be the result of a rise in one 1999; Miyazaki et al., 1998; Seyama et al.,
or several genotoxic environmental agents, 2003; Todoroki et al., 2000). We performed
causing cholangiocyte DNA damage. preoperative biliary drainage according to
At present, complete resection is the these criteria in our patients and obtained
only potentially curative therapy, but most a median preoperative bilirubin level of
of the patients with intra- and extrahepatic 2.1 mg/dl. Recovery of hepatic function
CC present with already advanced disease. after biliary decompression takes 4 weeks,
Given the small size of available studies so surgery should not be done too early.
and lack of randomized trials, there is no
established role for neoadjuvant and adju-
vant therapy. In the palliative setting, sys- PALLIATIVE BILE DUCT
temic chemotherapy (chemoradiation) has STENTING
not been clearly proven to prolong survival
significantly as discussed by Lazaridis and The goal of palliative treatment is to achieve
Gores (2005) in a review article. stable long-term biliary drainage in order to
improve quality of life and prolong survival.
Refractory obstructive cholestasis leads to
PREOPERATIVE BILIARY death owing to severe bacterial cholangitis
DRAINAGE or liver failure within 3–6 months. Because
surgical drainage procedures do not reveal
Liver dysfunction caused by obstructive any advantage to nonsurgical palliation
jaundice is a relevant risk factor in major according to studies such as by Kosuge
hepatic resections. There are no prospec- et al. (1999), biliary stenting is regarded as
tive randomized trials investigating the the palliative method of choice. Endoscopic
effect of preoperative biliary drainage approach for biliary stent insertion is pre-
on morbidity and mortality in jaundized ferred over an ultrasound-guided percuta-
patients before liver resection. In non- neous approach due to a lower complication
randomized studies such as by Cherqui rate. Furthermore, cholestasis was more
et al. (2000) no difference in mortality efficiently reduced and 30-day and overall
was observed between patients with or mortality rates were lower in a randomized
without biliary drainage. No difference study (Speer et al., 1987). Percutaneous
with respect to morbidity was seen in three transhepatic cholangiodrainage (PTCD)
studies and in our own series of 60 patients (Figure 35.1) should be used after failure
with hilar CC who underwent resection. of endoscopy and for anatomically dif-
On the other hand, higher complication ficult situations, i.e., after partial stomach
rates after biliary stenting were reported resection or gastrectomies, especially with
by other authors. Despite these inconclu- Y-Roux-anastomosis (Doctor et al., 1999).
sive data, the authors of the majority of Often in these cases the papilla Vateri can-
recent studies dealing with hilar CC recom- not be reached endoscopically. In selected
mend biliary drainage in order to decrease cases both methods can be combined as a
the serum bilirubin level below 2–5 mg/ “rendezvous-approach”. A new technique
dl before resection (Ebata et al., 2003; of endoscopic biliary-enteric bypass was
Hemming et al., 2005; Kawasaki et al., introduced by Burmester et al. (2003),
2003; Kondo et al., 2004; Kosuge et al., which comprises endosonographically
35. Hilar Cholangiocarcinoma: Photodynamic Therapy and Stenting 465
of biliary obstruction due to malignancy. (2003) concluded that these results showed
The incidence of cholangitis following that DoubleLayer™ stents have a longer
stent insertion was significantly lower with patency period than polyethylene stents.
10 F stents compared with 8 F stents (5 % Another new development, as described
vs. 34 %, p < 0.001). Stent survival until by Raju et al. (2006), is a 10 F “winged”
blockage was significantly longer for 10 F stent without a lumen to prevent stent occlu-
compared with 8 F (median 32 weeks vs. sion. Finally, our own group, Wiedmann
median 12 weeks, p < 0.001). et al. (2004b), studied in a small case
In addition, Kadakia and Starnes (1992) series, whether bacterial colonization and
retrospectively compared the efficacy and incrustration of stents can be avoided
complications of 10 F biliary stents with by using an external-internal Yamakawa-
11.5 F stents in the management of malig- endoprosthesis in combination with a sub-
nant and benign biliary tract diseases. cutaneously implanted port system. A total
They concluded that 10 F stents have the of three patients with bile duct obstruc-
same success rate and complication rate as tion (1 × recurrence of gastric cancer,
11.5 F stents in the management of biliary 1 × invasive gallbladder cancer, 1 × recur-
tract diseases; thus, a stent size larger than rence of bile duct cancer) were included.
10 F offers no significant advantage. In all patients open biliary drainage and
A new approach may be the use of prevention of cholangitis were achieved.
an Olympus DoubleLayer™ stent (DLS; Therefore, this method is currently being
perfluoro alkoxy, without sideholes). In a evaluated in a larger number of patients.
prospective randomized comparison of this Cost analyses according to studies by Prat
stent with the standard polyethylene stent et al. (1998) and Wagner et al. (1993) in
(with sideholes), including a total of 120 patients with middle and distal bile duct or
patients with jaundice due to malignant ampullary cancer showed that metallic stents
strictures of the middle to distal third of were advantageous in patients surviving
the common bile duct, 28 DLS patients more than 6 months, whereas a plastic stent
(47 %) and 17 polyethylene stent patients was advantageous in patients surviving 6
(29 %) died without clinical evidence of months or less (Figure 35.2). Self-expanding
stent occlusion after a mean of 114 and metal stents (SEMS) which can be inserted
105 days, respectively (p < 0.05). Twenty- endoscopically and percutaneously, have
six DLS patients (43 %) and 38 polyeth- the advantage of a higher patency rate with
ylene stent patients (63 %) had symptoms consecutive lower rates of reinterventions
of stent clogging after a mean of 144 and rehospitalizations in comparison to less
and 99 days, respectively (p < 0.05). Stent costly plastic stents (Deviere et al., 1988;
dysfunction (stent orifice impacted in the Wagner et al., 1993). The higher patency
bile duct or duodenal wall, stent migra- rate of metallic stents is related to the wider
tion) was recorded in 6 DLS patients lumen and better drainage of segmental bile
(10 %) and 5 polyethylene patients (8 %) duct branches (Figure 35.3). These data were
(n.s.). Kaplan-Meier analysis of DLS and confirmed by a single-center prospective
PE stent clogging-free survival showed a randomized controlled trial by Soderlund
significantly longer patency period with and Linder (2006) comparing the patency of
the DLS stents (p = 0.0005). Tringali et al. 10 F polyethylene stents and covered 30 F
a b c
Figure 35.2. Endoscopic view of a pigtail plastic endoprosthesis (a), a straight plastic stent (b) and a
metal stent (c), which jut out from the papilla Vateri into the duodenum (black arrows)
Figure 35.3. Different models of uncovered [Wallstent™, Ultraflex™, EndoCoil™ Endoflex™, and
Z-Stent™ (from left to right)] and covered [ePTFE-FEP (expanded polytretrafluoroethylene-fluorinated
ethylene propylene) Silicone, Polycaprolactone, Polyurethane (below)] metal stents
468 M. Wiedmann et al.
steel Wallstents™ in patients with nonresect- days in the covered group and after 319
able malignant common bile duct strictures. days in the uncovered Wallstent group (p >
As a result, median patency times were 1.8 0.05). Stent patency rates were 83, 78, 67,
and 3.6 months in the polyethylene stents and 54 % at 100, 200, 300, and 400 days,
and Wallstents™ groups, respectively (p = respectively, in the covered group and 83,
0.002). Interestingly, when overall costs of 66, 54, and 36 % in the uncovered group,
therapy, determined by adding the cost of which was not significantly different.
hospitalization to the costs of the stent, were However, it must be mentioned that this
compared, there was no difference between study was not a prospective randomized
both groups. The authors recommended study and the sample size was not large.
the more-effective SEMS in unresectable A prospective randomized study by
patients with malignant common bile duct De Palma et al. (2001) revealed a lower
strictures, who survive a median of 4.5 cholangitis rate after unilateral stent
months and less costly plastic stents for insertion in comparison to bilateral stent
patients who have distant metastases. insertion. Howewer, thus far general con-
Despite the higher patency rate of SEMS, sensus regarding unilateral versus bilateral
stent dysfuncion can be induced by tumor bile duct drainage has not been obtained
ingrowth, overgrowth or stent disloca- (Table 35.1). In general, at least 25 % of
tion. Stent dysfunction can be solved by liver tissue should be drained and con-
implantation of additional plastic or metal trast medium should be injected only
stents, laser treatment or balloon dilatation. in those bile duct segments that will be
Whether stent occlusion may be prevented finally drained. There is no need to drain
by covered SEMS or brachytherapy, is a certain liver lobe. It is better to drain the
currently unknown. A prospective ran- easiest accessable one. In case of concur-
domized study by Isayama et al. (2004) rent single-sided liver atrophy as a result
evaluating patients with distal malignant of tumor induced blood vessel occlusion,
bilary duct obstruction showed that a the primarily intact liver lobe has to be
polyurethane covered Diamond™-stent is drained. A recent study by Freeman and
superior to a non-covered metallic stent in Overby (2003) showed that MRCP-guided
terms of patency rate. However, cholecys- drainage of the widest dilated bile ducts
titis and pancreatitis rates were higher in can improve results of stent therapy.
the covered metallic stent group. Overall Despite controversial data with respect
survival was equivalent in both groups. to the benefit of unilateral or bilateral
In contrast, a recent study by Yoon biliary drainage, in studies by Liu et al.
et al. (2006) comparing polyurethane- (1998), Ducreux et al. (1992), Polydorou
covered Wallstents™ (that were used in et al. (1989), and Polydorou et al. (1991)
36 patients) with uncovered Wallstents™ successful endoscopic drainage rate with
(that were used in 41 patients) found no plastic stents was 41 % (20 % reduc-
significant difference between stent paten- tion of bilirubin within the first week),
cies. Cholecystitis occurred in 1 patient of 53 % (30 % reduction of bilirubin within
the covered Wallstent™ group but in none the first week), 74 % (30 % reduction of
of the uncovered Wallstent™ group. Stent bilirubin within the first month), and 80 %
occlusion occurred after a mean of 398 (30 % reduction of bilirubin within the
35. Hilar Cholangiocarcinoma: Photodynamic Therapy and Stenting 469
activation and thromboxane-induced vessel sure dose and light fluence rate, the oxygen
constriction and thrombus formation, prob- availability in the target, and finally the
ably mediated by nitric oxide, (3) cytosolic time interval between the administration of
and mitochondrial membrane damage with the photosensitizer and the light treatment.
cytochrome C release, bcl-2 depletion and Because PDT is a cold photochemical proc-
induction of apoptosis, and (4) release ess, there is no tissue heating, and connec-
of inflammatory mediators that induce tive tissues such as collagen and elastin are
T-lymphocyte mediated cellular immune largely unaffected.
response. The extent of photodamage and
cytotoxicity is multifactorial and depends Contraindications for Photodynamic
on the type of photosensitizer used, its Therapy
localization and administered dose, the Photodynamic therapy should not be per-
light source used, generating the light expo- formed in patients with acute porphyria,
472 M. Wiedmann et al.
poor kidney or liver function (creatinine > ative of pyropheophorbide (HPPH), benzo-
3 mg/dl, international normalized ratio of porphyrin derivatives (BPD), ring A-reduced
prothrombin time [INR] > 2.2), encasement monoacid benzoporphyrin derivative
or thrombosis of the main blood vessels, (BPDMA, Vertiporfin, Visudyne™), Sn eti-
leucopenia (leucocytes < 2000/cmm), throm- opurpurin (SnEt2, Rostaporphin, Purlytin™),
bocytopenia (< 50,000/cmm), and terminal meso-tetra (hydroxyphenyl) chlorins
tumor stage. [H2(THPC)] (Temoporfin, Foscan™), bac-
teriochlorin, and 5-aminolevulinic acid
Photosensitizers (ALA, Levulan™) are their representatives.
The second group consists of photosensi-
The ideal photosensitizer should be chemi- tizers based on isomeric porphyrins, like
cally pure and of known specific compo- porphycene, corrphycene, isoporphycene,
sition, should have a high quantum yield hemiporphycene, N-confused porphyrin,
for singlet oxygen production, a strong doubly N-confused porphyrin, and por-
absorption with high extinction coefficient phycene ATMPn (9-acetoxy-2,7,12,17-
ε at longer wavelength (red) region prefer- tetrakis-(β-methoxy-ethyl)-porphycene).
ably between 700–800 nm, and an excellent The third group consists of phthalocyanine
photochemical reactivity. It should also and naphthalocyanine based photosensitiz-
possess minimal dark toxicity and only be ers, like zinc phthalocyanine CGP 55847,
toxic in the presence of light, have prefer- sulphonated aluminium phthalocyanine
able retention by target tissue (tumor cells),(Photosense) and naphthalocyanines. The
be rapidly excreted from the body, be syn- fourth group consists of cationic photo-
thesizable from easy available precursors sensitizers, like Pt (II) complex of bis
and stable and easy to dissolve in the body’s (N,N-dimethylaminomethyl) deutero-
tissue fluids and be capable of formulation. porphyrin IX dimethyl ester, copper imi-
Starting from the development of the first nium salt of octaethylbenzochlorin, and
generation porphyrin based photosensitiz- unsymmetrically substituted benzonaph-
ers, hematoporphyrin derivative (HPD) and thopyrazines. Other potential candidates
the purified hematoporphyrin derivatives are expanded porphyrin based photosensi-
Photofrin™, Photosan™, and Photoheme™, tizers. Their representatives are sapphyrin,
a second generation of photosentizers,
vinylogous porphyrins, texaphyrins (e.g.,
have been developed and are now under
gadolinium (III) texaphyrin (XCYTRIN™)
trial. The first group consists of porphy- or Lutetium (III) texaphyrin (LUTRIN™) ),
rin, chlorin and bacteriochlorin derivatives. and core-modified expanded porphyrins
Boronated protoporphyrin, isohematopor- (e.g., ammonium salt of 5,10,15,20-tetrakis
phyrin, tetrasulfonated meso-tetraphenyl (meso-p-sulfonato phenyl)-25,27, 29-trithia
porphyrin (H2TPPS4), 5,10,15,20-tetrakis sapphyrin) (Table 35.2).
(4-sulphonato-phenyl)-21,23-dichalcogena-
porphyrin, o-,m- and p-isomers of tetra
(hydroxyphenyl)-porphyrin, picket fence Photodynamic Therapy for Palliation
porphyrins, monoaspartyl chlorin e6 (MACE, of Hilar Cholangiocarcinoma
Npe6), diaspartyl chlorin e6 (DACE), Interest in using PDT for the palliative
chlorin p6 lysyl derivative, hexylether deriv- treatment of advanced non-resectable CC
Table 35.2. Photosensitizers, potential indications and activation wavelengths.
Activation
wavelength
Sensitizer Trade name Potential indications (nm)
First Generation Photosensitizers
Purified hematoporphyrin Photofrin, Photosan, Head-, neck-, tracheobronchial, gynecological, esophageal, bladder, brain, gastric, 630
derivatives (HPD) and Photoheme bile duct cancer, skin basalioma, Barrett’s esophagus, psoriasis, Papilloma virus
Second Generation Photosensitizers
A. Porphyrin, Chlorin and Bacteriochlorin Derivatives
Boronated protoporphyrin BOPP Brain tumors 630
Monoaspartyl chlorin e6, (Npe6) MACE Skin cancer 654
Ring A-reduced monoacid benzoporphyrin Visudyne Non-melanoma skin cancer, age-related macular degeneration, rheumatoid arthritis, 689
derivative (BPDMA, Vertiporfin) bone marrow purging, Barett’s esophagus, psoriasis
Sn etiopurpurin (SnEt2, Rostaporfin) Purlytin Metastatic breast cancer, Kaposi’s sarcoma, prostatic cancer, brain, lung, 663
skin, head and neck cancer, age-related macular degeneration
Meso-tetra (hydroxyphenyl) chlorins Foscan Head and neck, upper aero-digestive tract cancers, gastric and prostate 652
[H2(THPC)] (Temoporfin) cancer
5-aminolevulinic acid (ALA) Levulan Actinic keratosis, basal cell carcinoma, squamous cell carcinoma, head and neck cancer, 635
gynecological tumors, Barrett’s esophagus, gastrointestinal tumors
Photodetection of bladder cancer, esophageal cancer, gastrointestinal cancer 410
5-ALA-methylesther Metvix Basal cell carcinoma, actinic keratosis 635
5-ALA-benzylesther Benzvix Gastrointestinal cancer 635
Hexylether derivative of Photochlor Basal cell carcinoma 665
pyropheophorbide (HPPH)
B. Phthalocyanine and Naphthalocyanine Based Photosensitizers
Zinc phthalocyanine CGP 55847 Squamous cell carcinoma 670
Sulphonated aluminium Photosense Skin, breast, lung, bladder, pancreatic, brain, gastrointestinal cancers 670
phthalocyanine
C. Expanded Porphyrin Based Photosensitizers
Gadolinium (III) texaphyrin XCYTRIN Brain metastases 700–780
Lutetium (III) texaphyrin LUTRIN, ANTRIN, Breast, cervical, prostate and brain tumors, angioplasty, age-related 732
OPTRIN macular degeneration
474 M. Wiedmann et al.
was given rise to by a case report, docu- the two monthly follow-ups. Quality of
menting the success of PDT performed via life indices improved dramatically and
percutaneous cholangioscopy in a single remained stable. Thirty-day mortality was
patient with incompletely resected bile 0 %, and median survival 439 days. Side
duct carcinoma (McCaughan et al., 1991). effects were limited to skin hyperpigmen-
This type of treatment caused a prolonged tation, but no acute phototoxicity in all
survival time of > 4 years in a tumor patients. One patient also reported fever
entity with a median survival of only 4–6 and pain, self-limited, for 2 weeks after
months. Ortner et al. (1998) performed the PDT (Table 35.1). Ortner’s work (Ortner,
first prospective non-randomized single- 2000) was subsequently expanded to a
arm study including nine patients with larger study of 21 patients (Table 35.3).
advanced Bismuth type III and IV hilar We performed our own study (Berr
CC, who showed no sufficient drainage et al., 2000b), including 23 patients with
after endoscopic stent insertion. Two days hilar CC (Bismuth type III, n = 2; type IV;
after intravenous application of porfimer n = 21) who were treated with a combina-
sodium (Photofrin™) at 2 mg/kg body tion of bile duct stenting and endocopi-
weight, intraluminal photoactivation at cally performed Photofrin™-PDT (Figure
630 nm was performed cholangioscopi- 35.5). In detail, we intubated the tumor
cally with laser diffusing fibers. After PDT, stenosis with a translucent ERC-cannula
bilirubin serum levels declined signifi- and inserted a 400 μm thin quartz fiber
cantly (p = 0.004) with no increase during with either a 2, 3, or 4 cm long cylindrical
duct stenosis. The median serum bilirubin sodium porfimer, and 48 h later a com-
value declined from 5.8 to 1.0 mg/dl. The mercially available cylindrical diffusing
median survival time was 119 days (52– laser fiber was inserted into an 8 F biliary
443 days). catheter equipped with a 0.038 in. side-
The therapeutic value of percutane- hole at its distal tip. After positioning
ous transhepatic PDT in patients with the 0.035 inch guidewire proximal to
advanced bile duct cancer was the subject the biliary stricture, the preloaded cath-
of a prospective trial by Shim et al. (2005). eter was advanced over the guidewire
The utility of intraductal ultrasonography by using the monorail technique. Laser
(IDUS) for the assessment of responses light was applied at a power of 400 mW/
and for regular follow-up after PDT was cm fiber for a total energy of 180 J/cm2.
also examined. PTCD was initiated before By using the preloaded biliary catheter,
PDT. Following dilation and maturation adequate positioning of the laser fiber
of the PTCD tract, percutaneous PDT was was achieved in all patients. A fracture of
performed. Intraluminal photoactivation the diffuser tip occurred during one of the
was carried out using percutaneous cholan- treatments. Two patients developed acute
gioscopy 2 days after intravenous applica- cholangitis and two patients experienced
tion of a hematoporphyrin derivative. All skin phototoxicity. The applicability of
patients were provided additionally with this technique was confirmed in recent
percutaneous bile duct drainage catheters studies by Harewood et al. (2005) and
after PDT. IDUS was conducted monthly Prasad et al. (2005) (Table 35.3). It was
to measure the thickness of the tumor concluded from the first study that (1)
mass before and after PDT. Twenty-four PDT treatment every 3 months is applica-
patients with advanced cholangiocarcino- ble, (2) patient selection for PDT need not
mas (Bismuth IIIa, n = 4; IIIb, n = 10; IV, to be based on response to primary biliary
n = 10) were treated with PDT. At 3 months stenting, and (3) the average cost of treat-
after PDT, the mean thickness of the tumor ment per patient would be US$10,337,
mass had decreased from 8.7 ± 3.7 to 5.8 ± which is based on direct medical costs
2.0 mm (p < 0.01). At 4 months after PDT, estimated from USA Medicare ambula-
the thickness of the mass had increased tory patient classification (ABC, 2003)
to 7.0 ± 3.7 mm. Quality of life indices plus professional fees for hospital-based
improved dramatically and remained sta- outpatient procedures, which can be com-
ble 1 month after PDT; the Karnofsky pared with the costs of management with
index increased from 39 to 58 points (p = biliary stenting alone. The conclusion
0.003). The 30-day mortality rate was 0 %, from the second study was that early
and the median survival time was 558 days treatment with PDT may lead to greater
(range 62–810 days). preservation of liver function based on
Another single-arm study by Rumalla a univariate and multivariate analysis,
et al. (2001) described a new method of showing a longer interval from diagnosis
applying photodynamic therapy in the to PDT treatment, which was significantly
biliary tract by using accessories avail- associated with shorter survival following
able in the United States (Table 35.3). PDT (HR 1.23, 1.02–1.47; p = 0.029 and
Patients were injected with 2 mg/kg of HR 2.38, 1.38–4.85; p = 0.0001).
35. Hilar Cholangiocarcinoma: Photodynamic Therapy and Stenting 477
The second prospective randomized trial Similar results have also been reported
by Zoepf et al. (2005) investigated the by Nanashima et al. (2004) who treated
effect of Photosan-3 based PDT plus stent- 8 patients with porfimer sodium PDT in
ing on survival time in 32 patients with an adjuvant setting. Five patients had ext-
advanced hilar CC. In the control group, rahepatic bile duct cancer, two had intra-
patients were treated with endoprostheses hepatic cholangiocarcinoma, and one had
but no PDT. PDT group and the control ampullary carcinoma. Cancer cells were
group were comparable in age, gender, microscopically detected in the stump of
performance status, bilirubin level, and the hepatic duct in six patients, and biliary
Bismuth stage. The median survival time stenosis caused by remnant tumor was
after randomization was 7 months for the observed in one patient. One patient had
control group and 21 months for the PDT tumor recurrence with occlusion of the bile
group (p = 0.01). In half of the initially duct. At 48 h prior to PDT, porfimer sodium
percutaneously treated patients, a change was injected intravenously. A pulse laser by
from percutaneous to transpapillary drain- an eximer dye laser (50–100 J/cm2) with a
age after PDT was possible. Four patients wavelength of 630 nm was applied through
showed infectious complications after PDT an endoscope to the hepatic stump or tumor
versus one patient in the control group. lesion. Marked destruction of the tumor and
ductal epithelium was observed on day 1
Photodynamic Therapy for Neoadjuvant after PDT. After PDT, four patients devel-
and Adjuvant Treatment of Hilar oped mild dermatitis, but no severe morbid-
Cholangiocarcinoma ity or mortality was noted. In patients who
underwent PDT for the stump, one patient
A new approach is the use of PDT for the showed distant metastasis at 31 months, and
treatment of hilar CC in a neoadjuvant four patients did not show tumor recurrence
setting because of the high tumor recur- at 17, 12, 12, and 6 months, respectively.
rence rate of up to 76 % after curative (R0) However, one of the eight patients died at
resection. Having accomplished a success- 2 months due to an unrelated cause. In two
ful treatment in a single case (Berr et al., patients with occlusion caused by tumor
2000a), we investigated the use of PDT growth, resolution of bile duct stenosis
prior to tumor resection in six patients and was noted on day 7. These patients showed
in one patient prior to liver transplantation in
reocclusion by tumor at 20 and 8 months.
a small pilot study (Wiedmann et al., 2003). In summary, neoadjuvant and adjuvant
In all patients, R0 resection was achieved. PDT in these two small studies were safe
Four patients developed minor surgical and useful options for a better survival
complications, even though the bilioenteric benefit in patients with bile duct cancer
anastomoses were sewn to PDT-treated bile undergoing surgical resection; however,
ducts. No viable tumor cells were found in further studies are required.
the inner 4 mm layer of the surgical speci-
mens. The PDT-pretreated epithelium of
Future Directions of Photodynamic
the tumor-free proximal resection margins
Therapy for Hilar Cholangiocarcinoma
exhibited only minimal inflammatory infil-
tration. The 1-year recurrence free survival The penetration depth of Photofrin™ and
rate was 83 %. Photosan™ is 4–4.5 mm, and exceeds the
35. Hilar Cholangiocarcinoma: Photodynamic Therapy and Stenting 479
Berr, F., Wiedmann, M., Tannapfel, A., Halm, U., 2003. Phase II study of photodynamic therapy
Kohlhaw, K. R., Schmidt, F., Wittekind, C., and metal stent as palliative treatment for nonre-
Hauss, J., and Mossner, J., 2000b. Photodynamic sectable hilar cholangiocarcinoma. Gastrointest.
therapy for advanced bile duct cancer: evidence Endosc. 57: 860–867.
for improved palliation and extended survival. Ebata, T., Nagino, M., Kamiya, J., Uesaka, K.,
Hepatology 31: 291–2988. Nagasaka, T., and Nimura, Y., 2003. Hepatectomy
Burmester, E., Niehaus, J., Leineweber, T., and with portal vein resection for hilar cholangiocar-
Huetteroth, T., 2003. EUS-cholangio-drainage cinoma: audit of 52 consecutive cases. Ann.
of the bile duct: report of 4 cases. Gastrointest. Surg. 238: 720–727.
Endosc. 57: 246–251. Freeman, M. L. and Overby, C., 2003. Selective
Chang, W. H., Kortan, P., and Haber, G. B., MRCP and CT-targeted drainage of malignant
1998. Outcome in patients with bifurcation hilar biliary obstruction with self-expanding
tumors who undergo unilateral versus bilateral metallic stents. Gastrointest. Endosc. 58:
hepatic duct drainage. Gastrointest. Endosc. 47: 41–49.
354–362. Harewood, G. C., Baron, T. H., Rumalla, A., Wang,
Cherqui, D., Benoist, S., Malassagne, B., Humeres, K. K., Gores, G. J., Stadheim, L. M., and de
R., Rodriguez, V., and Fagniez, P. L., 2000. Groen, P. C., 2005. Pilot study to assess patient
Major liver resection for carcinoma in jaundiced outcomes following endoscopic application of
patients without preoperative biliary drainage. photodynamic therapy for advanced cholan-
Arch. Surg. 135: 302–308. giocarcinoma. J. Gastroenterol. Hepatol. 20:
de Groen, P. C., Gores, G. J., LaRusso, N. F., 415–420.
Gunderson, L. L., and Nagorney, D. M., 1999. Hemming, A. W., Reed, A. I., Fujita, S., Foley, D.
Biliary tract cancers. N. Engl. J. Med. 341: P., and Howard, R. J., 2005. Surgical manage-
1368–1378. ment of hilar cholangiocarcinoma. Ann. Surg.
De Palma, G. D., Galloro, G., Siciliano, S., Iovino, 241: 693–699; discussion 699–702.
P., and Catanzano, C., 2001. Unilateral versus Isayama, H., Komatsu, Y., Tsujino, T., Sasahira, N.,
bilateral endoscopic hepatic duct drainage in Hirano, K., Toda, N., Nakai, Y., Yamamoto, N.,
patients with malignant hilar biliary obstruction: Tada, M., Yoshida, H., Shiratori, Y., Kawabe, T.,
results of a prospective, randomized, and control- and Omata, M., 2004. A prospective randomised
led study. Gastrointest. Endosc. 53: 547–553. study of “covered” versus “uncovered” diamond
Deviere, J., Baize, M., de Toeuf, J., and Cremer, stents for the management of distal malignant
M., 1988. Long-term follow-up of patients with biliary obstruction. Gut 53: 729–734.
hilar malignant stricture treated by endoscopic Kadakia, S. C. and Starnes, E., 1992. Comparison
internal biliary drainage. Gastrointest. Endosc. of 10 French gauge stent with 11.5 French gauge
34: 95–101. stent in patients with biliary tract diseases.
Doctor, N., Dick, R., Rai, R., Dafnios, N., Salamat, Gastrointest. Endosc. 38: 454–459.
A., Whiteway, H., Dooley, J., and Davidson, B. Kawasaki, S., Imamura, H., Kobayashi, A., Noike,
R., 1999. Results of percutaneous plastic stents T., Miwa, S., and Miyagawa, S., 2003. Results of
for malignant distal biliary obstruction following surgical resection for patients with hilar bile duct
failed endoscopic stent insertion and comparison cancer: application of extended hepatectomy
with current literature on expandable metal- after biliary drainage and hemihepatic portal
lic stents. Eur. J. Gastroenterol. Hepatol. 11: vein embolization. Ann. Surg. 238: 84–92.
775–780. Khan, S. A., Taylor-Robinson, S. D., Toledano,
Ducreux, M., Liguory, C., Lefebvre, J. F., Ink, O., M. B., Beck, A., Elliott, P., and Thomas, H. C.,
Choury, A., Fritsch, J., Bonnel, D., Derhy, S., 2002. Changing international trends in mortality
and Etienne, J. P., 1992. Management of malig- rates for liver, biliary and pancreatic tumours. J.
nant hilar biliary obstruction by endoscopy. Hepatol. 37: 806–183.
Results and prognostic factors. Dig. Dis. Sci. 37: Kondo, S., Hirano, S., Ambo, Y., Tanaka, E.,
778–783. Okushiba, S., Morikawa, T., and Katoh, H., 2004.
Dumoulin, F. L., Gerhardt, T., Fuchs, S., Scheurlen, Forty consecutive resections of hilar cholangio-
C., Neubrand, M., Layer, G., and Sauerbruch, T., carcinoma with no postoperative mortality and no
35. Hilar Cholangiocarcinoma: Photodynamic Therapy and Stenting 481
Seyama, Y., Kubota, K., Sano, K., Noie, T., Wagner, H. J., Knyrim, K., Vakil, N., and Klose,
Takayama, T., Kosuge, T., and Makuuchi, M., K. J., 1993. Plastic endoprostheses versus metal
2003. Long-term outcome of extended hemi- stents in the palliative treatment of malignant
hepatectomy for hilar bile duct cancer with no hilar biliary obstruction. A prospective and ran-
mortality and high survival rate. Ann. Surg. 238: domized trial. Endoscopy 25: 213–218.
73–83. Welzel, T. M., McGlynn, K. A., Hsing, A. W.,
Shim, C. S., Cheon, Y. K., Cha, S. W., Bhandari, O’Brien, T. R., and Pfeiffer, R. M., 2006. Impact
S., Moon, J. H., Cho, Y. D., Kim, Y. S., Lee, L. of classification of hilar cholangiocarcinomas
S., Lee, M. S., and Kim, B. S., 2005. Prospective (Klatskin tumors) on the incidence of intra- and
study of the effectiveness of percutaneous tran- extrahepatic cholangiocarcinoma in the United
shepatic photodynamic therapy for advanced States. J. Natl. Cancer Inst. 98: 873–875.
bile duct cancer and the role of intraductal ultra- West, J., Wood, H., Logan, R. F., Quinn, M., and
sonography in response assessment. Endoscopy Aithal, G. P., 2006. Trends in the incidence
37: 425–433. of primary liver and biliary tract cancers in
Soderlund, C. and Linder, S., 2006. Covered England and Wales 1971–2001. Br. J. Cancer.
metal versus plastic stents for malignant com- 94: 1751–1758.
mon bile duct stenosis: a prospective, rand- Wiedmann, M., Berr, F., Schiefke, I., Witzigmann,
omized, controlled trial. Gastrointest. Endosc. H., Kohlhaw, K., Mossner, J., and Caca, K.,
63: 986–995. 2004a. Photodynamic therapy in patients with
Speer, A. G., Cotton, P. B. and MacRae, K. D., non-resectable hilar cholangiocarcinoma: 5-year
1988. Endoscopic management of malignant follow-up of a prospective phase II study.
biliary obstruction: stents of 10 French gauge Gastrointest. Endosc. 60: 68–75.
are preferable to stents of 8 French gauge. Wiedmann, M., Caca, K., Berr, F., Schiefke, I.,
Gastrointest. Endosc. 34: 412–417. Tannapfel, A., Wittekind, C., Mossner, J., Hauss,
Speer, A. G., Cotton, P. B., Russell, R. C., Mason, J., and Witzigmann, H., 2003. Neoadjuvant pho-
R. R., Hatfield, A. R., Leung, J. W., MacRae, todynamic therapy as a new approach to treating
K. D., Houghton, J., and Lennon, C. A., 1987. hilar cholangiocarcinoma: a phase II pilot study.
Randomised trial of endoscopic versus percu- Cancer 97: 2783–2790.
taneous stent insertion in malignant obstructive Wiedmann, M., Dietrich, A., Mossner, J.,
jaundice. Lancet 2: 57–62. Witzigmann, H., and Caca, K., 2004b. Combined
Todoroki, T., Kawamoto, T., Koike, N., Takahashi, percutaneous transhepatic biliary drainage with
H., Yoshida, S., Kashiwagi, H., Takada, Y., port implantation for management of patients
Otsuka, M., and Fukao, K., 2000. Radical resec- with malignant biliary obstruction. Gastrointest.
tion of hilar bile duct carcinoma and predictors Endosc. 60: 117–120.
of survival. Br. J. Surg. 87: 306–313. Witzigmann, H., Berr, F., Ringel, U., Caca, K.,
Tringali, A., Mutignani, M., Perri, V., Zuccala, G., Uhlmann, D., Schoppmeyer, K., Tannapfel, A.,
Cipolletta, L., Bianco, M. A., Rotondano, G., Wittekind, C., Mossner, J., Hauss, J., and Wiedmann,
Philipper, M., Schumacher, B., Neuhaus, H., M., 2006. Surgical and Palliative Management
Schmit, A., Deviere, J., and Costamagna, G., and Outcome in 184 Patients With Hilar
2003. A prospective, randomized multicenter Cholangiocarcinoma: Palliative Photodynamic
trial comparing DoubleLayer and polyethylene Therapy Plus Stenting is Comparable to R1/R2
stents for malignant distal common bile duct Resection. Ann. Surg. 244: 230–239.
strictures. Endoscopy 35: 992–997. Yoon, W. J., Lee, J. K., Lee, K. H., Lee, W. J.,
van Berkel, A. M., Boland, C., Redekop, W. K., Ryu, J. K., Kim, Y. T., and Yoon, Y. B., 2006. A
Bergman, J. J., Groen, A. K., Tytgat, G. N., and comparison of covered and uncovered Wallstents
Huibregtse, K., 1998. A prospective randomized for the management of distal malignant biliary
trial of Teflon versus polyethylene stents for obstruction. Gastrointest. Endosc. 63: 996–1000.
distal malignant biliary obstruction. Endoscopy Zoepf, T., Jakobs, R., Arnold, J. C., Apel, D., and
30: 681–686. Riemann, J. F., 2005. Palliation of nonresectable
35. Hilar Cholangiocarcinoma: Photodynamic Therapy and Stenting 483
bile duct cancer: improved survival after pho- a new diode laser system. Am. J. Gastroenterol.
todynamic therapy. Am. J. Gastroenterol. 100: 96: 2093–2097.
2426–2430. Zoepf, T., Jakobs, R., Rosenbaum, A., Apel,
Zoepf, T., Jakobs, R., Arnold, J. C., Apel, D., D., Arnold, J. C., and Riemann, J. F., 2001b.
Rosenbaum, A., and Riemann, J. F., 2001a. Photodynamic therapy with 5-aminolevulinic
Photodynamic therapy for palliation of nonresect- acid is not effective in bile duct cancer.
able bile duct cancer – preliminary results with Gastrointest. Endosc. 54: 763–766.
A. Diagnosis
36
Splenic Metastases: Diagnostic Methods
Eva Compérat and Frédéric Charlotte
489
490 E. Compérat and F. Charlotte
PATHOGENESIS
The development of metastases in spe-
cific organs according to the type of pri-
mary source depends on anatomic factors,
the microenvironment of host tissue, and
the intrinsic characteristics of tumor cells
(Chambers et al., 2002). The pathogenesis
of splenic metastases is largely specula-
tive. The relative rarity of splenic metas-
tases might be explained by two main
causes: (1) mechanical factors impeding
Figure 36.1. Macroscopic findings: multiple large the splenic implantation of blood-borne
nodules of varying sizes in spleen parenchyma cancer cells, that is, the constant flow of
36. Splenic Metastases: Diagnostic Methods 491
blood through the spleen and the rhythmic various types of metastatic cells, are found
contraction of splenic capsule; the sharp not only in the preferential sites of metas-
angle of splenic artery branching from the tases such as the bone marrow, lung and
celiac artery preventing large clumps of lymph nodes, but also in the spleen where
tumor cells from passing through; the lack it plays a crucial role in lymphocyte hom-
of afferent lymphatic vessels limiting lym- ing (Chambers et al., 2002). Therefore,
phogenic metastases, and (2) inhibitory implantation of cancer cells in the splenic
effect of splenic microenvironment on the parenchyma may occur, but clinically
growth of metastatic cells (Berge, 1974). detectable metastases are rare because
The respective contribution of these factors the splenic microenvironment may not
to the resistance of the spleen to metastatic facilitate the growth of micrometastatic
involvement is still debated. The mechani- foci. This probably explains the contrast
cal theory was proposed at the time the between the high prevalence of splenic
dissemination from the primary tumor and micrometastases at autopsy in patients
the implantation of tumor cells in distant with multivisceral cancer and the rarity
organs were regarded as a late event in of clinically detectable metastases. This
the multistep process of cancer. However, suggests that mechanical factor seems to
the recent development of sensitive immu- play a minor role in explaining the rarity
nologic and molecular methods has per- of clinical splenic metastases.
mitted the detection of individual tumor
cells in regional lymph nodes, blood, or
distant organs at the time of primary tumor CLINICAL FEATURES
diagnosis for various histological types
of cancer (Pantel and Brakenhoff, 2004). Most splenic metastases are a part of
This supports the concept that micrometa- multivisceral metastatic disease. In this
static dissemination occurs, in fact, early context, breast, lung, ovarian, colorectal, and
in the course of malignant disease and is gastric carcinomas, and skin melanoma are
not affected by mechanical factors. This the most common primary sources (Berge,
is also in accordance with the results of 1974; Lam and Tang, 2000). Skin melanoma
experimental metastasis models using neo- has the highest rate of splenic metastases
plastic cells labeled with green fluorescent per primary tumor since more than 30 % of
protein (Goodison et al., 2003). patients with skin melanoma have splenic
From this recent advance in the knowl- metastasis at autopsy (Berge, 1974; Lam and
edge of metastatic process, it may be Tang, 2000). Splenic metastases also occur
assumed that the presence of disseminated as a solitary splenic mass, synchronous or
cancer cells in the spleen may not be metachronous to the primary tumor. To our
detected at the time of primary diagnosis knowledge, 93 well-documented cases of
by conventional methods. It is noteworthy solitary splenic metastases are known which
that the expression of chemokines in the are listed in Table 36.1. From diagnosis of
spleen does not constitute an obstacle to primary tumor to the discovery of solitary
cancer cell homing because stromal-cell- splenic metastasis it might take between
derived factor 1 (SDF-1) and the ligand 0 to 264 months with a median of 28
for CXC chemoreceptor-4 expressed by months. Colorectal and ovarian carcinomas
492 E. Compérat and F. Charlotte
are the most common sources. On the FDG-PET is to discriminate benign from
other hand, solitary splenic metastases malignant splenic masses, the latter is char-
from breast carcinoma (Barreca et al., 2001; acterized by an increase in FDG uptake.
Iype et al., 2002; Hoar et al., 2003) or skin Moreover, FDG-PET coupled with CT can
melanoma are rare. Exceptionally, a splenic also identify other tumor sites expressing
metastasis reveals a clinically occult primary as hypermetabolic foci which are unsus-
tumor (McGregor et al., 2003). Only one pected on conventional imaging but may
report described a solitary splenic metas- be more accessible for biopsy.
tasis from unknown primary cancer (Alici Serum level of tumor markers utilized
et al., 2003). in the follow-up of patient with cancer can
Splenic metastases are most often predict the appearance of solitary splenic
incidentally detected by medical imag- metastases before their clinical expression
ing techniques. When isolated, more than or radiological detection and decreases
60 % of splenic metastases are asympto- within normal limits after splenectomy
matic (Table 36.1). Splenic metastasis can (Dubecq-Princeteau et al., 1997). For
also be revealed by fatigue (Ishida et al., example, Compérat et al. (2007) have
1997), weight loss (Agha-Mohammadi and recently seen a solitary splenic metastasis
Calne, 2001), fever (Kugel et al., 2003), from prostate carcinoma occurring 5 years
abdominal pain (Carvalho et al., 1997; after a radical prostatectomy (Figure 36.3).
Agha-Mohammadi and Calne, 2001; Iype A progressive increase in the serum PSA
et al., 2002; Vyas, 2002; Mayayo et al.,
2003; Schmidt and Smith, 2004; Tas et al.,
2004;), splenomegaly (Sharpe et al., 1993;
Agha-Mohammadi and Calne, 2001; Alici
et al., 2003), anemia or thrombocytopenia
due to hypersplenism (Kugel et al., 2003;
Hadjileontis et al., 2004), and more rarely
by splenic rupture (Hoar et al., 2003;
Schmidt and Smith, 2004).
DIAGNOSTIC PROCEDURES
Splenic metastases can be detected by
ultrasonography or computed tomography
(CT) in the regular follow-up of patients
with cancer or in the work-up performed
at the time of any event related to cancer.
The increasing use of 18-fluorodeoxyglu-
cose (FDG)-positron emission tomography
(PET) has resulted in more patients with
asymptomatic metastases being identified
than previously found using conventional Figure 36.3. Solitary splenic metastasis from pros-
radiological techniques. The advantage of tate carcinoma, H&E × 4
494 E. Compérat and F. Charlotte
level was observed before the tumor could 0 % and 2 % and mainly affect patients
be detected by imaging technique. Serum with thrombocytopenia or vascular splenic
PSA level decreased within normal limits neoplasms.
after splenectomy and the patient remained
asymptomatic 13 months thereafter. This
case history along with other published DIFFERENTIAL DIAGNOSIS
similar cases suggests that splenic metas-
tases might result from the growth of early Despite improvement of medical imag-
blood-borne disseminated cancer cells ing techniques such as PET, the clini-
within the spleen after a period of clinical cal diagnosis of splenic metastases is
latency. This could explain why isolated largely dependent on patient history.
splenic metastases are not necessarily the Radiologically, metastatic disease fre-
precursor sign of active metastatic can- quently affects the spleen in a manner
cer at terminal stage because long term that is indistinguishable from primary
remission has been achieved in patients splenic lesions. A splenic mass without
treated with splenectomy alone (Agha- context of active cancerous disease is
Mohammadi and Calne, 2001). Further more suggestive of a primary splenic
follow-up studies of patients with solitary lesion such as lymphomas, vascular
splenic metastasis detected by imaging tumors, or infectious lesions. Conversely,
techniques such as PET, are needed to a previous history of malignant disease
better understand the significance of this constitutes an indication for the diagno-
clinical manifestation. sis of splenic metastasis. The search for
The finding of a splenic nodule is additional clinical findings such as lym-
regarded as metastatic in case of dissemi- phadenopathy or disease in other organs
nated cancerous disease, and splenectomy may aid in the diagnosis of lymphoma or
is generally not performed, as the diagno- metastatic disease. Rarer splenic lesions
sis relies on biopsy of a more accessible such as hamartoma and inflammatory
tumor site. Isolated splenic metastasis can pseudotumor can clinically and radiologi-
be diagnosed by fine-needle aspiration or cally mimic a metastatic tumor but there
percutaneous biopsy in place of splenec- is generally no confusion on histological
tomy. The use of these imaging-guided grounds.
techniques to diagnose splenic metastases Primary malignant lymphoma of the
has rarely been reported in the medical lit- spleen is defined as a disease limited to
erature. This may be attributed to the fear the spleen and splenic hilum. The main
of hemorrhagic complication because the histological types include large B-cell
spleen is highly vascular. However, these lymphoma, marginal zone B-cell lymphoma,
imaging-guided techniques are less aggres- mantle cell lymphoma, hairy-cell leuke-
sive than the surgical approach and permit mia, and gamma-delta T-cell lymphoma
the avoidance of unnecessary splenectomy (Mollejo et al., 2003). Large B-cell lym-
for the diagnosis of benign splenic lesions phoma can be mistaken for metastasis from
such as granulomatous disease. A success- undifferentiated carcinoma or melanoma.
ful diagnosis can be achieved in ∼ 90 % of Immunostaining with cytokeratins for car-
the cases. Complications range between cinoma, common leukocyte antigen for
36. Splenic Metastases: Diagnostic Methods 495
Mayayo, E., Blazquez, S., Gomez-Aracil, V., Sauri, vix in the spleen: case report. South Med. J.
A., and Martinez, S. 2003. Spleen metastasis 97:301–304.
from thyroid carcinoma. Report of a case with Pantel, K., and Brakenhoff, R.H. 2004. Dissecting
diagnosis by fine needle aspiration cytology. the metastatic cascade. Nat. Rev. Cancer 4:
Acta Cytol. 47:1116–1118. 448–456.
McGregor, D.H., Wu, Y., Weston, A.P., Mcanaw, Schmidt, B.J., and Smith, S.L. 2004. Isolated splenic
M.P., Bromfield, C., and Bhattatiry, M.M. 2003. metastasis from primary lung Adenocarcinoma.
Metastatic renal cell carcinoma of spleen diag- South. Med. J. 97:298–300.
nosed by fine-needle aspiration Spleen. Am. J. Sharpe, R.W., Rector, J.T., Rushin, J.M., Garvin,
Med. Sci. 326:51–54. D.F., and Cotelingam, J.D. 1993. Splenic metas-
Mollejo, M., Algara, P., Mateo, M., Menarguez, tasis in hairy cell leukemia. Cancer 71:2222–
J., Pascual, E., Fresno, M.F., Camacho, F., and 2226.
Piris, A. 2003. Large B-cell lymphoma present- Tas, F., Ustuner, Z., Buyukbabani, N., Tenekeci, N.,
ing in the spleen: Identification of different clin- and Topuz, E. 2004. Massive and isolated metas-
icopathologic conditions. Am. J. Surg. Pathol. tases to spleen of uveal malignant melanoma.
27: 895–902. Retina 24:170–172.
Naseem, M.S., Jan, H.A., Britton, K.E., and Tserkezoglou, A., Kontou, S., Hatjieleftheriou,
Nargund, V.H. 1998. Splenic metastasis from G., Nikolaidou, M.E., Plataniotis, G.,
adenocarcinoma of the prostate. Br. J. Urol. Apostolikas, N., and Magiakos, G. 2005.
82:597–598. Solitary parenchymal splenic recurrence
Okuyama, T., Oya, M., and Ishikawa, H. 2001. of ovarian adenocarcinoma: a case report
Isolated splenic metastasis of sigmoid colon and review of the literature. Anticancer Res.
cancer: a case report. Jpn. J. Clin. Oncol. 25:1471–1476.
31:341–345. Vyas, S.J., Chitaleb, A.R., and Deshpande, R.K.
Opocher, E., Santambrogio, R., Bianchi, P., Cioffi, 2002. Late splenic metastasis after curative
U., De Simone, M., Vellini, S., and Montorsi, resection for oesophageal carcinoma. Eur. J.
M. 2000. Isolated splenic metastasis from gastric Cardiothorac. Surg. 22:1011–1013.
carcinoma. value of CEA and CA 19–9 in early Yamanouchi, K., Ikematsu, Y., Waki, S., Kida, H.,
diagnosis: report of two case. Am. J. Clin. Oncol. Nishiwaki, Y., Gotoh, K., Ozawa, T., and
23:579–580. Uchimura, M. 2002. Solitary splenic metastasis
Pang, L.C. 2004. Solitary recurrent metastasis of from gastric cancer: report of a case. Surg. Today
squamous cell carcinoma of the uterine cer- 32:1081–1084.
Index
AASLD. See American Association for the Study Alcohol. See Fatty liver
of Liver Diseases guidelines Aldolase B (liver type), hepatitis C virus-
Abdomen related human hepatocellular
magnetic resonance imaging of, 21 carcinoma and, 348
T2-weighted images of, 22 Alpha-enolase, hepatitis C virus-related
2DT-FSE, 22 human hepatocellular carcinoma
lesion characterization with, 22–23 and, 347
single-shot echo planar, 22, 23 Alpha-fetoprotein (AFP), for hepatocellular
single-shot fast spin-echo, 22, 23 carcinoma screening, 138–139, 174
steady-state free precession, 22 Alternative-free response receiver operating
Abscess, characterization of, 167 characteristic (AFROC), 231
arterial phase, 167 American Association for the Study of Liver
parenchymal phase, 167 Diseases (AASLD) guidelines, 177,
portal vein phase, 167 262–263
typical features, 167 Angiography-assisted computed tomography,
ActiTest, 139, 140 36–37, 253
Adenoma, characterization of, 166–167 Angiosarcoma, characterization of, 171
arterial phase, 167 Antiangiogenesis therapy, evaluation of, future
atypical features, 167 perspectives of, 282–284
parenchymal phase, 167 APRI. See AST to platelet ratio index
portal vein phase, 167 Arginase 1, hepatitis C virus-related
typical features, 167 human hepatocellular carcinoma
Adenomatous hyperplasia, computed and, 348
tomography of, 255 Arterial portal microfistulas, contrast-enhanced
Adenosine triphosphate (ATP) synthase ultrasound and, 164–165
beta chain, hepatitis C virus-related Aspartate transaminase (AST)/alanine
human hepatocellular carcinoma transaminase (ALT) ratio, 139, 140
and, 347 AST/ALT. See Aspartate transaminase/alanine
ADI. See Agent detection imaging transaminase ratio
AFROC. See Alternative-free response receiver AST to platelet ratio index (APRI), 139, 140
operating characteristic ATP. See Adenosine triphosphate synthase beta
Agent detection imaging (ADI), 163 chain
contrast-enhanced, 151–153 Atypical adenomatous hyperplasia, computed
AI-700, 161 tomography of, 255
Albunex, 161 AXOR12, 325
499
500 Index
Colon/rectum, splenic metastases to, 492 enhanced findings, combined hepatocellular and
Colorectal cancer cholangiocarcinoma, 243–247
liver metastases and, 66–67 of localized fibrous tumors of liver, 17–19
detection of, 358–359 multidetector, methods of, 253
diagnostic hepatic ultrasound advantages, 357 of nodule-in-nodule hepatocellular
Doppler flow pattern, 359–363 carcinoma, 255
future advancements in, 363–364 spiral, for hepatocellular carcinoma, 265–266
grayscale echo pattern, 357–358 vascular endothelial growth factor, Western
postoperative follow-up ultrasound, 363 blotting measurement of, 202–209
ultrasound scanning technique for, Computed tomography during arterial portography
356–357 (CTAP), 37–39, 214, 253
liver metastases and, 18F-fluorodeoxyglucose Computed tomography during hepatic
positron emission tomography, 389–404 arteriography (CTHA), 37, 39, 253
abdominal cavity, 392–395 Contrast-enhanced agent detection imaging,
additional disease and, 398–399 151–153
anticancer systemic therapy, 401–402 Contrast-enhanced coded harmonic sonography,
benign increased activity of, 392, 397 150–151
change of management, 398–399 Contrast-enhanced color Doppler sonography,
contribution of, 393, 397 146–147
cryosurgery ablation, 403 Contrast-enhanced harmonic power Doppler
detection of, 389–391 sonography, 147–148
distant lymph nodes metastases, 395 Contrast-enhanced intraoperative ultrasonography
extra-abdominal organs, 395–397 (CE-IOUS), 63–64, 65–66, 176
extrahepatic metastases detection, 391–392 classification by, 66
hepatic arterial chemotherapy, 403–404 colorectal cancer, liver metastases and, 66–67
on patient’s management, 397–398 liver metastases, colorectal cancer and, 66–67
pelvic recurrence of rectal cancer, 393–395 tumor location by, 67–69
peritoneal metastases, 395 Contrast-enhanced power Doppler
physiologic activity of, 392, 397 sonography, 147
preoperative staging, 389 Contrast-enhanced pulse-inversion harmonic
radiofrequency ablation, 403 sonography, 148–150
regional therapy for, 401, 403 Contrast-enhanced sonography, hepatocellular
rising carcinoembryonic antigen, 400–401 carcinoma and, 153–156
survival impact, 399–400 Contrast-enhanced ultrasound (CEUS). See also
systemic therapy for, 401 Harmonic imaging; Nonlinear contrast
Combined hepatocellular and cholangiocarcinoma enhancement
(cHCC-CC) arterial portal microfistulas and, 164–165
differential diagnosis of, 246–247 for hepatocellular carcinoma, 264–265
enhanced computed tomographic findings, physical background of, 162–163
243–247 technological background of, 162–163
pathogenesis of, 242–243 Contrast Pulse Sequencing, 163
Computed tomography (CT) Cryoablation, 85
of adenomatous hyperplasia, 255 Cryosurgery ablation, colorectal cancer, liver
of advanced hepatocellular carcinoma, 253–254 metastases and, 18F-fluorodeoxyglucose
angiography, 36–37, 228–229 positron emission tomography, 403
angiography combined, 253 CT. See Computed tomography
of atypical adenomatous hyperplasia, 255 CTAP. See Computed tomography during arterial
contrast material fixed injection duration, 239 portography
contrast material fixed injection rate, 237–239 CTHA. See Computed tomography during hepatic
of early hepatocellular carcinoma, 254–255 arteriography
502 Index
Focal liver lesions (FLLs), nonlinear contrast Glucose-regulated 78 kDa protein (GRP78),
enhancement of (Continued) hepatitis C virus-related human
histiocytoma, characterization of, 171 hepatocellular carcinoma and, 347
inflammatory pseudo tumors, characterization Glutamine synthetase, hepatitis C virus-related
of, 167–168 human hepatocellular carcinoma and, 347
leiomyoma, characterization of, 171 Gradient-recalled echo (GRE), 184
liver metastases, characterization of, 168–171 Grayscale echo pattern, liver metastases,
arterial phase, 170, 171 colorectal cancer and, 357–358
hyper-vascular, 171 GRE. See Gradient-recalled echo
hypo-vascular, 170 GRP75. See Glucose-regulated 75 kDa protein
parenchymal phase, 170, 171 GRP78. See Glucose-regulated 78 kDa protein
portal vein phase, 170, 171
typical features, 170, 171 HAIC. See Hepatic arterial infusion
lymphoma, characterization of, 171 chemotherapy
percutaneous ablative treatment monitoring, HAP. See Hepatic arterial-dominant phase
174–176 scanning
regenerative nodule, characterization of, 166 Harmonic imaging, 160–161
arterial phase, 166 advantage of, 160
parenchymal phase, 166 disadvantage of, 160
portal vein phase, 166 HBV. See Hepatitis B virus
typical features, 166 HCC. See Hepatocellular carcinoma
sarcomas, characterization of, 171 HCV. See Hepatitis C virus
Focal nodular hyperplasia, characterization of, 165 HCV-HCC. See Hepatitis C virus-related human
arterial phase, 165 hepatocellular carcinoma
atypical features, 165 HDGF. See Hepatoma-derived growth factor
parenchymal phase, 165 Heat shock 60 kDa protein (HSP60), hepatitis
portal vein phase, 165 C virus-related human hepatocellular
typical features, 165 carcinoma and, 347
Forns index, 139, 140 Heat shock 70 kDa protein 1 (HSP70.1), hepatitis
FSE. See Fast spin echo C virus-related human hepatocellular
carcinoma and, 347
Gadobenate dimeglumine (Gd-BOPTA), 89, 227 Heat shock cognate 71 kDa protein (HSC70),
Gadolinium hepatitis C virus-related human
hepatocellular carcinoma, enhancement of, hepatocellular carcinoma and, 347
185–187 Hemangioma, characterization of
superparamagnetic iron oxide and, 187 arterial phase, 165
Gadolinium chelates, 39–40, 88 atypical features, 165
Gadoxetic acid (Gd-ethoxybenzyldiethylenetriami capillary, 164
nepentaacetic acid; Gd-EOB-DTPA), 89, cavernous, 164
94, 227 focal liver lesions, nonlinear contrast
Gallbladder carcinoma, 2-fluoro-2-deoxy-D- enhancement of, 164–165
glucose-positron emission tomography parenchymal phase, 165
for, 11–12 portal vein phase, 165
Gd-BOPTA. See Gadobenate dimeglumine typical features, 165
Gd-EOB-DTPA. See Gadoxetic acid Hepatic arterial-dominant phase (HAP)
Gd-ethoxybenzyldiethylenetriaminepentaacetic scanning, 237
acid. See Gadoxetic acid Hepatic arterial infusion chemotherapy (HAIC),
Glucose-regulated 75 kDa protein (GRP75), 287, 289–290
hepatitis C virus-related human Hepatic foregut cyst, non-neoplastic biliary cystic
hepatocellular carcinoma and, 347 lesions, 414–415
Index 505
Nonlinear contrast enhancement, of focal liver Non-neoplastic biliary cystic lesions, 412–416
lesions (Continued) biliary hamartoma, 415–416
arterial phase, 166 multicystic biliary hamartoma, 415–416
focal fatty change, 166 Von Meyenburg complexes, 415
focal fatty sparing, 165–166 hepatic foregut cyst, 414–415
parenchymal phase, 166 prebiliary cysts, 412–414
portal vein phase, 166 around extrahepatic bile ducts, 414
typical features, 166 in hepatic hilum, 412–414
focal nodular hyperplasia, characterization in large portal tracts, 412–414
of, 165 Nonvital parasitic cyst, characterization of, 166
arterial phase, 165 Nyquist criteria, 27
atypical features, 165
parenchymal phase, 165 Obesity. See Fatty liver
portal vein phase, 165 Oblique reconstruction images,
typical features, 165 cholangiocarcinoma, three-dimensional
hemangioma, characterization of, 164–165 intraductal ultrasonography, 432–433
arterial phase, 165 Oncologic patients, tumor arising in, 36
atypical features, 165 Optison, 161
capillary, 164 Ovary, splenic metastases to, 492
cavernous, 164
focal liver lesions, nonlinear contrast P53, role of, in malignant tumors of bile duct,
enhancement of, 164–165 443–444
parenchymal phase, 165 analysis of, 444
portal vein phase, 165 P63, role of, in malignant tumors of bile duct,
typical features, 165 444–448
hepatocellular carcinoma, characterization of analysis of, 447–448
arterial phase, 168 P73, role of, in malignant tumors of bile duct,
parenchymal phase, 168 444–448
portal vein phase, 168 analysis of, 447–448
typical features, 168 Palliative bile duct stenting, 464–469
histiocytoma, characterization of, 171 Pancreas mucinous cystic neoplasm
inflammatory pseudo tumors, characterization cystadenocarcinoma comparison, 419
of, 167–168 hepatobiliary cystadenoma comparison, 419
leiomyoma, characterization of, 171 Parenchymal dissection, in hepatocellular
liver metastases, characterization of, 168–171 carcinoma, 71–72
arterial phase, 170, 171 PEI. See Percutaneous ethanol injection
hyper-vascular, 171 Percutaneous ablative treatment monitoring,
hypo-vascular, 170 nonlinear contrast enhancement, of focal
parenchymal phase, 170, 171 liver lesions, 174–176
portal vein phase, 170, 171 Percutaneous ethanol injection (PEI), 153, 174
typical features, 170, 171 Percutaneous radiofrequency ablation therapy,
lymphoma, characterization of, 171 hepatocellular carcinoma after, 232–233
percutaneous ablative treatment monitoring, Percutaneous transhepatic cholangiography (PTC),
174–176 for cholangiocarcinoma, 10
regenerative nodule, characterization of, 166 Perfusion-weighted magnetic resonance
arterial phase, 166 imaging, 190
parenchymal phase, 166 Periodically Rotated Overlapping ParallEL
portal vein phase, 166 Lines with Enhanced Reconstruction
typical features, 166 (PROPELLER), 215–216
sarcomas, characterization of, 171 superparamagnetic iron oxide-enhanced DWI
Index 511