Liver Cancer

Methods of Cancer Diagnosis, Therapy, and Prognosis

Volume 5

For other titles published in this series, go to

www.springer.com/series/8172
Methods of Cancer Diagnosis,
Therapy, and Prognosis

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

ISBN 978-1-4020-9803-1 e-ISBN 978-1-4020-9804-8

Library of Congress Control Number: 2008930172

© 2009 Springer Science + Business Media B.V.

No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic,
mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception
of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the
purchaser of the work.

Printed on acid-free paper

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
Maria I. Altbach
Department of Radiology, University
of Arizona, 1501 N. Campbell Avenue,
Tucson, AZ 85724, USA
Tsutomu Araki
Department of Radiology, University
of Yamanashi, 1110 Simokato,
Chuo, Yamanashi, 409-3898, Japan
Rebecca Auer
Centre for Cancer Therapeutics,
Ottawa Health Research Institute,
501 Smyth Road, 7NE
Cancer Assessment Clinic, Ottawa,
ON, K1H 8L6, Canada
Kazuo Awai
Kumamoto University,
School of Medicine,
Department of Radiology,
1-1-1 Honjo, Kumamoto 860,
Japan
Yuji Baba
Kumamoto University, School of
Medicine, Department of Radiology,
1-1-1 Honjo, Kumamoto 860,
Japan
Oliver F. Bathe
Department of Surgery and Oncology,
University of Calgary, Tom Baker Cancer
Center, 1331 29th Street, N.W., Calgary,
AB T2N 4N2, Canada
Andrew K. Burroughs
Royal Free Hospital, Liver
Transplantation and Hepatobiliary
Medicine, Pond Street, London NW3
2QG, UK
Ann F. Chambers
Department of Medical Biophysics,
London Regional Cancer Program,
790 Commissioners Road East,
London, Ontario, Canada, N6A 4L6
Frédéric Charlotte
Service d’Anatomie Pathologie,
GH La Pitié-Salpetrière,
Boulevard de l’hôpital, 75013 Paris,
France
Chiung-Nien Chen
Departments of Oncology and Internal
Medicine, National Taiwan University
Hospital, 7, Chung-Shan South Road,
Taipei, Taiwan, 10002
vii
viii
Ann-Lii Cheng
Departments of Oncology and Internal
Medicine, National Taiwan University
Hospital, 7, Chung-Shan South Road,
Taipei, Taiwan, 10002
Byung Ihn Choi
Department of Radiology, Seoul National
University College of Medicine 28,
Yongon dong Chongno-gu, Seoul,
110-744, Korea
Eva Compérat
Service d’Anatomie Pathologie,
GH La Pitié-Salpetrière, Boulevard
de l’hôpital, 75013 Paris,
France
N. Davies
Royal Free Hospital, Liver
Transplantation and Hepatobiliary
Medicine, Pond Street,
London NW3 2QG, UK
Timm Denecke
Campus Virchow-Klinikum,
Charite-Universitatmedizin Berlin,
Klinik fur Strahlenheilkonde und PET
Zentrum Berlin, Augustenburger Platz 1,
Berlin 13353,
Germany
Yuman Fong
Department of Surgery, Memorial
Sloan-Kettering Cancer Center, Weill
Cornell Medical Center 1275 York Avenue,
New York, NY 10021, USA
Masamichi Goto
Department of Human Pathology,
Field of Oncology, Kagoshima University
Graduate School of Medical and
Dental Sciences, 8-35-1, Sakuragaoka,
Kagoshima, 890-8544,
Japan
Authors and Co-Authors of Volume 5
Kevin C. Graham
Department of Medical Biophysics,
London Regional Cancer Program, 790
Commissioners Road East, London,
Ontario, Canada, N6A 4L6
Kenichi Harada
Department of Human Pathology,
Kanazawa University Graduate School
of Medicine, Takaramachi 13-1,
Kanazawa 920-8640,
Japan
Michiyo Higashi
Department of Human Pathology,
Field of Oncology, Kagoshima University
Graduate School of Medical and
Dental Sciences, 8-35-1, Sakuragaoka,
Kagoshima, 890-8544,
Japan
Bert Hildebrandt
Campus Virchow-Klinikum,
Charite-Universitatmedizin Berlin,
Klinik fur Strahlenheilkonde und PET
Zentrum Berlin, Augustenburger Platz 1,
Berlin 13353, Germany
Seung-Mo Hong
Department of Surgery and Cancer
Biology, Vanderbilt University Medical
Center, 1255 Light Hall, 2215 Garland
Avenue, Nashville, TN 37232,
USA
Chiun Hsu
Departments of Oncology and Internal
Medicine, National Taiwan University
Hospital, 7, Chung-Shan South Road,
Taipei, Taiwan, 10002
Damien Huglo
Service de Chirurgie Digestive
et Transplantation, Hôpital Huriez,
59037 Lille Cedex, France
Authors and Co-Authors of Volume 5
Tomoaki Ichikawa
Department of Radiology, University
of Yamanashi, 1110 Simokato, Chuo,
Yamanhashi, 409-3898, Japan
Hiroko Ikeda
Department of Human Pathology,
Kanazawa University Graduate School
of Medicine, Takaramachi 13-1,
Kanazawa 920-8640, Japan
Kazuo Inui
Department of Internal Medicine,
Second University Hospital, Fujita Health
University School of Medicine, 3-6-10,
Otobashi, Nakagawa-ku,
Nagoya 454-8509, Japan
Masayuki Kanematsu
Department of Radiology Services,
Gifu University Hospital, 1-1 Yanagido,
Gifu 501-1194, Japan
Koichi Kawanaka
Kumamoto University, School
of Medicine, Department of Radiology,
1-1-1 Honjo, Kumamoto 860, Japan
Kwong-Ming Kee
Division of Hepatogastroenterology,
Department of Internal Medicine,
Kaohsiung Chang Gung Memorial
Hospital, 123 Taipei Road,
Niaosung 833, Kaohsiung, Taiwan
Kengo Yoshimitsu
Department of Clinical Radiology,
Graduate School of Medical Sciences,
Kyushu University, 3-1-1 Maidashi,
Higashi-ku, Fukuoka 812-8582, Japan
ix
Amir H. Khandani
Director, Nuclear Medicine Residency
and Fellowship Director, PET Services
Department of Radiology UNC
School of Medicine, B113 POB,
CB7510 Chapel Hill,
NC 27599-7510
Se Hyung Kim
Department of Radiology, Seoul National
University College of Medicine 28,
Yongon dong Chongno-gu, Seoul,
110-744, Korea
Shigeru Kiryu
Department of Radiology,
Institute of Medical Science,
University of Tokyo, 4-6-1
Shirokanedai, Minato-ku, Tokyo,
108-8639, Japan
Yasuhiro Kuramitsu
Yamaguchi University Graduate
School of Medicine, Department
of Biochemistry and Functional
Proteomics, 1-1-1 Minami-Kogushi,
Ube, Yamaguchi 755-8505,
Japan
James C. Lacefield
Department of Medical Biophysics,
London Regional Cancer Program,
790 Commissioners Road East, London,
Ontario, Canada, N6A 4L6
Francesca Lodato
Royal Free Hospital, Liver
Transplantation and Hepatobiliary
Medicine, Pond Street, London NW3
2QG, UK
x Authors and Co-Authors of Volume 5
Enrique Lopez-Hänninen
Campus Virchow-Klinikum,
Charite-Universitatmedizin Berlin, Klinik
fur Strahlenheilkonde und PET Zentrum
Berlin, Augustenburger Platz 1,
Berlin 13353, Germany
Sheng-Nan Lu
Division of Hepatogastroenterology,
Department of Internal Medicine,
Kaohsiung Chang Gung Memorial
Hospital, 123 Taipei Road,
Niaosung 833, Kaohsiung,
Taiwan
Junji Machi
405 N. Kuakini Street, Suite 601,
Honolulu, HI 96817, USA
Houman Mahallati
Department of Surgery and Oncology,
University of Calgary, Tom Baker Cancer
Center, 1331 29th Street, N.W.,
Calgary, AB T2N 4N2, Canada
Kwan Man
Center for the Study of Liver Disease and
Department of Surgery, The University
of Hong Kong, Queen Mary Hospital,
L9-55, Building 21, Sassoon Road,
Pokfulam, Hong Kong, Republic of China
Vincenzo Migaleddu
Sardinian Mediterranean Imaging Group,
via Gorizia no. 11, Sassari 07100,
Sardinia, Italy
Hironao Miyoshi
Department of Internal Medicine, Second
University Hospital, Fujita Health
University School of Medicine, 3-6-10,
Otobashi, Nakagawa-ku,
Nagoya 454-8509, Japan
Isabella Mosberger
Department of Clinical Pathology,
Medical University of Vienna,
Waehringer Guertel 18-20,
1090 Vienna, Austria
Thomas Moser
Service de Radiologie, Pavillon Clovis
Vincent, Hôpital Civil, 1 Place de
l’Hôpital, 6700 Strasbourg, France
Joachim Mössner
Department of Internal Medicine II,
University of Leipzig, Philipp-Rosenthal-
Str. 27, 04103, Leipzig,
Germany
Takamichi Murakami
Department of Radiology, Osaka
University Graduate School of Medicine
D1, 2-2, Yamadaoka, Suita-city, Osaka,
565 0871, Japan
Kohji Nagata
Department of Human Pathology,
Field of Oncology, Kagoshima
University Graduate School of
Medical and Dental Sciences,
8-35-1, Sakuragaoka, Kagoshima,
890-8544, Japan
Hideji Nakamura
Division of Heptobiliary and Pancreatic
Medicine, Department of Internal
Medicine, Hyogo College of Medicine,
Mukogawa-cho, 1-Ishinomiya, Hyogo
663-8501, Japan
Hironobu Nakamura
Department of Radiology, Osaka
University Graduate School of Medicine
D1, 2-2, Yamadaoka, Suita-city, Osaka,
565 0871, Japan
Authors and Co-Authors of Volume 5
Kazuyuki Nakamura
Yamaguchi University Graduate School
of Medicine, Department of Biochemistry
and Functional Proteomics, 1-1-1
Minami-Kogushi, Ube, Yamaguchi
755-8505 Japan
Koichi Nakamura
Department of Radiology,
Kanazawa University Graduate School of
Medicine, Takaramachi 13-1, Kanazawa
920-8640, Japan
Yasuni Nakanuma
Department of Human Pathology,
Kanazawa University Graduate School
of Medicine, Takaramachi 13-1,
Kanazawa 920-8640,
Japan
Kevin Tak-Pan Ng
Center for the Study of Liver Disease
and Department of Surgery,
The University of Hong Kong, Queen
Mary Hospital, L9-55, Building 21,
Sassoon Road, Pokfulam, Hong Kong,
Republic of China
Akihiro Nishie
Department of Clinical Radiology,
Graduate School of Medical Sciences,
Kyushu University, 3-1-1 Maidashi,
Higashi-ku, Fukuoka 812-8582,
Japan
Tereza S. Nogueira
Department of Radiology, Hôpital de
Hautepierre, CHU Strasbourg, Avenue
Molière, Strasbourg 67000, France
Kuni Ohtomo
Department of Radiology, Institute of
Medical Science, University of Tokyo,
4-6-1 Shirokanedai, Minato-ku,
Tokyo, 108-8639, Japan
xi
Hiromitsu Onishi
Department of Radiology, Osaka
University Graduate School of Medicine
D1, 2-2, Yamadaoka, Suita-city,
Osaka, 565 0871, Japan
Shinji Osada
Department of Radiology Services,
Gifu University Hospital, 1-1 Yanagido,
Gifu 501-1194, Japan
François-René Pruvot
Service de Chirurgie Digestive et
Transplantation, Hôpital Huriez,
59037 Lille Cedex, France
Sóren R. Rafaelsen
Department of Radiology, Vejle Hospital,
Kabbeltoft 25, DK-7100 Vejle, Denmark
Yasunori Sato
Department of Human Pathology,
Kanazawa University Graduate School
of Medicine, Takaramachi 13-1,
Kanazawa 920-8640, Japan
Katharina Schmid
Department of Clinical Pathology,
Medical University of Vienna, Waehringer
Guertel 18-20, 1090 Vienna, Austria
Richard C. Semelka
Department of Radiology Services,
Gifu University Hospital, 1-1 Yanagido,
Gifu 501-1194, Japan
Hui-Chuan Sun
Liver Cancer Institute and Zhongshan
Hospital, Fudan University,
Shanghai 200032, China
Kenichi Takayasu
National Cancer Center Hospital,
Department of Diagnostic Radiology,
Chuo ku, 5-1-1 Tsukiji, Tokyo 1040045,
Japan
xii
Shugo Tamada
Department of Human Pathology, Field
of Oncology, Kagoshima University
Graduate School of Medical and
Dental Sciences, 8-35-1, Sakuragaoka,
Kagoshima, 890-8544, Japan
Bachir Taouli
NYU Medical Center, Department of
Radiology, MRI 530 First Avenue,
New York, NY, 10016, USA
Yasuhiko Tomita
Division of Heptobiliary and Pancreatic
Medicine, Department of Internal
Medicine, Hyogo College of Medicine,
Mukogawa-cho, 1-Ishinomiya, Hyogo
663-8501, Japan
Guido Torzilli
Chief Liver Surgery Unit, Department
of General Surgery 3, Instituto Clinico
Humanitas IRCCS, Faculty of Medicine,
University of Milan, Via A. Manzoni, 56,
20089 Rozzano-Milano, Italy
Stéphanie Truant
Service de Chirurgie Digestive et
Transplantation, Hôpital Huriez, 59037
Lille Cedex, France
Po-Lin Tseng
Division of Hepatogastroenterology,
Department of Internal Medicine,
Kaohsiung Chang Gung Memorial
Hospital, 123 Taipei Road, Niaosung 833,
Kaohsiung, Taiwan
Giuseppe Virgilio
Sardinian Mediterranean Imaging Group,
via Gorizia no. 11, Sassari 07100,
Sardinia, Italy
Authors and Co-Authors of Volume 5
Jing-Houng Wang
Division of Hepatogastroenterology,
Department of Internal Medicine,
Kaohsiung Chang Gung Memorial
Hospital, 123 Taipei Road,
Niaosung 833, Kaohsiung, Taiwan
Marcus Wiedmann
Department of Internal Medicine II,
University of Leipzig, Philipp-Rosenthal-
Str. 27, 04103, Leipzig, Germany
Lauren A. Wirtzfeld
Department of Medical Biophysics,
London Regional Cancer Program, 790
Commissioners Road East, London,
Ontario, Canada, N6A 4L6
Helmut Witzigmann
Department of Surgery II, University
of Leipzig, Liebigstrasse 20a, 04103
Leipzig, Germany
Ching-Yee Oliver Wong
Positron Diagnostic Center and Medical
Cyclotron, Department of Nuclear
Medicine, Michigan State University,
William Beaumont Hospital, 3601 West
Thirteen Mile Road, Royal Oak, MI
58073, USA
Fritz Wrba
Department of Clinical Pathology,
Medical University of Vienna,
Waehringer Guertel 18-20, 1090 Vienna,
Austria
Yasuyuki Yamashita
Kumamoto University, School of
Medicine, Department of Radiology,
1-1-1 Honjo, Kumamoto 860,
Japan
Authors and Co-Authors of Volume 5 xiii

Suguru Yonezawa
Department of Human Pathology, Field
of Oncology, Kagoshima University
Graduate School of Medical and
Dental Sciences, 8-35-1, Sakuragaoka,
Kagoshima, 890-8544, Japan
Kenya Yoshida
Division of Heptobiliary and Pancreatic
Medicine, Department of Internal
Medicine, Hyogo College of Medicine,
Mukogawa-cho, 1-Ishinomiya, Hyogo
663-8501, Japan
Junji Yoshino
Department of Internal Medicine, Second
University Hospital, Fujita Health
University School of Medicine, 3-6-10,
Otobashi, Nakagawa-ku, Nagoya
454-8509, Japan
D. Yu
Royal Free Hospital, Liver
Transplantation and Hepatobiliary
Medicine, Pond Street,
London NW3 2QG, UK
Alex I. Zaika
Department of Surgery and Cancer
Biology, Vanderbilt University Medical
Center, 1255 Light Hall, 2215 Garland
Avenue, Nashville,
TN 37232, USA
Yoh Zen
Division of Pathology,
Kanazawa University Hospital,
Kanazawa, Japan
Preface
This volume presents a detailed survey of
various methodologies related to diagno-
sis, therapy, and prognosis of liver cancer,
including colorectal liver metastases and
biliary tract carcinomas, while the already
published Volumes 1, 2, 3, and 4 detail
similar aspects of breast, lung, prostate,
gastrointestinal, and colorectal cancer,
respectively.
Surgical resection is the standard therapy
for resectable liver disease, resulting in
5-year overall survival rates of 20–40%.
One the other hand, the median overall
survival of patients with unresectable liver
metastases does not exceed 18–20 months,
with a 5-year survival rate approaching
zero. In other words, there is virtually no
long-term survival. Both resectable and
unresectable liver cancers are discussed
in this volume. The method of selecting
patients for resection of hepatic colorectal
metastases is explained. The type of chem-
otherapy used for surgically unresectable
liver cancer is also presented.
In the current volume, attention is also
given to the role of molecular genetics and
biomarkers in the initiation and progression
of liver cancer. The application of a vari-
ety of imaging modalities, including PET,
SPECT, MRI, radial magnetic resonance
imaging, multidetector-row computed tom-
ography, helical computed tomography,
ultrasonography, and power Doppler ultra-
sound, for the prognosis and assessment of
liver cancer treatment (including HCC) and
liver metastases from colorectal cancer are
discussed in detail, as is the use of radiofre-
quency ablation in hepatic tumors.
Approximately 50% of colorectal cancer
(CRC) patients develop liver metastases
during the course of their disease, and
more than 50% of patients who die of CRC
have liver metastases at autopsy. Regional
lymph node (RLN) involvement in patients
with colorectal liver metastases is one
of the worst prognostic factors. Recent
studies indicate that for these patients,
combined liver resection and pedicular
lymphadenectomy can be recommended,
when RLN metastases respond to pre-
operative chemotherapy. In other words,
combined hepatectomy and lymphadenec-
tomy is justified for younger patients.
This volume includes explanations of liver
metastases from colorectal cancer using
ultrasound imaging, the methodology of
radiofrequency ablation for unresectable
liver metastases from colorectal cancer,
and preoperative staging of patients with
colorectal liver metastases.
xv
xvi
The hepatitis B and C viruses constitute
a major global health problem. It is esti-
mated that 180 million people, ∼3% of
the world’s population are infected with
these viruses (WHO). Most cases result in
chronic hepatitis, causing liver cirrhosis and
HCC. Spontaneous eradication of the virus,
however, does occur in ∼50% of acute
infections, which is associated with specific
immune response to the virus. Because
HCC is the most frequent form of liver can-
cer, it is discussed extensively here.
Biliary tract carcinomas are not very
common tumors, but middle-aged or eld-
erly people have poor prognosis. Although
palliative chemotherapy is commonly used
for these patients, there is no standard
nor very effective therapy for advanced
disease. Commonly used therapy includes
gemcitabine, mitomycin, fluoropyrimi-
dines, and platinum compounds, used alone
or in combination. Response rates with
these treatments range from 10–35%, and
median overall survival times vary between
5 and 12 months. Clinicopathological fea-
tures of biliary cystic tumors are discussed
in this volume. The role of Mucin 4 as a
prognostic factor for extrahepatic bile duct
carcinoma is also presented.
As in the previous volumes of this series,
each chapter is written by distinguished,
practicing clinicians/surgeons/pathologists
who provide detailed methodologies for
diagnosis and treatment of various forms
of liver cancer. This volume was written
by 78 oncologists representing 16 coun-
tries. Their practical experience highlights
Preface
their writings, which should build and fur-
ther the endeavors of the readers. The text
is divided basically into seven subheadings
for the convenience of the readers. It is my
hope that the current volume will join the
preceding volumes of this series for assist-
ing in the more complete understanding of
globally relevant cancer syndromes. There
exists a tremendous public demand on
the scientific community to address cancer
prevention, diagnosis, treatment and, hope-
fully, cures.
I am grateful to the contributors for their
promptness in accepting my suggestions,
and I appreciate their dedication and dil-
igent work in sharing their invaluable
knowledge with the public through this
series. Each chapter provides unique indi-
vidual, practical knowledge based on the
expertise and experience of the authors.
The chapters contain the most up-to-date
practical and theoretical information. It is
my hope that these handbooks will assist
in the more complete understanding of
at least come of the globally-encountered
cancer problems.
I am thankful to Dr. Dawood Farahi
and Dr. Kristie Reilly for recognizing
the importance of scholarship in an
institution of higher education and for
providing the resources for completing
this project. I appreciate receiving expert
help from Myrna Ortiz, Erin McNally
and Lina Builes in preparing this volume.
M.A. Hayat
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

2. Breast Cancer: Computer-Aided Detection

3. Sebaceous Carcinoma of the Breast: Clinicopathologic Features

4. Breast Cancer: Detection by In-Vivo Imaging of Angiogenesis

5. Breast and Prostate Biopsies: Use of Optimized High-Throughput

MicroRNA Expression for Diagnosis (Methodology)

6. Familial Breast Cancer: Detection of Prevalent High-Risk

Epithelial Lesions

7. Differentiation Between Benign and Malignant Papillary

Lesions of Breast: Excisional Biopsy or Stereotactic
Vacuum-Assisted Biopsy (Methodology)

8. Multicentric Breast Cancer: Sentinel Node Biopsy

as a Diagnostic Tool

9. Breast Cancer Recurrence: Role of Serum Tumor

Markers CEA and CA 15-3

10. Breast Cancer Patients Before, During or After Treatment:

Circulating Tumor Cells in Peripheral Blood Detected
by Multigene Real-Time Reverse Transcriptase-Polymerase
Chain Reaction

11. Breast Cancer Patients: Diagnostic Epigenetic

Markers in Blood

xix
xx Contents of Volumes 1, 2, 3, and 4

12. Breast Cancer Patients: Detection of Circulating Cancer

Cell-Related mRNA Markers with Membrane Array Method

13. Prediction of Metastasis and Recurrence of Breast Carcinoma:

Detection of Survivin-Expressing Circulating Cancer Cells

14. Node-Negative Breast Cancer: Predictive and Prognostic

Value of Peripheral Blood Cytokeratin-19 mRNA-Positive Cells

15. Breast and Colon Carcinomas: Detection with Plasma

CRIPTO-1

16. Breast Cancer Risk in Women with Abnormal Cytology

in Nipple Aspirate Fluid

17. Tissue Microarrays: Construction and Utilization

for Biomarker Studies

18. Systematic Validation of Breast Cancer Biomarkers

Using Tissue Microarrays: From Construction
to Image Analysis

19. Phyllodes Tumors of the Breast: The Role

of Immunohistochemistry in Diagnosis

20. Phyllodes Tumor of the Breast: Prognostic Assessment

Using Immunohistochemistry

21. Metaplastic Breast Carcinoma: Detection Using

Histology and Immunohistochemistry

22. Invasive Breast Cancer: Overexpression of HER-2 Determined

by Immunohistochemistry and Multiplex Ligation-Dependent
Probe Amplification

23. Operable Breast Cancer: Neoadjuvant Treatment

(Methodology)

24. Chemotherapy for Breast Cancer

25. Locally Advanced Breast Cancer: Role of Chemotherapy

in Improving Prognosis

26. Relevance of Dose-Intensity for Adjuvant Treatment

of Breast Cancer
Contents of Volumes 1, 2, 3, and 4 xxi

27. Advanced Breast Cancer: Treatment with Docetaxel/Epirubicin

28. Systemic Therapy for Breast Cancer: Using Toxicity

Data to Inform Decisions

29. Chemotherapy for Metastatic Breast Cancer Patients

Who Received Adjuvant Anthracyclines (An Overview)

30. Estrogen Receptor-Negative and HER-2/neu-Positive

Locally Advanced Breast Carcinoma: Therapy with
Paclitaxel and Granulocyte-Colony Stimulating Factor

31. Breast Cancer: Side Effects of Tamoxifen and Anastrozole

32. Breast Cancer: Expression of HER-2 and Epidermal

Growth Factor Receptor as Clinical Markers for Response
to Targeted Therapy

33. Young Breast Cancer Patients Undergoing

Breast-Conserving Therapy: Role of BRCA1 and BRCA2

34. Radiation Therapy for Older Women with Early Breast Cancer

35. Acute Side Effects of Radiotherapy in Breast Cancer

Patients: Role of DNA-Repair and Cell Cycle Control Genes
18
36. F-Fluorodeoxyglucose/Positron Emission Tomography
in Primary Breast Cancer: Factors Responsible for
False-Negative Results

37. Sentinel Lymph Node Surgery During Prophylactic

Mastectomy (Methodology)

38. Breast Conservation Surgery: Methods

39. Lymph Node-Negative Breast Carcinoma:

Assessment of HER-2/neu Gene Status as Prognostic Value

40. Multifocal or Multicentric Breast Cancer:

Understanding Its Impact on Management
and Treatment Outcomes

41. Are Breast Cancer Survivors at Risk for Developing

Other Cancers?
xxii Contents of Volumes 1, 2, 3, and 4

42. Distant Metastasis in Elderly Patients with

Breast Cancer: Prognosis with Nodal Status

43. Concomitant Use of Tamoxifen with Radiotherapy

Enhances Subcutaneous Breast Fibrosis in
Hypersensitive Patients

44. Malignant Phyllodes Tumor of the Breast:

Is Adjuvant Radiotherapy Necessary?

45. Locally Advanced Breast Cancer: Multidrug Resistance

46. Breast Cancer: Diagnosis of Recurrence

Using 18 F-Fluorodeoxyglucose-Positron Emission
Tomography/Computed Tomography

47. Role of Sentinel Lymph Node Biopsy in Ductal

Carcinoma In Situ: Diagnosis and Methodology

48. Breast Conservation Treatment of Early Stage Breast

Carcinoma: Risk of Cardiac Mortality

Volume 2
Part I General Methods and Overviews

1. Metabolic Transformations of Malignant Cells: An Overview

2. Detection of Recurrent Cancer by Radiological Imaging

3. Tumor Gene Therapy: Magnetic Resonance Imaging

and Magnetic Resonance Spectroscopy

4. Assessment of Gene Transfer: Magnetic Resonance Imaging

and Nuclear Medicine Techniques

5. Role of Mutations in TP53 in Cancer (An Overview)

6. Personalized Medicine for Cancer

7. Radiation Doses to Patients Using Computed Radiography,

Direct Digital Radiography and Screen-Film Radiography
Contents of Volumes 1, 2, 3, and 4 xxiii

8. Cancer Vaccines and Immune Monitoring (An Overview)

9. New Insights into the Role of Infection, Immunity, and Apoptosis

in the Genesis of the Cancer Stem Cell

10. Successful Cancer Treatment: Eradication of Cancer

Stem Cells

11. Overexposure of Patients to Ionizing Radiation: An Overview

Part II Lung Cancer

12. Lung Carcinoma

13. Extra-Pulmonary Small Cell Cancer: Diagnosis, Treatment,

and Prognosis

14. Magnetic Resonance Imaging of the Lung: Automated

Segmentation Methods

15. Peripheral Lung Lesions: Diagnosis Using Transcutaneous

Contrast-Enhanced Sonography

16. Small Pulmonary Nodules: Detection Using Multidetector-Row

Computed Tomography

17. Secondary Primary Cancer Following Chemoradiation

for Non-Small-Cell Lung Cancer

18. Advanced Non-Small Cell Lung Cancer: Second-Line

Treatment with Docetaxel

19. Non-Small Cell Lung Cancer with Brain Metastases:

Platinum-Based Chemotherapy

20. Non-Small Cell Lung Carcinoma: EGFR Gene Mutations

and Response to Gefitinib

21. Advanced Non-Small Cell Lung Carcinoma: Acquired

Resistance to Gefitinib

22. Prognostic Significance of [18F]-Fluorodeoxyglucose Uptake

on Positron Emission Tomography in Patients with Pathological
Stage I Lung Adenocarcinoma
xxiv Contents of Volumes 1, 2, 3, and 4

23. Non-Small Cell Lung Cancer: Prognosis Using

the TNM Staging System

24. Differentiation Between Malignant and Benign Pleural Effusions:

Methylation Specific Polymerase Chain Reaction Analysis

25. Pathological Distinction of Pulmonary Large Cell Neuroendocrine

Carcinoma from Small-Cell Lung Carcinoma Using
Immunohistochemistry

26. Differentiating Between Pleuropulmonary Desmoid Tumors

and Solitary Fibrous Tumors: Role of Histology and
Immunohistochemistry

27. Non-Small Cell Lung Cancer with Brain Metastasis:

Role of Epidermal Growth Factor Receptor Gene Mutation

Part III Prostate Cancer

28. Prostate Carcinoma

29. The Role of Intermediary Metabolism and Molecular Genetics

in Prostate Cancer

30. Array-Based Comparative Genomic Hybridization in

Prostate Cancer: Research and Clinical Applications

31. Prostate Cancer: Role of Vav3 Overexpression in

Development and Progression

32. Prostate Cancer: Detection and Monitoring Using Mitochondrial

Mutations as a Biomarker

33. Prognostic Markers in Prostate Carcinoma

34. Prostate Cancer: Detection of Free Tumor-Specific DNA

in Blood and Bone Marrow

35. Prostate Carcinoma: Evaluation Using Transrectal Sonography

36. Prostate Cancer: 16β-[18F]Fluoro-5α-Dihydrotesterone(FDHT)

Whole-Body Positron Emission Tomography

37. Effects of Standard Treatments on the Immune

Response to Prostate Cancer
Contents of Volumes 1, 2, 3, and 4 xxv

38. Vinorelbine, Doxorubicin, and Prednisone

in Hormone Refractory Prostate Cancer

39. Locally Advanced Prostate Cancer Biochemical Recurrence

After Radiotherapy: Use of Cyclic Androgen Withdrawal
Therapy

Volume 3

Part I Gastrointestinal Cancers

1. Introduction: Gastrointestinal Cancer

2. Metastatic Gastrointestinal Cancer: Safety of Cisplatin

Combined with Continuous 5-FU Versus Bolus 5-FU
and Leucovorin (Methodology)

3. Gastrointestinal Cancer: Endoscopic Submucosal

Dissection (Methodology)

4. Gastrointestinal Epithelial Neoplasms: Endoscopic

Submucosal Dissection (Methodology)

5. Inoperable Abdomino-Pelvic Tumors: Treatment

with Radio-Frequency Ablation and Surgical Debulking

6. Gastrointestinal Neuroendocrine Tumors:

Diagnosis Using Gastrin Receptor Scintigraphy

Part II Esophageal Cancer

7. Distal Esophagus: Evaluation with 18F-FDG PET/CT

Fusion Imaging

8. Endoscopic Ultrasound and Staging of Esophageal Cancer

9. Esophageal Cancer: Role of RNASEN Protein

and microRNA in Prognosis

10. Esophageal Cancer: Initial Staging

xxvi Contents of Volumes 1, 2, 3, and 4

Part III Gastric Cancer

11. Automated Disease Classification of Colon and Gastric

Histological Samples Based on Digital Microscopy
and Advanced Image Analysis

12. Early Gastric Cancer: Prediction of Metachronous

Recurrence Using Endoscopic Submucosal Dissection
(Methodology)

13. Helicobacter pylori-Infected Neoplastic Gastric Epithelium:

Expression of MUC2 as a Biomarker

14. Gastric Cancer: Role of Intestinal Metaplasia

by Histochemical Detection Using Biopsy Specimens

15. Gastric Cancer: Antitumor Activity of RUNX3

16. Early Gastric Cancer: Laparoscopic Gastrectomy (Methodology)

17. Gastric Cancer: Overexpression of Hypoxia-Inducible

Factor 1 as a Prognostic Factor

Part IV Pancreatic Cancer

18. Pancreatic Cancer: Hepatoma-Derived Growth Factor

as a Prognostic Factor

19. Pancreatic Cancer: 18F-Fluorodeoxyglucose Positron

Emission Tomography as a Prognostic Parameter

20. Imaging and Pathologic Findings of Peculiar Histologic

Variants of Pancreatic Endocrine Tumors

21. Periampullary Adenocarcinoma: Diagnosis and Survival

After Pancreaticoduodenectomy

22. Unresectable Locally Advanced Pancreatic Cancer:

Concurrent Chemotherapy

Index
Contents of Volumes 1, 2, 3, and 4 xxvii

Volume 4
Part I Colorectal Cancer

1. Introduction: Colorectal Cancer

2. Poorly Differentiated Colorectal Adenocarcinoma: (Methodology)

3. Colorectal Cancer: Immunohistochemical Diagnosis with Heterogenous

Nuclear Ribonucleoprotein K

4. Metastases and Recurrence of Colorectal Cancer: Diagnostic Role

of Immunoscintigraphy

5. Colorectal Cancer Diagnosis Using DNA Levels in Blood and Stool

6. Colorectal Carcinoma: Identification of MicroRNAs Using Real-Time

Polymerase Chain Reaction

7. Colorectal Cancer: Optimization of the Combination of 5-Flouroracil

and Irinotecan

8. Detection of Abdominal Abscesses After Colorectal Surgery:

Ultrasonography, Computed Tomography, and Gallium Scan

9. Antimetastatic Therapy in Colorectal Cancer: Role of Tumor Cell

Matrix Metalloproteinase 9 (Methodology)

10. Endoscopic Resection of Early Colorectal Tumours: Novel Diagnostic

and Therapeutic Techniques

11. Role of Stromal Variables in Development and Progression of Colorectal

Cancer

12. Quantitative Assessment of Colorectal Cancer Perfusion: Perfusion

Computed Tomography and Dynamic Contrast-Enhanced Magnetic
Resonance Imaging

13. Colorectal Cancer: Positron Emission Tomography

14. Prognostic Significance of Protein Markers in Colorectal Cancer

Stratified by Mismatch Repair Status

15. Colorectal Cancer: Lactate Dehydrogenase (LDH) Activity

as a Prognostic Marker
xxviii Contents of Volumes 1, 2, 3, and 4

Part II Colon Cancer

16. Detection of Tumor Cells in Lymph Nodes of Colon Cancer Patients Using
Real-Time Quantitative Reverse Transcription-Polymerase Chain Reaction

17. Colon Cancer: Laparoscopic Surgery

18. Sentinal Node-Based Immunotherapy of Colon Cancer

Part III Rectal Cancer

19. Rectal Cancer: Preoperative Staging Using Endorectal Ultrasonography

(Methodology)

20. Rectal Cancer: Spectral Imaging and Immunohistochemistry

of Thymidylate Synthase

21. Cancer of the Rectum: Abdominoperineal and Sphincter-Saving Resections

22. Chemoradiation for Rectal Cancer

23. Resectable Rectal Cancer: Preoperative Short-Course Radiation

24. Preoperative Chemoradiotherapy Allows for Local Control in

Rectal Cancer, but Distant Metastases Remain an Unsolved Problem

25. Locally Advanced Rectal Cancer: Combined Chemotherapy During

Preoperative Radiation Therapy

Part IV Colorectal Liver Metastases

26. Colorectal Cancer Liver Metastases: Neoadjuvant Therapy

with Bevacizumab

27. Colorectal Liver Metastases: Radiofrequency Ablation

Part V Anal Cancer

28. Anal Squamous Cell Carcinomas: Diagnosis Using p63

Immunohistochemistry

29. Anorectal Melanoma: Prediction of Outcome Based

on Molecular and Clinicopathologic Features
Contents

Authors and Co-Authors of Volume 5 .................................................................. vii

Preface ..................................................................................................................... xv

Introduction ............................................................................................................ xvii

Contents of Volumes 1, 2, 3, and 4 ........................................................................ xix

Part I Liver Cancer

A. Diagnosis
1. Applications of Positron Emission Tomography
in Liver Imaging: An Overview ..................................................................... 5
Amir H. Khandani
Introduction ................................................................................................... 5
Metastatic Liver Disease ............................................................................... 7
Cholangiocarcinoma ..................................................................................... 10
Gall Bladder Carcinoma ............................................................................... 11
Hepatocellular Carcinoma ............................................................................ 12
Therapy Monitoring ...................................................................................... 14
References ..................................................................................................... 15

2. Localized Fibrous Tumor of the Liver: Imaging Features .......................... 17

Thomas Moser and Tereza S. Nogueira
References ..................................................................................................... 19

xxix
xxx Contents

3. A Radial Magnetic Resonance Imaging Method

for Imaging Abdominal Neoplasms ............................................................... 21
Maria I. Altbach
Introduction ................................................................................................... 21
Techniques for T2-Weighted Imaging .......................................................... 22
Lesion Characterization with T2-Weighted Imaging .................................... 22
Radial Magnetic Resonance Imaging Methods:
An Alternative for Reducing Motion Artifacts
and Improving Image Quality ................................................................... 24
Radial Fast Spin-Echo, A New Alternative for T2
Mapping of the Liver ................................................................................ 27
Conclusions ................................................................................................... 30
References ..................................................................................................... 30

4. Liver: Helical Computed Tomography

and Magnetic Resonance Imaging................................................................. 33
Yuji Baba, Yasuyuki Yamashita, Kazuo Awai,
and Koichi Kawanaka
Introduction ................................................................................................... 33
Dynamic Computed Tomography Using Multi-Detector
Computed Tomography............................................................................. 33
Tumors Arising in Liver Cirrhosis ................................................................ 35
Tumor Arising in Noncirrhotic Liver or in Oncologic Patients .................... 36
Computed Tomography Angiography .......................................................... 36
Angiography-Assisted Computed Tomography........................................ 36
Computed Tomography During Arterial Portography .................................. 37
Computed Tomography During Hepatic Angiography ................................. 39
Magnetic Resonance Imaging ....................................................................... 39
Dynamic Magnetic Resonance Imaging ....................................................... 39
Superparamagnetic Iron Oxide ..................................................................... 40
References ..................................................................................................... 42

Part II Resectable Liver Cancer

A. Diagnosis
5. Selection of Patients for Resection of Hepatic Colorectal Metastases:
18F-Fluorodeoxyglucose/Positron Emission Tomography .......................... 49
Rebecca Auer and Yuman Fong
Introduction ................................................................................................... 49
Positron Emission Tomography Scanning as a Staging
Modality to Complement Conventional Imaging ..................................... 50
Positron Emission Tomography Scan for the Detection
of Extrahepatic Disease............................................................................. 50
Contents xxxi

Positron Emission Tomography Scan

for the Detection of Intrahepatic Disease.................................................. 51
Value of Positron Emission Tomography Correlated
to Prognostic Clinical Risk Score (CRS) .................................................. 52
Positive Impact of Positron Emission Tomography
on the Management of Patients with Liver Metastases ............................. 53
Negative Impact of Positron Emission Tomography
on the Management of Patients with Liver Metastases ............................. 54
Outcome of Patients Selected for Hepatic Resection Following
[18F]Fluorodeoxyglucose-Positron Emission Tomography
(FDG-PET) ............................................................................................... 54
Standard Uptake Value and Predicting Prognosis
or Response to Therapy ............................................................................ 55
Positron Emission Tomography Scanning
for Surveillance and Follow-Up ................................................................ 57
The Role of Positron Emission Tomography
in Evaluating an Elevated Carcinoembryonic Antigen (CEA) ................. 57
The Role of Positron Emission Tomography
in Follow-Up Post Hepatic Resection ....................................................... 58
Algorithm to Include Positron Emission Tomography
Scanning in Work-Up of Hepatic Colorectal Metastases.......................... 58
Conclusion .................................................................................................... 59
References ..................................................................................................... 59

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

Part III Unresectable Liver Cancer

A. Treatment
7. Intraoperative Magnetic Resonance Imaging
for Radiofrequency Ablation of Hepatic Tumors......................................... 81
Oliver F. Bathe and Houman Mahallati
Introduction ................................................................................................... 81
Radiofrequency Ablation as a Treatment for Liver Tumors ......................... 81
xxxii Contents

Principles of Radiofrequency Ablation ..................................................... 81

Indications ................................................................................................. 83
Outcomes Associated with Radiofrequency Ablation .............................. 84
Alternative Ablative Techniques ............................................................... 85
Problems Associated with Monitoring the Ablative Endpoint ..................... 86
Impedence and Temperature ..................................................................... 86
Ultrasound ................................................................................................. 86
Defining the Extent of Thermal Injury by Magnetic
Resonance Imaging ................................................................................... 86
Magnetic Resonance Imaging Characteristics of Thermal Injury ............ 86
Enhancing the Definition of Extent of Thermal Injury ............................. 88
Potential Utility of Functional Magnetic Resonance Imaging.................. 89
Intraopertive Magnetic Resonance Imaging as an Adjunct
to Radiofrequency Ablation ...................................................................... 90
Rationale ................................................................................................... 90
Technical Developments ........................................................................... 90
Potential Indications.................................................................................. 92
Conduct of Intraoperative Magnetic Resonance Imaging
for Radiofrequency Ablation ................................................................ 92
Future Developments ................................................................................ 96
References ..................................................................................................... 98

8. Surgically Unresectable and Chemotherapy-Refractory

Metastatic Liver Carcinoma: Treatment with Yttrium-90
Microsphere Followed by Assessment with Positron
Emission Tomography .................................................................................... 103
Ching-Yee Oliver Wong
Introduction ................................................................................................... 103
Assessment Using Positron Emission Tomography...................................... 104
Y-90 Microsphere Radioembolization .......................................................... 105
Clinical Results ............................................................................................. 109
Illustrations ................................................................................................... 111
References ..................................................................................................... 112

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

Unresectability of Tumors ........................................................................ 120

Preoperative and Intraoperative Evaluation .............................................. 121
Radiofrequency Ablation Methods ........................................................... 121
Postoperative Follow-Up........................................................................... 122
Statistical Analyses ................................................................................... 122
Short-Term Results ................................................................................... 123
Long-Term Results.................................................................................... 123
Current Role of Radiofrequency Ablation .................................................... 124
Future Perspective ......................................................................................... 128
References ..................................................................................................... 129

Part IV Hepatocellular Carcinoma

A. Diagnosis
10. Screening with Ultrasonography of Patients
at High-Risk for Hepatocellular Carcinoma:
Thrombocytopenia as a Valid Surrogate of Cirrhosis ............................... 137
Sheng-Nan Lu, Jing-Houng Wang, Kwong-Ming Kee,
and Po-Lin Tseng
Introduction ................................................................................................. 137
Epidemiology of Hepatocellular Carcinoma and Liver Cirrhosis .......... 137
Benefit of HCC Screening ...................................................................... 138
Surrogate Tests for Liver Cirrhosis and Fibrosis .................................... 139
Ultrasonographic HCC Screening on the Thrombocytopenic Adult ...... 141
References ................................................................................................... 142

11. Hepatocellular Carcinoma: Contrast-Enhanced Sonography .................. 145

Byung Ihn Choi and Se Hyung Kim
Introduction ................................................................................................. 145
Physics of Microbubbles ............................................................................. 146
Contrast-Enhanced Color Doppler Sonography ......................................... 146
Contrast-Enhanced Power Doppler Sonography ........................................ 147
Contrast-Enhanced Harmonic Power Doppler Sonography ....................... 147
Contrast-Enhanced Pulse-Inversion Harmonic Sonography ....................... 148
Contrast-Enhanced Coded Harmonic Sonography ..................................... 150
Contrast-Enhanced Agent Detection Imaging ............................................ 151
Low Mechanical Index Imaging ................................................................. 153
Contrast-Enhanced Sonography in Assessing
the Therapeutic Response of Hepatocellular Carcinomas ...................... 153
References ................................................................................................... 156
xxxiv Contents

12. Focal Liver Lesion: Nonlinear Contrast-Enhanced

Ultrasound Imaging ...................................................................................... 159
Vincenzo Migaleddu and Giuseppe Virgilio
Introduction ................................................................................................. 159
Harmonic Imaging: Basic Principles .......................................................... 160
Physical and Technological Background .................................................... 162
Nonlinear Contrast Enhancement of Focal Liver Lesions .......................... 163
(A) Characterization of Focal Liver Lesions .......................................... 163
(B) Detection of Focal Liver Lesions...................................................... 171
(C) Monitoring of Percutaneous Ablative Treatment ............................. 174
Conclusion .................................................................................................. 176
References ................................................................................................... 177

13. Hepatocellular Carcinoma: Magnetic Resonance Imaging ...................... 183

Bachir Taouli
Introduction ................................................................................................. 183
Magnetic Resonance Imaging Technique ................................................... 183
Magnetic Resonance Imaging Appearance
of Hepatocellular Carcinoma .................................................................. 184
Gadolinium Enhancement of Hepatocellular Carcinoma ........................... 186
Magnetic Resonance Imaging Tissue-Specific Contrast Agents ................ 187
Combined Use of Superparamagnetic
Iron Oxide and Gadolinium .................................................................... 187
Atypical Appearances of Hepatocellular Carcinoma .................................. 188
New Perspectives ........................................................................................ 189
References ................................................................................................... 190

14. Expression of Vascular Endothelial Growth Factor

in Hepatocellular Carcinoma: Correlation
with Radiologic Findings .............................................................................. 193
Masayuki Kanematsu, Richard C. Semelka, and Shinji Osada
Introduction ................................................................................................... 193
History of Radiologic Diagnosis of Hepatocellular Carcinoma ................... 194
Development of Hepatocellular Carcinoma, Its Vasculature,
and Vascular Endothelial Growth Factor Expression ............................... 196
Correlation of Vascular Endothelial Growth Factor Expression
and Radiologic Findings in Hepatocellular Carcinoma
in Prior Reports ......................................................................................... 199
Our Previous Research with Immunohistochemistry.................................... 200
Magnetic Resonance and Computed Tomographic
Data Correlated to Vascular Endothelial Growth
Factor Measured Using Western Blotting ................................................. 202
Summary ....................................................................................................... 209
References ..................................................................................................... 210
Contents xxxv

15. Detection of Small Hepatic Lesions:

Superparamagnetic Oxide-Enhanced
Diffusion-Weighted T2 FSE Imaging .......................................................... 213
Shigeru Kiryu and Kuni Ohtomo
Introduction ................................................................................................. 213
Superparamagnetic Iron Oxide-Enhanced
Magnetic Resonance Imaging ................................................................. 213
Diffusion-Weighted Imaging for Suppression
of Signals from Hepatic Vessels ............................................................. 214
Periodically Rotated Overlapping Parallel Lines
with Enhanced Reconstruction Technique .............................................. 215
SPIO-Enhanced DWI T2 FSE Imaging Using Propeller ............................ 216
References ................................................................................................... 218

16. Diagnosis of Hepatocellular Carcinoma:

Multidetector-Row Computed Tomography
and Magnetic Resonance Imaging............................................................... 221
Hiromitsu Onishi, Takamichi Murakami, and Hironobu Nakamura
Introduction ................................................................................................. 221
Multidetector-Row Helical Computed Tomography................................... 221
Multidetector-Row Helical Computed Tomography
Scanning Technique ................................................................................ 222
Contrast Administration for the Dynamic
Multidetector-Row Helical Computed Tomography Study .................... 222
Administration Dose of Contrast Material.............................................. 222
Iodine Concentration of Contrast Material ............................................. 223
Injection Flow Rate and Duration of Contrast Material ......................... 223
Optimal Scanning Delay ............................................................................. 223
Scanning Timing ......................................................................................... 224
Image Processing ........................................................................................ 224
Magnetic Resonance Imaging ..................................................................... 225
Magnetic Resonance Scanning Technique.................................................. 225
Contrast-Enhanced Dynamic Magnetic Resonance Imaging ..................... 226
Tissue-Specific Contrast-Enhanced Magnetic Resonance Imaging............ 226
Computed Tomography Imaging Features
of Hepatocellular Carcinoma .................................................................. 227
Computed Tomographic Angiography........................................................ 228
Magnetic Resonance Imaging Features
of Hepatocellular Carcinoma .................................................................. 229
Superparamagnetic Iron Oxide-Enhanced Magnetic
Resonance Imaging Feature of Hepatocellular Carcinoma .................... 230
Staging of Hepatocellular Carcinoma ......................................................... 231
Accuracy for the Diagnosis of Hepatocellular Carcinoma ........................ 232
xxxvi Contents

Hepatocellular Carcinoma After Transcatherter

Arterial Chemoembolization................................................................... 232
Hepatocellular Carcinoma After Percutaneous
Radiofrequency Ablation Therapy .......................................................... 232
Limitations and Prospects ........................................................................... 233
Summary ..................................................................................................... 233
References ................................................................................................... 233

17. Hepatocellular Carcinoma: Effect of Injection

Rate/Injection Duration of Contrast Material
on Computed Tomography........................................................................... 237
Tomoaki Ichikawa and Tsutomu Araki
Introduction ................................................................................................. 237
Fixed Injection Rate and Injection Duration of Contrast Material ............. 237
Fixed Injection Rate of Contrast Material .............................................. 238
Fixed Injection Duration of Contrast Material........................................ 239
References ................................................................................................... 239

18. Detection of Combined Hepatocellular

and Cholangiocarcinomas: Enhanced Computed Tomography ............... 241
Akihiro Nishie and Kengo Yoshimitsu
Introduction ................................................................................................. 241
Pathogenesis ................................................................................................ 242
Enhanced Computed Tomographic Findings .............................................. 243
References ................................................................................................... 247

19. Hepatocellular Carcinoma and Adenomatous

Hyperplasia (Dysplastic Nodules): Dynamic Computed
Tomography and a Combination of Computed Tomography
and Angiography ........................................................................................... 249
Kenichi Takayasu
Introduction ................................................................................................. 249
Classification of Nodular Hepatocellular Lesions ...................................... 249
Multistep Progression of Hepatocarcinogenesis ......................................... 250
Needle Biopsy ............................................................................................. 252
Methods of Multidetector Computed Tomography .................................... 253
Methods of a Combination of Computed
Tomography and Angiography ............................................................... 253
CT Images of Advanced Hepatocellular Carcinoma .................................. 253
CT Images of Early Hepatocellular Carcinoma .......................................... 254
Computed Tomography Images of Adenomatous
Hyperplasia and Atypical Adenomatous Hyperplasia ............................ 255
Computed Tomography Images of Nodule-in-Nodule
Hepatocellular Carcinoma ...................................................................... 255
Contents xxxvii

Natural Outcome of Hypo-Attenuating Nodular Lesions ........................... 256

Critical Consideration to Treat Hypoattenuating Lesions........................... 257
References ................................................................................................... 257

20. Hepatocellular Cancer in Cirrhotic Patients:

Radiological Imaging .................................................................................... 261
Francesca Lodato, N. Davies, D. Yu, and Andrew K. Burroughs
Introduction ................................................................................................. 261
The EASL Consensus Statements and AASLD Guidelines ....................... 262
Surveillance for HCC: Radiological Techniques ........................................ 263
Ultrasonography (US), Doppler-Ultrasonography,
Power Doppler-Ultrasonography and
Contrast-Enhanced Ultrasonogrpahy ...................................................... 264
Spiral Computed Tomography .................................................................... 265
Magnetic Resonance Imaging ..................................................................... 267
The Problem of Small Nodules ................................................................... 268
Conclusions ................................................................................................. 270
References ................................................................................................... 271

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

22. Perfusion Scintigraphy with Integrated Single

Photon Emission Computed Tomography/Computed
Tomography in the Management of Transarterial
Treatment of Hepatic Malignancies ............................................................ 287
Timm Denecke, Bert Hildebrandt, and Enrique Lopez-Hänninen
Introduction ................................................................................................. 287
Current Status of Hepatic Arterial Chemotherapy
and Radioembolization ........................................................................... 287
Intraarterial Chemotherapy ..................................................................... 287
Transarterial Radioembolization ............................................................. 289
xxxviii Contents

Technique of Transarterial Treatment ......................................................... 289

Hepatic Arterial Infusion Chemotherapy ................................................ 289
Radioembolization .................................................................................. 290
Visualization of Perfusion Territories of Hepatic
Intraarterial Catheters in Planning and Control
of Transarterial Treatment................................................................... 291
Imaging Techniques ................................................................................ 291
Single Photon Emission Computed Tomography
with Integrated Computed Tomography for Port
Perfusion Scintigraphy ........................................................................ 291
Image Analysis........................................................................................ 293
Therapeutic Consequences...................................................................... 297
Discussion ................................................................................................... 298
References ................................................................................................... 300

23. Postoperative Interferon Alpha Treatment of Patients

with Hepatocellular Carcinoma: Expression of p48
Using Tissue Microarray .............................................................................. 303
Hui-Chuan Sun
Introduction ................................................................................................. 303
Materials and Methods............................................................................ 304
Tissue Microarray and Immunohistochemistry ...................................... 304
Scoring of P48 Immunohistochemistry .................................................. 304
Statistical Analysis .................................................................................. 305
Results ......................................................................................................... 305
Clinicopathological Data ........................................................................ 305
Survival ................................................................................................... 306
Prognostic Factors for Disease-Free Survival
and Overall Survival in Group 1 ......................................................... 306
Discussion ................................................................................................... 307
References ................................................................................................... 309

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

Reverse Transcription-Polymerase Chain Reaction................................ 314

Western Blot............................................................................................ 315
Statistical Analysis .................................................................................. 315
Results ......................................................................................................... 315
Six1 Expression in HCC Cell Lines........................................................ 315
Six1 Expression in HCC Patients and Normal Donors ........................... 317
Six1 Protein Expression Correlated with Advanced Tumor Stage ......... 317
Six1 Protein Expression Correlated with Poor Survival ......................... 318
Discussion ................................................................................................... 319
References ................................................................................................... 322

25. Hepatocellular Carcinoma:

KiSS-1 Overexpression as a Prognostic Factor .......................................... 325
Katharina Schmid, Isabella Mosberger, and Fritz Wrba
Introduction ................................................................................................. 325
Materials ..................................................................................................... 327
Methods....................................................................................................... 327
Tissue Preparation ................................................................................... 327
Immunohistochemistry ........................................................................... 328
Evaluation of Immunohistochemical Result ........................................... 328
Results and Discussion ............................................................................... 328
Patients and Tissue Samples ................................................................... 328
References ................................................................................................... 330

26. Hepatocellular Carcinoma: Prognosis

Using Hepatoma-Derived Growth Factor
Immunohistochemistry ................................................................................. 333
Hideji Nakamura, Kenya Yoshida, and Yasuhiko Tomita
Hepatoma-Derived Growth Factor .............................................................. 333
Developmentally Regulated Expression of Hepatoma-Derived
Growth Factor ..................................................................................... 334
Role in Hepatocarcinogenesis ................................................................. 334
Role in Cancer Progression and Angiogenesis ....................................... 335
Immunohistochemical and Analytical Methods ......................................... 336
Materials ................................................................................................. 336
Method .................................................................................................... 337
Evaluation of Hepatoma-Derived Growth Factor
Expression in Hepatocellular Carcinoma................................................ 337
Prognostic Significance of Hepatoma-Derived
Growth Factor in Hepatocellular Carcinoma .......................................... 338
References ................................................................................................... 340
xl Contents

27. Hepatitis C Virus-Related Human

Hepatocellular Carcinoma: Predictive Markers
Using Proteomic Analysis (Methodology) ................................................... 343
Yasuhiro Kuramitsu and Kazuyuki Nakamura
Hepatitis C Virus-Related Human Hepatocellular
Carcinoma (HCV–HCC)......................................................................... 343
Proteomics................................................................................................... 343
Proteomics for Hepatocellular Carcinoma Tissues ..................................... 344
Proteomics for Sera from Hepatocellular Carcinoma Patients ................... 344
Proteomics for Hepatitis C Virus-Related Hepatocellular
Carcinoma Tissues .................................................................................. 345
Preparation of HCV–HCC Tissue Samples ............................................ 345
Two-Dimensional Gel Electrophoresis (2-DE) ....................................... 345
Sodium Dodecyl Sulfate-Polyacrilamide Gel
Electrophoresis (SDS-PAGE) ............................................................. 345
Image Analysis of the Gels ..................................................................... 346
Mass Spectrometry Analysis................................................................... 346
Amino Acid Sequencing by Liquid Chromatography–
Tandem Mass Spectrometry (LC-MS/MS) ......................................... 346
Proteins Whose Expression Was Increased in HCV–HCC Tissues ........ 346
Proteins With Decreased Expression ...................................................... 347
Proteomics for Auto-Antibodies in the Sera
of Hepatocellular Carcinoma Patients .................................................... 348
References ................................................................................................... 349

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

29. Preclinical Liver Metastases: Three-Dimensional

High-Frequency Ultrasound Imaging ......................................................... 369
Kevin C. Graham, Lauren A. Wirtzfeld, James C. Lacefield,
and Ann F. Chambers
Introduction ................................................................................................. 369
Method: High-Frequency Ultrasound ......................................................... 373
Three-Dimensional Imaging and Volume Calculation................................ 374
High-Frequency Ultrasound Imaging
of Preclinical Liver Metastases ............................................................... 376
Other Applications of High-Frequency Ultrasound
Imaging to Preclinical Cancer Research ................................................. 379
Future Developments – Ultrasound Contrast Agents ................................. 380
Conclusions ................................................................................................. 383
References ................................................................................................... 383

30. Colorectal Liver Metastases:

18
F-Fluorodeoxyglucose-Positron Emission Tomography ......................... 387
Stéphanie Truant, Damien Huglo, and François-René Pruvot
Introduction ................................................................................................. 387
Principle of FDG-PET and General Pitfalls................................................ 388
Colorectal Liver Metastases: Yield of 18F-Fluoro-
Deoxyglucose-Positron Emission Tomography ...................................... 389
Preoperative Staging of Patients
with Colorectal Liver Metastases ........................................................ 389
Detection of Liver Metastases................................................................. 389
Detection of Extrahepatic Metastases ..................................................... 391
Abdominal Cavity ................................................................................... 392
Extra-Abdominal Organs ........................................................................ 395
Impact of FDG-PET Findings on Patient’s Management ....................... 397
Additional Disease and Change of Management .................................... 398
Survival Impact ....................................................................................... 399
Postoperative Follow-Up After Resection of Colorectal Liver
Metastases and Rising Carcinoembryonic Antigen (CEA) ................ 400
FDG-PET for Monitoring the Response to Systemic
or Regional Therapy of Colorectal Liver Metastases ......................... 401
Anticancer Systemic Therapy ................................................................. 401
New Targeted Therapies (Monoclonal Antibodies) ................................ 402
Regional Therapies ................................................................................. 403
Hepatic Arterial Chemotherapy .............................................................. 403
Other Regional Therapies ....................................................................... 404
Conclusion .................................................................................................. 404
References ................................................................................................... 404
xlii Contents

Part VI Biliary Cancer

A. Diagnosis
31. Biliary Cystic Tumors: Clinicopathological
Features .......................................................................................................... 411
Yasuni Nakanuma, Hiroko Ikeda, Yasunori Sato,
Kenichi Harada, Koichi Nakamura, and Yoh Zen
Introduction ................................................................................................. 411
Anatomical Classification of the Biliary Tree............................................. 412
Non-Neoplastic Biliary Cystic Lesions ...................................................... 412
Peribiliary Cysts ...................................................................................... 412
Hepatic Foregut Cyst .............................................................................. 414
Biliary Hamartoma.................................................................................. 415
Biliary Cystic Neoplasm ............................................................................. 416
Hepatobiliary Cystadenoma and Cystadenocarcinoma .......................... 417
Intraductal Papillary Neoplasm of the Bile Duct (IPNB) ....................... 419
Other Types of Neoplastic Cystic Diseases ............................................ 426
References ................................................................................................... 426

32. Cholangiocarcinoma: Intraductal Sonography.......................................... 429

Kazuo Inui, Hironao Miyoshi, and Junji Yoshino
Introduction ................................................................................................. 429
Methods....................................................................................................... 429
Instruments.............................................................................................. 429
Insertion Methods ................................................................................... 430
Indications ............................................................................................... 431
Images and Clinical Utility ......................................................................... 432
Intraductal Sonography ........................................................................... 432
Three-Dimensional Intraductal Sonography ........................................... 432
Discussion ................................................................................................... 434
Conclusions ................................................................................................. 436
References ................................................................................................... 436

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

Role of p53 in Malignant Tumors of the Bile Duct .................................... 443

Analysis of p53 Mutations ...................................................................... 444
Roles of p73 and p63 in Malignant Tumors of the Bile Duct ..................... 444
Analyses of p63 and p73 in Tumors ....................................................... 447
References ................................................................................................... 448

34. Extrahepatic Bile Duct Carcinoma:

Mucin 4, a Poor Prognostic Factor .............................................................. 451
Michiyo Higashi, Shugo Tamada, Kohji Nagata,
Masamichi Goto, and Suguru Yonezawa
Introduction ................................................................................................. 451
Mucin Characteristics ................................................................................. 451
MUC4 Mucin .............................................................................................. 452
Antibodies Against MUC4 ......................................................................... 453
MUC1 Mucin .............................................................................................. 453
Antibodies Against MUC1 ......................................................................... 455
Methods....................................................................................................... 455
Tissue Sample Management ................................................................... 455
Immunohistochemistry ............................................................................... 458
References ................................................................................................... 458

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

Part VII Splenic Cancer

A. Diagnosis
36. Splenic Metastases: Diagnostic Methods .................................................... 489
Eva Compérat and Frédéric Charlotte
Introduction ................................................................................................. 489
Epidemiology .............................................................................................. 489
Pathologic Features ..................................................................................... 489
Pathogenesis ................................................................................................ 490
Clinical Features ......................................................................................... 491
Diagnostic Procedures ................................................................................ 493
Differential Diagnosis ................................................................................. 494
References ................................................................................................... 496

Index ........................................................................................................................ 499

A. Diagnosis
1
Applications of Positron Emission
Tomography in Liver Imaging:
An Overview
Amir H. Khandani
INTRODUCTION
While positron emission tomography (PET)
has been used for several decades in the
research setting, its clinical use has grown
substantially in the past decade. This is a
quantitative physiological imaging modal-
ity utilizing positron emitters such as
Fluorine-18, Oxygen-15, Nitrogen-13 and
Carbon-11. The fact that several of these
nuclides are components of common bio-
logical molecules makes PET particularly
suitable to visually capture different bio-
logic pathways; however, the short physical
half-life of Oxygen-15 (2 min), Nitrogen-13
(10 min), and Carbon-11 (20 min) limits
their utilization to the centers with an on-
site cyclotron. Fluorine-18 has a physical
half-life of 110 min and can be synthesized
in commercially operating cyclotrons and
transported to remote PET facilities. The
most commonly used radiotracer is 2-[18F]
fluoro-2-deoxy-D-glucose (FDG).
Tumor imaging with FDG is based on
the principle of increased glucose metabo-
lism of cancer cells. Like glucose, FDG
is taken up by the cancer cells through
facilitative glucose transporters (Glut).
These transporters are glycoproteins, and
so far, 12 isoforms have been identified
in different organs. Normal hepatocytes
express Glut2, Glut9, and Glut10 (Joost
and Thorens, 2001). Expression of Gluts,
predominantly Glut1 and Glut3, is signifi-
cantly increased in most malignancies.
Once in the cell, glucose or FDG is
phosphorylated by hexokinase to glucose-
6-phospate or FDG-6-phosphate, respec-
tively. Expression of hexokinase and its
affinity/functional activity for phosphory-
lation of glucose or FDG is often higher in
cancer cells compared to normal cells; hex-
okinase II is predominantly expressed in
cancer cells. Glucose-6-phosphate travels
further down the glycolytic or oxidative
pathways to be metabolized, in contrast to
FDG-6-phosphate, which cannot go further,
and cannot be metabolized. In normal cells,
glucose-6-phospate or FDG-6-phosphate
can be dephosphorylated and exit the cells.
In many cancer cells, however, expression
of glucose-6-phosphatase is often signifi-
cantly decreased, and therefore, glucose-
6-phospate or FDG-6-phosphate is only
minimally dephosphorylated and remains
in large part within the cell. On the other
hand, because FDG-6-phosphate cannot
be metabolized, it is trapped in the cell as a
polar metabolite, and can be visualized by
PET. In normal liver cells, the concentration
5
6 A.H. Khandani
of glucose-6-phosphatase is high, which
causes rapid clearance of FDG from the
liver. This may account for the mild inten-
sity of the normal liver on FDG-PET.
Furthermore, the interindividual as well as
the intraindividual variation of the hepatic
uptake appears to be minimal (Rydberg
et al., 2005). Therefore, while comparing
PET scans of different patients or of the
same patient at different occasions, liver
uptake can be used as the reference; first,
the PET scans to be compared should be
displayed with the liver uptake on all the
scans being the same, then the areas of
interest can be compared to each other.
The radioisotope portion of the molecule
used in PET imaging emits a positron (a
positively charged electron), which travels
a distance of a few millimeters in tissue
before it collides with a negatively charged
electron. This collision annihilates the
entire mass of the positron and electron into
two photons (gamma rays) with energy of
511 KeV each. These two photons travel at
the speed of light in exactly opposite direc-
tions (180° apart). Coincidental detection
of these two photons by two oppositely posi-
tioned detectors in the PET scanner results
in images with a much higher resolution
compared to the conventional, single pho-
ton nuclear medicine studies and presents
the possibility of quantitative measurement
of the tracer uptake in a lesion of interest.
PET CT allows the fusion of the metabolic
information on PET with the anatomic
information on CT. PET CT has been
proven to increase the diagnostic accuracy
compare to stand-alone PET. In PET CT,
the patient undergoes a CT scan, followed
by a PET scan without changing the patient
position. PET for most oncologic indica-
tions (whole body PET) is acquired from
the base of the skull through the upper
thighs. The CT portion (transmission scan)
of the PET CT is acquired within seconds
while the PET acquisition time for each
bed position (~ 15 cm) is several minutes;
the total PET acquisition time in newer
machines is 10–25 min.
In addition to delivering anatomic infor-
mation, the CT portion of PET CT is used
to measure the attenuation of the X-ray
photons traveling through the patient in
order to produce the so-called “attenuation
map” and correct the PET data for tissue
attenuation (attenuation corrected PET).
During PET acquisition, photons from
structures deep in the abdomen or pelvis
are attenuated more compared to those
from the superficial structures and the
chest. Therefore, the intensity of uptake
in the deeper structures may be underesti-
mated on the “non-attenuation corrected”
PET images. The intensity of uptake in
the deeper structures is normalized to the
intensity of uptake in the superficial struc-
tures on the attenuation-corrected PET
images. While correction of the PET data
for tissue attenuation is indispensable,
misalignment between PET and CT can
cause mislocalization of lesions on the
fused PET-CT images. This may be due
to changes in position of a body part (e.g.,
neck or legs) or due to physiologic changes
in the position of an organ (i.e., respiratory
movement) between the transmission (CT)
and the emission scan (non-attenuation
corrected PET). As the degree of misalign-
ment and resulting mislocalization can be
significant, one has to be cautious when
interpreting the attenuation-corrected PET
CT images or using PET CT images for
radiation therapy planning. The magnitude
of this misalignment can be assessed by
1. Applications of Positron Emission Tomography in Liver Imaging: An Overview
fusing the non-attenuation corrected PET
images with CT, which can be performed
on any PET review stations. In case of sig-
nificant misalignment, the non-attenuation
corrected PET images should be reviewed
without fusion with CT and the metabolic
findings on PET should be correlated side
by side with the anatomic findings on CT.
At its introduction, the CT portion of the
PET CT was acquired without contrast and
used solely for attenuation correction and
anatomic localization of the PET findings.
With the development of multislice CT
machines and their introduction to PET CT
scanners, the CT portion of the PET CT
studies began to be performed as diagnostic
CT with IV and oral contrast, as needed.
Because the density of IV and oral contrast
used in the diagnostic CT can produce arti-
facts with overestimation of FDG uptake in
areas of high contrast density, the protocol
for CT has to be modified with the general
approach of dilution of the contrast medium.
Alternatively, one can acquire a low or inter-
mediate dose unenhanced CT first to be
used as attenuation map, followed by emis-
sion scan; after which the IV enhanced CT
should be performed. The IV enhanced CT
scan may be performed from the entire body
or only from the area of interest in order to
limit the patient’s exposure to radiation.
METASTATIC LIVER DISEASE
Zimmerman et al. (2002a) studied expres-
sion of Glut1 in hepatic metastases origi-
nating from different primaries, and found
that Glut1 was overexpressed in hepatic
metastases of 3 of 5 lung primaries, 7
of 11 pancreatic primaries, 7 of 9 colon
primaries, 2 of 7 breast primaries, 2 of 2
7
squamous cell primaries, 1 of 3 biliary
tract primaries, and none of neuroendo-
crine primaries that were examined.
Metastatic disease accounts for > 90% of
malignant liver lesions. Often, the presence
of liver metastases is the main determi-
nant of survival and guides the therapeutic
strategy, particularly in patients with color-
ectal cancer. Among the colorectal cancer
patients that die of the disease, 70% have
liver metastases and a significant 10% have
liver-only disease. Even in those patients
with metastases to multiple organ sites, the
extent of liver disease remains the primary
determinant of survival. Therefore, the
timely diagnosis of the metastatic disease
to the liver is of paramount importance.
The sensitivity of FDG-PET in detect-
ing hepatic metastases from different pri-
maries is > 90%. Delbeke et al. (1998)
studied the diagnostic value of FDG-PET
in hepatic metastases and detected all 66
metastatic lesions originating from various
primaries including the colon, pancreas,
esophagus, sarcoma, and parotid. Similar
results with an overall greater sensitivity
of PET compared to spiral CT have been
reported by other groups (Hustinx et al.,
1998). Computed tomography can be read
to achieve a higher sensitivity, but it will
be at the expense of specificity, as many
CT positive lesions are false-positives.
This was shown in a study by Marom
et al. (1999) in metastatic lung cancer,
where in a prospective study of 100 patients
nearly twice as many lesions in the liver
were identified by CT as by PET; however,
all of the incremental lesions identified
by CT were false-positives. Comparing
magnetic resonance imaging (MRI) and
PET, Yang et al. (2003) reviewed studies
of 30 patients with histopathologically
8 A.H. Khandani

proven (# 27) or clinically suspected (# 3) detect additional hepatic metastases. This

hepatic metastases from non-hepatic pri-
maries. The sensitivity, specificity, and
positive and negative predictive values on
MRI were 85.7%, 100%, 100%, and 89%,
respectively, compared to 71%, 93.7%,
90.9%, and 79% on PET. The difference
between the two methods was not sta-
tistically significant (Yang et al., 2003).
A meta-analysis of the literature on detec-
tion of hepatic metastases from colorectal,
gastric, and esophageal cancers by ultra
sound (US), CT, MRI, and PET revealed
that in studies with a specificity > 85%, the
mean weighted sensitivity was 55% (95%
CI: 41, 68) for US, 72% (95% CI: 63, 80)
for CT, 76% (95% CI: 57, 91) for MRI,
and 90% (95% CI: 80, 97) for PET. Results
of pairwise comparison between imaging
modalities demonstrated a greater sensi-
tivity of PET than US (P = .001), CT (P =
.017), or MRI (P = .055). The conclusion
was that at an equivalent specificity, PET
is the most sensitive noninvasive imag-
ing modality for the diagnosis of hepatic
metastases from colorectal, gastric, and
esophageal cancers (Kinkel et al., 2002).
In cases of known solitary hepatic metas-
tasis diagnosed by CT or MRI, PET can
is of particular importance in preoperative
evaluation of solitary hepatic metastasis
as detection of additional lesions often
changes the management. In the case of
suspected recurrent cancer based on rising
tumor markers or other parameters, FDG-
PET is more sensitive than CT or MRI for
discovering hepatic metastases with the
potential of detecting disease earlier than
CT or MRI, so that metastatic disease is
more amenable to curative resection.
False-positive PET findings for malig-
nancy in the liver can occur due to acute
infection (intrahepatic abscess, penetrating
gallbladder empyema) or chronic inflam-
matory processes (granulomas). Since most
of these conditions can be excluded by clin-
ical history and review of the transmission
CT and/or available dedicated CT or MRI,
a focal area of increased uptake in the liver
on PET that does not represent one of the
above conditions, has to be considered as
malignant until proven otherwise. Positron
emission tomography is of particular ben-
efit in case of negative or indeterminate CT
or MRI findings for malignancy (Figure
1.1). Hemangiomas and adenomas of the
liver are not FDG avid.

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
False-negative PET for hepatic metas-
tases can be due to limitations of spatial
resolution. It should be considered, how-
ever, that many of the PET publications
were based on non-attenuation corrected
images, which may have lower sensitivity,
especially for lesions deeper in the liver;
such lesions generally appear much less
intense on non-attenuation corrected than
on attenuation corrected PET images.
Generally, the role of PET in detecting
subcentimeter lesions should be redefined,
considering the higher spatial resolution
and sophisticated image correction and
reconstruction algorithms of current PET
machines. Further improvements in image
resolution from currently 0.6–0.8 cm to a
few millimeters can be expected with the
development of small surface area crystal
elements in combination with alternative
position-sensitive photomultiplier tubes,
and the implementation of depth of inter-
action measurements.
False-negative FDG-PET can occur also
due to underestimation of uptake, mislo-
calization of foci as well as recent comple-
tion of chemotherapy. The latter is likely
associated with significant decrease in the
number of the cells or their glucose metab-
olism after chemotherapy. No definitive
information is available in the literature on
the time interval after completion of chem-
otherapy during which PET can be false-
negative. Based on experience at large
centers, this time interval is ~ 4–6 weeks
for high-grade lymphoma. However, no
definite statement can be made regard-
ing the time interval after chemotherapy
during which liver metastases may not be
visualized on PET.
Underestimation of uptake of malignant
lesions causing false-negative findings on
PET can occur due to physiologic move-
9
ment of the liver during the emission
scan. The liver is an upper abdominal
organ that moves with respiratory move-
ment of the diaphragm. Emission scans
are acquired over several minutes during
which hepatic lesions are in a repetitive
cranio-caudal pendulous movement. The
respiratory excursion of the liver is
~ 10–30 mm. Therefore, it is conceivable
that cranial and caudal portions of small
lesions are registered only during a portion
of the acquisition time, and their uptake is
underestimated, so that they appear less
intense on images than they really are.
The degree of this underestimation is vari-
able, and particularly in the case of a sub-
centimeter lesion, this may even lead to
non-visualization of the lesion. One way
to overcome this problem is to increase
the target-to-background count ratio by
increasing the acquisition time. This can
be done by increasing the acquisition time
in the mid- and upper-abdomen while
acquiring the whole body scan, if this is
possible on the PET machine. Another
solution would be to acquire a second
PET scan from mid- and upper-abdomen,
consisting of a transmission scan followed
by an emission scan with longer acquisi-
tion time, once the whole body scan is
completed (“Liver View”). A different and
preferable approach to solve this problem
is Respiratory Gating, in which case only
emission data collected in certain parts of
the respiratory cycle are used for image
reconstruction, resulting in better visuali-
zation of small lesions with the disadvan-
tage of longer acquisition time.
A changed position of the liver between
the emission and transmission scan can cause
misalignment between these two images,
resulting in mislocalization of pulmonary
or colonic foci into the liver or vice versa.
10
This is especially of concern when trans-
mission CT is acquired during full inspira-
tion and the PET scan is acquired during
shallow breathing, but appears to be less
of a problem if the transmission CT is
performed during shallow breathing or in
stand-alone PET units where a transmis-
sion source such as Ge-68 is used for
attenuation correction. In the case of sus-
pected misalignment, the non-attenuation
corrected PET images should be reviewed
for corrected localization of the lesion.
Given the higher sensitivity of PET com-
pared to CT or MRI in detecting hepatic
metastasis, it is conceivable that PET will
be increasingly employed for preoperative
staging of malignant tumors with poten-
tial of hepatic metastasis. Presently, CT
with IV contrast or MRI is still needed for
adequate evaluation of anatomic resecta-
bility of liver metastases since an assess-
ment of the venous anatomy in the liver is
critical for defining hepatic lobar anatomy.
However, with the availability of PET-
multislice CT, one can expect that in the
future, PET and contrast CT will be per-
formed in one setting as the preoperative
imaging study of choice in patients under-
going resection of hepatic metastases.
CHOLANGIOCARCINOMA
Cholangiocarcinoma originates from the
epithelial cells of the biliary tract. After
hepatocellular carcinoma, cholangiocarci-
noma is the second most common primary
tumor of the liver, accounting for ~ 5–30%
of the primary hepatic malignancies. Biliary
obstruction with jaundice is the most com-
mon presenting clinical feature in hilar
cholangiocarcinoma while it is uncom-
A.H. Khandani
mon in peripheral cholangiocarcinoma.
The diagnosis of cholangiocarcinoma has
been based on clinical picture, laboratory
values, radiologic imaging, and histol-
ogy, although the latter is often inconclu-
sive in differentiating between this tumor
and metastatic adenocarcinoma. Work up
generally consists of MRI, magnetic reso-
nance cholangiopancreatography (MRCP),
CT, endoscopic retrograde cholangiopan-
creatography (ERCP), and percutaneous
transhepatic cholangiography (PTC).
Overall, the prognosis of this tumor is
dismal with 5-year survival as low as 15%;
however, improved survival of 20–30% at 5
years are achievable with portal or arterial
embolization followed by trisegmentec-
tomy. Therefore, preoperative assessment
for hepatic and extra-hepatic metastases is
likely of prognostic value.
Glut1 is not expressed in normal bile
duct, but has been described to be strongly
expressed in cholangiocarcinoma (Roh
et al., 2004; Zimmerman et al., 2002b).
Overall, cholangiocarcinoma appears to
be highly FDG avid. Delbeke et al. (1998)
evaluated eight patients with cholangiocar-
cinoma: all lesions demonstrated intense
FDG uptake (Delbeke et al., 1998). Hilar
and extrahepatic cholangiocarcinoma,
however, appears to be less intense on
FDG-PET than the peripheral cholangi-
ocarcinoma, which may be associated with
smaller size and/or higher mucin content
of the hilar tumors compared to the periph-
eral ones. Peripheral tumors accounts for
~ 10% of all cholangiocarcinoma, and
often has a characteristic central photo-
penia on FDG-PET, which corresponds
to the central core of fibrotic tissue and
desmoplastic reaction provoked by the
neoplastic cells; on contrast enhanced CT
or MRI, this is evident by early moderate
1. Applications of Positron Emission Tomography in Liver Imaging: An Overview
peripheral enhancement followed by pro-
gressive and concentric filling.
FDG-PET is helpful in detecting unsus-
pected extra-hepatic metastatic disease
from cholangiocarcinoma. Kim et al.
(2003) discovered three unsuspected
extra-abdominal lymph nodes and three
unsuspected lung metastases in a group
of 21 patients with this tumor. Kato et
al. (2002) reported 100% specificity for
regional nodal involvement on FDG-PET.
Especially in cases of peripheral cholan-
giocarcinoma, PET should be considered
to evaluate the patient for extrahepatic
metastases: peripheral cholangiocarcinoma
attains a large size before it becomes clini-
cally apparent, since it does not obstruct
the central biliary system. Also, extrahe-
patic metastases are commonly present at
the time of diagnosis. Also, if the histology
is inconclusive in differentiating between
cholangiocarcinoma and metastatic adeno-
carcinoma from an extrahepatic primary,
FDG-PET may help to make this dif-
ferentiation by detecting the extrahepatic
primary site.
GALL BLADDER CARCINOMA
Gallbladder carcinoma is the most com-
mon malignant tumor of the biliary tract;
this is a highly fatal disease that presents
in 80–95% of cases as adenocarcinoma.
The diagnosis of gallbladder carcinoma
is rarely made preoperatively; in most
cases, gallbladder carcinoma presents as
an incidental finding at the time of chole-
cystectomy for presumed benign etiol-
ogy. In post-cholecystectomy setting, in
patients with T1a (tumor invades mucosa),
simple cholecystectomy is regarded as
adequate treatment, but these represent
11
only ∼15% of all gallbladder carcinoma
cases. Patients with T1b tumors (tumor
invades muscularis) and beyond should
be further evaluated for spread of disease
as T1b tumors are associated with lymph
node involvement in 15% of cases vs.
2.5% of cases with T1a disease.
Kim et al. (2002) examined 71 cases of
gallbladder carcinoma and demonstrated
expression of Glut1 in 37 cases (52.1%).
Preoperatively, FDG-PET has been used
to distinguish malignant from benign gall-
bladder lesions with a sensitivity of 75%
(six of eight patients), and specificity of
87.5% (seven of eight patients) (Koh et al.,
2003); however, this will likely be of lim-
ited clinical use, since most patients are
status post cholecystectomy at the time of
the diagnosis of gallbladder carcinoma.
Anderson et al. (2004) used PET to
assess possible local residual and meta-
static disease after cholecystectomy. This
method detected local residual disease in
seven of nine cases (78%), and remote
metastases in two of three cases. Positron
emission tomography, however, detected
only three of six cases with carcinomatosis.
This imaging modality can potentially be
used to assess for local residual and meta-
static disease within hours to days after
incidental detection of gallbladder carci-
noma at cholecystectomy. There has been
some concern that non-specific uptake in
the early post-surgical period may disturb
image interpretation. Although generally
surgical interventions can cause increased
FDG uptake in the surgical bed, there have
been no systematic studies or publications
concerning FDG uptake in a post-operative
setting. Additionally, one has to consider
that the degree of local inflammatory reac-
tion varies between different tissues and
also depends on the type of intervention.
12 A.H. Khandani

Generally, the decision to perform PET

after a surgical intervention, should strongly
take into account the clinical factors, such
as T-stage in case of gallbladder carcinoma,
and not merely the post-operative time
interval. Furthermore, the interpretation of
PET images should not be only based on the
intensity of uptake, but should also take into
account the pattern of uptake (round/local-
ized/multiple for malignancy vs. diffuse
for inflammation). Lastly, however, asses-
sing for distant metastases is likely to be
independent of the factors influencing the
FDG in the surgical bed. In author’s experi-
ence, FDG-PET is very useful for local and
distant staging of patients with gallblad-
der carcinoma detected at cholecystectomy
(Figure 1.2).

HEPATOCELLULAR
CARCINOMA
Hepatocellular carcinoma (HCC) or
hepatoma develops through malignant
transformation of hepatocytes and is com-
mon in the setting of chronic liver changes
and cirrhosis. This cancer is the most
common primary epithelial malignancy of
the liver, accounting for ~ 80% of malig-
nant epithelial neoplasms of the liver.
Hepatocellular carcinoma has been showing
an increasing incidence in Western coun-
tries due to increasing frequency of hepa-
titis B and C viral infection. The diagnosis
is based on screening risk populations with
measurements of serum AFP and liver
ultrasound. Magnetic resonance imaging,
CT, and lipiodol angiography with follow-
up CT are used in inconclusive cases to
establish the diagnosis. Biopsy is only
performed on patients in whom the radio-
logical diagnosis cannot be made.
Figure 1.2. Patient status post laparoscopic chole-
cystectomy with T2 gallbladder cancer. PET was
requested for staging and indicated metastases in
the porta hepatis, paraaortic (arrow heads) and par-
atracheal lymph nodes (arrow)
Facilitative glucose transporters do not
appear to be overexpressed in HCC as often
as in other malignant tumors. Zimmerman
et al. (2002b) and Roh et al. (2004) studied
expression of Glut1 in 35 and 22 cases of
HCC, respectively; the first group reported
two positive cases and the second group
reported one positive case. Delbeke et al.
(1998) examined a series of 23 patients
with HCC. In visual assessment, differ-
entiation of tumor from normal liver was
definitive in 13 patients, equivocal (mildly
increased compared to normal liver) in
1. Applications of Positron Emission Tomography in Liver Imaging: An Overview
3 patients, and poor (same or less intense
compared to normal liver) in 7 patients.
The sensitivity of FDG-PET for HCC is
~ 50–60%. There appears to be some asso-
ciation between histologic differentiation
of HCC and FDG uptake, with poorly dif-
ferentiated tumors being more intense on
FDG-PET, likely explained by enzymology
of HCC. Concentration of glucose-6-
phospatase is high in normal liver, causing
rapid clearance of glucose-6-phospate or
FDG-6-phosphate from hepatocytes with
consequent mildly intense appearance of
liver on PET. The enzymology of well-
differentiated HCC resembles that of the
normal liver, possibly explaining mild FDG
uptake or non-visualization of these tumors
on PET. Less differentiated HCC tumors
have lower levels of glucose-6-phosphatase
and higher levels of hexokinase, possi-
bly causing intense FDG uptake of these
tumors on PET (Torizuka et al., 1995).
Also, there appears to be some association
between FDG uptake and tumor-volume
doubling time as well as between FDG
uptake and tumor size. Trojan et al. (1999)
visualized all six tumors larger than 5 cm
in a series of 14 HCC tumors. Therefore,
FDG PET could possibly be used to assess
the effect of treatment in larger and less
differen-tiated HCC. High FDG uptake in
HCC has been associated with overexpres-
sion of mRNA levels for several markers,
such as VEGF of aggressive tumor behavior
(Lee et al., 2004).
Detection of extrahepatic FDG avid
metastases originating from HCC has also
been reported; especially in cases of less-
differentiated HCC, metastases appear
to be more FDG avid (Ho et al., 2003;
Trojan et al., 1999). Although more data
are needed to establish the clinical role
of FDG-PET in HCC, based on author’s
13
experience, whole body FDG-PET is very
helpful to assess the malignant poten-
tial of a hepatic lesion of unknown pri-
mary because FDG avidity of the lesion
increases the likelihood of its malignant
nature. Furthermore, simultaneous visu-
alization of FDG avid foci outside of the
liver can be viewed as almost consistent
with malignant etiologyof the liver lesion.
Positron emission tomography may also
visualize an easily accessible biopsy site
outside of the liver. Also in cases of known
HCC with clinically suspected extrahe-
patic metastasis, FDG-PET can be used
to confirm this suspicion. Nonetheless, a
negative FDG-PET scan in such patients
does not exclude the possibility of HCC.
[1-11C] acetate (11-C Acetate) is a
short-lived PET tracer with higher uptake
in well-differentiated hepatomas compared
to poorly differentiated ones. The physical
half-life of 11-C is 20 min. Several possible
pathways of acetate incorporation into
tumor metabolism have been postulated;
the dominant one appears to be partici-
pation in free fatty acid synthesis. Ho
et al. (2003) studied a series of 57 patients
with liver lesions, consisting of 39 cases
of HCC (group I), 13 cases of non-HCC
malignancies (three cases of cholangi-
ocarcinomas and ten cases of hepatic
metastases from colon, breast, lung and
carcinoid primaries, group II) as well as
5 cases of benign liver lesions (2 cases of
cavernous hemangioma, 2 cases of FNH
and 1 case of adenoma, group III). In group
I, there was a “complementing” sensitivity
for HCC on FDG and 11-C Acetate-PET,
as these two tracers together gave a sensi-
tivity of 100% for detecting HCC. Most of
the well-differentiated HCC lesions were
intense on 11-C Acetate scan with mild
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)

In conclusion, FDG-PET has an impor- Einstein College of Medicine, for critical

tant role in determining if there are metas-
tases to the liver and whether disease has
spread beyond the liver. This information
is critical for planning surgical resections
of liver metastases. While FDG-PET can
fail to detect many well differentiated
hepatocellular carcinomas, it does detect
many of the poorly differentiated ones,
and other PET tracers show promise for the
detection of better differentiated hepato-
mas. While low volume cholangiocarci-
nomas can escape detection by FDG-PET,
higher volume lesions are well detectable.
Similarly, gallbladder carcinoma is gener-
ally well detected by FDG-PET. The ability
of FDG-PET to quantitatively estimate
metabolic rates makes it an important tool for
therapy monitoring. Thus, with increas-
ingly broad applications for FDG-PET
imaging, it is expected that FDG-PET
(and PET/CT) of the liver will play a
growing and increasingly important role
in detecting and monitoring treatment of
tumors involving the liver.
Acknowledgement. The author would like
to thank M. Donald Blaufox, Chairman,
Department of Nuclear Medicine, Albert
review of the manuscript.
REFERENCES
Anderson, C.D., Rice, M.H., Pinson, C.W.,
Chapman, W.C., Chari, R.S., and Delbeke, D.
2004. Fluorodeoxyglucose PET imaging in
the evaluation of gallbladder carcinoma and
cholangiocarcinoma. J. Gastrointest. Surg. 8:
90–97.
Delbeke, D., Martin, W.H., Sandler, M.P.,
Chapman, W.C., Wright, J.K. Jr., and Pinson,
C.W. 1998. Evaluation of benign vs malignant
hepatic lesions with positron emission tomogra-
phy. Arch. Surg. 133: 510–515.
Ho, C.L., Yu, S.C., and Yeung, D.W. 2003. 11C-
acetate PET imaging in hepatocellular carci-
noma and other liver masses. J. Nucl. Med. 44:
213–221.
Hustinx, R., Paulus, P., Jacquet, N., Jerusalem, G.,
Bury, T., and Rigo, P. 1998. Clinical evaluation
of whole-body 18F-fluorodeoxyglucose positron
emission tomography in the detection of liver
metastases. Ann. Oncol. 9: 397–401.
Joost, H.G., and Thorens, B. 2001. The extended
GLUT-family of sugar/polyol transport facilita-
tors: nomenclature, sequence characteristics, and
potential function of its novel members. Mol.
Membr. Biol. 18: 247–256.
Kato, T., Tsukamoto, E., Kuge, Y., Katoh, C.,
Nambu, T., Nobuta, A., Kondo, S., Asaka, M.,
and Tamaki, N. 2002. Clinical role of (18)
F-FDG PET for initial staging of patients with
16
extrahepatic bile duct cancer. Eur. J. Nucl. Med.
Mol. Imaging 29: 1047–1054.
Kim, Y.J., Yun, M., Lee, W.J., Kim, K.S., and Lee,
J.D. 2003. Usefulness of 18F-FDG PET in intra-
hepatic cholangiocarcinoma. Eur. J. Nucl. Med.
Mol. Imaging 30: 1467–1472.
Kim, Y.W., Park, Y.K., Yoon, T.Y., and Lee,
S.M. 2002. Expression of the GLUT1 glu-
cose transporter in gallbladder carcinomas.
Hepatogastroenterology 49: 907–911.
Kinkel, K., Lu, Y., Both, M., Warren, R.S., and
Thoeni, R.F. 2002. Detection of hepatic metas-
tases from cancers of the gastrointestinal tract
by using noninvasive imaging methods (US, CT,
MR imaging, PET): a meta-analysis. Radiology
224: 748–756.
Koh, T., Taniguchi, H., Yamaguchi, A., Kunishima, S.,
and Yamagishi, H. 2003. Differential diagnosis
of gallbladder cancer using positron emission
tomography with fluorine-18-labeled fluoro-
deoxyglucose (FDG-PET). J. Surg. Oncol. 84:
74–81.
Lee, J.D., Yun, M., Lee, J.M., Choi, Y., Choi, Y.H.,
Kim, J.S., Kim, S.J., Kim, K.S., Yang, W.I.,
Park, Y.N., Han, K.H., Lee, W.J., Yoo, N., Lim,
S.M., and Park, J.H. 2004. Analysis of gene
expression profiles of hepatocellular carcinomas
with regard to 18F-fluorodeoxyglucose uptake
pattern on positron emission tomography. Eur. J.
Nucl. Med. Mol. Imaging 3: 1621–1630.
Marom, E.M., McAdams, H.P., Erasmus, J.J.,
Goodman, P.C., Culhane, D.K., Coleman, R.E.,
Herndon, J.E., and Patz, E.F. Jr. 1999. Staging
non-small cell lung cancer with whole-body
PET. Radiology 212: 803–809.
Roh, M.S., Jeong, J.S., Kim, Y.H., Kim, M.C., and
Hong, S.H. 2004. Diagnostic utility of GLUT1
in the differential diagnosis of liver carcinomas.
Hepatogastroenterology 5: 1315–1318.
A.H. Khandani
Rydberg, J., Khandani, A.H., Ivanovic, M., and
McCartney, W.H. 2005. Comparison of Liver,
Mediastinum, and Muscle as References for
Background Activity in 18-F FDG PET-CT
Imaging. RSNA Annual Meeting: Abstract # LPL
15-04-P.
Torizuka, T., Tamaki, N., Inokuma, T., Magata, Y.,
Sasayama, S., Yonekura, Y., Tanaka, A.,
Yamaoka, Y., Yamamoto, K., and Konishi, J.
1995. In vivo assessment of glucose metabolism
in hepatocellular carcinoma with FDG-PET.
J. Nucl. Med. 36: 1811–1817.
Trojan, J., Schroeder, O., Raedle, J., Baum, R.P.,
Herrmann, G., Jacobi, V., and Zeuzem, S. 1999.
Fluorine-18 FDG positron emission tomography
for imaging of Hepatocellular carcinoma. Am. J.
Gastroenterol. 94: 3314–3319.
Tsuda, M., Rikimaru, H., Majima, K., Yamada, T.,
Saito, H., Ishibashi, T., Takahashi, S., Miyachi,
H., Endoh, M., and Yamada, S. 2003. Time-
related changes of radiofrequency ablation lesion
in the normal rabbit liver: findings of magnetic
resonance imaging and histopathology. Invest.
Radiol. 38: 525–531.
Yang, M., Martin, D.R., Karabulut, N., and Frick,
M.P. 2003. Comparison of MR and PET imaging
for the evaluation of liver metastases. J. Magn.
Reson. Imaging 17: 343–349.
Zimmerman, R.L., Burke, M., Young, N.A.,
Solomides, C.C., and Bibbo, M. 2002a.
Diagnostic utility of Glut-1 and CA 15-3 in
discriminating adenocarcinoma from hepatocel-
lular carcinoma in liver tumors biopsied by fine-
needle aspiration. Cancer 96: 53–57.
Zimmerman, R.L., Fogt, F., Burke, M., and
Murakata, L.A. 2002b. Assessment of Glut1
expression in cholangiocarcinoma, benign
biliary lesions and hepatocellular carcinoma.
Oncol. Rep. 9: 689–692.
2
Localized Fibrous Tumor
of the Liver: Imaging Features
Thomas Moser and Tereza S. Nogueira
Also known as solitary fibrous tumors,
or localized fibrous mesothelioma, local-
ized fibrous tumors (LFTs) can occur
in numerous different locations including
most organs and soft tissues. The arche-
type of this entity has been described
in the pleura, where it accounts for 5%
of tumors. On the other hand, LFT of
the liver is a very infrequent tumor with
no more than 30 cases reported in the
English language literature. Hepatic LFT
usually manifests with vague abdominal
complaints or even mildly abnormal liver
blood tests. However, some patients remain
asymptomatic and the tumor represents an
incidental finding. It generally attains large
size before discovery. Patients may also
present with hypoglycemia as a paraneo-
plastic manifestation. Definitive diagnosis
of LFT reposes on histology coupled with
the characteristic immunological labeling
for CD34.
Imaging studies may sometimes suggest
the diagnosis. As for any hepatic tumor,
multiphasic acquisition, either at computed
tomography (CT) scan or magnetic reso-
nance imaging (MRI), is required. It means
that after intravenous injection of contrast
media, images are obtained during maxi-
mal hepatic artery and portal and hepatic
veins opacification (~30 and 70 s following
the beginning of injection, respectively).
Occasionally, delayed images are also
obtained (at 5 min or more). Contrast uptake
within hepatic lesions is analyzed with
respect to these different phases, enabling
tumor characterization. For example, hyper-
vascular tumors enhance simultaneously to
the hepatic artery. Hypovascular tumors
disclose only mild or absent enhancement
during the portal phase. Lesions contain-
ing a fibrous contingent may enhance only
during the delayed phase. Such delayed
enhancement is believed to correspond
to interstitial contrast diffusion through
collagen fibers.
On imaging studies, LFTs usually present
as large well limited masses that could be
partially or entirely capsulated. Most are
heterogeneous lesions, sometimes disclos-
ing cystic spaces. Calcification and necrosis
are infrequent. On unenhanced CT, they are
essentially hypodense to the surrounding
parenchyma. On unenhanced MRI, they
are hypointense on T1-weighted images,
whereas T2-weighted images show
hypo- and hyperintense areas with some
of them appearing fluid-like. After intra-
venous contrast injection, two principal
enhancement types have been described
so far.We summarize them as the hyper-
vascular and delayed enhancement patterns.
17
18 T. Moser and T.S. Nogueira

In their observations, Fuksbrumer et al.
(2000) depicted large solitary lesions dem-
onstrating early contrast uptake and het-
erogeneous enhancement with areas of
differential uptake and washout. These
findings suggestive of a hypervascu-
lar tumor are also consistent with most
descriptions of pleural LFTs. In the series
of the Armed Force Institute of Pathology,
heterogeneous contrast enhancement was
observed in 62% of cases (Rosado-de-
Christenson et al., 2003). Conversely, we
recently reported the case of a LFT of the
liver presenting with delayed enhancement
pattern (Moser et al., 2005). After intrave-
nous contrast administration, a progressive
enhancement began during arterial and
portal phases and became marked on both
CT and MRI delayed images. This large
tumor also showed multiple areas of cystic
degeneration (Figure 2.1).
These two separate imaging patterns
are well correlated with histological find-
ings. Localized fibrous tumors are com-
posed of spindle cells intermixed with
collagen bundles. Most commonly, tumor
cells are disposed haphazardly in an
arrangement known as the storiform pat-
tern. Alternating hyper- and hypocellular
areas are present in the same tumor.
Prominent vessels are generally associated
with the cellular proliferation, determining
the hemangiopericytoma-like vascular pat-
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
in all nine cases reported by Moran et al.
(1998) as well as in those cases reported by
Fuksbrumer et al. (2000). In fact, it repre-
sents the most common microscopic pattern
described so far and corresponds to the most
common hypervascular pattern of contrast
uptake where heterogeneous enhancement
seems to result from the varying cellularity
observed throughout the tumor. However,
the hemangiopericytoma-like vascular pat-
tern was absent in our study. Rather, it was
a poorly vascularized tumor with an inter-
stitium almost exclusively composed of
collagen fibers. Therefore, its enhancement
dynamic at CT and MR was consistent with
a collagen rich interstitium where contrast
gradually diffuses between the fibers.
Most hepatic LFTs are benign lesions.
However, a malignant variant with a more
aggressive behavior such as recurrence
and distant metastasis has been described.
Malignancy can be suspected solely in
the presence of numerous cellular atypias
on histology, or with the identification of
distant metastases (Chan, 1997). Thus,
from the imaging viewpoint, LFTs should
be included in the differential diagnosis
of any extensive and well limited hepatic
mass. A delayed enhancement pattern has
been demonstrated for a wide range of
fibrous containing hepatic lesions, such
as fibrolamellar hepatocellular carcinoma,
cholangiocarcinoma, or sclerosing heman-
gioma. Localized fibrous tumor should
also be included in the differential diag-
19
nosis of other rare mesenchymal hepatic
tumors, particularly fibrous tissue contain-
ing sarcomas such as fibrosarcoma and
malignant fibrous histiocytoma. Hepatic
fibrosarcoma also manifests as a large
mass and possible hypoglycemia. It con-
tains an abundant collagenous stroma with
spindle cells, which are in this case char-
acteristically arranged in a herringbone
pattern. Moreover, immunological labe-
ling for CD34 is negative.
To summarize, imaging can only sug-
gest the diagnosis of LFT in rare cases,
but thorough analysis of vascularity often
brings clues about histological content.
REFERENCES
Chan, J.K. 1997. Solitary fibrous tumour–every-
where, and a diagnosis in vogue. Histopathology
31:568–576.
Fuksbrumer, M.S., Klimstra, D., and Panicek, D.M.
2000. Solitary fibrous tumor of the liver: imaging
findings. Am. J. Roentgenol. 175:1683–1687.
Moran, C.A., Ishak, K.G., and Goodman, Z.D.
1998. Solitary fibrous tumor of the liver: a clin-
icopathologic and immunohistochemical study
of nine cases. Ann. Diagn. Pathol. 2:19–24.
Moser, T., Nogueira, T.S., Neuville, A., Riehm, S.,
Averous, G., Weber, J.C., and Veillon, F. 2005.
Delayed enhancement pattern in a localized
fibrous tumor of the liver. Am. J. Roentgenol.
184:1578–1580.
Rosado-de-Christenson, M.L., Abbott, G.F.,
McAdams, H.P., Franks, T.J., and Galvin, J.R.
2003. From the archives of the AFIP: local-
ized fibrous tumor of the pleura. Radiographics
23:759–783.
3
A Radial Magnetic Resonance Imaging
Method for Imaging Abdominal
Neoplasms
Maria I. Altbach
INTRODUCTION
Abdominal magnetic resonance imaging
(MRI) has progressed rapidly during the
last 12 years. The development of new
hardware and software has allowed the
implementation of faster and better pulse
sequences, and MRI is now used routinely
in the clinic as a complement to, or as
the method of choice over other imaging
techniques such as ultrasound (US) and
computed tomography (CT). Magnetic
resonance imaging combines good spatial
resolution, lack of radiation, non-inva-
siveness, and three-dimensional imaging
capability. This method provides useful
information for the detection and charac-
terization of focal liver lesions that cannot
be diagnosed with US or CT (Robinson,
2000). In particular for small lesions
(< 1 cm2), studies have shown that the
specificity and sensitivity of MRI is better
than US and CT (Ward et al., 2005).
The standard MRI protocol for liver
and most abdominal imaging applications
consists of acquiring T1-weighted, T2-
weighted, and contrast-enhanced images.
All these methods are used for the detec-
tion of lesions but each has a different
role in terms of lesion characterization.
T1-weighted images acquired with lipid and
water in- and out-of-phase, respectively,
are used primarily for the characteriza-
tion of neoplasms such as adenomas,
focal nodular hyperplasia, regenerative
nodules, well-differentiated hepatocellular
carcinoma, nodular regenerative hyperpla-
sia, renal clear cell carcinoma, and angi-
omyolipoma (Ichikawa and Araki, 1999;
Outwater et al., 1998).
Metastatic lesions, however, cannot be
distinguished from the most common
benign lesions (cysts and hemangiomas)
with T1-weighted methods because the
two types of lesions have the same sig-
nal characteristics (i.e., hypointense signal
relative to liver). Differentiation of metas-
tasis, and other primary malignancies,
from hemangiomas and cysts is based on
T2-weighted images (Ito et al., 1997).
Because this is a crucial aspect of the
diagnosis, T2-weighted imaging is a very
important part of the abdominal MRI
examination. Contrast-enhanced MRI is
also important for characterization of liver
lesions. It has been shown that the pattern
of enhancement after the administration
of a gadolinium-based contrast agent is
different for malignancies, hemangiomas,
and cysts, and that the combined use of
21
22
T2-weighted and contrast-enhancement
data increases the sensitivity and spe-
cificity for detecting and characterizing
lesions compared to each technique alone
(Pawluk et al., 1999).
TECHNIQUES FOR
T2-WEIGHTED IMAGING
Initially, T2-weighted images of the abdo-
men were acquired with two-dimensional
Fourier transform (2DFT) spin-echo
(SE) methods (McFarland et al., 1994).
Conventional T2-weighted SE imaging
suffers from long acquisition times (5–
10 min) and artifacts caused by respiratory
motion, cardiac motion, pulsatile flow, and
peristalsis. Thus, the technqiue has been
replaced by fast spin-echo (FSE) methods.
Multi-shot 2DFT-FSE techniques are used
for data acquisition during quiet breath-
ing (using respiratory trigerring) or during
periods of suspended respiration. Both res-
piratory-trigerred and breath-hold 2DFT-
FSE methods yield images with good
spatial resolution and signal-to-noise ratio
(SNR) and the method was shown to be an
improvement over SE acqusitions (Keogan
et al., 1996; Gaa et al., 1996). Despite
this improvement, however, the quality of
images obtained with multi-shot 2DFT-
FSE is still affected by ghosting and blur-
ring caused by pulsatile flow and residual
breathing motion caused by improper res-
piratory trigerring or breath holding.
As an altenative to multi-shot 2DFT
FSE, single-shot fast spin-echo (SSFSE)
is a faster imaging technique that has
become very popular in abdominal imag-
ing (Tang et al., 1997). With SSFSE, the
acquisition of data for one section of the
liver takes < 1 s providing motion-free
M.I. Altbach
data (Semelka et al., 1996). The speed of
the method, however, is at the expense of
image bluriness and loss of small object
detectability in soft tissue (Constable
and Gore, 1992), which primarily affects
the detection of small malignant tumors
(Coates et al., 1998). Other fast imaging
techniques such as single-shot echo pla-
nar (SSEPI) and steady-state free pre-
cession (SSFP) have also been proposed
for obtaining motionless images of the
abdomen. Single-shot echo planar offers
a better lesion-to-liver contrast-to-noise
ratio (CNR) than SSFSE (Namimoto
et al., 1997), but the technique is limited
in spatial resolution and is affected by
artifacts caused by changes in magnetic
susceptibility. Steady-state free preces-
sion offers good spatial resolution but the
CNR for malignant lesions is poor (Jung
et al., 2004; Herborn et al., 2003). Images
acquire with SSFP are also affected by
artifacts caused by magnetic field inho-
mogeneities (Bangerter et al., 2004).
An inherent problem of all T2-weighted
imaging methods is that the hyperintense
signal from flow can mask or mimic the
signal of lesions.
LESION CHARACTERIZATION
WITH T2-WEIGHTED
IMAGING
Since the early developments of abdomi-
nal MRI it was found that malignant
and benign lesions have different signal
characteristics in T2-weighted images.
Numerous studies have been undertaken
to evaluate the best method to characterize
liver neoplasms with T2-weighted imaging
(Reimer et al., 1996; Yu et al., 1998). The
performance of the various T2-weighted
3. A Radial Magnetic Resonance Imaging Method for Imaging Abdominal Neoplasms
techniques and imaging parameters was
investigated using qualitative approaches
(i.e., observers visually comparing signal
intensities between a moderate and heavily
T2-weighted images) as well as quantita-
tive approaches (i.e., observers measuring
signal intensity and/or CNR of lesions
relative to liver). Although reports indi-
cated that the quantitative measurements
yielded a better characterization of liver
lesions, these measurements require
significant operator interaction and are
not practical for clinical use (Ito et al.,
1997). Consequently, in current clinic
practice, lesion characerization based on
T2-weighted imaging is performed by
visually comparing the signal intensity
of lesions between moderate and heavily
T2-weighted images.
Several studies have demonstrated that
a better quantitative approach for liver
lesion charaterization is to directly meas-
ure the T2 value by acquiring data at two
or more echo time (TE) points and fitting
the signal intensities to a single exponen-
tial decay. Initially, T2 measurements in
the upper abdomen were mainly obtained
with dual-echo 2DFT-SE methods with
the drawback of long acquisition times
(10–20 min) and errors due to motion
(i.e., signal displacement along the phase-
encoding direction and motion-induced
volume averaging). Thus, although one of
these studies reported a 97% accuracy for
discriminating malignancies from heman-
giomas (McFarland et al., 1994), the others
show significant overlap between these two
type of lesions (Ohtomo et al., 1988). To
reduce the effect of motion in T2 measure-
ments, Goldberg et al. (1993) used SSEPI
to measure T2 in 46 liver neoplasms and
show that hemangiomas and cysts can
be discriminated from malignancies with
23
100% accuracy. The SSEPI T2 measure-
ment used data at five TE time points to
improve the accuracy of the T2 estimate,
but required the acquistion of data over
several breath hold periods. This made
the experiment both time-consuming and
dependent on the ability of the subject
to perform reproducible breath holds in
order to avoid errors due to misregistra-
tion between TE data sets. The low spatial
resolution of SSEPI and its sensitivity to
susceptibility artifacts limit the technique
even further.
Abe et al. (2000) used SSFSE to cal-
culate T2 at two TE time points within a
breath hold, but showed significant overlap
between malignancies and benign lesions.
Olcott et al. (1999) measured T2 values in
the liver with a pulse sequence designed
to acquire data from four TE time points
with reduced errors due to radiofrequency
pulse imperfections. Although they dem-
onstrated a good separation between
malignant and benign lesions for a set
of 24 lesions, their method is also time-
consuming. Two recent studies based dual-
echo respiratory-trigerred 2DFT-FSE, a
technique that is also sensitive to motion,
reported that the accuracy of separating
malignancies from hemangiomas based
on T2 is 92–94% (Cieszanowski et al.,
2002; Kim et al., 2005). In summary,
the current methods for measuring T2
values in the body have at least one of
the following problems: long imaging
times, motion-induced errors, misregistra-
tion of images acquired in different breath
holds, low spatial resolution, and/or low
number of measured points on the relaxa-
tion curve. These technical difficulties not
only reduce the accuracy of these methods
but also make them impractical for routine
clinical use.
24
RADIAL MAGNETIC
RESONANCE IMAGING
METHODS: AN
ALTERNATIVE FOR
REDUCING MOTION
ARTIFACTS AND IMPROVING
IMAGE QUALITY
Motion-dependent phase shifts accumu-
lated within the time domain data (i.e., the
Fourier or k-space data) cause artifacts in
MR images. If k-space data are sampled in
a rectilinear grid, as in the case of 2DFT-
MRI, motion artifacts are manifested as
blurring and replicated ghost patterns along
one direction in the image, i.e., the direction
of the phase encoding imaging gradient.
If k-space data are sampled on a polar grid,
as in radial MRI, motion artifacts are reduced
due to two fundamental reasons (Glover
and Pauly, 1991). First, motion artifacts
are spread in two dimensions making the
intensity of the artifact less severe. Second,
in polar data, there is a higher density of
sampled points in the center of k-space
(in Cartesian data the sampling density is
uniform) which averages data and reduces
phase errors due to motion. The insensitiv-
ity of radial MRI to motion has been widely
demonstrated. Glover and Pauly (1991)
showed that radial SE methods produce
abdominal images with less motion-induced
artifacts than 2DFT-SE methods. Rasche et
al. (1994) showed that radial FSE images of
the abdomen acquired within a breath hold
have less ghosting and blurring than images
acquired with breath-hold 2DFT-FSE meth-
ods. The robustness of radial methods to
motion have also been demonstrated for
lung (Kauczor and Kreitner, 1999), cardiac
(Schaeffter et al., 2001), and diffusion-
weighted MRI (Trouard et al., 1999).
M.I. Altbach
Recently, a multi-shot radial FSE tech-
nique was developed for T2-weighted
imaging of the liver (Altbach et al., 2002).
The technique yields motion-insensitive
abdominal images with high spatial reso-
lution. One of the highligts of the tech-
nique is that a small amount of diffusion
weighting can be incorporated to sup-
press the bright signal from blood vessels
improving the visualization of lesions.
Examples of images acquired with the
radial FSE and the conventional 2DFT-
FSE method are shown in Figure 3.1. The
images in Figure 3.1a were acquired dur-
ing free breathing using respiratory-trig-
gered 2DFT-FSE and radial FSE methods.
Note that the two small metastatic lesions
seen in the radial FSE image (arrows) are
masked in the 2DFT-FSE image by flow
and motion artifacts. The images in Figure
3.1b were acquired with 2DFT-FSE and
radial FSE in a breath hold. Again in this
example, flow and motion artifacts impair
the visualization of lesions in the 2DFT-
FSE image. Lesions are better depicted in
the radial FSE image (arrow heads). The
importance of flow suppression is quite
evident in this example, where the small
metastatic lesion in the lower right portion
of the liver is only distinguishable in the
radial FSE image due to improved flow
suppression.
Examples of images acquired with radial
FSE and the conventional SSFSE method
in a breath hold are shown in Figure 3.2.
The two small metastatic lesions indicated
by the arrows in Figure 3.2a are only seen
in the radial FSE image but not seen in the
SSFSE image due to the intrinsic bluriness
of the latter. The effect of blurriness in
SSFSE is also seen in the images shown
in Figure 3.2b where lesions are more
conspicuous in the radial FSE image than
3. A Radial Magnetic Resonance Imaging Method for Imaging Abdominal Neoplasms 25

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
the advantage of flow suppression via dif-
fusion weighting in radial FSE is shown in
Figure 3.2c for a patient with very small
metastatic lesions (as reference the lesion
highlighted by the arrow is 5 mm in diam-
eter). In the SSFSE image the signal from
lesions cannot be distinguished from the
signal from blood vessels. This is not the
case in the radial FSE image where the
signal from flow is suppressed leaving the
bright signal from lesions present.
These qualitative findings were cor-
roborated by a statistical evaluation of
image quality for a set of 16 patients with
various types of liver neoplasms (Altbach
et al., 2002). In the study, images were
acquired with radial FSE and the conven-
tional 2DFT-FSE and SSFSE methods.
The results of the study indicated that
radial FSE performed significantly better
in terms of organ and lesion definition
than SSFSE or 2DFT-FSE, In particular,
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
malignancies (22 metastases, 1 hepatoma,
1 hepatocellular carcinoma) and 19 benign
lesions (6 hemangiomas and 12 cysts) cor-
responding to 28 patients showed that the
T2 value determined from radial FSE data
alone can discrimate hemangiomas and
cysts from malignancies (Figure 3.5).
An important parameter in the determina-
tion of T2, based on the radial method pre-
sented above, is the radius at which all TE
data are included. This radius determines
the accuracy of T2 estimates; the larger
the radius the more accurate the estimates,
in particular for small lesions. This radius
depends on the total number of radial lines
collected, and recently it was demon-
strated that for a full data set of 512 radial
lines the error in T2 estimates for lesions
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
0.8–1.0 cm the error is less than 4% (Altbach
et al., 2005). It should be pointed out that
if the collection of sufficient radial lines is
by the breath hold period, techniques were
only half of the radial lines collected and
the other half is calculated and (Barger
et al., 2002; Bydder et al., 2002) can be
used to render the data needed for accurate
T2 estimation.
As any other technique that uses a single
exponential fit to estimate T2, the method
30 M.I. Altbach

methods used for imaging the upper abdo-

men. The radial FSE method does not
suffer from the characteristic bluriness or
lack of spatial resolution of single-shot
methods. The radial FSE method does
not have the motion artifacts of multi-shot
2DFT-FSE methods. The robustness of
the technique to motion allows the use
of a small amount of diffusion weighting
to suppress the signal from flow, which
improves the visualization of lesions.
A major advantage of radial FSE is that
T2-weighted images as well as a T2 map
can be obtained from a single k-space data
set. The T2 maps obtained from radial
FSE data have high-spatial resolution as
well as adequate TE time resolution. T2
maps are generated from data acquired
in a single breath hold, thus errors due
Figure 3.5. T2 values for malignant and the most to motion-induced volume averaging
common benign liver lesions (i.e., hemangiomas and/or misregistration are minimized.
and cysts) obtained from radial FSE data using the The method is fast and can be implemented
algorithm presented in Figure 3.3. The T2 axis is as part of a regular clinical MR examina-
displayed on a logarithmic scale.
tion. This novel methodolgy may proved
to be very valuable for the detection and
presented here does not take into account characterization of liver neoplasms.
partial volume effects (i.e., when the signal
from liver is within the volume occupied
by the lesion). Partial volume may result Acknowledgement. This work was supported
in the misclassification of small hemangi- by NIH grant: CA099074.
omas and cysts as malignancies. However,
since the radial FSE method yields 8–16 REFERENCES
points along the T2 curve, it is possible to
use a double-exponential fit to calculate Abe, Y., Yamashita, Y., Tang, Y., Namimoto T., and
Takahashi, M. 2000. Calculation of T2 relaxa-
the correct T2 value when more than one tion time from ultrafast single shot sequences
type of tissue is present within a voxel. for differentiation of liver tumors: comparison of
echo-planar, HASTE, and spin-echo sequences.
Radiat. Med. 18:7–14.
Altbach, M.I., Outwater, E.K., Trouard, T.P.,
CONCLUSIONS Krupinski, E.A., Theilmann, R.J., Stopeck, A.T.,
Kono, M., and Gmitro, A.F. 2002. Radial fast
In summary, the radial FSE method over- spin-echo method for T2-weighted imaging
comes most of the major problems encoun- and T2 mapping of the liver. J. Magn. Reson.
tered with conventional T2-weighted MRI Imaging 16:179–189.
3. A Radial Magnetic Resonance Imaging Method for Imaging Abdominal Neoplasms
Altbach, M.I., Bilgin, A., Li, Z., Clarkson, E.W.,
and Gmitro, A.F. 2005. Processing of radial fast-
spin echo data for obtaining T2 estimates from
a single K-space data set. Magn. Reson. Med.
54:549–559.
Bangerter, N.K., Hargreaves, B.A., Vasanawala,
S.S., Pauly, J.M., Gold, G.E., and Nishimura,
D.G. 2004. Analysis of multiple-acquisition
SSFP. Magn. Reson. Med. 51:1038–1047.
Barger, A.V., Block, W.F., Toropov, Y., Grist, T.M.,
and Mistretta, C.A. 2002. Time resolved con-
trast-enhanced imaging with isotropic resolu-
tion and broad coverage using an undersampled
3D projection trajectory. Magn. Reson. Med.
48:297–305.
Bydder, M., Larkman, D.J., and Hajnal, J.V. 2002.
Generalized SMASH imaging. Magn. Reson.
Med. 47:160–170.
Cieszanowski, A., Szeszkowski, W., Golebiowski,
M., Bielecki, D.K., Grodzicki, M., and
Pruszynski, B. 2002. Discrimination of benign
from malignant hepatic lesions based on their
T2-relaxation times calculated from moderately
T2-weighted turbo SE sequence. Eur. Radiol.
12:2273–2279.
Coates, G.G., Borrello, J.A., McFarland, E.G.,
Mirowitz, S.A., and Brown, J.J. 1998. Hepatic
T2-weighted MRI: a prospective comparison of
sequences, including breath-hold, half-Fourier
turbo spin echo (HASTE). J. Magn. Reson.
Imaging 8:642–649.
Constable, R.T., and Gore, J.C. 1992. The loss of
small objects in variable TE imaging: implica-
tions for FSE, RARE, and EPI. Magn. Reson.
Med. 28:9–24.
Gaa, J., Hatabu, H., Jenkins, R.L., Finn J.P., and
Edelman, R.R. 1996. Liver masses: replace-
ment of conventional T2-weighted spin-echo
MR imaging with breath-hold MR imaging.
Radiology 200:459–464.
Glover, G.H., and Pauly, J.M. 1992. Projection
reconstruction techniques for reduction of
motion effects in MRI. Magn. Reson. Med.
28:275–289.
Goldberg, M.A., Hahn, P.F., Saini, S., Cohen,
M.S., Reimer, P., Brady, T.J., and Mueller, P.R.
1993. Value of T1 and T2 relaxation times from
echoplanar MR imaging in the characterization
of focal hepatic lesions. Am. J. Roentgenol. 160:
1011–1017.
31
Herborn, C.U., Vogt, F., Lauenstein, T.C., Goyen, M.,
Debatin, J.F., and Ruehm, S.G. 2003. MRI
of the liver: can True FISP replace HASTE?
J. Magn. Reson. Imaging 17:190–196.
Ichikawa, T., and Araki, T. 1999. Fast magnetic
resonance imaging of liver. Eur. J. Radiol.
29:186–210.
Ito, K., Mitchell, D.G., Outwater, E.K., Szklaruk, J.,
and Sadek, A.G. 1997. Hepatic lesions: discrimi-
nation of nonsolid, benign lesions from solid,
malignant lesions with heavily T2-weighted
fast spin-echo MR imaging. Radiology 204:
729–737.
Jung, B., Krombach, G.A., Gunther, R.W., and
Buecker, A. 2004. Is postcontrast trueFISP imag-
ing advantageous? Invest. Radiol. 39:517–523.
Kauczor, H.U., and Kreitner, K.F. 1999. MRI
of the pulmonary parenchyma. Eur. Radiol.
9:1755–1764.
Keogan, M.T., Spritzer, C.E., Paulson, E.K., Paine,
S.S., Harris, L., Dahlke, J.L., and MacFall, J.R.
1996. Liver MR imaging: comparison of respira-
tory triggered fast spin echo with T2-weighted
spin-echo and inversion recovery. Abdom.
Imaging 21:433–439.
Kim, Y.H., Saini, S., Blake, M.A., Harisinghani, M.,
Chiou, Y.Y., Lee, W.J., Yu, J.S., and Hahn, P.F.
2005. Distinguishing hepatic metastases from
hemangiomas: qualitative and quantitative
diagnostic performance through dual echo res-
piratory-triggered fast spin echo magnetic reso-
nance imaging. J. Comput. Assist. Tomogr. 29:
571–579.
McFarland, E.G., Mayo-Smith, W.W., Saini, S.,
Hahn, P.F., Goldberg, M.A., and Lee, M.J.
1994. Hepatic hemangiomas and malignant
tumors: improved differentiation with heavily
T2-weighted conventional spin-echo MR imag-
ing. Radiology 193:43–47.
Namimoto, T., Yamashita, Y., Sumi, S., Tang, Y.,
and Takahashi, M. 1997. Focal liver masses:
characterization with diffusion-weighted echo-
planar MR imaging. Radiology 204:739–744.
Ohtomo, K., Itai, Y., Yoshikawa, K., Kokubo, T.,
and Iio M. 1988. Hepatocellular carcinoma and
cavernous hemangioma: differentiation with MR
imaging. Efficacy of T2 values at 0.35 and 1.5 T.
Radiology 168:621–623.
Olcott, E.W., Li, K.C., Wright, G.A., Pattarelli,
P.P., Katz, D.S., Ch’en, I.Y., and Daniel, B.L.
32
1999. Differentiation of hepatic malignancies
from hemangiomas and cysts by T2 relaxation
times: early experience with multiply refocused
four-echo imaging at 1.5 T. J. Magn. Reson.
Imaging 9:81–86.
Outwater, E.K., Blasbalg, R., Siegelman, E.S., and
Vala, M. 1998. Detection of lipid in abdominal
tissues with opposed-phase gradient-echo images
at 1.5 T: techniques and diagnostic importance.
Radiographics 18:1465–1480.
Pawluk, R.S., Tummala, S., Brown, J.J., and
Borrello, J.A. 1999. A retrospective analysis
of the accuracy of T2-weighted images and
dynamic gadolinium-enhanced sequences in the
detection and characterization of focal hepatic
lesions. J. Magn. Reson. Imaging 9:266–273.
Rasche, V., Holz, D., and Schepper, W. 1994.
Radial turbo spin echo imaging. Magn. Reson.
Med. 32:629–638.
Reimer, P., Rummeny, E.J., Wissing, M., Bongartz,
G.M., Schuierer, G., and Peters, P.E. 1996.
Hepatic MR imaging: comparison of RARE
derived sequences with conventional sequences
for detection and characterization of focal liver
lesions. Abdom. Imaging 21:427–432.
Robinson, P.J. 2000. Imaging liver metastases:
current limitations and future prospects, Br. J.
Radiol. 73:234–241.
Schaeffter, T., Weiss, S., Eggers, H., and Rasche, V.
2001. Projection reconstruction balanced fast
field echo for interactive real-time cardiac imag-
ing. Magn. Reson. Med. 46:1238–1241.
M.I. Altbach
Semelka, R.C., Kelekis, N.L., Thomasson, D.,
Brown, M.A., and Laub, G.A. 1996. HASTE
MR imaging: description of technique and pre-
liminary results in the abdomen. J. Magn. Reson.
Imaging 6:698–699.
Song, H.K., and Dougherty, L. 2000. K-space
weighted image contrast (KWIC) for contrast
manipulation in projection reconstruction MRI.
Magn. Reson. Med. 44:825–832.
Tang, Y., Yamashita, Y., Namimoto, T., Abe, Y.,
and Takahashi, M. 1997. Liver T2-weighted
MR imaging: comparison of fast and conven-
tional half-Fourier single-shot turbo spin-echo,
breath-hold turbo spin-echo, and respiratory-
triggered turbo spin-echo sequences. Radiology
203:766–772.
Trouard, T.P., Theilmann, R.J., Altbach, M.I., and
Gmitro, A.F. 1999. High-resolution diffusion
imaging with DIFRAD-FSE (diffusion-weighted
radial acquisition of data with fast spin-echo)
MRI. Magn. Reson. Med. 42:11–18.
Ward, J., Robinson, P.J., Guthrie, J.A., Downing, S.,
Wilson, D., Lodge, J.P., Prasad, K.R., Toogood,
G.J., and Wyatt, J.I. 2005. Liver metastases in
candidates for hepatic resection: comparison of
helical CT and gadolinium- and SPIO-enhanced
MR imaging. Radiology 237:170–180.
Yu, J.S., Kim, K.W., Kim, Y.H., Jeong, E.K.,
and Chien, D. 1998. Comparison of multishot
turbo spin echo and HASTE sequences for
T2-weighted MRI of liver lesions. J. Magn.
Reson. Imaging 8:1079–1084.
4
Liver: Helical Computed Tomography
and Magnetic Resonance Imaging
Yuji Baba, Yasuyuki Yamashita, Kazuo Awai, and Koichi Kawanaka
INTRODUCTION
Both computed tomography (CT) and mag-
netic resonance (MR) imaging have made
major contributions to the evaluation of
liver morphology, blood flow and detection
and characterization of liver neoplasms.
Studies in numerous institutions have
established that both CT and MR imag-
ing are highly accurate in the diagnosis of
hepatic pathology.
In recent years, there has been a major
revolution in these imaging modalities,
i.e., multi-detector computed tomography
and fast MR imaging. These advanced
imaging technologies as applied in liver
imaging are discussed in this section.
DYNAMIC COMPUTED
TOMOGRAPHY USING
MULTI-DETECTOR
COMPUTED TOMOGRAPHY
Ultrafast imaging has been made possi-
ble by the development of multi-detector
computed tomography (MDCT). There
are many advantages of multi-detector
computed tomography for the evaluation
of the liver disease. First, because imaging
with MDCT is so rapid, the liver can
be imaged as a volume data during
a single breath hold. Second, both the
hepatic arterial-dominant and portal
venous-dominant phases of enhancement
can be visualized. Numerous studies have
demonstrated improved detection of hyper-
vascular masses when such dual-phase
hepatic CT is used (Baron et al., 1996;
Bonaldi et al.,1995).
In our current practice, multiple phases
of the whole liver are imaged, including
the precontrast, arterial, portal venous and
equilibrium phase, or some of these phases
are selected for scanning. With the recent
development of MDCT technologies, we
can obtain double arterial phases (early
arterial phases and late arterial phases) in
a single breath-hold. Hepatic scanning can
be completed during the interval of peak
hepatic contrast enhancement. However,
because the liver can generally be imaged
within a single breath-hold with MDCT,
respiratory misregistration between adja-
cent scan slices is eliminated.
In addition, partial volume averaging
of small lesions with the normal liver is
reduced. This may improve the accuracy
of the evaluation of small focal lesions.
Isotropic volume data acquisition enables
33
34 Y. Baba et al.

multiplanar reconstruction of the image. liver contrast for hypervascular primary

With single-detector helical CT, the entire
liver can be reliably imaged within 20-s.
With MDCT, on the other hand, a true
hepatic arterial phase occurs in the 10-s
time frame after the arrival of contrast in
the abdominal aorta. A double-pass hepatic
scan is performed in a single breath-hold.
The first pass is a pure arterial phase and
the second pass a late arterial or portal
venous inflow phase. Maximum tumor-to-
and metastatic neoplasms occurs with the
second pass. The first pass provides opti-
mum data for hepatic CT arteriography
in pretransplant and preresection patients.
Multiplanar imaging is useful to evaluate
the relationship of hepatic tumors to major
vascular structures. The porta hepatis is
anatomically complex, and multiplanar
display of reconstructed images may also
be useful in this region (Figure 4.1).

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

TUMORS ARISING IN LIVER Many primary hepatic neoplasms and

CIRRHOSIS
In patients with liver cirrhosis, dysplastic
nodules and hepatocellular carcinomas
frequently develop. The diagnosis of hepa-
tocellular carcinoma is extremely impor-
tant in this subset of patients. The majority
of hepatocellular carcinomas are hyper-
vascularbecause of the prominent blood
supply from the hepatic artery, and MDCT
using a bolus injection of contrast material
has played an important role in the detec-
tion of hypervascular hepatocellular carci-
nomas (Hosoki et al., 1982; Bluemke et al.,
1994). These lesions are typically rounded
or oval in configuration and exhibit a mass
effect on adjacent portal or hepatic veins in
the portal venous phase.
hypervascular metastases can become
isoattenuating during the portal venous
phase. Because of their hypervascularity,
such neoplasms are better imaged during
the arterial-phase of hepatic enhancement.
Multi-detector computed tomography pro-
vides the important potential to image
the entire liver in the multiple phases of
enhancement. For visualization of hyper-
vascular hepatocellular carcinomas, a rapid
injection rate is preferred to maximize the
arterial-phase enhancement with appropri-
ateadjustments made to the scan timing.
Tumors are usually more conspicuous in
the arterial phase and equilibrium phase.
Two other arterial-phase patterns are
typically associated with hepatocellular
carcinoma. One is a mosaic pattern with

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
alternating areas of hyperintense enhance-
ment and areas of necrosis (Figure 4.2).
These lesions may demonstrate a peripher-
ally enhancing pseudocapsule, often best
visualized during the equilibrium phase.
Necrotic lesions, regardless of their histology,
will demonstrate poor arterial-phase enhan-
cement. They are best evaluated with portal
venous-phase or equilibrium images.
When we should start scanning? In our
institute, the start of arterial-phase scan is
determined by monitoring the density of
aorta. The peak enhancement time of aorta
varies from patient to patient, and cannot
be estimated from body weight or other
physical data.
TUMOR ARISING IN
NONCIRRHOTIC LIVER OR
IN ONCOLOGIC PATIENTS
Most metastatic tumors, having a relatively
poor vascular supply, are not well detected
during arterial-phase. Such tumors, termed
hypovascular tumors, are among the most
common tumors metastatic to the liver,
including adenocarcinomas from primary
tumors such as colon, pancreas, and gastric
carcinomas. A uniformly enhancing rim on
the periphery of a lesion strongly suggests
metastatic disease. This finding is also
present with hepatic abscesses, although
the clinical presentation of the two enti-
ties will be quite different. Hypervascular
tumors are seen best during the hepatic
arterial phase of enhancement as hyperat-
tenuating masses compared with the sur-
rounding liver (Hollett et al., 1995).
For most hypovascular tumors, the dif-
ference in attenuation between the liver
and the tumor is maximized during the
portal phase of hepatic enhancement.
Y. Baba et al.
A subset of so-called hypervascular tumors
in the liver behaves as though the arterial
blood supply of the tumors was greater
than that of the liver.
Computed tomography imaging dur-
ing the equilibrium phase of enhance-
ment results in decreased tumor detection.
Contrast from the interstitial spaces of the
liver diffuses into the liver tumor, which
is enhanced (Figure 4.3). The combina-
tion of declining hepatic attenuation and
increasing attenuation of the tumor mass
may result in the tumor becoming nearly
isoattenuating to the normal liver and not
detectable.
COMPUTED TOMOGRAPHY
ANGIOGRAPHY
Accurate knowledge of the hepatic arterial
anatomy in a transplant recipient is helpful
to the surgeon before liver transplantation
(Winter et al., 1995). With a thin slice
thickness MDCT data set during a single
breath fold period, CT angiography of the
hepatic artery can be reconstructed, with
promising results (Figure 4.4). Computed
tomography portography can also be
reconstructed with the data set obtained
during the portal venous phase.
Angiography-Assisted Computed
Tomography
Hepatic CT that uses contrast material
administered through an indwelling arte-
rial catheter is the most sensitive technique
for detecting liver lesions (Heiken et al.,
1989; Matsui et al., 1985; Matsui, 1988).
However, because it is an invasive proce-
dure, this technique is not used for routine
hepatic screening, but to evaluate selected
4. Liver: Helical Computed Tomography and Magnetic Resonance Imaging 37

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.,
1994). The reported sensitivity of CTAP
for lesion detection is among the highest
of the available radiologic techniques,
ranging from 81% to 91% (Heiken et al.,
1989; Deflandre et al., 1994; Bluemke
et al., 1995). Rapid scanning during injec-
tion is important because recirculation of
contrast to the systemic circulation results
in diminished contrast differences between
the neoplasm and normal parenchyma.
Recirculation of contrast to the aorta is
evident 35-s following the start of the con-
trast injection, indicating the need for rapid
scanning. Multi detector computed tomog-
raphy is an excellent method for perform-
ing rapid hepatic scanning. This method
results in high levels of liver parenchymal
and vascular enhancement that are readily
maintained over the duration of the scan.
Computed tomography during arterial
portography may be particularly problem-
atic in patients with cirrhosis, not only
because of fibrosis and altered hepatic
architecture but also due to altered hemo-
dynamics.
In patients with portal hypertension, portal
venous blood may be shunted away from the
4. Liver: Helical Computed Tomography and Magnetic Resonance Imaging
liver through portosystemic collateral ves-
sels, limiting hepatic parenchymal enhance-
ment. In such patients, CTHA may provide
more useful information than CTAP.
COMPUTED TOMOGRAPHY
DURING HEPATIC
ANGIOGRAPHY
Computed tomography during hepatic
angiography is performed with the cath-
eter in the hepatic artery [13, 14]. This
technique is superb in analyzing neovascu-
larity of hepatic neoplasms. On CT images
obtained with this technique, malignant
hepatic neoplasms appear as peripherally
or homogeneously enhancing hyperdense
masses compared with the normal hepatic
parenchyma (Irie et al., 1995; Chezmar
et al., 1993), although these tumors may
appear hypoattenuating on conventional
liver CT. In early hepatocellular carcino-
mas developed in a precancerous adenom-
atous hyperplasia, enhancement is seen in
most areas of the mass, indicating portal-
dominant supply of the lesion, consistent
with precancerous atypical adenomatous
hyperplasia on CTAP images. Small foci
with absence of portal blood flow show
arterial-dominant blood supply on CTHA.
MAGNETIC RESONANCE
IMAGING
With the advent of the fast magnetic reso-
nance (MR) imaging technique, this has
become the modality of choice in some
patients for imaging of the liver. An advan-
tage of MR imaging compared with CT is
its excellent tissue contrast and multiplanar
39
imaging capability. The presence of vari-
ous kinds of contrast agents, including
extracellular agents, hepatocyte-specific
agents, and reticuloendothelial system-
specific agents, also has some advantages.
Because of the numerous choices of imag-
ing sequences and parameters, selection
and optimization of pulse sequences is a
critical issue in MR imaging of the liver.
The current standard MR imaging exam-
ination of the liver includes a T1-weighted
sequence, a T2-weighted sequence and a
contrast-enhanced dynamic sequence.
The most frequent contrast administration
approach is the use of a nonspecific extra-
cellular contrast agent, gadolinium chelate,
as a rapid-bolus injection with serial imag-
ing using a gradient echo sequence as
dynamic MR imaging. In several circum-
stances, tissue-specific contrast agents such
as superparamagnetic iron oxide (SPIO)
are used not only for tumor detection but
also for tumor characterization.
DYNAMIC MAGNETIC
RESONANCE IMAGING
With recent advances in fast MR imaging
technology, the entire liver can be imaged
in a breath-hold with a spoiled gradient
echo sequence. The sequence is repeated
several times after administration of gado-
linium chelates (Figure 4.5). It is usually
advisable to use an in-phase echo time on
spoiled gradient echo images for contrast-
enhanced studies in order to avoid confus-
ing variations in signal intensity based on
fatty infiltration of the liver and black-ring
phase cancellation artifacts that will mask
capsular-based disease.
In current clinical practice, gadolinium
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

study as iodinated contrast agents on CT visually similar to iodine contrast enhance-

(Hamm et al., 1994; Yamashita et al.,
1994, 1996). Gadolinium chelates, that
are nonspecific extracellular space agents,
are distributed initially within the intra-
vascular compartment and then diffuse
rapidly throughout the extracellular space
in a distribution similar to that of iodinated
contrast agents on CT. These agents cause
T1 shortening of adjacent water molecules,
which results in an increase in signal inten-
sity on T1-weighted images. Therefore, the
pattern of enhancement with gadolinium is
ment on CT images, and a similar phase of
enhancement of the liver is seen as in CT.
SUPERPARAMAGNETIC
IRON OXIDE
Super paramagnetic iron oxide (e.g., feru-
moxides) is a crystalline compound that is
taken up selectively by the Kupffer cells
of the liver. It markedly shortens the T2
relaxation, resulting in a decrease in signal
4. Liver: Helical Computed Tomography and Magnetic Resonance Imaging 41

from normal liver tissue. Because the agent
is not taken up by tumors, hepatic masses
stand out as hyperintense foci in the low-
signal intensity liver on proton density
or T2/T2*-weighted images (Figure 4.6).
Studies performed at 0.6 and 1.5 T have
demonstrated an increased tumor detec-
tion rate (Tsang et al., 1988; Ros et al.,
1995), and other studies have shown
increased lesion-to-liver contrast (Stark
et al., 1988; Winter et al., 1993) compared
with unenhanced imaging. Although most
people may agree that this agent is effec-
tive in detecting metastatic tumors, its role
in detecting hepatocellular carcinomas is
questionable. The role of SPIO in char-
acterizing focal liver lesions is limited.
Some studies reported SPIO uptake in
lesions containing Kupffer cells, such as
focal nodular hyperplasia, and in lesions
in which blood pooling occurs, such as
hemangioma, which also shows a decrease
in signal intensity on T2-weighted images.
Some well-differentiated hepatocellular
carcinomas and adenomatous hyperplasias
may also show accumulation of this agent
within the tumor. Therefore, in early hepa-
tocellular carcinomas that developed in a
precancerous adenomatous hyperplasia,
a signal decrease is seen in most areas of
atypical adenomatous hyperplasia, indicat-
ing the presence of Kupffer cells in this
area. Small foci of signal hyperintensity
within the lesion may be present, indicat-
ing the absence of Kupffer cells, consistent
with malignant foci developed in atypical
adenomatous hyperplasia. Delineation of
the clinical role of iron-containing hepatic
contrast agents for tumor characterization
requires further testing.
In conclusion, with modern imaging
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.

MR imaging, tumor detection is possible rial portography, and MR imaging. Radiology

in a noninvasive manner. Lesion charac- 171:47–51.
Hollett, M.D., Jeffrey, R.B.J., Nino-Murcia, M.,
terization is also possible in the majority
Jorgensen, M.J., and Harris, D.P. 1995. Dual-
of cases. Three-dimensional evaluation is phase helical CT of the liver: Value of arterial
feasible with fast imaging techniques such phase scans in the detection of small (1.5 cm)
as MDCT. With tissue-specific MR con- malignant hepatic neoplasms. Am. J. Roentgenol.
trast agents such as SPIO, our capability 164:879–884.
is enhanced not only with respect to lesion Hosoki, T., Chatani, M., and Mori, S. 1982. Dynamic
computed tomography of hepatocellular carci-
detection but also lesion characterization.
noma. Am. J. Roentgenol. 139:1099–1106.
We hope that these imaging techniques are Irie, T., Takeshita, K., Wada, Y., Kusano, S.,
properly applied in clinical practice. Terahata, S., Tamai, S., Hatsuse, K., Aoki, H.,
and Sugihara, Y. 1995. CT evaluation of hepatic
REFERENCES tumors: Comparison of CT with arterial portog-
raphy, CT with infusion hepatic arteriography,
Baron, R.L., Oliver, J., Dodd, G., Nalesnik, M., and simultaneous use of both techniques. Am. J.
Holbert, B.L., and Carr, B.V. 1996. Hepatocellular Roentgenol. 164:1407–1412.
carcinoma: Evaluation with biphasic, contrast- Matsui O. 1988. Hepatic metastasis detection:
enhanced, helical CT. Radiology 199:505–511. Comparison of three CT contrast enhancement
Bluemke, D.A., Urban, B., and Fishman, E.K. methods (letter). Radiology 168:283–284.
1994. Spiral CT of the liver: Current applica- Matsui, O., Takashima, T., Kadoya, M., Ida, M.,
tions. Semin. Ultrasound CT MR 15:107–121. Suzuki, M., Kitagawa, K., Kamimura, R., and
Bluemke, D.A., Soyer, P.A., Chan, B.W., Bliss, Inoue, K. 1985. Dynamic computed tomography
D.F., Calhoun, P.S., and Fishman, E.K. 1995. during arterial portography: The most sensitive
Spiral CT during arterial portography: Technique examination for small hepatocellular carcino-
and applications. Radiographics 15:623–637. mas. J. Comput. Assist. Tomogr. 9:19–24.
Bonaldi, V.M., Bret, P.M., Reinhold, C., and Matsui, O., Takahashi, S., Kadoya, M., Suzuki, M.,
Atri, M. 1995. Helical CT of the liver: Value Hirose, J., Kameyama, T., Choto, S., Konishi, H.,
of an early hepatic arterial phase. Radiology Ida, M., and Yamaguchi, A. 1987. Liver metas-
197:357–363. tases from colorectal cancers: Detection with
Chezmar, J.L., Bernardino, M.E., Kaufman, S.H., CT during arterial portography. Radiology 165:
and Nelson, R.C. 1993. Combined CT arterial 65–69.
portography and CT hepatic angiography for Ros, P.R., Freeny, P.C., Harms, S.E., Seltzer, S.E.,
evaluation of the hepatic resection candidate. Davis, P.L., Chan, T.W., Stillman, A.E., Muroff,
Work in progress. Radiology 189:407–410. L.R., Runge, V.M., and Nissenbaum, M.A. 1995.
Deflandre, M.F., Vilgrain, V., Zins, M., Najmark, D., Hepatic MR imaging with ferumoxides: A mul-
Arrive, L., Vullierme, M.P., and Nahum, H. ticenter clinical trial of the safety and efficacy in
1994. Nontumorous attenuation changes on CT the detection of focal hepatic lesions. Radiology
arterial portography. J. Comput. Assist. Tomogr. 196:482–488.
18:761–767. Stark, D.D., Weissleder, R., Elizondo, G., Hahn,
Hamm, B., Thoeni, R.F., Gould, R.G., Bernardino, P.F., Saini, S., Todd, L.E., Wittenberg, J., and
M.E., Luening, M., Saini, S., Mahfouz, A.E., Ferrucci, J.T. 1988. Superparamagnetic iron
Taupits, M., and Wolf, K.J. 1994. Focal liver oxide: Clinical application as a contrast agent
lesions: Characterization with nonenhanced and for MR imaging of the liver. Radiology 168:
dynamic contrast material-enhanced MR imag- 297–301.
ing. Radiology 190:417–423. Tsang, Y.M., Stark, D.D., Chen, M.C., Weissleder,
Heiken, J.P., Weyman, P.J., and Lee, J.K.T. 1989. R., Wittenberg, J., and Ferrucci, J.T. 1988.
Detection of focal hepatic masses: Prospective Hepatic micrometastases in the rat: Ferrite-
evaluation with CT, delayed CT, CT during arte- enhanced MR imaging. Radiology 167:21–24.
4. Liver: Helical Computed Tomography and Magnetic Resonance Imaging
Winter, T.C., Freeny, P.C., Nghiem, H.V., Mack, L.A.,
Patten, R.M., Thomas, C.R., and Elliott, S. 1993.
MR imaging with i.v. superparamagnetic iron
oxide: Efficacy in the detection of focal hepatic
lesions. Am. J. Roentgenol. 161:1191–1198.
Winter, T.C., Freeny, P.C., Nghiem, H.V.,
Hommeyer, S.C., Barr, D., Croghan, A.M.,
Coldwell, D.M., Althans, S.J., and Mack, L.A.
1995. Hepatic arterial anatomy in transplantation
candidates: Evaluation with three-dimensional
CT arteriography. Radiology 195:363–370.
43
Yamashita, Y., Yamamoto, H., Arakawa, A.,
Matsukawa, T., Hatanaka, Y., and Takahashi, M.
1994. Differential diagnosis of liver neoplasms:
Role of dynamic MR imaging. Radiology
193:59–65.
Yamashita, Y., Mitsuzaki, K., Yi, T., Ogata, I.,
Nishiharu, T., Urata, J., and Takahashi, M. 1996.
Small hepatocellular carcinoma in patients with
chronic liver damage: Prospective comparison of
detection with dynamic MR imaging and helical
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
INTRODUCTION
The selection of appropriate patients for
surgical resection of colorectal liver metas-
tases is a central clinical dilemma. Using
conventional imaging, up to one quarter
of patients scheduled for surgical resection
will be determined to have unresectable or
metastatic disease at the time of surgery.
Of the patients who do undergo a resection
with curative intent, ~ 60% will develop a
recurrence within 3 years, either within the
liver or in an extrahepatic location (Bines
et al., 1996). It is clear that a more effective
preoperative staging modality is needed to
avoid unsuccessful or ineffective surgeries.
Positron emission tomography (PET)
using [18F]fluorodeoxyglucose (FDG) is
emerging as a promising diagnostic modal-
ity in metastatic colorectal cancer. [18F]
fluorodeoxyglucose-positron emission
tomography (FDG-PET) has the poten-
tial to differentiate between patients
with resectable and unresectable hepatic
colorectal cancer metastases and therefore
improve patient selection for surgery with
a positive effect on surgical treatment
outcomes, including survival. Positron
emission tomography may improve the
survival time in patients undergoing
hepatic resection for metastatic colorectal
cancer for three reasons. First, if preop-
erative workup with PET scanning detects
more extrahepatic disease than conven-
tional imaging alone, surgery would not be
undertaken in these patients, and the group
of patients undergoing surgery would have
a more favorable prognosis; a concept
similar to the Will-Rogers effect, which is
used to describe stage migration. Second,
if hepatic recurrences are detected ear-
lier in patients screened with PET scan-
ning as compared to conventional imaging
alone, surgery would be performed earlier
and survival, as defined from the time of
metastasectomy, would be longer, a con-
cept termed lead-time bias. Third, if hepatic
recurrences are detected earlier with PET
scan, surgery might be undertaken prior
to the development of additional systemic
micrometastases. Based on tumor biology,
this results in a true improvement in survival.
The benefit of PET scanning is most
likely related to the ability to find more
extrahepatic disease. This review will focus
on the available evidence for the ability of
PET scanning to improve the selection of
patients for hepatic resection, and thereby
improve the prognosis of resected patients.
Topics to be outlined include the sensitivity
49
50
and specificity of PET scans to detect
intrahepatic and extrahepatic lesions. The
correlation of PET scans with the preop-
erative clinical risk score (CRS), and the
impact on surgical management. These
data will be used to argue that the ability
to detect and eliminate patients with unre-
sectable metastatic colorectal cancer may
also translate into improved survival post-
hepatic resection. The benefit of PET dur-
ing follow-up, particularly in patients with
an undiagnosed elevation in carcinoem-
bryonic antigen (CEA) or a suspected
recurrence post-hepatectomy will also be
discussed. Finally, a preoperative staging
algorithm that incorporates the available
information regarding PET scanning will
be presented.
POSITRON EMISSION
TOMOGRAPHY SCANNING
AS A STAGING MODALITY
TO COMPLEMENT
CONVENTIONAL IMAGING
Although there are no published rand-
omized controlled trials on the use of FDG-
PET in patients with metastatic colorectal
cancer to the liver, a number of prospective
and retrospective studies have attempted to
evaluate the ability of PET scan to improve
the staging of patients in the preoperative
workup for hepatic resection of colorectal
metastases. All of these studies included a
cohort of patients who presented for evalu-
ation of suspected colorectal cancer recur-
rence to the liver based on conventional
imaging, such as computed tomography
(CT) or, in some cases, an elevated CEA
level. These studies compare the relative
value of PET scans in staging a patient for
hepatic resection; in the replacement of CT
R. Auer and Y. Fong
scanning (i.e., blinded to the CT result)
or the incremental value of PET scan fol-
lowing CT. These studies have attempted
to define a role for PET scanning in this
specific cohort of patients. A meta-analysis
of these studies has also been published
providing a concise summary of the data
available to date (Wiering et al., 2005).
A more general overview on the use of PET
scanning in screening patients with colon
cancer for local and distant recurrences can
be found in other chapters in this volume.
POSITRON EMISSION
TOMOGRAPHY SCAN
FOR THE DETECTION OF
EXTRAHEPATIC DISEASE
Extrahepatic disease is the most significant
operative finding related to adverse long-
term survival (Rosen et al., 1992). The
existence of extrahepatic disease should
be considered a contraindication for hepa-
tectomy, aside from isolated portal lymph
node metastases (Jaeck et al., 2002) or
limited resectable pulmonary metastases
(DeMatteo et al., 1999).
The ability of PET scans to detect previ-
ously unrecognized extrahepatic disease is
the most compelling reason for use in the
preoperative setting for patients with
hepatic metastases. Numerous studies have
demonstrated the superiority of PET scans
(Wiering et al., 2005) as compared to
conventional CT scans in this setting. The
sensitivity of a PET scan is 91.5% (range
from 63% to 100%) compared to 60.9%
(range 25–86%) (Wiering et al., 2005).
Whereas CT scans have a specificity of
95.4% as compared to 91.9% with PET
(Wiering et al., 2005). [18F]fluorodeoxy-
glucose-positron emission tomography
5. Selection of Patients for Resection of Hepatic Colorectal Metastases 51

(FDG-PET) is particularly helpful in within the liver compared to extrahepatic

nodal involvement, differentiating malig- disease as normal hepatocytes contain a
nant nodes from post-surgical change, and high hexokinase activity and therefore
evaluating malignancy of indeterminate incorporate a significant amount of 18FDG
pulmonary nodules (Truant et al., 2005). (Fong et al., 1999a). Conventional preop-
Evaluation of PET scan alone without a erative imaging to detect hepatic metas-
corresponding CT scan is associated with a tases includes CT and CT-portography.
number of drawbacks, particularly relating In prospective, blinded studies comparing
to the poor resolutions of PET imaging FDG-PET to conventional CT scans in
patients with colon cancer being evaluated
(Selzner et al., 2004). It is difficult to evalu-
ate the exact location of FDG uptake. This for hepatic resection, the detection of liver
may result in misinterpretation of a lesion metastases was equivalent or superior with
as intrahepatic, when it is actually extrahe- conventional CT scans (~ 80%) as com-
patic or vice-versa. In addition, obtaining pared to PET scans (65–80%) (Truant et al.,
histological confirmation of malignancy 2005). When directly compared to CT por-
in an area of increased FDG uptake may tography in a similar population, CT portog-
be impossible without anatomical infor- raphy had a higher or equivalent sensitivity
mation. In one study following blinded (97% versus 90% with PET), but a lower
interpretation of the CT and PET scans, a specificity, resulting in a lower accuracy
rereview of the CT scans was performed with CT portography (76–80%) as com-
for those patients with discordant and pared to PET scan (92–93%) (Fong et al.,
positive FDG-PET results. Of seven PET 1999b). When evaluated in conjunction
positive/CT negative lesions, four could with CT portography, PET scans appear
be detected on rereview, emphasizing the to be sensitive at distinguishing between
importance of combined reading of CT and benign and malignant intrahepatic lesions
PET images in daily practice (Ruers et al., based on the magnitude of the standard-
2002). The combined fusion technique of ized uptake value (SUV) (Figure 5.2). In
PET/CT will best address the shortcomings two separate analyses of the magnitude
of both modalities, however, at present the of SUV in hepatic colorectal metastases
CT resolution in the combined modality is (total of 144 lesions), the SUV for all but
not comparable to that of a standard CT. 1 malignant lesions was > 3.5, and only 2
benign lesion had an SUV > 3.5 (Delbeke
et al., 1997).
POSITRON EMISSION When compared directly to intraopera-
TOMOGRAPHY SCAN tive ultrasound and surgical pathology,
FOR THE DETECTION OF PET scans have a sensitivity of 71–78%
INTRAHEPATIC DISEASE with the probability of lesion detection
directly correlated with lesion size (Rohren
The utility of PET scans to evaluate the et al., 2002). Positron emission tomogra-
number and distribution of lesions in the phy scanning will detect between 25–40%
liver and therefore help select patients with of lesions ≤ 1 cm and 90% of lesions ≥
surgically resected disease has also been 3 cm (Rohren et al., 2002). The assess-
studied (Figure 5.1). Such scanning is gen- ment of lobar disease distribution was also
erally thought to be less sensitive for lesions found to be discordant between PET scans
52
Figure 5.1. A patient with bilobar, multiple hepatic
metastases treated with neoadjuvant chemotherapy
of FOLFOX with good response and shrinking of
all the lesions. There are still residual tumors in the
liver by CT scan
and surgical findings in 13% of patients,
frequently underestimating the extent of
disease (Rohren et al., 2002). The authors
of this study concluded that PET scans
lack the anatomic definition and sensi-
tivity, particularly with small lesions, to
replace surgical exploration when evaluat-
ing resectability in patients with colorectal
liver metastases (Rohren et al., 2002).
The added value of intraoperative ultra-
sound may be limited when patients are
evaluated preoperatively with PET scan.
In a study of 47 patients who underwent
surgical exploration for curative resec-
tion of hepatic colorectal metastases, the
positive predictive value of FDG-PET was
93% as compared to 87% for intraopera-
tive sonography (Rydzewski et al., 2002).
More importantly, the findings of the intra-
operative ultrasound led to a change in the
R. Auer and Y. Fong
Figure 5.2. A patient with bilobar, multiple hepatic
metastases treated with neoadjuvant chemotherapy
of FOLFOX with good response and shrinking of
all the lesions. 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
clinical management of only one patient
(2%) (Rydzewski et al., 2002).
VALUE OF POSITRON
EMISSION TOMOGRAPHY
(PET) CORRELATED TO
PROGNOSTIC CLINICAL RISK
SCORE (CRS)
A number of clinical and pathological
parameters have been found to be inde-
pendent predictors of prognosis by robust
multivariate analysis (Fong et al., 1999a).
5. Selection of Patients for Resection of Hepatic Colorectal Metastases
These parameters include: (1) the presence
of extrahepatic disease at laparotomy, (2)
a positive resection margin, (3) nodal
metastases from primary cancer, (4) a short
disease-free interval, (5) largest tumor
> 5 cm, (6) more than one liver metasta-
sis, and (7) a CEA over 200 ng/ml. Using
the last five criteria, a preoperative CRS
system was created with each positive
criterion counting as one point. This CRS
is a simple, easily remembered staging
system for classifying patients with liver-
exclusive metastatic colorectal cancer.
The presence of any one of these charac-
teristics was still associated with a 5-year
survival of 24–34%, and therefore no single
criterion can be considered a contraindica-
tion to resection (Fong et al., 1999a). It is
the total score out of 5 that is highly predic-
tive of outcome. A score of 2 or less places
a patient in a good prognostic group, for
whom resection is ideal. For scores of 3 or
4, outcome is less favorable and patients
should be considered for aggressive tri-
als of adjuvant therapy. For a score of 5,
long-term disease-free survivors are rarely
encountered and resections in this high risk
group should be accompanied by trials of
adjuvant therapy (Fong et al., 1999a).
This CRS is also now being used to
help select the extent and sophistication
of preoperative assessment, including the
use of FDG-PET. The ability of the CRS
to predict the yield of a preoperative PET
scan was demonstrated in a subset of 63
patients presenting with a first occur-
rence of hepatic colorectal metastases
(Schussler-Fiorenza et al., 2004). Among
patients with a CRS of 0, no patient
had extrahepatic disease detected by PET
and 57% showed a false-positive reading
(Schussler-Fiorenza et al., 2004). Among
patients with a CRS of ≥ 1, 14% had
53
evidence of additional disease detected
by PET (Schussler-Fiorenza et al., 2004).
The authors concluded that patients with a
CRS of 0 should not undergo preoperative
evaluation with a PET scan because the
yield is too low to justify the expense and
morbidity associated with false-positive
results (Schussler-Fiorenza et al., 2004).
POSITIVE IMPACT OF
POSITRON EMISSION
TOMOGRAPHY ON THE
MANAGEMENT OF PATIENTS
WITH LIVER METASTASES
Positron emission tomography (PET)
scans have repeatedly shown the ability to
alter clinical treatment plans when used
for preoperative staging in patients with
hepatic colorectal metastases. In a well
designed prospective study evaluating the
added benefit of CT/PET scan to CT scan
alone, 76 patients with hepatic colorectal
metastases were imaged preoperatively
(Selzner et al., 2004). In this study a pre-
PET treatment plan was prospectively
defined and compared to the treatment
plan outlined after consideration of the
results of the CT/PET scan (Selzner et al.,
2004). In 60 patients (79%) PET/CT did
not change the therapeutic strategy, and
in 10 patients (13%) considered resect-
able by CT alone, the PET/CT demon-
strated extensive extrahepatic disease,
representing a contraindication to surgery
(Selzner et al., 2004). In the remaining
six patients (8%) the surgical strategy
and extent of resection was changed
because the PET/CT demonstrated pre-
viously unrecognized positive nodes in
the hepatoduodenal ligament (Selzner
et al., 2004). Taken together, PET/CT
54 R. Auer and Y. Fong

provided information that significantly OUTCOME OF PATIENTS

altered the surgical treatment plan in SELECTED FOR HEPATIC
> 20% of the patients, including the RESECTION FOLLOWING [18F]
avoidance of morbidity associated with
an exploratory laparotomy/laparoscopy in
FLUORODEOXYGLUCOSE-
> 10% of patients (Selzner et al., 2004). POSITRON EMISSION
These results compare favorably to other TOMOGRAPHY (FDG-PET)
similarly designed studies where clinical
management was directly altered in Survival following hepatic resection for
20–58% of patients (Truant et al., 2005). colorectal liver metastases depends on the
stringency of the patient selection crite-
ria. In order to gain a long-term survival
NEGATIVE IMPACT OF benefit, surgical candidates must have
POSITRON EMISSION hepatic metastases that can be resected
TOMOGRAPHY ON THE with a negative margin and no evidence
MANAGEMENT OF PATIENTS of extrahepatic disease. As PET scanning
WITH LIVER METASTASES appears to be more sensitive at detect-
ing extrahepatic recurrences, this should
Although PET scans clearly have an advan- eliminate patients unlikely to benefit from
tage over conventional CT scans by avoid- surgery and improve the overall survival of
ing unnecessary laparotomy in unresectable patients who do undergo curative resection
patients, the potentially harmful impact of of hepatic colorectal metastases.
PET scans should not be overlooked. The Although there are no studies that
negative impact may result from a false- directly compare the survival results of
negative, or possibly more disconcerting, contemporary patients who underwent
a false-positive. The false-negative rate for resection with or without a preoperative
detection of extrahepatic disease appears to FDG-PET scan, one group published sur-
range from 0% to 17% (Truant et al., 2005), vival results in a cohort of patients who
and was generally explained by tumors < were evaluated with preoperative PET
1 cm in size, proximity of tumors to adja- scans and compared the results to his-
cent lesions, preoperative chemotherapy, or torical controls (Fernandez et al., 2004;
hypocellular mucinous tumors. The sensitiv- Strasberg et al., 2001). In their first pub-
ity of PET scans appears to be particularly lication in 2001, Strasberg et al. (2001)
limited for peritoneal metastases with a sen- found that 43 patients with metastatic
sitivity as low as 63% in one series (Truant colorectal cancer were referred for hepatic
et al., 2005). The false-positive rate ranged resection after conventional tumor stag-
from 0% to 6% (Truant et al., 2005) and was ing with a CT scan. FDG-PET identified
secondary to inflammation, infection or a additional cancer in ten patients, in whom
secondary malignancy, such as lymphoma. surgery was contraindicated in six (14%).
Although the false-positive rate is small, it Laparotomy was performed in 37 patients
is not negligible. In patients in whom this with a curative resection undertaken in all
information would represent a contraindica- but 2 patients (95%), each of whom had a
tion to surgery, a biopsy to confirm meta- single small focus of extrahepatic disease
static disease is usually warranted. (Strasberg et al., 2001). The disease-free
5. Selection of Patients for Resection of Hepatic Colorectal Metastases
survival estimate at 3 years was 40%,
higher than the previously reported results
of 15–28% (Strasberg et al., 2001).
In a follow-up study, 100 patients who
underwent complete hepatic resection for
curative intent were evaluated by preoperative
FDG-PET and CT scans, and were followed
for a median of 31 months (Fernandez et al.,
2004). This study did not include patients
who underwent an exploratory laparotomy
but were found to be unresectable; therefore,
it is not possible to comment on the ability of
PET scans to prevent unnecessary laparoto-
mies from this study. The Kaplan-Meier esti-
mate of 5-year overall survival in this group
of patients was 58.6% (95% CI of 46–72%)
which compares favorably to the results
of 19 previously reported series where the
estimated 5-year overall survival was 30%
(range 12–41%) (Fernandez et al., 2004).
One criticism of this study is that in com-
paring to historical controls, the influences
of improved surgical techniques and new
chemotherapeutic agents which might have
contributed to the improved results are not
considered. The authors acknowledge the
potential for these confounding variables but
state that only 4% of the patients received
irinotecan as adjuvant treatment and that
oxaliplatin was not yet available by the close
of the study (Fernandez et al., 2004).
It is interesting to note that in this study
none of the previously established prog-
nostic factors relating to the hepatic metas-
tasis, including the components of the
CRS, were significant predictors of worse
outcome (Fernandez et al., 2004). This
suggests that PET scanning represents a
surrogate for previously identified clinical
variables that make up the CRS. This
observation may be related to the fact that
PET scans preferentially exclude patients
with multiple, bilateral, and synchronous
tumors because such patients are more
55
likely to have extrahepatic disease detected
by FDG-PET and therefore considered
unresectable (Strasberg et al., 2001). In
the subsequent analysis of survival follow-
ing hepatic resection among this selected
PET negative population, there is loss of
significance for the prognostic indicators,
reflecting the added information obtained
by preoperative PET scanning.
STANDARD UPTAKE
VALUE AND PREDICTING
PROGNOSIS OR RESPONSE
TO THERAPY
A fundamental metabolic difference
between normal and malignant cells
relates to the cells primary energy source.
Whereas a normal cell derives most of its
energy from oxidative phosphorylation, a
rapidly growing tumor cell derives more
than half of its energy from glycolysis,
with only a minor amount from the citric
acid cycle (Nakashima et al., 1984). The
glycolytic intermediates are used not only
to provide energy, but also to intermedi-
ate in the synthesis of DNA/RNA and
membrane lipids. The glycolytic capacity
of tumor cells has been demonstrated in a
number of studies to correlate with lack
of differentiation of these cells (Pauwels
et al., 2000), and when cells undergo
oncogenic transformation, their glycolytic
activity increases. Many human tumors
including colon cancers, have demon-
strated increased expression and activity
of several glucose transporters (GLUT 1
and 3) and the key enzyme hexokinase
(reviewed in Pauwels et al., 2000).
Positron emission tomography scan-
ning has the ability to give functional
information about a tumor and may be
able to predict the prognosis or response
56 R. Auer and Y. Fong

to treatment on this basis. In order to of the pretreatment value) (Findlay et al.,

assess the response to 5-FU based chem-
otherapy, FDG uptake on a PET scan was
measured in colorectal liver metastases
before and at 2 and 4 weeks after treat-
ment and compared to tumor dimension
measured on a CT scan performed at
12 weeks (Findlay et al., 1996). In this
study the pretreatment SUV did not cor-
relate with response, but at 4 weeks the
responding lesions had a significantly
greater reduction in metabolism than
non-responding lesions (99% versus 67%
1996). In patients receiving FOLFOX as
a second-line treatment of hepatic color-
ectal metastases, a reduction in SUV
was similarly predictive of a > 1-year
survival (Dimitrakopoulou-Strauss et al.,
2003). These findings are in agreement
with the ability of PET to predict patho-
logical response to chemotherapy and
survival in patients receiving neoadju-
vant treatment in rectal cancer (Guillem
et al., 2004) and in a number of other
malignancies (Young et al., 1999).

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
The ability of chemotherapy to decrease
tumor FDG uptake must be taken into
consideration when evaluating a patient
for surgery following neoadjuvant treat-
ment (Figures 5.1–5.3). A recent pro-
spective study evaluating the detection of
colorectal hepatic metastases found that in
patients treated with neoadjuvant chemo-
therapy, 37% of lesions were missed by
FDG-PET, as compared to 23% of lesions
in patients who received no preoperative
chemotherapy (Akhurst et al., 2005). In
addition, while no malignant hepatic lesion
> 1.2 cm was missed in the later group, a
lesion as large as 3.2 cm was missed in
the former group (Akhurst et al., 2005).
The phosphorylating activity of hexoki-
nase in the tumor was also significantly
lower in patients treated preoperatively
with chemotherapy, providing a rationale
for the decrease in sensitivity of FDG-PET
(Akhurst et al., 2005). Although the disap-
pearance of a tumor by FDG-PET may
indicate a good prognosis, it is clear from
this study that it cannot be equated to a
pathological complete response.
POSITRON EMISSION
TOMOGRAPHY SCANNING
FOR SURVEILLANCE AND
FOLLOW-UP
The role of PET scanning in screening all
colon cancer patients for recurrence and
metastases is discussed in more detail in
other chapters in this volume. There are
two areas, however, unique to patients
being evaluated for hepatic colorectal
metastases that deserve mention; patients
with an elevated CEA and no lesion on
conventional imaging, and patients being
followed-up post-hepatic resection.
57
THE ROLE OF POSITRON
EMISSION TOMOGRAPHY IN
EVALUATING AN ELEVATED
CARCINOEMBRYONIC
ANTIGEN (CEA)
Elevation of serum CEA levels may be the
earliest sign of colorectal cancer recur-
rence and, therefore, CEA is used for
postoperative surveillance. Conventional
diagnostic imaging does not lead to locali-
zation of the site of recurrence in as many
as 8% of patients resulting in the delay of
a potentially curative surgery (Flamen et
al., 2001). Several studies have evaluated
the additive value of PET scan in patients
with an increase in CEA during surveil-
lance and normal or equivocal findings on
CT scan (Flamen et al., 2001). In the larg-
est study, 50 patients with elevated CEA
during postoperative surveillance and a
normal or equivocal conventional diagnos-
tic workup underwent FDG-PET imaging
(Flamen et al., 2001). A total of 15 patients
were determined to have hepatic metas-
tases by histology or long-term follow-
up, and among this group PET detected
12 lesions, but only 5 were ultimately
resected for cure (Flamen et al., 2001).
The PET scan also detected three lesions
in the liver that were not malignancies
(20% false-positive) (Flamen et al., 2001).
These findings are in agreement with other
smaller studies and when taken together
can indicate that in the subset of patients
with an elevated CEA and no or equivocal
findings on CT scan, ~ 10% of the entire
cohort will have resectable hepatic metas-
tases detected by PET (Flanagan et al.,
1998). Overall, PET in this setting has a
positive predictive value of close to 90%
with false-positives in 10–40% of patients
(Flanagan et al., 1998). These studies
58
however, fail to document the number of
invasive procedures, such as biopsies that
PET scans either avoid or mandate, and the
associated morbidity and costs. Moreover,
these studies cannot claim that a PET scan
is a superior test to evaluate for recurrent
disease in the setting of an elevated CEA
during surveillance, because these studies
evaluated PET with full knowledge of the
results of conventional imaging.
THE ROLE OF POSITRON
EMISSION TOMOGRAPHY
IN FOLLOW-UP POST
HEPATIC RESECTION
The most common site for recurrence after
resection of hepatic colorectal metastases
is the liver, and it is the sole site of recur-
rence in up to 40% of cases (Taylor, 1996).
In the absence of extrahepatic disease and
in a patient with adequate hepatic reserve,
a repeat hepatectomy may be considered.
In approximately one third of patients the
recurrence will be amenable to further
resection with a 5-year overall survival
range from 31–41% (Bozzetti et al., 1992).
Positron emission tomography/com-
puted tomography (PET/CT) imaging may
be superior to conventional CT in estab-
lishing the diagnosis of intrahepatic recur-
rence in patients with prior hepatectomy.
In a prospective study of the impact of
FDG-PET on treatment in patients with
colorectal liver metastases, a subset of
18 patients with previous hepatic resec-
tion had a suspected local recurrence by
CT scan (Selzner et al., 2004). A total of
nine patients had an intrahepatic recur-
rence, resulting in specificity of 50% for
CT scans (Selzner et al., 2004). By com-
parison, all nine patients were positive by
FDG-PET (PET sensitivity 100%) and
R. Auer and Y. Fong
only one patient had a false-positive (PET
specificity 89%) (Selzner et al., 2004). In
another blinded study, PET had a diag-
nostic accuracy of 86% in differentiat-
ing intrahepatic postsurgical change from
malignant recurrence, as compared to 45%
for CT portography (Delbeke et al., 1997).
The significant difference is the result of
more false-positives with CT portography.
ALGORITHM TO INCLUDE
POSITRON EMISSION
TOMOGRAPHY SCANNING
IN WORK-UP OF HEPATIC
COLORECTAL METASTASES
The preoperative staging prior to hepatic
resection of colorectal metastases must be
meticulous to avoid unexpected findings or
obstacles at the time of laparotomy. A sys-
tematic approach to the colorectal cancer
patient who is referred for consideration of
hepatic resection can facilitate a thorough
patient work-up and minimize unnecessary
investigations. Patients should initially be
evaluated with a high quality CT scan of the
chest abdomen and pelvis to identify hepatic
metastases as well as local recurrence and
lung metastases. In patients with an equiv-
ocal finding on CT scan or a persistently
elevated CEA and a normal CT scan, PET
scanning is warranted and may delineate
the site of metastasis and/or demonstrate
additional extrahepatic disease. In patients
with a documented suspicious lesion in
the liver on a CT scan, a PET scan is
best reserved for patients with a CRS ≥ 2
because these patients will have the highest
incidence of extrahepatic disease that may
impact on the clinical management.
As previously mentioned, PET scans may
have a false-negative rate as high as 17%
(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
N., Blumgart, L.H., and Larson, S.M. 1999b.
Utility of 18F-FDG positron emission tom-
ography scanning on selection of patients for
resection of hepatic colorectal metastases. Am.
J. Surg. 178:282–287.
Guillem, J.G., Moore, H.G., Akhurst, T., Klimstra,
D.S., Ruo, L., Mazumdar, M., Minsky, B.D.,
Gollub, M.J., Saltz, L., Wong, W.D., and Larson,
S. 2004. Sequential preoperative fluorodeoxyglu-
cose-positron emission tomography assessment
of response to preoperative chemoradiation: a
means for determining long term outcomes of
rectal cancer. J. Am. Coll. Surg. 199:1–7.
Jaeck, D., Nakano, H., Bachellier, P., Inoue, K.,
Weber, J.C., Oussoultzoglou, E., Wolf, P., and
Chenard-Neu, M.P. 2002. Significance of hepatic
pedicle lymph node involvement in patients with
colorectal liver metastases: a prospective study.
Ann Surg. Oncol. 9:430–438.
Jarnagin, W.R., Conlon, J., Bodniewicz, J.,
Dougherty, E., DeMatteo, R.P., Blumgart, L.H.,
and Fong, Y. 2001. A clinical scoring system
predicts the yield of diagnostic laparoscopy
in patients with potentially resectable hepatic
colorectal metastases. Cancer 91:1121–1128.
Nakashima, R.A., Paggi, M.G., and Pederson, P.L.
1984. Contributions of glycolysis and oxidative
phosphorylation to adenosine 5′-triphosphate
production in AS-30D hepatoma cells. Cancer
Res. 44:5702–5706.
Pauwels, E.K., Sturm, E.J., Bombardieri, E.,
Cleton, F.J., and Stokkel, M.O. 2000. Positron-
emission tomography with [18F]fluorodeoxy-
glucose. Part I. Biochemical uptake mechanism
and its implications for clinical studies. J. Can.
Res. Clin. Oncol. 126:549–559.
Rohren, E.M., Paulson, E.K., Hagge, R., Wong,
T.Z., Killius, J., Clavien, P.A., and Nelson, R.C.
2002. The role of F-18 FDG positron emission
tomography in preoperative assessment of the
liver in patients being considered for curative
resection of hepatic metastases from colorectal
cancer. Clin. Nucl. Med. 27:550–555.
Rosen, C.B., Nagorney, D.M., Taswell, H.F.,
Helgeson, S.L., Illstrup, D.M., Van Heerden,
J.A., and Adson, M. 1992. Perioperative blood
transfusion and determinants of survival after
liver resection for metastatic colorectal carci-
noma. Ann. Surg. 216:492–505.
Ruers, T.J., Langenhoff, B.S., Neeleman, N., Jager,
G.J., Strijk, S., Wobbes, T., Corstens, F.H., and
R. Auer and Y. Fong
Oyen, W.J. 2002. Value of positron emission
tomography with [F-18]fluorodeoxyglucose in
patients with colorectal liver metastases: a pro-
spective study. J. Clin. Oncol. 20:388–395.
Rydzewski, B., Dehdashti, F., Gordon, B.A.,
Teefey, S.A., Strasberg, S.M., and Siegel, B.A.
2002. Usefulness of intraoperative sonography
for revealing hepatic metastases from colorec-
tal cancer in patients selected for surgery after
undergoing FDG PET. AJR. Am. J. Roentgenol.
178:353–358.
Schussler-Fiorenza, C.M., Mahvi, D.M.,
Niederhuber, J., Rikkers, L.F., and Weber,
S.M. 2004. Clinical risk score correlates with
yield of PET scan in patients with colorec-
tal hepatic metastases. J. Gastrointest. Surg.
8:150–157.
Selzner, M., Hany, T.F., Wildbrett, P., McCormack,
L., Kadry, Z., and Clavien, P.A. 2004. Does the
novel PET/CT imaging modality impact on the
treatment of patients with metastatic colorectal
cancer to the liver? Ann. Surg. 204:1027–1034.
Strasberg, S.M., Dehdashti, F., Siegel, B.A.,
Drebin, J.A., and Linehan, D. 2001. Survival
of patients evaluated by FDG-PET before
hepatic resection for metastatic colorectal car-
cinoma: a prospective database study. Ann.
Surg. 233:293–299.
Taylor, I. 1996. Liver metastases from colorectal
cancer: lessons from past and present clinical
studies. Br. J. Surg. 83:456–460.
Truant, S., Huglo, D., Hebbar, M., Ernst, O.,
Steinling, M., and Pruvot, F.R. 2005. Prospective
evaluation of the impact of [18F]fluoro-2-deoxy-
D-glucose positron emission tomography of
resectable colorectal liver metastases. Br. J.
Surg. 92:362–369.
Wiering, B., Krabbe, P.F., Jager, G.J., Oyen, W.J.,
and Ruers, T.J. 2005. The impact of fluor-18-
deoxyglucose-positron emission tomography in
the management of colorectal liver metastases.
Cancer 104:2658–2670.
Young, H., Baum, R., Cremerius, U., Herholz,
K., Hoekstra, O., Lammertsma, A., Pruim, J.,
and Price, P. 1999. Measurement of clinical
and subclinical tumour response using [18F]-
fluorodeoxyglucose and positron emission
tomography: review and 1999 ORTC recommen-
dations. European Organization for Research
and Treatment of Cancer (EORTC) PET Study
Group. Eur. J. Cancer 35:1773–1782.
B. Treatment
6
Ultrasonography During Liver Surgery
Guido Torzilli
INTRODUCTION
Since the late 1980s, it has been evident
that hepatic surgery performed using the
conventional approach, carries a very high
risk for patients’ survivals due to the high
rate of postoperative liver failure (Lin,
1979). In particular, the coexistence of
liver cirrhosis in most cases with hepato-
cellular carcinoma (HCC) has a consider-
able adverse effect on the surgical results
with 6% postoperative mortality, which is
not negligible (Liu et al., 2004). For this
reason and for the broadening of ultra-
sound-guided percutaneous therapies, the
surgical treatment of HCC has loosen its
role of first choice treatment. Furthermore,
the good results with interstitial therapies
in the treatment of HCC has induced the
extension of their use also for those patients
with colorectal cancer (CRC) liver metas-
tases (Solbiati et al., 2001). However, the
imaging techniques have been introduced
also as aids for surgeons in performing liver
resection. In fact, at the end of the 1970s,
intraoperative ultrasonography (IOUS) in
hepatic surgery in the case of liver cirrhosis
started (Makuuchi et al., 1980, 1998). Now,
liver resections can be carried out with no
mortality even if cirrhosis is associated
(Torzilli et al., 1999a), combining the
needs for oncological radicality and
liver parenchyma sparing, and this goal
is mainly achievable because of IOUS
(Torzilli et al., 2005a, 2006). Recently, the
demonstration of feasibility and efficacy
of contrast-enhanced ultrasonography per-
formed intraoperatively (CE-IOUS) has
furthermore stressed the relevance of IOUS
guidance during liver surgery (Torzilli
et al., 2004a, b, 2005b).
Technical Aspects
For a proper IOUS, high frequency echo-
probes (7.5–10 MHz) are necessary and
they should have a flat shape to allow their
management in deep and narrowed spaces.
For this purpose, T-shaped probes, inter-
digital probes, and micro-convex probes
are available. Main factors for probe
selection are its volume, its stability and
the wideness of ultrasonographic scan-
ning window: the best probe should be
small, thin in width and short in transverse
length, stable and with a wide ultrasono-
graphic scanning window. In this sense
the micro-convex probe represents the
best compromise among all these require-
ments. Indeed, the T-shaped probe is more
stable but has a lower ratio between lateral
length and ultrasonographic scanning win-
dow than the micro-convex probe.
For CE-IOUS we initially used a convex
3–6 MHz frequency and 1.88–3.76 MHz
harmonic frequency transducer was
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
cedures by IOUS in their series. However,
the impact of IOUS on the operative deci-
sion making, when compared with those
of preoperative imaging techniques, has
decreased to 4–7% (Jarnagin et al., 2001;
Cerwenka et al., 2003; Sahani et al., 2004).
However, the problem of the impact of
IOUS on the operative decision making
depends on two main factors: the surgical
policy of each specific team, and the type
of tumor. Indeed, the low rates shown in
the latest reports (Jarnagin et al., 2001;
Cerwenka et al., 2003; Sahani et al.,
2004) are also partially motivated by the
surgeon’s surgical policy: in fact, since a
considerable number of patients under-
went major hepatectomies, new nodules
detected by IOUS in the same hemiliver
would not have modified the surgical strat-
egy. In our experience major hepatectomies
are carried out in the minority of patients
(Torzilli et al., 2005a, 2006) because of the
extensive use of the IOUS-guidance for
achieving parenchymal sparing resections,
so that detection of new nodules is more
suitable for changing the surgical strategy.
Conversely, depending on the type of the
tumor, liver exploration is focused mainly
on the detection of new lesions as in the
case of CRC liver metastases rather than
both on the detection and on the differen-
tiation, as in the case of HCC. Indeed, for
this last tumor, which generally develops
in the cirrhotic liver the risk nowadays is
to overestimate the tumor stage with IOUS.
In fact, in the case of patients with HCC
in a cirrhotic liver, except for those nod-
ules with mosaic ultrasonographic pattern
which are malignant in 84% of cases, only
24–30% of hypoechoic nodules (Figure
6.1), and 0–18% of those hyperechoic are
neoplasm (Kokudo et al., 1996; Takigawa
et al., 2001). To overcome this problem
65
even biopsy does not seem to be adequate.
The only nodule which can be easily dif-
ferentiated intraoperatively from a HCC
or a liver metastases is the small hemangi-
oma, which is often discovered primarily
at IOUS: it has a typical ultrasonographic
pattern and, moreover, when compressed
changes its size and appearance. Further
improvement in differential diagnosis of
liver nodules with IOUS may be expected
by the introduction and diffusion of the
intraoperative use of the last generation
contrast agents.
Contrast Enhanced Intraoperative
Ultrasonography
More recently the introduction of CE-IOUS
has set the rate of modified operative deci-
sion-making in 30–40% of cases (Torzilli
et al., 2004a, b, 2005b).
Tumor vascularity as a criterion for dif-
ferentiating the regenerative or dysplastic
nodules from the HCC correlate well with
the histological evidence of a progressive
increase in unpaired arteries from dys-
plastic to neoplastic nodules in a cirrhotic
liver (Roncalli et al., 1999). Certainly, the
pattern of vascular enhancement is not suf-
ficient for differentiating malignant from
non malignant nodules in a cirrhotic liver
with 100% specificity. However, CE-US
provides differential diagnosis of FLL
with a 95% specificity (Quaia et al.,
2004); of course, it must be considered
that this last rate was referred to another
type of lesion in comparison to the target
of the present study. Indeed, the intraop-
erative exploration takes profit from the
higher resolution of the ultrasonography in
direct contact with the liver. Therefore, the
need of differentiating nodules detected at
IOUS is mostly focused on lesions smaller
than 1 cm: for these nodules, the vascularity
66
as criterion for differential diagnosis is
less specific. However, some improve-
ments compared with conventional IOUS
could be expected. In our preliminary
experience, CE-IOUS provided remark-
able findings, either by adding information
on nodular vascularity in patients with
HCC, or by detecting nodules that were
not visible at IOUS, in patients with CRC
liver metastases (Torzilli et al., 2004b,
2005b). Focusing the attention on patients
operated for HCC our preliminary experi-
ence showed encouraging results (Torzilli
et al., 2004a): however, they were related
to a very limited number of patients, and
data did not meet the expectancies shown
by optimistic findings with 100% specifi-
city. Therefore, we classified the CEIOUS
pattern based on our preliminary experi-
ence and on what was reported in literature
about the CE-US (Quaia et al., 2004).
Classification was as follows:
(A) If an hypoechoic or hyperechoic nod-
ule showed at CEIOUS a full enhance-
ment (hyperechoic) in the arterial
phase and became hypoechoic in the
delayed portal and late phases (A1)
or it remained hypoechoic with thin
vessels detectable into the nodule in the
arterial phase and delayed phase (A2)
or did not show any early enhance-
ment but remained hypoechoic in the
delayed phases with peripheral and/or
intralesional neovascularization (A3),
it was considered malignant and then
removed.
(B) If an hypoechoic or hyperechoic
nodule had no early enhancement at
CEIOUS and showed a hysoechoic
or slightly hypoechoic pattern in all
the phases in comparison with the
surrounding liver parenchyma, it was
considered a non neoplastic lesion.
G. Torzilli
These last were not removed, and
their benignity was confirmed after at
least 6 months of follow-up if there
was no tumor growth or they were
still not detectable at spiral CT.
(C) If a new nodule showed a mosaic pat-
tern at IOUS, it was considered malig-
nant (Takigawa et al., 2001), and then
removed, independently from the pat-
tern of enhancement at CE-IOUS.
Specificity of this classification criteria
is in our experience 69% (Torzilli et al.,
2007) which is certainly closer to the
true performance of CEIOUS, rather than
the results obtained at the initial experi-
ence (Torzilli et al., 2004a). This value is
probably not that high, especially when
compared with that reported for CE-US
(Quaia et al., 2004). However, as we men-
tioned earlier the small size of the lesions
targeted for CEIOUS study could explain
this discrepancy: for these tiny nodules
the neovascularity as criterion for differ-
entiation between malignant and benign
lesions has limits which are independ-
ent from the method we use for studying
them. Therefore, CEIOUS can be helpful
in a certain percentage of nodules but not
in all. In this perspective the rate of 69%
of specificity is encouraging as it means
that we can provide proper information
with this new technique in seven out of ten
lesions we detect at the time of laparotomy.
Further improvements in this sense could
come from the introduction of hepto-
specific contrast spents for CE-IOUS.
Colorectal Cancer Liver Metastases
Intraoperative ultrasonography is still
the most accurate diagnostic technique
for detecting focal liver lesions (FLLs)
(Sahani et al., 2004). However, during
surgery for CRC liver metastases, IOUS
6. Ultrasonography During Liver Surgery 67

has a sensitivity of just 82% (Machi et al.,
1991), and as a consequence may miss
nodules less than 1 cm in diameter: this is
particularly evident in those patients who
undergo surgery after chemotherapy. In
these patients, liver fore sites have similar
echo-pattern with the surrounding liver
parenchyma.
In our experience, CE-IOUS confirmed
its improvement of IOUS sensitivity in
detecting new small fore sites reducing
the risk of down-staging the disease and
enhancing the rate of treatment with cura-
tive intent: indeed 9% of patients had
new lesions discovered only by CEIOUS
(Figure 6.3a, b) (Torzilli et al., 2008a).
The association of IOUS and CEIOUS
increased the sensitivity from 88 to 93%.
Additionally, the better visualization of
the margins of the main lesion in 19% of
patients with CRC liver metastases helps
the surgeon to better define the resection
area and the proper dissection plane ruling
out the risk of tumor exposure, and thus
enhancing the treatment radicality.
Tumor Location
IOUS allows an accurate three-dimen-
sional reconstruction of the relation-
ship between the tumor and the portal
branches, and hepatic veins: this is a fun-
damental step in the definition of the
proper surgical strategy. Indeed, surgical
decision making should be obtained hav-
ing portal branches and hepatic veins as
landmarks to reduce the risk of major
morbidity and mortality.
The definition of tumor-vessels relation-
ship is relevant for planning the type of
resection and for this purpose we classi-
fied its appearance at IOUS. Based on this
classification each category corresponds to
a specific operation.
Based on IOUS findings, relationship with
intrahepatic vascular structures were classi-
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

1. Portal branch or hepatic vein in contact with vessel wall discontinuation at

with a capsulated HCC without vessel
wall discontinuation at IOUS: vascular
resection is not associated, and just
enucleation at the level of the vascular
contact is performed (Figure 6.4a).
2. Portal branch or hepatic vein separated
by a thin layer of liver parenchyma (at
least 1 mm) from CRC liver metastases:
vascular resection is not associated.
3. Portal branch or hepatic vein in contact
with a HCC with undefined margins
without vessel wall discontinuation
at IOUS: vascular resection is associ-
ated.
4. Portal branch or hepatic vein in contact
with a CRC liver metastases without
vessel wall discontinuation at IOUS:
vascular resection is associated.
5. Portal branch or hepatic vein in contact
with HCC or CRC liver metastases
IOUS: vascular resection is associated
(Figure 6.4b).
6. Portal branch in contact with a HCC or
CRC liver metastases with or without
vessel wall discontinuation but with
proximal bile duct dilation at IOUS:
vascular resection is associated.
Extension of the hepatectomy is always
considered for the parenchyma fed by an
infiltrated portal branch (categories 3–6).
Inversely, in case of infiltration of a hepatic
vein, an extension of the resection to the
whole liver parenchyma theoretically
drained by this vein is considered only
if there are no accessory hepatic veins
at IOUS, and color-Doppler IOUS shows
hepatofugal blood flow in the feeding portal
branch once the hepatic vein is clamped.
IOUS tumor-vessel classification and
the related surgical policy has proven that

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
in selected patients it is possible to get
close to the tumor burden without increas-
ing the risk of incomplete removal, and as
a consequence of local recurrence (Torzilli
et al., 2005a, 2006). In practice, this evi-
dence means that with IOUS guidance it is
possible to perform conservative but radi-
cal hepatectomies also in complex pres-
entations, and then to enlarge the surgical
indications. With this approach we have
limited the rate of major hepatectomies to
8% in patients with tumors involving one
or more hepatic veins close to their caval
confluence, without performing any vas-
cular reconstruction (Torzilli et al., 2006).
These results not only underline how the
IOUS guidance allows operation other-
wise not feasible, but also because this
approach reduces the rate of major hepa-
tectomies, and in particular of right hemi-
hepatectomy (Torzilli et al., 2008b). The
real need for interventions such as preop-
erative portal vein embolization which are
adopted to secure from liver failure after
major removal of liver parenchyma can be
discussed.
RESECTION GUIDANCE
Hepatocellular Carcinoma
Systematic Segmentectomy
In a cirrhotic patient, the liver volume to
be resected must be determined with par-
ticular care with the purpose to associate
surgical radicality and non-cancerous liver
parenchyma sparing. Liver function tests
and liver volumetry on CT scans help in
this decision. Tumor dissemination from
the main lesion through the portal branches
cannot be detected with certainty by the
pre- and intraoperative imaging modalities
69
(Makuuchi et al., 1991). As a consequence
of that, some authors consider that the
resected specimen should comprehend at
least the portal area, which includes the
lesion (Makuuchi et al., 1991). This is
impossible to correctly identified without
the aid of IOUS, especially in a cirrhotic
liver where there are generally wide varia-
tions and abnormalities in the distributions
of the portal branches (Makuuchi et al.,
1991). For this purpose the systematic seg-
mentectomy was devised in the early 1980s
(Makuuchi et al., 1980), and its peculiar-
ity consists originally in the dye injection
into the feeding portal branch. We have
recently established alternatives to this
approach for tumors located in the left
hemiliver which will be also described.
Puncture of the Portal Branch
After the accomplishment of liver mobili-
sation and exploration with IOUS, segmen-
tal staining is carried out by puncturing
the portal branch near the target point to
be divided during dissection (Makuuchi
et al., 1991). The portal branch is punctured
under echo-guidance, through a free-hand
technique or with a proper device, and
then 3–5 ml of indigo-carmine dye are
injected into the vessel. The stained area
becomes evident on the liver surface and
it is marked with the electrocautery. If the
nodule is located between two adjacent
segments, two portal branches afferent to
the area must be punctured and injected.
In this case, the deepest and most dorsal
vessel must be punctured at first to prevent
air bubbles contained in the dye from dis-
turbing the ultrasonographic detection and
puncture of the other branch. The portal
branch is punctured 1–2 cm distally from
its origin to prevent the dye from reflexing
and the direction and velocity of the infusion
70
are controlled at IOUS. To prolong the
staining, the hepatic artery at the hilum is
clamped before the portal branch puncture.
When there are numerous and thin vessels
which should be punctured or in the case
of tumor thrombus in the segmental portal
branch of the segment that should be
removed, the dye is injected in the portal
branches afferent to the adjacent segment:
this is the so-called counterstaining tech-
nique (Takayama et al., 1991).
Hooking of the Portal Branch
There are four segmental portal branches
to segment which are generally divided
in two groups: those for the superior and
those for the inferior portion but the most
common branching pattern can be rec-
ognized in just half of the cases. These
branches rather than punctured under
IOUS guidance, can be approached by
dissecting the umbilical portion: once
exposed, the vessel can be encircled with
a suture and pulled under IOUS control to
verify if it is the branch to S4 inferior or
not; then the proper portalbranch can be
ligated and divided and the discoloured
area which will appear on the liver sur-
face should correspond to the S4 inferior
which can be marked with the electro-
cautery to proceed with the liver dissec-
tion: this is a peculiar application of the
so-called hooking technique (Torzilli and
Makuuchi, 2001). Furthermore, the sub-
segment 4 superior could be resected by
clamping the portal branch to subseg-
ment 4 inferior, as it is identified with the
earlier described hooking technique; the
discoloured subsegment 4 inferior cau-
dally, the plan at IOUS which includes
the middle hepatic vein laterally, and that
marked by the falciform ligament medi-
ally delimitates the area to be resected.
G. Torzilli
Compression of the Portal Branch
In addition to the hooking technique we
have proposed a new technique to perform
subsegmental resections, initially just for
segments 2 and 3, without clamping the
hepatic artery and puncturing the feeding
portal branches (Torzilli and Makuuchi,
2004). Indeed, once the feeding portal
branch has been identified at IOUS, it is
compressed using the IOUS probe by one
side of the liver and the finger by the oppo-
site side: in this way it is possible to induce
a transient ischemia of the subsegmental
portion of the liver distally to the com-
pression site. This portion can be marked
with the electrocautery, the compression
is released, and the subsegmentectomy is
carried out. This technique is simple, fast,
not invasive and reversible. It is possible
to modify the site of compression and the
resection volume allows the sizing of the
resection in function of the tumor features
and the status of the background liver.
Together with the S4 subsegmentectomies,
this operation represents an alternative to
the puncture technique which is not as
reproducible as these two last methods
(Torzilli and Makuuchi, 2001, 2004).
Limited Resection
After the identification of the tumor, the
surgeon under IOUS control can mark
with the electrocautery the border of the
lesion on the surface of the liver just above
the nodule. To carry out this maneuver the
flat and thin tip of the electrocautery is
positioned between the probe and the liver
surface: this maneuver results in a shadow
at the IOUS image which runs deeply just
below the electrocautery. In this way it is
possible to define the position of the elec-
trocautery with the tumor edge and con-
sequently to mark with the electrocautery
6. Ultrasonography During Liver Surgery 71

itself the nodule profile on the liver surface

and to select the safer edge for the inci-
sion. The adequacy of the marked edge HCC
can be furthermore checked with IOUS as
the air trapped between the probe and the
irregular surface of the demarcation line
drawn with the electrocautery on the liver L
surface can be visualized at IOUS.
Another way to draw precisely on the
liver surface with the aid of IOUS involves
the tumor edge carrying out using the finger
tips. With the probe positioned on the liver
surface the surgeon’s finger-tip pushes on
the opposite side and its profile is visual- Figure 6.5. At IOUS the dissection line can be
well visualized (arrows), and the planned direction
ized at IOUS: as a consequence relations
(transparent arrows) allow radical removel of the
between the finger-tip and the tumor edge HCC nodule
can be precisely estimated and the resection
area can be marked on the liver surface.
Parenchymal Dissection
L
The main advantage provided by the resec- MTX
tion guidance accomplished with the aid of
IOUS is the modification of the traditional
way to dissect the liver tissue which was
done on vertical planes to avoid the tumor
exposure on the cut surface. IOUS allows
to follow in real-time the dissection plane,
to put it constantly in relation with the
tumor edge, and to modify its direction
when needed: this is because it is possible
to visualize on the IOUS image the dis-
section plane which appears as an echoic
line due to the entrapment of air bubbles
and clots between the faced cut surfaces Figure 6.6. At IOUS the dissection plane (arrows)
(Figure 6.5). If the dissection plane is not has been modified, because its original direction
(transparent arrow) would have led to enter into the
clearly visible, it can be better visualized by
to metastatic nodule (MTX)
inserting a gauze or a specifically devised
silicon gauze between the faced cut sur-
faces. These techniques allow the surgeon (Figure 6.6). In this way it is possible to
to keep the proper dissection plane: an early carry out a rounded trajectory of the dis-
recognition of a wrong dissection plane section plane around the tumor avoiding its
permits the surgeon to modify it properly exposure, and allowing the surgeon to spare
and to avoid a possible tumor exposure important vascular structures: this results
72
in more conservative but radical treatments
and in a lower rate of major hepatectomies.
Furthermore, in those patients in which
major resections are needed, IOUS allows
better achievement of the proper dissection
plane, which should run along the hepatic
vein to be fully anatomic.
The artifacts which allow the dissec-
tion plan at IOUS can sometimes mask
structures such as portal branches which
should be ligated or conversely respected.
For this reason, to better visualise the
targeted point where the portal branch
should be divided, the so-called hooking
technique has been devised (Torzilli et al.,
1999b). When the Glissonian sheath is
exposed and skeletonized, it is encircled
with a stitch, which is visualized by IOUS
as an echogenic spot with a posterior
shadow. Then under sonographic control,
the stitch, hooking the exposed vessel, is
gently pulled up, which stretches the portal
branch slightly and the traction point is
demonstrated clearly by IOUS. If the
exposed portal branch is not clearly visible
because it has collapsed, the portal triad is
unclamped to enable it to fill with blood
and then it is visualised better by IOUS. If
the target site is correct, the portal branch
is ligated and divided and segmentec-
tomy is completed under IOUS guidance;
conversely, if the exposed vessel was not
the targeted one, it is spared and useless
sacrifice of further liver parenchyma is
avoided.
A practical example in which the hook-
ing technique is used is during ventral or
dorsal subsegmentectomy of segment 8.
The portal trunk to this segment may show
bifurcation in its dorsal branch and ventral
trunk just close to the origin of the portal
vessel to segment 5. In this situation, there
is the risk of ligating and dividing the
G. Torzilli
portal branch of segment 5 instead of the
planned subsegmental branch of segment
8 and then, necrosis of the segment 5 may
occur. The hooking technique under IOUS
control enables the identification of the
branch, which was encircled, and then the
surgeon can decide with certainty whether
to ligate it. This is useful also in the case of
tumor thrombus in major portal branches.
In this situation, once the portal branch is
skeletonized, it is encircled with a stitch
and, under IOUS control, the stitch, is
gently pulled up: this traction stretches
the portal branch slightly and the traction
point is demonstrated clearly by IOUS. If
the traction point is not at the level of the
tumor thrombus it is possible to ligate the
portal branch and proceed with the liver
resection being sure that thrombus will not
migrate because of surgical manipulation
(Torzilli et al., 2005c).
During liver dissection the back-flow
bleeding from the hepatic veins is an
important source of blood loss, and it
is one of the most important factors in
determining the short-term and long-term
patient’s outcome. Therefore, limiting
the backflow bleeding from the hepatic
veins is a priority in liver resections. We
described a ultrasound-guided technique
for backflow bleeding control from the
RHV during right-sided liver resection.
The technique is very simple: once the
right surface of the extrahepatic RHV is
exposed to allow its compression by the
surgeon’s finger-tips, the effectiveness of
finger compression is checked by IOUS,
and color-Doppler.
Post-resectional Control
There are possibilities given by IOUS after
nodule removal: one is the “water bath”
technique which consists of the real-time
6. Ultrasonography During Liver Surgery
control of the proper resection of the tar-
geted nodule verifying its complete inclu-
sion in the specimen once removed from
the liver (Makuuchi, 1987); the second is
done by checking the cut surface refilled
with saline to avoid the artifacts generated
by the residual air bubbles and clots.
The aforementioned methods guarantee,
whenever possible, anatomical and limited
resection: this has consequences either for
the effectiveness of the surgical treatment
or for its safety. Patients with HCCs who
underwent IOUS-guided subsegmentec-
tomies had better survivals (Makuuchi
et al., 1991; Hasegawa et al., 2005), and
those with HCCs in contact with intrahe-
patic vessels, once removed under IOUS-
guidance sparing the involved veins did
not show local recurrences (Torzilli et al.,
2005a, 2006). Additionally, this kind of
surgery allows radical surgical treatment
of HCC without mortality (Torzilli et al.,
1999a). Procedures, which are not IOUS-
guided, led to dangerous and useless major
resection or incomplete operations.
Liver Metastases
At present, tumor clearance can be achieved
in most cases of patients undergoing liver
resection for metastatic disease with sur-
vival benefits even in the presence of
multiple fore sites or vascular infiltration
(Minagawa et al., 2000). To achieve this
standard, most of the merits of conserva-
tive and radical resections should prob-
ably be addressed to the extensive use of
IOUS. In fact, similar to the case of HCC,
IOUS is not only a simple, accurate diag-
nostic modality, but it is a real surgical
device which allows a proper hepatic
resection. In the case of CRC liver
metastases there is no need for anatomical
73
resections, but it is mandatory to obtain the
complete tumor clearance: for this purpose
IOUS is essential.
Definition of the Resection Area
Definition of the resection area once a non-
anatomical resection has been chosen is
similar for both resection for HCC and CRC
liver metastases. In both these conditions,
the surgeon’s target should be to determine
a resection volume which could result in a
flat and regular cut surface to minimize the
risk of postoperative collections.
Liver Dissection and Post-resectional Control
The same concepts discussed for HCC
regarding IOUS-guided dissection plane
maintenance and IOUS post-resectional
control are applied in the case of resection
for liver metastases.
The complete surgical clearance of the
tumor tissue, even for multiple liver metas-
tases, could be justified assuming that
every single nodule is secondary to the
primary tumor rather than an expression
of the intrahepatic diffusion of the first
metastatic lesion. For this purpose several
authors showed that the occurrence of
satellite nodules around the main meta-
static lesion are rare and wedge resection
even with tumor-free margin less than
0.5–1 cm in thickness but without expo-
sure of the tumor on the cut surface is jus-
tified (Kokudo et al., 2002; Pawlik et al.,
2005). Therefore, the goal to be achieved
is the margin clearance: in this sense IOUS
is fundamental to guarantee the complete
removal of the lesions. The other goal
is the conservativeness of the treatment
which should avoid useless parenchyma
sacrifice and then safer procedures. Also
in this sense, IOUS with the techniques
described above allows the accomplishment
74
of radical but limited resection introduc-
ing the aforementioned concept of the
safe rounded dissection plane. With this
approach, it has been possible to carry
out resections without mortality, with no
residual gross tumor on the surgical stump
and with a relatively low rate of major
hepatectomies (21.3%) (Minagawa et al.,
2000). These results were obtained in
spite of the high mean number of nodules
resected per patient (3.2) and its relatively
large mean size (4.5 cm).
This is a further confirmation that liver
resection is an imaging-guided procedure
and as every interventional imaging-guided
procedure, its features are the highest ther-
apeutic efficacy combined with the mini-
mal invasiveness. With IOUS aid, it is now
possible to carry out surgical procedures
which are safe, oncologically radical and
conservative for the liver function. Mostly
because of these reasons, surgery can be
still considered the treatment of choice for
most liver metastases, even in the case of
the diffusion of other treatments such as
the thermal ablation which is impaired by
a 39% tumor relapse rate in the first year of
follow-up (Solbiati et al., 2001).
In conclusion, IOUS has acquired pro-
gressive importance in the field of sur-
gical resection in cirrhotic and normal
liver, either in the diagnosis and staging
or strictly in the surgical technique, and
now it can be considered a fundamental
tool for hepatobiliary and other surgi-
cal procedures (Makuuchi et al., 1998).
Presently, we can affirm that liver resec-
tion is an imaging-guided procedure and
as every interventional imaging-guided
procedure, its features include the highest
therapeutic efficacy combined with mini-
mal invasiveness. IOUS should be a famil-
iar instrument for the hepatic surgeon.
G. Torzilli
The American College of Surgeons has
recently recognized the need for surgeons
of a specific training in US, meanwhile
dedicated monographs have been pub-
lished almost simultaneously in America
and in Europe (Machi and Staren, 2004;
Torzilli et al., 1997). The way for a wider
diffusion of ultrasound in surgeons’ prac-
tice has definitely been opened.
REFERENCES
Cerwenka, H., Raith, J., Bacher, H., Werkgartner,
G., el-Shabrawi, A., Kornprat, P., and Mischinger,
H.J. 2003. Is intraoperative ultrasonography
during partial hepatectomy still necessary
in the age of magnetic resonance imaging?
Hepatogastroenterology 50: 1539–1541.
Couinaud, C. 1957. Le foie, etude anatomiques et
chirurgicales. Paris, Masson.
Hasegawa, K., Kokudo, N., Imamura, H.,
Matsuyama, Y., Aoki, T., Minagawa, M., Sano,
K., Sugawara, Y., Takayama, T., and Makuuchi,
M. 2005. Prognostic impact of anatomic resec-
tion for hepatocellular carcinoma. Ann. Surg.
242: 252–259.
Jarnagin, W.R., Bach, A.M., Winston, C.B., Hann,
L.E., Heffernan, N., Loumeau, T., DeMatteo,
R.P., Fong, Y., and Blumgart, L.H. 2001. What is
the yield of intraoperative ultrasonography dur-
ing partial hepatectomy for malignant disease? J.
Am. Coll. Surg. 192: 577–583.
Kane, R.A., Hughes, L.A., Cua, E.J., Steele,
G.D., Jenkins, R.L., and Cady, B. 1994. The
impact of intraoperative ultrasonography on
surgery for liver neoplasms. J. Ultrasound
Med. 13: 1–6.
Kojiro, M. 2004. Focus on dysplastic nodules and
early hepatocellular carcinoma: an Eastern point
of view. Liver Transpl. 10 (2 Suppl 1): 3–8.
Kokudo, N., Bandai, Y., Imanishi, H., Minagawa,
M., Uedera, Y., Harihara, Y., and Makuuchi,
M. 1996. Management of new hepatic nodules
detected by intraoperative ultrasonography during
hepatic resection for hepatocellular carcinoma.
Surgery 119: 634–640.
Kokudo, N., Miki, Y., Sugai, S., Yanagisawa, A., Kato,
Y., Sakamoto, Y., Yamamoto, J., Yamaguchi, T.,
6. Ultrasonography During Liver Surgery
Muto, T., and Makuuchi, M. 2002. Genetic and
histological assessment of surgical margins in
resected liver metastases from colorectal carci-
noma: minimum surgical margins for successful
resection. Arch. Surg. 137: 833–840.
Lin, T.Y. 1979. Resectional therapy for primary
malignant hepatic tumors. In: International
Advances in Surgical Oncology. Murphy, GP.
ed., New York, Alan R Liss, pp. 25–54.
Liu, C.L., Fan, S.T., Lo, C.M., Wong, Y., Ng, I.O.,
Lam, C.M., Poon, R.T., and Wong, J. 2004.
Abdominal drainage after hepatic resection is
contraindicated in patients with chronic liver
diseases. Ann. Surg. 239: 194–201.
Machi, J., and Staren, E.D. 2004. Ultrasound for
Surgeons. New York, Lippincott-Williams &
Wilkins, 2nd Edition.
Machi, J., Isomoto, H., Kurohiji, T., Yamashita, Y.,
Shirouzu, K., Kakegawa, T., Sigel, B., Zaren,
H.A., and Sariego, J. 1991. Accuracy of intra-
operative ultrasonography in diagnosing liver
metastasis from colorectal cancer: evaluation
with postoperative follow-up results. World J.
Surg. 15: 551–556.
Makuuchi, M. 1987. Abdominal Intraoperative
Utrasonography. Tokyo/New York, lgaku-Shoin.
Makuuchi, M., Yamazaki, S., and Hasegawa, H.
1980. Ultrasonically guided liver surgery. Jpn. J.
Ultrasonics Med. 7: 45–49.
Makuuchi, M., Takayama, T., Kosuge, T., Yamazaki,
S., Yamamoto, J., Hasegawa, H., and Takayasu,
K. 1991. The value of ultrasonography for hepatic
surgery. Hepatogastroenterology 38: 64–70.
Makuuchi, M., Torzilli, G., and Machi, J. 1998.
History of intraoperative ultrasonography.
Ultrasound Med. Biol. 24: 1229–1242.
Minagawa, M., Makuuchi, M., Torzilli, G.,
Takayama, T., Kawasaki, S., Kosuge, T.,
Yamamoto, J., and Imamura, H. 2000. Extension
of the frontiers of surgical indications in the
treatment of liver metastases from colorectal
cancer: long-term results of our experience. Ann.
Surg. 231: 487–499.
Pawlik, T.M., Scoggins, C.R., Zorzi, D., Abdalla,
E.K., Andres, A., Eng, C., Curley, S.A., Loyer,
E.M., Muratore, A., Mentha, G., Capussotti, L.,
and Vauthey, J.N. 2005. Effect of surgical mar-
gin status on survival and site of recurrence after
hepatic resection for colorectal metastases. Ann.
Surg. 241: 715–722.
75
Quaia, E., Calliada, F., Bertolotto, M., Rossi, S.,
Garioni, L., Rosa, L., and Pozzi-Mucelli, R.
2004. Characterization of focal liver lesions with
contrast-specific US modes and a sulfur hexaflu-
oride-filled microbubble contrast agent: diag-
nostic performance and confidence. Radiology
232: 420–430.
Roncalli, M., Roz, E., Coggi, G., Di Rocco, M.G.,
Bossi, P., Minola, E., Gambacorta, M., and
Borio, M. 1999. The vascular profile of regen-
erative and dysplastic nodules of the cirrhotic
liver: implications for diagnosis and classifica-
tion. Hepatology 30: 1174–1178.
Sahani, D.V., Kalva, S.P., Tanabe, K.K., Hayat,
S.M., O’Neill, M.J., Halpern, E.F., Saini, S.,
and Mueller, P.R. 2004. Intraoperative US in
patients undergoing surgery for liver neoplasms:
comparison with MR imaging. Radiology 232:
810–814.
Solbiati, L., Livraghi, T., Goldberg, S.N., Ierace,
T., Meloni, F., Dellanoce, M., Cova, L., Halpern,
E.F., and Gazelle, G.S. 2001. Percutaneous
radiofrequency ablation of hepatic metastases
from colorectal cancer in 117 patients. Radiology
221: 159–166.
Takayama, T., Makuuchi, M., Watanabe, K.,
Kosuge, T., Takayasu, K., Yamazaki, S., and
Hasegawa, H. 1991. A new method for mapping
hepatic subsegment: counterstaining identifica-
tion technique. Surgery 109: 226–229.
Takigawa, Y., Sugawara, Y., Yamamoto, J., Shimada,
K., Yamasaki, S., Kosuge, T., and Makuuchi,
M. 2001. New lesions detected by intraoperative
ultrasound during liver resection for hepatocellular
carcinoma. Ultrasound Med. Biol. 27: 151–156.
Torzilli, G., and Makuuchi, M. 2001. Ultrasound-
guided liver subsegmentectomy: the peculiarity
of segment 4. J. Am. Coll. Surg. 193: 706–708.
Torzilli, G., and Makuuchi, M. 2004. Ultrasound-
guided finger compression in liver subsegmen-
tectomy for hepatocellular carcinoma. Surg.
Endosc. 18: 136–139.
Torzilli, G., Olivari, N., Livraghi, T., and Di Candio,
G. 1997. Ecografia in chirurgia: modalità diag-
nostiche e terapeutiche. Milano, Poletto Editore.
Torzilli, G., Makuuchi, M., Inoue, K., Takayama,
T., Sakamoto, Y., Sugawara, Y., Kubota, K.,
and Zucchi, A. 1999a. No-mortality liver resec-
tion for hepatocellular carcinoma in cirrhotic
and non-cirrhotic patients: is there a way?
76
A prospective analysis of our approach. Arch.
Surg. 134: 984–992.
Torzilli, G., Takayama, T., Hui, A.M., Kubota, K.,
Harihara, Y., and Makuuchi, M. 1999b. A new
technical aspect of ultrasound-guided liver sur-
gery. Am. J. Surg. 178: 341–343.
Torzilli, G., Olivari, N., Moroni, E., Del Fabbro,
D., Gambetti, A., Leoni, P., Montorsi, M., and
Makuuchi, M. 2004a. Contrast-enhanced intra-
operative ultrasonography in surgery for hepato-
cellular carcinoma in cirrhosis. Liver Transpl. 10
(2 Suppl 1): 34–38.
Torzilli, G., Del Fabbro, D., Olivari, N., Calliada,
F., Montorsi, M., and Makuuchi, M. 2004b.
Contrast-enhanced ultrasonography during liver
surgery. Br. J. Surg. 91: 1165–1167.
Torzilli, G., Montorsi, M., Donadon, M., Palmisano,
A., Del Fabbro, D., Gambetti, A., Olivari, N.,
and Makuuchi, M. 2005a. “Radical but con-
servative” is the main goal for ultrasonography-
guided liver resection: prospective validation of
this approach. J. Am. Coll. Surg. 201: 517–528.
Torzilli, G., Del Fabbro, D., Palmisano, A.,
Donadon, M., Bianchi, P., Roncalli, M., Balzarini,
L., and Montorsi, M. 2005b. Contrast-enhanced
intraoperative ultrasonography during hepatec-
tomies for colorectal cancer liver metastases. J.
Gastrointest. Surg. 9: 1148–1153.
Torzilli, G., Montorsi, M., Gambetti, A., Del
Fabbro, D., Donadon, M., Bianchi, P., Olivari,
G. Torzilli
N., and Makuuchi, M. 2005c. Utility of the
hooking technique for cases of major hepatec-
tomy. Surg. Endosc. 19: 1156–1157.
Torzilli, G., Montorsi, M., Del Fabbro, D.,
Palmisano, A., Donadon, M., and Makuuchi
M. 2006. Ultrasonographically guided surgical
approach to liver tumours involving the hepatic
veins close to the caval confluence. Br. J. Surg.
93: (10)1238–1246.
Torzilli, G., Palmisano, A., Del Fabbro, D.,
Marconi, M., Donadon, M., Spinelli, A., Bianchi,
PP., Montorsi, M., 2007. Contrast-enhanced
intraoperative ultrasonograpy during surgery for
hepatocellular carcinoma in liver cirrhosis: is it
useful or useless? A prospective cohort study
of our experience. Ann. Surg. Oncol. 14(4):
1347–1355.
Torzilli, G., Botes, F., Procopio, F., Donadon, M.,
Balzarini, L., Lutman, F., Calliada, F., Montorsi,
M. 2008a. Use of contrast-enhanced intraop-
erative ultrasonograpy during liver surgery for
colorectal cancer liver metastases. Its impact
on operative outcome. Analysis of prospective
Cohort study. Eu. J. Cancer 6: 16–23.
Torzilli, G., Donadon, M., Marconi, M., Botes,
F., Palmisano, A., Del Fabbro, D., Procopio,
F., Montorsi, M. 2008b. Systematic extended
right posterior sectionectomy: a safe and effec-
tive alternative to right hepatectomy. Ann. Surg.
247(4): 603–611.
A. Treatment
7
Intraoperative Magnetic Resonance
Imaging for Radiofrequency Ablation
of Hepatic Tumors
Oliver F. Bathe and Houman Mahallati
INTRODUCTION
Curative resection represents the best thera-
peutic alternative for colorectal liver metas-
tases and primary liver tumors (Abdalla
et al., 2004; Llovet, 2005). For functioning
metastatic neuroendocrine tumors, resec-
tion for the purpose of debulking is also
a therapeutic consideration (Veenendaal
et al., 2006). However, in some patients,
resection is not feasible because of ana-
tomic constraints, insufficient hepatic func-
tional reserve, or other safety limitations.
Radiofrequency ablation (RFA) represents
an alternative therapy, but there are limita-
tions to this treatment modality. It is best
for small lesions that can be fully ablated
by a single placement of the probe (Berber
et al., 2005; Komorizono et al., 2003). Larger
lesions require multiple probe placements
to fully encompass the volume of the target
lesion (Figure 7.1). Unfortunately, ultra-
sound does not clearly delineate the full
extent of the burn and it cannot distinguish
necrotic from viable tissue (Solbiati et al.,
2004). In fact, bubbles occupy the general
region of the heated liver and, if multiple
heat applications are required, the bub-
bles produced by the first burns have often
dissipated by the time the last burns have
been completed. This further impairs the
surgeon’s ability to estimate the extent of
tissue necrosis. Thus, RFA, while a suitable
adjunct in some instances, has technical
limitations that limit its use in the setting
of more extensive lesions.
Intraoperative magnetic resonance imag-
ing (iMRI) represents a potential solution
to the limitations of RFA. Magnetic reso-
nance imaging provides excellent soft tis-
sue contrast and resolution, and it provides
clear visualization of vascular structures.
It detects hepatic tumors with high sensi-
tivity, and is known to be useful for fol-
lowing lesions after RFA (Aliberti et al.,
2004; Limanond et al., 2003). In view of
this, iMRI may be a useful tool for opti-
mizing the application of RFA or other
similar ablative adjuncts.
RADIOFREQUENCY ABLATION
AS A TREATMENT FOR
LIVER TUMORS
Principles of Radiofrequency Ablation
Radiofrequency ablation is a localized
thermal treatment that enables tumor
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-
frequency energy, a high frequency alter-
nating current is passed through the tip of
an electrode into the tissue surrounding
the electrode. The ions within the tissue
are compelled to move in the direction
of the alternating current, causing fric-
tional heating of the tissue. At tempera-
tures above 60°C, intracellular proteins
become denatured and cell death occurs.
Radiofrequency ablation treatments typi-
cally result in local tissue temperatures in
7. Intraoperative Magnetic Resonance Imaging for Radiofrequency Ablation of Hepatic Tumors
excess of 100°C, which produces coagu-
lative necrosis of the tumor tissue and sur-
rounding hepatic parenchyma.
The temperature is greatest near the tips
of the electrode array and falls rapidly
with increasing distance from the elec-
trode. Only small cylindrical regions of
coagulative necrosis are possible when
using simple monopolar needle electrodes.
Modern RFA systems typically consist of
an array of electrodes that deploy in an
umbrella formation. The result is a sphere
of tissue destruction with some viable
cells closer to the periphery of the sphere.
Because the electrode shaft is insulated,
the tract is spared of thermal injury, mak-
ing transcutaneous administration of RFA
possible.
Indications
In general, RFA is indicated in patients
with primary hepatic malignancies and
metastatic liver tumors in the absence
of extrahepatic disease, in whom resec-
tion is not feasible. Reasons for ineligi-
bility for resection might include poor
general medical condition (precluding a
major operation attendant with potentially
considerable blood loss, under general
anesthesia) or insufficient hepatic func-
tional reserve.
Radiofrequency ablation can be admin-
istered percutaneously if lesions are clearly
visible by transcutaneous ultrasound, ena-
bling one to position the probe into the
lesion. Another prerequisite for this approach
is that the lesion is easily accessible by
that route. Lesions that are poorly situated
for the percutaneous approach include
those directly adjacent to the gallbladder or
another hollow viscus, or those lying high
up near the dome of the liver. Percutaneous
83
RFA is typically best for smaller lesions
(i.e., ≤ 3 cm), in the absence of coagulopathy
or thrombocytopenia.
In patients who can tolerate a general
anesthetic, operative (open) RFA is cho-
sen in some circumstances. Larger lesions
typically require multiple deployments
to ablate the entire tumor area and they
are frequently best treated by open RFA.
Lesions that cannot be easily accessed
percutaneously may be treated by an open
procedure. Finally, RFA may be utilized
in conjunction with resection in instances
of bilateral hepatic disease (Abdalla et al.,
2004; Elias et al., 2005).
Hepatocellular carcinoma most fre-
quently occurs in patients with underlying
cirrhosis. Therefore, in this situation, treat-
ment options available to any individual
depend on the degree of liver dysfunction.
The best outcomes have been reported in
patients undergoing liver transplantation
(Llovet, 2005), but the limited supply of
donor livers precludes its widespread use.
Therefore, liver transplantation is typi-
cally reserved for patients with impaired
liver function (i.e., Child B or C) in
whom there is a limited disease burden
(Llovet, 2005). In practical terms, there
is frequently a delay before an organ is
available, so (percutaneous) RFA can be
utilized as a temporizing measure, to avoid
tumor growth beyond the limits of criteria
for transplantation (Lu et al., 2005). Other
treatments that must also be considered
include transarterial chemoembolization
and ethanol ablation.
Colorectal liver metastases infreque-
ntly occur in the setting of cirrhosis,
so hepatic functional reserve is a lesser
concern. However, in patients pre-
treated with chemotherapy, some degree of
hepatic dysfunction may occur secondary
84
to chemotherapy-associated steatohepatitis
(Karoui et al., 2006). Resection is the best
option if all disease can be removed, with
recent reports of 5-year survivals up to
58% (Pawlik et al., 2005). Patients typi-
cally selected for RFA include those who
cannot tolerate surgery; those who have
prohibitive anatomy or small hepatic rem-
nants from previous resections; and those
with bilateral disease (in which case RFA
may be used in conjunction with resection)
(Abdalla et al., 2004; Elias et al., 2005).
There are fewer data on the utility of
RFA in non-colorectal liver metastases.
The exception is in patients with liver
metastases from neuroendocrine tumors.
These tumors are classically more indolent
than most malignancies. Moreover, if they
are functional, greater disease burden is
attendant with worse symptoms that are
more difficult to treat medically. Therefore,
even if it is impossible to ablate all grossly
visible disease, RFA is frequently consid-
ered as a debulking procedure to provide
palliation (Veenendaal et al., 2006).
In each of the above situations, treat-
ment options are diverse and decisions
on the most appropriate treatments are
complex. It is therefore imperative that all
options be considered by a multidiscipli-
nary group that includes hepatic surgeons,
Table 7.1. Outcomes following radiofrequency ablation for hepatocellular carcinoma.
# Patients
Series (# lesions)
Livraghi et al., 1999 42 (52)
Pearson et al., 1999 34
Machi et al., 2001 18 (70)
Bowles et al., 2001 25
Kosari et al., 2002 (16)
Shibata et al., 2002 36 (48)
Bleicher et al., 2003 21
Lencioni et al., 2003 52 (69)
O.F. Bathe and H. Mahallati
hepatologists, medical oncologists, and
interventional radiologists.
Outcomes Associated
with Radiofrequency Ablation
There is a paucity of data on long-term
outcomes following RFA. There are few
randomized controlled trials that com-
pare long-term survivals following RFA
to those seen in resection or other ablative
techniques. What has been the focus of
studies on RFA is the local control rate.
Technical success rates and local recur-
rence rates vary widely in the literature
for colorectal liver metastases and hepa-
tocellular carcinoma (Tables 7.1 and 7.2).
Because technical factors are likely to be
responsible for suboptimal local control,
it is this outcome that is most likely to be
influenced by iMRI.
A number of factors have been identified
as important determinants of local control.
Tumor size appears to be the most impor-
tant factor associated with local recurrence
following RFA, regardless of the histology
(Bleicher et al., 2003; Bowles et al., 2001;
Kosari et al., 2002; Machi et al., 2001;
Mulier et al., 2005). The local recurrence
rate following an open (surgical) approach
is superior to that seen when lesions are
Incomplete Local
ablation (%) recurrence (%) Survival
10
2.9
4.3
14.7 24 months (median)
6.3
4 12
22
9 5.7 98% (2 years)
7. Intraoperative Magnetic Resonance Imaging for Radiofrequency Ablation of Hepatic Tumors
Table 7.2. Outcomes following radiofrequency ablation for colorectal liver metastases.
# Patients
Series (# lesions)
Pearson et al., 1999 46
Solbiati et al., 2001 117 (179)
Machi et al., 2001 (130)
Bowles et al., 2001 39
Kosari et al., 2002 76
Bleicher et al., 2003 59
Aloia et al., 2006 30
(all solitary)
Abdalla et al., 2004 57 (110)
ablated via a percutaneous route (Mulier
et al., 2005). In a meta-analysis, local con-
trol was best for neuroendocrine metastases
(Mulier et al., 2005). In that same study,
local recurrence rates for HCC and colorec-
tal metastases were similar. While several
investigators have demonstrated that prox-
imity of a lesion next to a large vessel is
associated with an increased risk of local
recurrence (Bowles et al., 2001; Machi et al.,
2001) due to a heat sink effect, a meta-
analysis failed to demonstrate the predic-
tive significance of this factor (Mulier
et al., 2005). As a whole, these data sug-
gest that a percutaneous approach should
be reserved for small, easily accessible
lesions. Moreover, adjunctive techniques
should be considered when treating large
lesions near major blood vessels by RFA.
Alternative Ablative Techniques
Cryoablation and ethanol injection are two
alternatives that are becoming less fre-
quently used. RFA is better tolerated than
cryoablation, as there are few if any sys-
temic side effects, and most complications
are minor (de Baere et al., 2003; Pearson
et al., 1999). In HCC, RFA outcomes
appear slightly superior to those seen in
percutaneous ethanol injection (Lencioni
85
Incomplete Local
ablation (%) recurrence (%) Survival
4.3
2 39.1 46% (3 years)
9.2
30.8 25 months (median)
6.6
18.3
37 57% (3 years)
27% (5 years)
9 22% (4 years)
et al., 2003). There are currently two
competing thermal ablation technologies:
microwave coagulation therapy (MCT)
and laser-induced thermotherapy (LITT).
Microwave coagulation therapy involves
ultrasound-guided placement of an elec-
trode into a lesion, followed by microwave
treatment. Laser-induced thermotherapy
involves placement of a laser catheter into
a lesion and heat photocoagulation using a
Nd-YAG laser. Each of these technologies
can be applied during laparotomy or by
percutaneous techniques.
In general, MCT and LITT are advanta-
geous in that multiple electrodes or cath-
eters can be simultaneously placed to ablate
larger or more complex liver lesions. Success
rates (i.e., complete ablation rates and local
recurrence rates in primary or metastatic
tumors) are reportedly similar with each of
these technologies and with RFA. However,
there are few direct comparisons between
RFA and LITT or MCT. In one randomized
controlled trial, RFA and MCT were found
to be equally safe; RFA and microwave
ablation were approximately equivalent in
their ability to completely ablate tumor, but
microwave ablation on average took less
time to administer (Shibata et al., 2002).
More such comparisons will be essential
for progress in this field.
86
PROBLEMS ASSOCIATED
WITH MONITORING
THE ABLATIVE ENDPOINT
So far few clinically applicable and reli-
able methods have been devised to moni-
tor the effectiveness of RFA, as reflected
by the complete necrosis of a given lesion.
Conventional methods of evaluating the
completion of a thermal treatment include
monitoring for target temperatures or sud-
den increases in impedance, as well as
estimating the extent of injury with ultra-
sound. While each of these methods of
estimating the adequacy of RFA is suf-
ficient for small lesions, their utility is DEFINING THE EXTENT
limited when treating larger lesions.
Impedance and Temperature
Temperature and impedance have each
been used as surrogate endpoints for an
ablation applied to any given region (Arata
et al., 2001; Frich et al., 2004). Indeed,
for small lesions, a sudden increase in
tissue impedance (“roll-off”) is a significant
predictor of local control after RF ablation
(Arata et al., 2001). However, when mul-
tiple deployments are required to cover a
larger lesion (Figure 7.1), impedance and
temperature become less relevant end-
points, and it becomes important to be
able to gauge whether the whole lesion
has been ablated. Anatomic correlates of
tissue kill therefore become necessary.
Ultrasound
Ultrasound is typically utilized to tar-
get a given lesion. Unfortunately, ultra-
sound does not clearly delineate the full
extent of the burn, and it cannot distin-
guish necrotic from viable tissue (Solbiati
et al., 2004). Rather, hyperechoic bubbles
O.F. Bathe and H. Mahallati
occupy the general region of the heated
liver. When ablating larger lesions, where
multiple deployments are necessary, the
utility of ultrasound for estimating extent
of the thermal injury becomes even more
limited. That is, the bubbles produced by
the first burns have often dissipated by the
time the last burns have been completed,
further impairing the surgeon’s ability
to estimate the extent of tissue necrosis.
Ultrasound, therefore, represents a poor
tool for estimating the extent of the RFA
thermal injury.
OF THERMAL INJURY BY
MAGNETIC RESONANCE
IMAGING
Magnetic Resonance Imaging
Characteristics of Thermal Injury
Magnetic resonance imaging is an excel-
lent imaging modality for assessment
of hepatic lesions (Martin and Semelka,
2001; Noone et al., 2004). It has a number
of advantages over CT imaging, such as
lack of ionizing radiation, less toxicity
with commonly used contrast agents, and
greater tissue contrast resolution for liver
lesions (Kirchin and Runge, 2003; Martin
and Semelka, 2001). Magnetic resonance
imaging has proven to be particularly use-
ful for the evaluation of the adequacy of
thermal ablation and for monitoring for
recurrence. Indeed, MR images reflect
complete histopathological destruction of
tissue, although tumor has not been spe-
cifically analyzed (Lee et al., 2001).
Most modern MR hepatic imaging
protocols rely on a combination of T1
and T2 weighted images, with or without
7. Intraoperative Magnetic Resonance Imaging for Radiofrequency Ablation of Hepatic Tumors 87

administration of contrast agents. Using

the combined information from noncon-
trast T1 and T2 signal intensities and the
dynamic enhancement pattern of lesions
over time, lesions can be characterized
with great accuracy (Noone et al., 2004).
Excellent image quality requires that
motion artefacts from breathing be mini-
mized. This is usually achieved by having
patients suspend respiration for 20–25 s
while image data are acquired. Recent
advances enable the acquisition of high
quality images with patients breathing
freely (Asbach et al., 2005).
Thermal ablative treatments have been
shown to change the MR properties (T1
and T2 relaxation times) of tissues. Hepatic
parenchyma surrounding a site of ther-
mal injury initially contains inflammatory
changes, which gradually subside to be
replaced by granulation tissue (Goldberg
et al., 2000; Kuszyk et al., 2000). These Figure 7.2. Post-ablation changes 2 months after
changes manifest as an ill-defined rim RFA. The T1 opposed phase image (a) and the STIR
of mild signal increase on T2 weighted (T2 weighted) image (b) demonstrate hemorrhagic
contents within the RFA bed (white arrows)
images and on arterial phase enhance-
ment, which can be intense (Dromain
et al., 2002; Kuszyk et al., 2000). Over
time, the intensity and thickness of rim images, and liquefactive necrosis appears
enhancement decrease; these changes usu- as high signal (Kuszyk et al., 2000).
ally disappear by 6 months (Braga and Defining complete responses follow-
Semelka, 2005; Goldberg et al., 2000; ing RFA is unlike the assessment of the
Kuszyk et al., 2000). degree of response for any other type of
The ablated region similarly shows dis- antineoplastic therapy. That is, WHO and
tinct changes (Figure 7.2). In the first RECIST criteria for response to cancer
week after therapy, it is typically filled therapy rely on a decrease in size of lesions
with hemorrhagic debris from liquefactive (Therasse et al., 2000). On the other hand,
or coagulative necrosis (Dromain et al., the aim of RFA and other thermal ablative
2002; Kuszyk et al., 2000). Non-contrast techniques is to produce an area of ther-
T1 weighted images typically show a mocoagulation greater than or equal to the
high signal area with some heterogeneity, size of the tumor. Therefore, conventional
depending on the stage of hemorrhage. size criteria are not appropriate measures
Coagulative necrosis typically appears as of response to RFA (Dromain et al., 2002).
low signal on non-contrast T2 weighted Early series showed that contrast enhanced
88 O.F. Bathe and H. Mahallati

CT and MRI are useful modalities for

assessing the size of the ablated area.
The non-enhancing region corresponded
to within 2 mm of the size of the area
of coagulative necrosis (Goldberg et al.,
2000). Therefore, MRI clearly defines the
completeness of necrosis and the extent of
the necrotic injury.
Subtraction images are also useful
adjuncts for MRI. Specifically, subtrac-
tion images in which precontrast (mask)
images are subtracted from postcontrast
(enhanced) images can better demonstrate
areas of enhancement. Such images have
been shown to be useful in assessing
hepatic lesions (Yu et al., 2005). In our
experience, application of subtraction
imaging greatly improves one’s ability to
determine the extent of thermal injury and
the characteristics of enhancement around
post-RFA beds.
Magnetic resonance imaging is similarly Figure 7.3. Local recurrence of hepatocellular car-
useful for monitoring lesions for local recur- cinoma following radiofrequency ablation. Axial
rence (Figure 7.3). It is slightly superior to and coronal reformatted post-contrast (Gd-DTPA)
CT in its sensitivity for detecting recur- subtraction images show nodular enhancement at
the superior peripheral margin of the RFA bed of a
rence (Dromain et al., 2002). In the first large lesion (white arrows)
few months after RFA, size criteria are not
particularly helpful. Rather, the margins of
the RFA bed must be evaluated. A smooth
internal contour and a smooth, thin rim-like Enhancing the Definition of Extent
enhancement of the margins are expected of Thermal Injury
around an ablated cavity. The presence of Magnetic resonance contrast agents are
nodular enhancement at the periphery of essential for the accurate definition of the
the ablation zone, or any protuberance dis- extent of thermal injury. However, conven-
torting the smooth inner margin are signs tional MR contrast agents are not ideal in
of tumor recurrence (Braga and Semelka, the setting of iMRI. Therefore, the utility
2005; Goldberg et al., 2000; Kuszyk et al., of newer contrast agents must be explored
2000). After the first few months, one may in the context of iMRI. Liver MRI is most
expect a progressive reduction in the size of often performed with contrast agents that
the ablative lesion (Dromain et al., 2002). are extracellular space agents (Hyslop
Any increase in size of the ablative zone is et al., 2005). These are comprised of solu-
considered a sign of recurrence (Braga and tions of gadolinium (Gd) chelates that
Semelka, 2005; Limanond et al., 2003). are rapidly and almost entirely cleared
7. Intraoperative Magnetic Resonance Imaging for Radiofrequency Ablation of Hepatic Tumors 89

by renal excretion. Enhancement with

such agents reflects tissue vascularity,
flow dynamics, and perfusion (Weinmann
et al., 1991). Liver MRI scans are typically
performed with three postcontrast phases:
arterial, portal venous, and equilibrium (or
interstitial) phases. Conventional extracel-
lular space agents are useful in determin-
ing enhancement of post-RFA beds and
they are helpful for the detection of tumor
recurrence (Dromain et al., 2002; Kuszyk
Figure 7.4. Hepatocellular carcinoma following
et al., 2000). administration of a hepatocyte-specific contrast
Tissue-specific agents for hepatic imag- agent. With Gd-EOB-DTPA (Primovist), hepato-
ing include hepatocyte-specific and retic- cytes are T1 hyperintense and tumor is hypointense
ulo-endothelial system-specific agents. on this opposed phase T1 weighted image
Early hepatocyte-specific agents were
based on manganese (e.g., MnDPDP),
have functioning hepatocytes (Huppertz
but they are not in widespread clinical use
et al., 2004). If a lesion does not contain
(Hyslop et al., 2005). While these agents
functioning hepatocytes, it will be hypoin-
provided excellent visualization of the bil-
tense on T1W images compared to func-
iary tree and hepatic parenchyma, the lack
tioning hepatocytes that have prolonged
of dynamic contrast enhancement infor-
signal increase because of uptake of Gd
mation limits their clinical utility.
compounds (Figure 7.4). Gd-BOPTA
Recently, a new class of contrast agents
results in maximal hepatic parenchymal/
has been introduced that has the properties
biliary enhancement at approximately
of both extracellular space agents as well as
1–2 h after injection and Gd-EOB-DTPA
hepatocyte-specific agents. These include
has the highest liver-to-lesion contrast at
gadobenate dimeglumine (Gd-BOPTA)
20–45 min after injection (Huppertz et al.,
and gadoxetic acid (Gd-ethoxybenzyl-
2004; Hyslop et al., 2005). In the case
diethylenetriaminepentaacetic acid;
of Gd-EOB-DTPA, despite the fact that
Gd-EOB-DTPA). Both contrast agents are
contrast between lesion and normal liver
distributed in the extracellular space in
is maximal at 20–45 min, this temporal
the early stages after injection and offer
window has been shown to persist for at
dynamic contrast enhancement data, and
least 90–120 min. The use of these contrast
both are ultimately taken up by function-
agents in the context of monitoring ther-
ing hepatocytes with subsequent biliary
mal damage from RFA has been relatively
excretion. Gd-BOPTA has 3–5% biliary
unexplored.
excretion and approximately 50% of
Gd-EOB-DTPA is excreted in the bile
(Hyslop et al., 2005). The hepatocyte- Potential Utility of Functional
specific uptake results in T1 shortening of Magnetic Resonance Imaging
the liver parenchyma signal and results in While there are good MRI criteria for
increased conspicuity of lesions that do not complete tumor ablation, functional MR
90
imaging may also be useful in supporting
the lack of significant residual tumor after
ablation. The technology exists to monitor
intratumoral temperature by MRI (Dick
et al., 2003; Puls et al., 2003; Weidens-
teiner et al., 2004). While functional MR
techniques have been used in neuroimaging
for some time, it is only relatively recently
that these techniques have become appli-
cable in abdominal imaging. Magnetic
resonance spectroscopy has been used to
evaluate patients with prostate cancer and
renal cell carcinoma (Hyslop et al., 2005;
Katz-Brull et al., 2005). Kamel et al.
(2006) used functional MR techniques as
an adjunct to conventional MR imaging to
assess response to chemotherapy (TACE)
in patients with HCC. Early results are
promising, and MR imaging has the unique
advantage of offering a combination of
anatomic as well biochemical informa-
tion in assessing tumor response. These
combined approaches may offer improved
sensitivity and specificity in assessing
response.
INTRAOPERATIVE
MAGNETIC RESONANCE
IMAGING AS AN ADJUNCT
TO RADIOFREQUENCY
ABLATION
Rationale
As described earlier, RFA is a promis-
ing technology that enables ablation of
lesions that may otherwise not be resect-
able. However, technical limitations limit
the efficacy of the treatment. Local failure
rates and local recurrence rates are unac-
ceptably high, particularly in larger lesions
that are difficult to access. Ideally, local
O.F. Bathe and H. Mahallati
recurrence rates should approach those
seen after resection.
Lesions are often difficult to visualize
in their entirety with diagnostic modali-
ties other than MRI. This is especially
true for some neuroendocrine tumors and
some HCCs. Monitoring the extent of
the ablation is similarly difficult, as the
ablated region is immediately filled with
hyperechoic bubbles. The difficulty is
compounded in large lesions, where mul-
tiple overlapping RFA applications must
be utilized to cover a lesion larger than
the probe (Figure 7.1). In this instance,
treatment is planned such that the zones
of necrosis overlap, to ensure complete
destruction of the tumor. However, during
the latter RFA applications, the bubbles
disappear in the initially ablated regions,
making it impossible to keep track of the
total extent of the burn. The area of the
burn is not always consistent, particularly
if there is a nearby large vessel which
induces a heat sink. Finally, the surgeon
cannot always accurately reach the desired
target with the RFA probe. All of these
obstacles can limit the surgeon’s capability
to completely ablate any lesion, particu-
larly with adequate margins (Mulier et al.,
2005). Improved intraoperative imaging
may help to promptly identify local treat-
ment failures, reducing the need for repeat
treatments and, perhaps, reducing local
recurrence rates.
Technical Developments
The first intraoperative MRI systems were
relatively recently described (Jolesz and
Blumenfeld, 1994; Schenck et al., 1995),
and operative interventions guided by
these systems were pioneered by the neu-
rosurgical community. The first systems
7. Intraoperative Magnetic Resonance Imaging for Radiofrequency Ablation of Hepatic Tumors
consisted of 2 weak magnets (i.e., ≤ 0.5
tesla (T) ), which were arranged in a verti-
cal biplanar donut configuration (Schenck
et al., 1995). The more recent iterations of
these systems permitted excellent vertical
and lateral exposure between magnets.
Subsequently, horizontal low-field sys-
tems were developed (Steinmeier et al.,
1998), but patients had to be transported
between an operating theater and an imag-
ing room. In this way, MR incompatibility
of surgical instruments was no longer an
issue, but patient safety was suboptimal
due to the need for transportation during
the procedure. Using such systems, virtu-
ally any anatomical region could be easily
reached for interventional procedures, as
the patient could be placed supine, prone
or lateral along the longitudinal axis of
the magnets. However, such systems had
limited imaging capabilities and resolution
due to the low magnetic field strength.
Stronger magnetic field strength increa-
ses signal-to-noise ratio, enhances image
resolution, and accelerates image acquisi-
tion. Therefore, “high-field” imaging sys-
tems incorporating stronger magnets were
developed (Sutherland et al., 2002; Truwit
and Hall, 2001). These 1.5T systems pro-
duced image quality that was equivalent to
diagnostic scanners. Moreover, the higher
magnetic field strength enabled physi-
ologic MR imaging, such as perfusion
MR, spectroscopic MR, and functional
MR. However, because of the enhanced
magnetic field strength, it was not possible
to perform procedures that were moni-
tored “real-time” by MRI. Rather, the sys-
tems were limited to taking static images
before, during and after surgery.
At our institution, the neurosurgeons
developed a unique iMRI system that is
comprised of an operating theatre in which
91
the 1.5T magnet remains retracted except
when it is utilized for imaging (Sutherland
et al., 2002). The use of a movable magnet
is more “patient focused” than previously
developed systems. That is, the operating
table is fixed, the patient is positioned
once, and the magnet moves to the patient,
as opposed to transporting the patient to
the magnet. When the magnet is retracted,
conventional surgical instruments can be
utilized behind the 5-Gauss line. This
patient-focused approach is safer in that
it minimizes the anesthetic risk associated
with patient transport, and it minimizes
disruption of the surgical procedure.
Recently, we have demonstrated the fea-
sibility of doing complex open abdominal
operations using this system (Bathe et al.,
2006). Major hepatic surgery would not be
readily feasible in a high-field open magnet
system, given the equipment requirements
and access constraints. Our ability to do
surgery with a 1.5T magnet was facilitated
by the portability of the magnet. The one
drawback of a closed unit such as the one
we use is that it does not allow real-time
monitoring of the ablation or changes in
intralesional temperature, which is fea-
sible for percutaneous procedures (Dick
et al., 2003; Puls et al., 2003). Therefore,
a moveable, retractable system is essential
for safely performing high-field imaging
in the setting of open intracavitary surgery,
but real-time information is not available.
Others have described minimally inva-
sive intraabdominal and hepatic proce-
dures in low magnetic field systems (i.e.,
0.2–0.5T) (Dick et al., 2003; Morikawa
et al., 2002). Typically, these systems
employ open magnets that enable access
for percutaneous procedures. The imag-
ing limitations imposed by such low-field
systems limit their use in hepatic surgery.
92
Moreover, with a 1.5T magnet temperature
mapping of the thermal burn is possible
(Weidensteiner et al., 2004), which poten-
tially provides additional important intra-
operative information. Recently, a report
demonstrated that percutaneous RFA was
possible through a closed high-field (1.5T)
system (Mahnken et al., 2004), but open
intracavitary surgery with such a strong
magnet has not so far been reported out-
side of our own experience.
Potential Indications
Patients currently selected for iMRI as an
adjunct to surgery are those in whom it is
anticipated that RFA (with or instead of
resection) is a likelihood and that iMRI
might be helpful in making intraopera-
tive decisions. Specific indications include
the following situations: (1) large lesions
(≥ 3 cm) that will be treated by RFA; (2)
recurrent lesions in which RFA will play
an essential role; (3) lesions that are vis-
ible on MRI; (4) lesions located at difficult
sites, such as near potential heat sinks,
close to the gallbladder or any other hol-
low viscus, or up near the diaphragm.
The present closed iMRI systems utilize
magnets with a small bore. Therefore,
large patients who did not fit into a closed
magnet system cannot undergo this proce-
dure, until larger bore magnets are devel-
oped. Open systems may be less limited in
this regard, although it is essential that any
lesions targeted by this approach are vis-
ible in the therapeutic magnet. Frequently,
in patients with a larger body habitus,
lesions may be suboptimally visible if
they lie outside of the magnetic “sweet
spot”. Patients who had ferrous prostheses
or other ferrous foreign bodies that would
normally be a contraindication to diagnostic
O.F. Bathe and H. Mahallati
MRI are similarly not eligible for this type
of approach. Finally, because intraopera-
tive cardiac monitoring is impaired by the
iMRI system, patients with significant cor-
onary artery disease are currently excluded
from these procedures.
Conduct of Intraoperative
Magnetic Resonance Imaging
for Radiofrequency Ablation
The iMRI system that is used at the
University of Calgary has been described
in previous reports (Bathe et al., 2006;
Sutherland et al., 2002). Briefly, the sys-
tem is based on a moveable 1.5T mag-
net designed and constructed by Magnex
UK. The system enables the positioning
of the patient and, because the magnet
enters and exits the room as required,
the patient remains the focal point. The
intubated patient is positioned on a MR
compatible table that is bolted to the
floor, with the anesthetic machine at his/
her foot. The hands must be tucked to
the side to accommodate the patient in the
magnet. An inflatable positioning device
is inserted under the right flank to elevate
the patient’s right side slightly, enabling
a more complete right subcostal incision,
which is essential for open hepatic surgery.
The anesthetic tubes are passed along the
left flank to the feet, where the anesthetic
equipment resides as far from the magnet
as feasible.
Following induction of anesthesia, base-
line images are obtained and compared to
a diagnostic MRI taken before the date of
surgery. Sequence parameters are set to
minimize breath-hold time and maximize
coverage. Time to repetition (TR), echo
time (TE), flip angle, echo train lengths
(ETL), and slice thickness are all adjusted
7. Intraoperative Magnetic Resonance Imaging for Radiofrequency Ablation of Hepatic Tumors
to optimize imaging of the liver in each
patient. All patients are imaged with res-
piration suspended and monitored by the
attending anesthesiologist. Axial precon-
trast 2D T1 weighted gradient echo (GRE)
images are obtained with or without fat
saturation, and axial 2D fast spin echo
(FSE) T2 weighted images with fat satura-
tion are obtained. An intravenous bolus
of 0.15 mmol/kg gadopentetate dimeg-
lumine (Gd-DTPA; Magnevist; Berlex
Laboratories, Wayne, NJ) is then admin-
istered, and postcontrast T1 weighted
fat saturated images are acquired, with
parameters identical to the precontrast T1
weighted fat saturated images. All target
lesions are visualized, measured, and
localized relative to nonmobile extrahe-
patic landmarks (e.g., vertebral body and
liver capsule). In particular, it is important
to measure the distances from these non-
mobile landmarks to the tumor edges. The
mobile magnet is then retracted, allowing
the safe introduction of normal surgical
instruments to the surgical field.
Following a laparotomy, the liver is
mobilized as necessary to permit access
to the targeted hepatic segments. An intra-
operative ultrasound is performed, and
decisions regarding the final operative
approach are made. It is feasible using
this system to do a resection, if that is
considered appropriate. RFA is performed
for any remaining lesions, using as large a
probe as required to cover as much of the
tumor field as possible. Currently, probes
up to 5 cm are available. If it is expected
that the probe will cover the entire diam-
eter of the tumor in a single deployment,
the probe is initially guided with the aid of
ultrasound to the center of the tumor. Once
the probe tip is at the desired region of the
tumor, the tines are deployed and a final
93
check of probe position is made with ultra-
sound. The protocol for increases in elec-
trical power delivered to the probe is set
by the manufacturer, and this is dependent
on probe size. Larger probes require more
power to achieve the target temperature
throughout the volume of the burn region.
Where multiple deployments are required
to cover the entire tumor volume, it is best
to do the first deployments at the least
accessible part of the tumor. Usually, that
means targeting the most posterior and
superior aspects of the lesion. Thereafter,
the probe is pulled back 2 cm, the tines
are redeployed, and another thermal abla-
tion cycle is applied. This is repeated
until it is estimated that the entire lesion
has been ablated. It is best if the regions
ablated overlap, to ensure that no areas
are exposed to insufficient heat to induce
coagulative necrosis. The latter ablations
are more difficult to target accurately with
ultrasound, as the lesion fills with bubbles
after the first ablation. Therefore, if it is
anticipated that multiple deployments will
be required, it is essential to plan probe
placements prior to initiating the ablation
for each lesion.
In general, we take an aggressive
approach to thoroughly burn each of the
target lesions. For lesions that are close to
the gallbladder, it is essential to remove
the gallbladder to avoid perforation sec-
ondary to thermal injury. For lesions that
are located near the center of the liver, care
must be taken to avoid a thermal injury
to the bile ducts. Injury to the bile ducts
may be manifested postoperatively by a
bile leak, a bile duct stricture, or complete
disruption of the bile ducts; such compli-
cations add considerable morbidity to the
procedure. If necessary, the bile ducts can
be infused with cooled saline via the cystic
94
duct, to reduce the chance of this type of
injury. All targeted lesions are resected
and/or ablated prior to MRI assessment of
the margins.
After resection and/or RFA, all instru-
ments are counted; metal objects are
removed from the field, and the operating
field is covered with protective drapes.
An intradissectional image is obtained,
to ensure complete disappearance of the
target lesion(s) with adequate margins. We
strive to burn the entire area of the tumor
plus a margin of normal tissue of at least
5 mm. This can be estimated by measur-
ing the distance from selected nonmo-
bile neighboring structures to the various
margins of the burn and comparing these
distances to those measured before the
surgery, where the edge of the tumor was
mapped. If necessary, more ablative treat-
ments are applied, and further confirma-
tory images are taken.
Digital subtraction images consisting of
a mathematical subtraction of T1-weighted
data acquired before and after IV contrast
facilitate estimation of the burn extent
and determination of whether viable
tumor remains. Such images pronounce
enhanced structures like residual tumor:
when the center of the lesion is a signal
void (i.e., black), this confirms that there
is no residual tumor (Figures 7.5D, E). In
our early work, following administration
of Gd-DTPA, in cases where repeat burns
were needed, repeat imaging was limited
by dose (Bathe et al., 2006).
Most recently we have employed
Gd-EOB-DTPA as a contrast agent (Figure
7.5). The prolonged hepatocyte uptake of
this agent offers an extended temporal
window which can be used to advan-
tage in the intraoperative imaging. The
90–120 min temporal window offered by
O.F. Bathe and H. Mahallati
Gd-EOB-DTPA allows visualization of
the lesion prior to RFA and precise deline-
ation of RFA margins, even after multiple
time-consuming RFA deployments.
When using Gd-EOB-DTPA as the con-
trast agent, it is administered 20–120 min
prior to taking intradissectional images.
If possible, the same dose of contrast
is utilized for preoperative imaging and
intradissectional imaging. The preabla-
tion size of the lesion is easily determined
on T1-weighted imaging by measuring
the hypo-intense lesion diameters against
the hyperintense hepatic parenchyma. The
ablation zone can then be measured on T1
and T2-weighted images, ensuring that
it exceeds the size of the original lesion.
In our limited experience, on delayed
(90 min) post-Gd-EOB-DTPA images, the
ablation bed is much more hyperintense
than hepatic parenchyma (Figure 7.5B).
This corresponds to the hemorrhagic and
necrotic products that are known to fill
the recently ablated region (Braga and
Semelka, 2005). The use of Gd-EOB-
DTPA eliminates the need for multiple
doses of contrast agent with multiple RFA
attempts. In addition, subtraction images
are no longer required to estimate the
extent of thermal injury.
Several technical limitations with our
system have become apparent during our
development of this technology for hepatic
surgery. First, since our iMRI system was
primarily engineered for brain and spine
surgery, the magnetic isocenter (“sweet
spot”) was posterior to the liver. Therefore,
more anterior lesions were difficult to
see because of poor signal-to-noise ratio.
A modified RF coil is required to address
this problem. Second, lesions in segment
IV of the liver can be obscured by the
aortic “ghost” artefact. If the lesion is veiled
7. Intraoperative Magnetic Resonance Imaging for Radiofrequency Ablation of Hepatic Tumors 95

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
Figure 7.6. Lesions in segment IV of the liver
(white arrow) can be obscured by the aortic “ghost”
artefact. If the lesion is veiled by aortic ghost, then
tilting the patient effectively separates the target
lesion from the artefact.
by aortic ghost, then tilting the patient
effectively separates the target lesion from
the artefact (Figure 7.6). For this reason,
we routinely place air bags beneath each
flank, enabling repositioning with less
trouble. Third, we have found that image
quality is unacceptable in large patients.
Not only is it difficult to fit these patients
into the magnet, it is impossible to image
the anterior aspect of their liver. Each of
these limitations can be addressed with
specific design changes to the magnet (see
below).
Future Developments
Image-guided ablation is an important
alternative to resection. However, to
ensure that its therapeutic effectiveness
better resembles that of resection, specific
technological developments will be neces-
sary. This is particularly important for the
treatment of larger lesions, which have so
far infrequently been treated using these
techniques.
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
rapid, high quality imaging with combined
morphologic and functional information.
Systems must be built with a stronger
magnetic field (e.g., ≥ 3.0T), as this ena-
bles faster image acquisition, greater reso-
lution, and better functional imaging. On
the other hand, MRI using stronger mag-
nets is associated with some disadvantages
such as increased susceptibility artefacts,
increased energy deposition, and altered
T1 properties of tissues (Hyslop et al.,
2005; Lauzon et al., 2006). In the case of
intraoperative MR, the increased suscep-
tibility artefacts warrant special attention.
The presence of air/tissue interfaces in the
surgical field can lead to signal loss from
susceptibility and may decrease the image
quality, negating the advantages of high
field imaging. Sequences and systems will
need to be optimized to minimize artefacts
in order to gain the maximum potential
from higher field intra-operative systems.
Systems should ideally offer targeting
capabilities and real-time monitoring of
therapies such as RFA. Magnetic resonance
has been used to monitor RF ablation in
real-time in animal models (Lepetit-Coiffe
et al., 2006), but this is not yet feasible in
high-field systems so far described in clin-
ical use. Real-time MR monitoring of RFA
would be greatly enhanced by robotics,
7. Intraoperative Magnetic Resonance Imaging for Radiofrequency Ablation of Hepatic Tumors
which would allow the remote insertion
of instruments into the lesion (McBeth
et al., 2004). Robots are indefatigable,
have superior spatial resolution and excel-
lent geometric accuracy. The development
of an MR-compatible robot would there-
fore facilitate real-time monitoring of RFA
of liver lesions.
Finally, iMRI could benefit from the
development of novel contrast agents.
While conventional extracellular space
agents are useful in determining enhance-
ment of post-RFA beds and potential tumor
recurrence (Dromain et al., 2002; Kuszyk
et al., 2000), in our experience, the tem-
poral window offered by such agents and
the dose limitations (Kirchin and Runge,
2003) limit their utility in the intraopera-
tive setting. Ideally, a contrast agent that
has greater retention and better definition
of the target lesion relative to the extent of
thermal injury would be useful. Gd-EOB-
DTPA fulfills these requirements, but it
may not be as good for lesions containing
functioning hepatocytes, such as hepatic
adenoma or well differentiated hepatocel-
lular carcinoma. Tumor-specific contrast
agents may be useful. Necrosis-avid agents
and high molecular weight intravascular
agents have been used in animal models
to assess response to RFA therapy (Kim
et al., 2005; Ni et al., 2006), but we have
not yet tested them in the clinical context.
Development of Alternative Ablative
Techniques
Radiofrequency ablation has some dis-
advantages, when utilized in the context
of MRI. If real-time monitoring of the
ablation is to be done using MRI, then it
is essential to utilize MR-compatible tita-
nium probes, which are quite expensive.
Moreover, RFA produces electromagnetic
97
noise, impairing the acquisition of quality
images when monitoring the progression
of the burn. To address the latter issue,
Zhang et al. (1998) have described a
switching circuit to reduce this noise.
While we have only had experience with
RFA, MCT and LITT could be explored
in the context of MR-guided ablation of
liver lesions. The MRI characteristics of
ablation zones produced by each of these
methods are unlikely to differ significantly
from those produced by RFA (Braga and
Semelka, 2005). MR-guided insertion of
microwave electrodes and real-time moni-
toring of the ablation by MRI has success-
fully been performed (Morikawa et al.,
2002). Laser-induced thermotherapy may
be particularly suited for MR imaging
because laser fibers and catheter systems
are not affected by and do not interfere with
MR signal (Puls et al., 2003). Puls et al.
(2003) recently described a new miniatur-
ized irrigated laser catheter system that
can be inserted percutaneously under MR
guidance and that allows monitoring of
therapy in a closed high-field MR imager.
In conclusion, radiofrequency ablation
may occasionally cause incomplete tumor
necrosis. Moreover, the local recurrence
rates following RFA of liver tumors are
high, especially in larger tumors that are
difficult to access. One contributing fac-
tor to this poor local control rate is the
difficulty in monitoring thermal ablations
of larger tumors in which multiple RFA
deployments are required. Additionally,
the margins of ablation may be inadequate.
Intraoperative MRI enables the surgeon
to better estimate the technical success of
RFA as well as the size of the ablative mar-
gins. However, it is still unknown whether
iMRI changes local recurrence rates.
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
comparison of gadoxetic acid-enhanced MR
images with intraoperative findings. Radiology
230:266–275.
Hyslop, W.B., Balci, N.C., and Semelka, R.C.
2005. Future horizons in MR imaging. Magn.
Reson. Imaging Clin. N. Am. 13:211–224.
Jolesz, F.A., and Blumenfeld, S.M. 1994.
Interventional use of magnetic resonance imag-
ing. Magn. Reson. Q. 10:85–96.
Kamel, I.R., Bluemke, D.A., Eng, J., Liapi, E.,
Messersmith, W., Reyes, D.K., and Geschwind,
J.F. 2006. The role of functional MR imaging in
the assessment of tumor response after chem-
oembolization in patients with hepatocellular
carcinoma. J. Vasc. Interv. Radiol. 17:505–512.
Karoui, M., Penna, C., Amin-Hashem, M., Mitry,
E., Benoist, S., Franc, B., Rougier, P., and
Nordlinger, B. 2006. Influence of preoperative
chemotherapy on the risk of major hepatec-
tomy for colorectal liver metastases. Ann. Surg.
243:1–7.
Katz-Brull, R., Rofsky, N.M., Morrin, M.M.,
Pedrosa, I., George, D.J., Michaelson, M.D.,
Marquis, R.P., Maril, M., Noguera, C., and
Lenkinski, R.E. 2005. Decreases in free choles-
terol and fatty acid unsaturation in renal cell car-
cinoma demonstrated by breath-hold magnetic
resonance spectroscopy. Am. J. Physiol. Renal
Physiol. 288:F637–F641.
Kim, T.J., Moon, W.K., Cha, J.H., Goo, J.M.,
Lee, K.H., Kim, K.H., Lee, J.W., Han, J.G.,
Weinmann, H.J., and Chang, K.H. 2005. VX2
carcinoma in rabbits after radiofrequency abla-
tion: comparison of MR contrast agents for
help in differentiating benign periablational
enhancement from residual tumor. Radiology
234:423–430.
Kirchin, M.A., and Runge, V.M. 2003. Contrast
agents for magnetic resonance imaging: safety
update. Top Magn. Reson. Imaging 14:426–435.
Komorizono, Y., Oketani, M., Sako, K., Yamasaki,
N., Shibatou, T., Maeda, M., Kohara, K.,
Shigenobu, S., Ishibashi, K., and Arima, T.
2003. Risk factors for local recurrence of small
hepatocellular carcinoma tumors after a single
session, single application of percutaneous radi-
ofrequency ablation. Cancer 97:1253–1262.
Kosari, K., Gomes, M., Hunter, D., Hess, D.J.,
Greeno, E., and Sielaff, T.D. 2002. Local, intra-
hepatic, and systemic recurrence patterns after
99
radiofrequency ablation of hepatic malignancies.
J. Gastrointest. Surg. 6:255–263.
Kuszyk, B.S., Boitnott, J.K., Choti, M.A., Bluemke,
D.A., Sheth, S., Magee, C.A., Horton, K.M.,
Eng, J., and Fishman, E.K. 2000. Local tumor
recurrence following hepatic cryoablation: radi-
ologic-histopathologic correlation in a rabbit
model. Radiology 217:477–486.
Lauzon, M.L., Mahallati, H., and Frayne, R. 2006.
Time-efficient breath-hold abdominal MRI at
3.0 T. Am. J. Roentgenol. 187:649–657.
Lee, J., Lee, J., Kim, S., Kim, C., and Mun, W.
2001. MR imaging-histopathologic correlation
of radiofrequency thermal ablation lesion in a
rabbit liver model: observation during acute and
chronic stages. Korean J. Radiol. 2:151–158.
Lencioni, R., Allgaier, H.-P., Cioni, D., Olchewski, M.,
Deibert, P., Crocetti, L., Frings, H., Laubenberger,
J., Zuber, I., Blum, H., and Bartolozzi, C. 2003.
Small hepatocellular carcinoma in cirrhosis: ran-
domized comparison of radio-frequency thermal
ablation versus percutaneous ethanol injection.
Radiology 228:235–240.
Lepetit-Coiffe, M., Quesson, B., Seror, O., Dumont,
E., Le Bail, B., Moonen, C.T., and Trillaud, H.
2006. Real-time monitoring of radiofrequency
ablation of rabbit liver by respiratory-gated
quantitative temperature MRI. J. Magn. Reson.
Imaging 24:152–159.
Limanond, P., Zimmerman, P., Raman, S.S.,
Kadell, B.M., and Lu, D.S. 2003. Interpretation
of CT and MRI after radiofrequency ablation
of hepatic malignancies. Am. J. Roentgenol.
181:1635–1640.
Livraghi, T., Goldberg, S.N., Lazzaroni, S., Meloni,
F., Solbiati, L., and Gazelle, G.S. 1999. Small
hepatocellular carcinoma: treatment with radio-
frequency ablation versus ethanol injection.
Radiology 210:655–661.
Llovet, J.M. 2005. Updated treatment approach
to hepatocellular carcinoma. J. Gastroenterol.
40:225–235.
Lu, D.S., Yu, N.C., Raman, S.S., Lassman, C.,
Tong, M.J., Britten, C., Durazo, F., Saab, S.,
Han, S., Finn, R., Hiatt, J.R., and Busuttil, R.W.
2005. Percutaneous radiofrequency ablation of
hepatocellular carcinoma as a bridge to liver
transplantation. Hepatology 41:1130–1137.
Machi, J., Uchida, S., Sumida, K., Limm, W.M.,
Hundahl, S.A., Oishi, A.J., Furumoto, N.L.,
100
and Oishi, R.H. 2001. Ultrasound-guided radi-
ofrequency thermal ablation of liver tumors:
percutaneous, laparoscopic, and open surgical
approaches. J. Gastrointest. Surg. 5:477–489.
Mahnken, A., Buecker, A., Spuentrup, E.,
Krombach, G., Henzler, D., Gunther, R., and
Tacke, J. 2004. MR-guided radiofrequency abla-
tion of hepatic malignancies at 1.5 T: initial
results. J. Magn. Reson. Imaging 19:342–348.
Martin, D.R., and Semelka, R.C. 2001. Imaging of
benign and malignant focal liver lesions. Magn.
Reson. Imaging Clin. N. Am. 9:785–802.
McBeth, P.B., Louw, D.F., Rizun, P.R., and
Sutherland, G.R. 2004. Robotics in neurosur-
gery. Am. J. Surg. 188:68S–75S.
Morikawa, S., Inubushi, T., Kurumi, Y., Naka, S.,
Sato, K., Tani, T., Yamamoto, I., and Fujimura,
M. 2002. MR-guided microwave thermocoagu-
lation therapy of liver tumors: initial clinical
experiences using a 0.5 T open MR system. J.
Magn. Reson. Imaging 16:576–583.
Mulier, S., Ni, Y., Jamart, J., Ruers, T., Marchal,
G., and Michel, L. 2005. Local recurrence after
hepatic radiofrequency coagulation: multivariate
meta-analysis and review of contributing factors.
Ann. Surg. 242:158–171.
Ni, Y., Chen, F., Mulier, S., Sun, X., Yu, J.,
Landuyt, W., Marchal, G., and Verbruggen,
A. 2006. Magnetic resonance imaging after
radiofrequency ablation in a rodent model of
liver tumor: tissue characterization using a
novel necrosis-avid contrast agent. Eur. Radiol.
16:1031–1040.
Noone, T.C., Semelka, R.C., Chaney, and D.M.,
Reinhold, C. 2004. Abdominal imaging studies:
comparison of diagnostic accuracies resulting
from ultrasound, computed tomography, and
magnetic resonance imaging in the same indi-
vidual. Magn. Reson. Imaging 22:19–24.
Pawlik, T.M., Scoggins, C.R., Zorzi, D., Abdalla,
E.K., Andres, A., Eng, C., Curley, S.A., Loyer,
E.M., Muratore, A., Mentha, G., Capussotti, L.,
and Vauthey, J.N. 2005. Effect of surgical mar-
gin status on survival and site of recurrence after
hepatic resection for colorectal metastases. Ann.
Surg. 241:715–724.
Pearson, A.S., Izzo, F., Fleming, R.Y., Ellis, L.M.,
Delrio, P., Roh, M.S., Granchi, J., and Curley,
S.A. 1999. Intraoperative radiofrequency abla-
tion or cryoablation for hepatic malignancies.
Am. J. Surg. 178:592–599.
O.F. Bathe and H. Mahallati
Puls, R., Stroszczynski, C., Gaffke, G., Hosten, N.,
Felix, R., and Speck, U. 2003. Laser-induced
thermotherapy (LITT) of liver metastases:
MR-guided percutaneous insertion of an MRI-
compatible irrigated microcatheter system using
a closed high-field unit. J. Magn. Reson. Imaging
17:663–670.
Schenck, J.F., Jolesz, F.A., Roemer, P.B., Cline,
H.E., Lorensen, W.E., Kikinis, R., Silverman,
S.G., Hardy, C.J., Barber, W.D., and Laskaris,
E.T. 1995. Superconducting open-configuration
MR imaging system for image-guided therapy.
Radiology 195:805–814.
Shibata, T., Iimuro, Y., and Yamamoto, Y. 2002. Small
hepatocellular carcinoma: comparison of radio-
frequency ablation and percutaneous microwave
coagulation therapy. Radiology 223:331–337.
Solbiati, L., Livraghi, T., Goldberg, S.N., Ierace,
T., Meloni, F., Dellanoce, M., Cova, L., Halpern,
E.F., and Gazelle, G.S. 2001. Percutaneous
radio-frequency ablation of hepatic metastases
from colorectal cancer: long-term results in 117
patients. Radiology 221:159–166.
Solbiati, L., Ierace, T., Tonolini, M., and Cova, L.
2004. Guidance and monitoring of radiofrequency
liver tumor ablation with contrast-enhanced
ultrasound. Eur. J. Radiol. 51:S19–S23.
Steinmeier, R., Fahlbusch, R., Ganslandt, O.,
Nimsky, C., Buchfelder, M., Kaus, M., Heigl,
T., Lenz, G., Kuth, R., and Huk, W. 1998.
Intraoperative magnetic resonance imaging with
the magnetom open scanner: concepts, neurosur-
gical indications, and procedures: a preliminary
report. Neurosurgery 43:739–747.
Sutherland, G., Kaibara, T., and Louw, D. 2002.
Intraoperative MR at 1.5 tesla - experience and
future directions. Acta Neurochir. 85:21–28.
Therasse, P., Arbuck, S.G., Eisenhauer, E.A.,
Wanders, J., Kaplan, R.S., Rubinstein, L.,
Verweij, J., vanGlabbeke, M., van Oosterom, A.T.,
Christian, M.C., and Gwyther, S.G. 2000. New
guidelines to evaluate the response to treatment in
solid tumors. J. Natl. Cancer Inst. 92:205–216.
Truwit, C., and Hall, W. 2001. Intraoperative MR
systems: high-field approaches. Neuroimaging
Clin. N. Am. 11:645–650.
Veenendaal, L.M., Rinkes, I.H., Lips, C.J., and van
Hillegersberg, R. 2006. Liver metastases of neu-
roendocrine tumours; early reduction of tumour
load to improve life expectancy. World J. Surg.
Oncol. 4:35.
7. Intraoperative Magnetic Resonance Imaging for Radiofrequency Ablation of Hepatic Tumors
Weidensteiner, C., Kerioui, N., Quesson, B.,
Senneville, B.D., Trillaud, H., and Moonen, C.
2004. Stability of real-time MR temperature
mapping in healthy and diseased human liver. J.
Magn. Reson. Imaging 19:438–446.
Weinmann, H., Schuhmann-Giampieri, G., Schmitt-
Willich, H., Vogler, H., Frenzel, T., and Gries, H.
1991. A new lipophilic gadolinium chelate as a
tissue-specific contrast medium for MRI. Magn.
Reson. Med. 22:233–237.
101
Yu, H., Cheng, J., Lai, K., Lin, C., Lo, G., Lin, C.,
Hsu, P., Chan, H., Lo, C., Tsai, W., and Chen,
W. 2005. Factors for early tumor recurrence
of single small hepatocellular carcinoma after
percutaneous radiofrequency ablation therapy.
World J. Gastroenterol. 11:1439–1444.
Zhang, Q., Cheung, Y.-C., Lewin, J., and
Duerk, J. 1998. A method for simultaneous
RF ablation and MRI. J. Magn. Reson. Imaging
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
INTRODUCTION
Detection of metastatic cancer to the liver
with positron emission tomography (PET)
or computed tomography (CT) usually
signifies an ominous sign for patients
with colorectal cancer (Fernandez et al.,
2004; Arulampalam et al., 2001) and
other solid cancers (Silvestri et al., 2003;
Eubank et al., 2002). Surgically unresectable
and chemotherapy refractory metastatic
cancer to the liver often represents an
end-stage event. For these patients, a wide
variety of local and regional therapies
have been developed from the experience
of treating hepatocellular carcinoma for
the purpose of providing local control
of liver tumors while sparing essential
vascular structures, liver parenchyma,
and adjacent organs (Rilling and Drooz,
2002). Yttrium-90 glass and resin micro-
spheres have been shown to be chemically
durable for in vivo delivery of β-radiation
(Drbe and Day, 1993). Recently hepatic
intraarterial infusion of Yttrium-90 glass
or resin microspheres has been introduced
for palliating unresectable hepatocellu-
lar carcinoma (Dancey et al., 2000; Ho
et al., 1997). Their use in palliation of
unresectable liver metastases from solid
cancers has been shown to be promising
(Wong et al., 2004, 2005). Two commer-
cial products of these microspheres are
available in glass matrix (20–30 μm, 3.6 g/
dl, 2,500 Bq/particle, THERA-Sphere®,
MDS Nordion, Ottawa, Canada) or resin
bound (20–60 μm, 1.6 g/dl, 50 Bq/particle,
SIR-Spheres®, Sirtex Medical, New South
Wales, Australia) forms (Wong et al.,
2006). Both the glass or resin microspheres
are introduced via catheterization of the
hepatic artery that supplies the majority of
the blood. Tumors and the microspheres
are trapped in the tumor capillary bed
where they exert a local radio-therapeutic
effect. Yttrium-90, a pure beta-emitter
with a physical half-life of 64.2 h, decays
to stable zirconium-90. The average energy
of the beta emissions from the Yttrium-90
is 0.9367 MeV, with an average penetra-
tion range of 2.5 mm in tissue (Dancey
et al., 2000).
103
104
ASSESSMENT USING
POSITRON EMISSION
TOMOGRAPHY
The traditional anatomic imaging by
computed tomography (CT) or magnetic
resonance imaging (MRI) is insensitive in
monitoring response compared to meta-
bolic imaging (Wong et al., 2002). Prior
studies (Wong et al., 2002, 2004, 2005)
addressed the feasibility of using PET
for assessing and quantifying metabolic
response after intraarterial Y-90 micro-
spheres. These studies justify the use of
the novel technique by using the objective
and quantitative property intrinsic to PET
imaging in evaluating tumor load reduc-
tion following Y-90 microsphere treatment
(Wong et al., 2004, 2005).
Positron emission tomography (PET) or
recently combined PET/CT imaging has
been widely used in oncologic evaluation
by taking advantage of diverse capabilities
among cell types in metabolizing serum
glucose with 2-deoxy-2-[fluorine-18]
fluoro-D-glucose (18F-FDG) as its analog
tracer. The modeling of glucose metabolic
rate (MRglc) with 18 F-FDG (Huang
et al., 1980; Phelps et al., 1979) follows
the solid foundation from C-14 deoxyglu-
cose (Sokoloff et al., 1977):
MRglc = {[Glc]/LC}{K1k3/(k2 + k3)}
where the rate constants, K1 and k2 are
the forward and reversed transport across
the capillary/cellular membrane, k3 is the
phosphorylation of FDG to FDG-6-P in
cells and LC is a lumped constant that
accounts for the transport and phospho-
rylation difference between FDG and
glucose, assuming a small dephosphory-
lation rate constant (k4) for FDG-6-P.
C.-Y.O. Wong
A standard uptake value (SUV) has been
introduced and defined by:
SUV = Tumor activity/Dose per body
mass
This SUV is often quoted for daily clini-
cal quantification of metabolic activity.
Assuming negligible free 18F-FDG at the
time of imaging, it has been theorized that
SUV is related to MRglc (Huang, 2000):
MRglc = G x [Glc] x SUV
where G is a product of LC and a subject
independent proportional constant relating
the term of dose per body weight to the
time integral of total plasma FDG activity
(Huang, 2000). Thus, the SUV is a semi-
quantitative measurement of tumor metab-
olism from PET or PET/CT imaging. It has
no dimensions when both the tumor activity
and dose per body mass are expressed in
the same unit as μCi/g. The percentage
change in total liver SUV from all axial
slices of the liver has been shown to cor-
relate well with visual evaluations of meta-
bolic response (Wong et al., 2004, 2005).
The reduction of metabolism is related to
the tumor load reduction and correlates
with reduction in tumor size measured by
CT or MRI (Wong et al., 2002, 2007). The
quantitative assessment total liver SUV
reduction after treatment has also been
shown to correlate with visual assessment
of decreased metabolism by PET in the
tumor lesions, while the background
metabolism of the liver remains more
or less the same (Wong et al., 2004). As
the reduced tumor metabolism after Y-90
treatment parallels the reduction of tumor
markers if available (Wong et al., 2004)
without significant permanent changes
in transaminases reflecting liver injury
(Wong et al., 2004, 2006), the reduction of
8. Surgically Unresectable and Chemotherapy-Refractory Metastatic Liver Carcinoma
metabolism assessed by functional FDG-
PET or PET/CT imaging is related to
tumor load reduction but not to the hepato-
cellular death.
Y-90 MICROSPHERE
RADIOEMBOLIZATION
The radioisotope, Yttrium-90 (Y-90) is
usually attached to some form of micro-
spheres for its brachytherapeutic and/or
radioembolising effects. The resin-bound
Y-90 microspheres have ~ 40–60 times
more particles than the glass matrix Y-90
microsphere per GBq of Y-90 (Wong
et al., 2006). As there are differences
in the size and radioactivity per particle
among different forms of microspheres,
the metabolic response of Y-90 micro-
spheres may be due to varying degrees of
the embolic effect of the particle number,
the local radiation effect that is imparted,
or both. Irrespective of the underlying
tumor killing mechanism, PET will assess
the presence of viable tumor cells and thus
the final outcome of the treatment.
Candidates for radioembolization can
have either unresectable primary or meta-
static hepatic disease with liver-dominant
tumor burden and a life expectancy greater
than 3 months (Kennedy et al., 2007).
Other eligibility criteria for Yttrium-90
treatment are age older than 18 years
and good (Eastern Cooperative Oncology
Group) performance status (less or equal
to 2) (Wong et al., 2004; Salem et al.,
2005). Absolute contraindications to Y-90
microsphere treatment include pretreat-
ment Tc-MAA scan demonstrating the
potential of > 30 Gy radiation exposure to
the lung or flow to the gastrointestinal tract
105
that cannot be corrected by catheter tech-
niques. It is important that liver injection
of macroaggregated albumin (MAA) is
delivered with flow rates and catheter posi-
tion similar to the anticipated Y-90 infu-
sion rate and catheter position (Kennedy
et al., 2007). Patients are also excluded
from treatment if they are pregnant, had
limited hepatic reserve or irreversibly ele-
vated bilirubin levels (serum creatinine >
2.0 mg/dl or serum bilirubin > 3.0 mg/dl),
compromised portal vein (unless selective
or superselective radioembolization can
be performed), prior radiation therapy
involving the liver, hematologic disorders
(absolute granulocyte count < 1,500/ml,
platelet count < 75,000/ml), any con-
traindications to angiography or selective
visceral catheterization, significant extra-
hepatic disease representing an immi-
nent life-threatening outcome, pulmonary
insufficiency, active uncontrolled infec-
tion or any other significant underlying
medical or psychiatric illness.
Treatment planning and administra-
tion of Yttrium-90 microspheres are per-
formed generally according to standard
published guidelines (Wong et al., 2004,
2005; Kennedy et al., 2007). Dual phase
liver CT is first performed within 30
days before the embolization procedure
and 5-mm images are usually reviewed
with special attention to the volume
of the liver, degree and distribution
of tumor infiltration by experienced
interventional/body radiologists. Once
a patient is deemed a suitable candidate
for Yttrium-90 microsphere treatment, a
visceral arteriogram with intra-hepatic
arterial 99mTc-MAA (macroaggregated
albumin) scan will be performed to
define the intraarterial distribution of
106 C.-Y.O. Wong

particles and to evaluate the amount of
pulmonary and gastrointestinal flow.
All treated patients are then evaluated
to confirm an acceptable risk of shunting
of Yttrium-90 microsphere to the lungs
(< 30 Gy) (Wong et al., 2004, 2005). In
patients with hepatic tumors, a portion of
the arterial supply may bypass the capil-
lary bed and drain directly into venous
system. The microspheres that are not
trapped in the liver will be shunted via the
heart and deposited in the lungs. Thus, a
lung shunting measurement is essential
before the treatment to obtain a unique
shunting fraction (F) for the given physi-
ology and anatomy of the hepatic tumors.
Prior to treatment, planar anterior 99mTc
MAA scans (Figure 8.1) are obtained
within 30 min after intrahepatic arterial
injection of 150 MBq 99mTc MAA during
angiography to include the lungs, liver,
and stomach (Figure 8.2). Recently, single
photon emission computed tomography
(SPECT) or combined SPECT/CT is also
utilized after planar scintigraphic imag-
ing to define more precisely the hepatic
distribution of MAA (Figure 8.3), which
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

Figure 8.4. Spreadsheet for calculating lung shunting fraction (F = 2.4%)

distribution within the normal liver and
metastatic tumor lesions. Lung shunting
(F) is assessed using intra-hepatic arterial
99m
Tc MAA images [F = 100% × lung
counts/(lung + liver + stomach counts)]
(Wong et al., 2004). Lung activity < 10%
is considered insignificant for any dose
or activity adjustment (Figure 8.4). The
lung fraction, even in the setting of F
being > 10%, is used to adjust treatment
dose or activity to optimize the radiation
dose given to each lobe of liver or the
whole liver while maintaining radiation
exposure to lung to be < 30 Gy. The treat-
ment dose needs to be further verified by
the experienced physicists according to
the formulae outlined later. The method
of lobar treatment for Y-90 microsphere
radioembolization has recently been used
by some investigators (Wong et al., 2004,
2005), which differs from the whole liver
approach reported for hepatocellular
carcinoma (Dancey et al., 2000; Herba
et al., 1988) and metastatic colorectal
cancer (Gray et al., 1992) by other inves-
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
To prevent any possible radiation pneu-
monitis, the cumulative radiation dose to
lungs is also limited to a maximum of
30 Gy. The activity prescribed may also
be reduced if the hepatic function is com-
promised by as much as 30% (Kennedy
et al., 2007). All these considerations
are balanced against the assessment of
whether the final adjusted radiation dose
or activity will be enough to deliver to
treatment response on the target lesions
based on the following formula (Ho et al.,
1997; Wong et al., 2004).
Delivered Dose (Gy) =Delivered Activity (GBq) x (1 - F) x 50
Liver Lobe Mass (kg)
where the lung shunt fraction (F) is
obtained as the ratio of counts in the
lungs to the total counts in the abdo-
men and lungs, as measured by intra-
hepatic arterial 99mTc-MAA scan. After
the appropriate activity has drawn into
the delivery device, Yttrium-90 micro-
spheres are injected into a percutaneous
catheter inserted via the femoral artery that
directed to the targeted liver lobe or the
entire liver. The administration of Y-90 is
done by slow intraarterial infusion at low
pressure under strict aseptic conditions,
fluoroscopic guidance, institutional radia-
tion safety guidelines, and the manufac-
turer’s instructions. Treatment is usually
delivered on an outpatient basis. Patients
are observed in recovery room for 4–6 h
for stability of vital signs, bleeding com-
plications or signs or symptoms of acute
post-embolization syndrome. Patients can
be discharged home when they are clini-
cally stable and without complaints. Liver
function tests are monitored on all patients,
before and after treatment.
Patients with bilobar disease, if not
yet treated by whole liver approach, can
109
be scheduled for treatment of the oppo-
site lobe 30–60 days following initial
treatment. To determine eligibility for
retreatment, patients have repeat labora-
tory tests including hepatic function panel
and cross-sectional imaging, similar to the
first treatment. At the time of retreatment
evaluation, consideration of the total dose
to the lungs is based on the cumulative
dose over all prior treatments so as not to
exceed a total estimated cumulative dose
of 30 Gy to the lungs.
CLINICAL RESULTS
Prior data have suggested that PET is
useful in assessing therapy for hepatic
colorectal metastases (Ho et al., 1997;
Selzner and Hany, 2004). Positron
emission tomography has changed the
conventional management of patients with
liver metastases due to its enhanced ability
to detect recurrent or metastatic lesions
over CT (Meta et al., 2001; Rohren et al.,
2002). The former has also been known to
be more accurate than the latter in detect-
ing liver metastases (Wong et al., 2002;
Valk et al., 1999; Ogunbiyi et al., 1997).
A prior prospective study on Y-90 glass
microsphere has shown that it is feasible
to quantify reduction of hepatic tumor
metabolism objectively after Y-90 glass
microsphere treatment for unresectable
metastatic colorectal cancer to the liver
(Wong et al., 2004). The total SUV of the
entire axial slices of liver agrees well with
visual evaluation. The PET findings are
also correlated with serum tumor marker,
suggesting that the reduction of metabo-
lism is visually, quantitatively, and bio-
logically related to the tumoricidal effects
110
of Y-90 rather than hepatocellular death
(Wong et al., 2004).
The same technique of analysis of
PET or PET/CT can be used for the two
different Y-90 agents for treating hepatic
metastases from various solid cancers.
The resin-based agent contains more par-
ticles per radioactivity of Y-90 (40–80
microsphere than glass microspheres (1.2
million/3 GBq or 50 times on average less
than resin microsphere), which causes
both radiation and embolization (radi-
oembolization) for its therapeutic effects
(Wong et al., 2006; Kennedy et al., 2007).
The radioactivity of Y-90 glass micro-
sphere is 2,500 Bq per particle, which is
50 times more than that of resin based
microsphere. Thus, more particles will be
required if the Y-90 resin microsphere is
used to achieve the same treatment dose.
It is not currently well known whether the
embolic effect will impede the local radia-
tion effect of Y-90 by the potential reduc-
tion of oxygen free radicals (Koukourakis,
2001). However, even though treatment
with Y-90 resin-based microspheres may
involve the combination of significantly
lower radiation and significantly greater
embolization when compared to glass
microspheres, it is evident from a recent
comparison that the radiation compo-
nent imparts a greater therapeutic effect
than the embolization component and the
embolic effect does not appear to impede
significantly the local radiation effect
(Wong et al., 2006).
One advantage rather than limitation of
the Y-90 microsphere treatment is related to
the heterogeneous presentation of patients.
By the nature of metastasis, no single
patient will be identical. Even though it is
virtually impossible to quantify exactly the
tumor burden in terms of number, size and
C.-Y.O. Wong
metabolism or variable radiosensitivity
of tumors, the preferential delivery of Y-90
to the tumors suggests possible similar
results in treating these metastatic liver
cancers by either of the two available
agents at a comparable safety level (Wong
et al., 2006).
Anatomical imaging by CT is insensi-
tive in assessing tumor response by sim-
ply measuring the change in the product
or sum of orthogonal longest diameters
due to the presence of necrosis, edema,
hemorrhage, or cystic changes (Wong
et al., 2002). Given the global nature of
tumor markers in the presence of extra-
hepatic metastatic cancers, PET appears
to be an excellent adjunct in assessing
response following regional treatment of
liver metastases (Wong et al., 2004, 2005;
Barker et al., 2005). Therefore, the quan-
titative application of PET or combined
PET/CT scan is useful for assessing the
response after Y-90 glass or resin-based
microsphere treatment. Positron emission
tomography can assess residual hepatic
disease and the status of extra-hepatic
metastases for potential additional salvage
chemotherapy, additional Y-90 intra-arte-
rial brachytherapy or other interventional
radiological procedures (radiofrequency
ablation, chemo- or particle-embolization).
Although a few patients may suffer from
permanent hyperbilirubinemia, the majority
of the treated patients generally feel better
after treatment, which parallels the find-
ings of PET on the reduction in metastatic
load in the liver (Wong et al., 2006).
The available results appear to be in
agreement with the survival bene-
fit imparted to patients with colorectal
metastases described previously by other
investigators (Van Hazel et al., 2004).
Positron emission tomography allows the
8. Surgically Unresectable and Chemotherapy-Refractory Metastatic Liver Carcinoma 111

assessment of response following lobar or ILLUSTRATIONS

whole liver treatment, guiding the deci-
sion whether to proceed with treatment of
the untreated lobe, retreatment or other
additional treatments. Moreover, the per-
centage change of total liver SUV by PET
reflects the metastatic load reduction by
Y-90 microspheres in a quantitative way.
In conclusion, there is significant reduc-
tion of hepatic metastatic tumor load
(metabolism) as evaluated objectively by
PET after Y-90 radioembolization for
the treatment of unresectable metastatic
disease to the liver. Y-90 microsphere
therapy provides encouraging and safe
results by arresting progression of meta-
static cancer to the liver as evident by
decreasing tumor metabolism. While the
cost-effectiveness of PET imaging has
been documented in colorectal cancers
(Park et al., 2001), the application to this
type of liver-directed therapy will require
further investigation.
Sixty-three year-old female had a history
of moderately differentiated infiltrating
adenocarcinoma of the cecum diagnosed
20 months ago. The tumor was removed
by right hemicolectomy and was
found to be 2.2 × 2.0 × 0.3 cm with
invasion through serosal surface, lymphatic
involvement and extension into peri-
nodal fat. Her baseline CEA was 1.4 ng/dl.
The patient also received adjuvant chemo-
therapy. Thirteen months after surgery, the
CEA became 4.2 ng/dl, which gradually
increased to 30.1 and 70.2 ng/dl within
the next 3 and 6 months, respectively. At
this point, she was found to have hepatic
metastasis, measuring 2.4 × 2.1 and 3.3 ×
2.9 cm by CT. She was not a surgical can-
didate and underwent evaluation for Y-90
resin microsphere treatment (Figures 8.1–
8.5). Post-operative CT did not show much
change in the size, measuring respectively

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
Herba, M.J., Illescas, F.F., Thirlwell, M.P., Boos,
G.J., Rosenthall, L., Atri, M., and Bret, P.M.
1988. Hepatic malignancies: Improved treat-
ment with intraarterial Yttrium-90. Radiology
169:311–314.
Ho, S., Lau, W.Y., Leung, T.W.T., Chan, M.,
Johnson, P.J., and Li, A.K.C. 1997. Clinical
evaluation of the partition model for estimating
radiation doses from Yttrium-90 microspheres
in the treatment of hepatic cancer. Eur. J. Nucl.
Med. 24:293–298.
Huang, S.C. 2000. Anatomy of SUV. Nucl. Med.
Biol. 27:643–646.
Huang, S.C., Phelps, M.E., Hoffman, E.J., Sideris,
K., Selin, C.J., and Kuhl, D.E. 1980. Noninvasive
determination of local cerebral metabolic rate of
glucose in man. Am. J. Physiol. 238:E69–E82.
Kennedy, A., Nag, S., Salem, R., Murthy, R., McEwan,
A.J., Nutting, C., Benson, A., Espat, J., Bilbao,
J.I., Sharma, R.A., Thomas, J.P., and Coldwell,
D. 2007. Recommendations for radioemboliza-
tion of hepatic malignancies using Yttrium-90
microsphere brachytherapy: A consensus panel
report from the radioembolization brachytherapy
oncology consortium. 1. Int. J. Radiat. Oncol.
Biol. Phys. 68:13–23.
Koukourakis, M.I. 2001. Tumour angiogenesis
and response to radiotherapy. Anticancer Res.
21:4285–4300.
Meta, J., Seltzer, M., Schiepers, C., Silverman, D.H.,
Ariannejad, M., Gambhir, S.S., Phelps, M.E.,
Valk, P., and Czernin, J. 2001. Impact of 18F-FDG
PET on managing patients with colorectal cancer:
The referring physician’s perspective. J. Nucl.
Med. 42:586–590.
Ogunbiyi, O.A., Flanagan, F.L., Dehdashti, F.,
Siegel, B.A., Trask, D.D., Birnbaum, E.H.,
Fleshman, J.W., Read, T.E., Philpott, G.W., and
Kodner, I.J. 1997. Detection of recurrent and
metastatic colorectal cancer: Comparison of
positron emission tomography and computed
tomography. Ann. Surg. Oncol. 4:613–620.
Park, K.C., Schwimmer, J., Shepherd, J.E., Phelps,
M.E., Czernin, J.R., Schiepers, C., and Gambhir,
S.S. 2001. Decision analysis for the cost-effective
management of recurrent colorectal cancer. Ann.
Surg. 233:310–319.
Phelps, M.E., Huang, S.C., Hoffman, E.J., Selin,
C.E., and Kuhl, D.E. 1979. Tomographic meas-
urement of regional cerebral glucose metabolic
113
rate in man with (F-18) fluorodeoxyglucose:
Validation of method. Ann. Neurol. 6:371–388.
Rilling, W.S., and Drooz, A. 2002. Multidisciplinary
management of hepatocellular carcinoma. J.
Vasc. Interv. Radiol. 13(Suppl 9):S259–S263.
Rohren, E.M., Paulson, E.K., Hagge, R., Wong,
T.Z., Killius, J., Clavien, P.A., and Nelson, R.C.
2002. The role of F-18 FDG positron emission
tomography in preoperative assessment of the
liver in patients being considered for curative
resection of hepatic metastases from colorectal
cancer. Clin. Nucl. Med. 27:550–555.
Salem, R., Lewandowski, R.J., Atassi, B., Gordon,
S.C., Gates, V.L., Barakat, O., Sergie, Z., Wong,
C.Y., and Thurston, K.G. 2005. Treatment of
unresectable hepatocellular carcinoma with use
of 90Y microspheres (TheraSphere): Safety,
tumor response, and survival. J. Vasc. Interv.
Radiol. 16:1627–1639.
Selzner, M., and Hany, T.F. 2004. Does the novel
PET/CT imaging modality impact on the treat-
ment of patients with metastatic colorectal cancer
of the liver? Ann. Surg. 240:1027–1036.
Silvestri, G.A., Tanoue, L.T., Margolis, M.L., Barker,
J., and Detterbeck, F. 2003. The noninvasive
staging of non-small cell lung cancer: The guide-
lines. Chest 123(Suppl 1):147S–156S.
Snyder, W.S., Ford, M.R., Warner, G.G., and
Watson, S.B. 1975–1976. Absorbed dose per unit
cumulated activity for selected radionuclides and
organs. MIRD Pamphlet No. 11:S. New York:
Society of Nuclear Medicine.
Sokoloff, L., Reivich, M., Kennedy, C., Des Rosiers,
M.H., Patlak, C.S., Pettigrew, K.D., Sakurada,
O., and Shinohara, M. 1977. The [C-14] deoxy-
glucose method for the measurement of local
cerebral glucose utilization: Theory, procedure,
and normal values in the conscious and anesthe-
tized albino rat. J. Neurochem. 28:897–916.
Stubbs, R.S., Cannan, R.J., and Mitchell, A.W.
2001. Selective internal radiation therapy with
90-Yttrium microspheres for extensive color-
ectal liver metastases. J. Gastrointest. Surg.
5:294–302.
Valk, P.E., Abella-Columna, E., Haseman, M.K.,
Pounds, T.R., Tesar, R.D., Myers, R.W., Greiss,
H.B., and Hofer, G.E. 1999. Whole-body PET
imaging with F-18 fluorodeoxyglucose in man-
agement of recurrent colorectal cancer. Arch.
Surg. 134:503–511.
114
Van Hazel, G., Blackwell, A., Anderson, J., Price,
D., Moroz, P., Bower, G., Cardaci, G., and Gray, B.
2004. Randomised phase 2 trial of SIR-Spheres
plus Fluorouracil/Leucovorin chemotherapy ver-
sus Fluorouracil/Leucovorin chemotherapy alone
in advanced colorectal cancer. J. Surg. Oncol.
88:78–85.
Wong, C.Y., Salem, R., Raman, S., Gates, V.L., and
Dworkin, H.J. 2002. Evaluating 1 2 9 0 Y- g l a s s
microsphere treatment response of unresectable
colorectal liver metastases by [18F]FDG PET: A
comparison with CT or MRI. Eur. J. Nucl. Med.
Mol. Imaging 29:815–820.
Wong, C.Y., Salem, R., Qing, F., Wong, K.T., Barker,
D., Gates, V., Lewandowski, R., Hill, E.A.,
Dworkin, H.J., and Nagle, C. 2004. Metabolic
response after intraarterial 90Y-glass micro-
sphere treatment for colorectal liver metastases:
Comparison of quantitative and visual analyses
by 18F-FDG PET. J. Nucl. Med. 45:1892–1897.
Wong, C.Y., Qing, F., Savin, M., Campbell, J.,
Gates, V.L., Sherpa, K.M., Lewandowski, R.J.,
C.-Y.O. Wong
Nagle, C., and Salem, R. 2005. Reduction of
metastatic load to liver after intraarterial hepatic
Yttrium-90 radioembolization as evaluated by
[18F]fluorodeoxyglucose positron emission
tomographic imaging. J. Vasc. Interv. Radiol.
16:1101–1106.
Wong, C.Y., Savin, M., Sherpa, K.M., Qing, F.,
Campbell, J., Gates, V.L., Lewandowski, R.J.,
Cheng, V., Thie, J., Fink-Bennett, D., Nagle,
C., and Salem, R. 2006. Regional yttrium-90
microsphere treatment of surgically unresectable
and chemotherapy refractory metastatic liver
carcinoma: A quantitative comparison of tumor
response with glass and resin based agents.
Cancer Biother. Radiopharm. 21:305–313.
Wong, C.Y., Schmidt, J., Bong, J.S., Chundru, S.,
Kestin, L., Yan, D., Grills, I., Gaskill, M., Cheng,
V., Martinez, A.A., and Fink-Bennett, D. 2007.
Correlating metabolic and anatomic responses
of primary lung cancers to radiotherapy by
combined F-18 FDG PET-CT imaging. Radiat.
Oncol. 2:18.
B. Prognosis
9
Unresectable Liver Metastases from
Colorectal Cancer: Methodology and
Prognosis with Radiofrequency Ablation
Junji Machi
INTRODUCTION
Liver metastases from colorectal cancer,
although it is a stage 4 disease, has more
treatment options and has better prognosis
than liver metastases from other cancers.
Multimodality treatment is required for
colorectal liver metastases, including local
and systemic treatment. Systemic treatment
is mainly chemotherapy, that is described
in detail in other chapters of this book.
Local treatment includes surgical resec-
tion, ablation, and hepatic artery infusion
therapy. Surgical resection is the estab-
lished curative local therapy; however,
only 10–20% of patients with colorectal
liver metastases have resectable tumors at
the time of diagnosis. Resection cannot
be performed even in patients with liver
only metastases because of various reasons
including multiple bilobar tumors, invasion
or close proximity to major blood vessels
or bile ducts, limited liver function, or high
surgical risk due to comorbidity.
Ablation therapy was introduced initi-
ally in the hope of achieving local control
of unresectable liver tumors; however,
more lately it has been used for curative
intent. Chemical ablation such as ethanol
injection is effective for hepatomas, but
not for metastatic tumors. Cryoablation
was extensively investigated in 1980s to
early 1990s. Although less invasive than
resection and applicable for unresect-
able metastases, cryoablation is associated
with various complications and usually
requires laparotomy. Thermal ablation is a
newer modality, which can be achieved by
various methods: microwave, laser, and
high-intensity ultrasound are under inves-
tigation, but sufficient clinical data are
not available. On the other hand, radiofre-
quency ablation (RFA) has been increas-
ingly used during the last decade, and has
almost replaced cryoablation.
Since its introduction, RFA has been
widely investigated for both primary and
metastatic liver tumors. For metastatic
tumors, RFA has been used mainly for
unresectable colorectal cancer metastases,
although neuroendocrine or other meta-
static tumors have also been ablated. This
is based on the assumption that RFA can
provide a benefit to patients with unresect-
able colorectal metastases. A large volume
of literature on RFA for metastatic liver
tumor is available, and its safety, repeat-
ability, and local control efficacy have
been well demonstrated for metastatic liver
tumors (Berber et al., 2002; Bilchik et al.,
117
118
1999; Bowles et al., 2001; Curley et al.,
1999; Hellman et al., 2002; Iannitti et al.,
2002; Jiao et al., 1999; Machi et al., 2001;
Parikh et al., 2002; Siperstein et al., 2000;
Solbiati et al., 1997). However, as described
by Seidenfeld et al. (2002), many ques-
tions remain unanswered regarding the
current and future role of RFA in the treat-
ment of liver metastases: a comparison of
RFA to alternative methods, efficacy of
combining RFA with other methods such
as chemotherapy, comparison of different
RFA approaches (percutaneous, laparo-
scopic, and open surgical), and long-term
outcome after RFA. To determine the role
of RFA for metastatic liver tumors, long-
term follow-up data are essential but have
rarely been reported until recently.
Moreover, while series published to date
have generally included patients with prog-
ressive disease after chemotherapy, no one
has described and compared survival after
RFA given prior to the first chemotherapy
regimen for metastatic disease (i.e., “first-
line”), and survival after RFA when given
following disease progression subsequent
to chemotherapy (i.e., “second-line”).
Failure to describe the impact of RFA in
relation to chemotherapy in distinct stages
of disease progression has hampered the
ability to estimate incremental survival
benefit gained by RFA over the expected
survival on chemotherapy alone.
We have prospectively acquired data of
patients undergoing RFA for unresectable
metastatic liver tumors since 1997. In this
chapter, first I describe the methodology
of RFA. Next, I describe our study, a part
of which was recently published (Machi
et al., 2006). The purpose of this study
is to assess the long-term outcome of
the patients based on a specific grade of
unresectability and in relation to its timing
J. Machi
opposite first- and second-line chemo-
therapy, and to identify prognostic factors
associated with survival. We focused on
colorectal cancer metastases, although
noncolorectal metastases were also treated
by RFA during the study period. Finally,
I describe the current role and the future
perspective of RFA.
METHODOLOGY OF
RADIOFREQUENCY
ABLATION
Most of RFA therapies are performed
under ultrasound guidance. Computed
tomography-guided RFA is possible, but
it is much cumbersome. Because of real-
time feature, ultrasound is a much better
imaging method to guide and monitor
RFA. The basic ultrasound guidance tech-
nique for needle, cannula, or probe place-
ment for RFA is the same as other needle
guidance under ultrasound.
Radiofrequency ablation treatment of
abdominal organs such as tumors of the
liver can be performed percutaneously,
laparoscopically, or open surgically (Machi
et al., 2001). Advantages and disadvan-
tages of each approach are as follows:
1. Percutaneous Advantages: least invasive,
possible out-patient procedure, needle
guidance system (+), possible local
anesthesia
Disadvantages: less accurate cancer
staging, some areas not accessible, risk of
thermal injury to adjacent organs (bowel,
diaphragm, gallbladder, etc.)
2. Laparoscopic Advantages: less invasive,
better cancer staging with laparoscopic
ultrasound, short recovery, possible con-
current colorectal surgery
9. Unresectable Liver Metastases from Colorectal Cancer
Disadvantages: technically demanding,
lack of needle guidance system
3. Open Surgical Advantages: better cancer
staging with intraoperative ultrasound,
needle guidance system (+), inflow
occlusion possible, possible concurrent
colorectal surgery
Disadvantages: most invasive, longer
recovery
Interventional radiologists can perform
percutaneous RFA, while surgeons can
use either approach. Different approaches
should be used appropriately depending
on the condition of patients and tumors
to be ablated (also see the next section).
Percutaneous ultrasound-guided RFA is
usually performed using a needle guid-
ance system because the distance to the
tumor is usually longer. Laparoscopic
ultrasound-guided RFA needs to be per-
formed by a freehand technique because
of lack of a needle guidance system with
laparoscopic ultrasound transducer. Open
surgical RFA can be performed either by
a freehand technique or using a needle
guidance system depending on the size,
depth, and location of tumors. Once open
surgical approach is used, any tumors in
the liver can be ablated except for tumors
abutting or in close proximity to the main
right or left hepatic ducts. Ablation of
these ducts causes a bile duct stricture or
obstruction, which is a major complica-
tion of RFA. Hand-assisted laparoscopic
ultrasound-guided RFA can be performed,
taking advantages of both laparoscopic
and open surgical approaches (Machi
et al., 2002).
In the United States, there are several
companies manufacturing RFA devices.
RITA Medical System (AngioDynamics,
Inc.) and Radio Therapeutics (Boston
Scientific, Inc.) use electrode-cannulas
119
with multiple retractable curved electrode-
needles (arrays), while Radionics uses a
cool-tip single needle or a triple needle
cluster. The size of cannulas or needles
ranges from 18-gauge to 14-gauge.
Under ultrasound guidance, cannulas
or needles are inserted into the liver and
then into the tumor. It is important to
visualize the tips of needles and to under-
stand their locations before ablation starts
in order to obtain an adequate margin of
ablation. Depending on the size of the
tumor and the achievable size of abla-
tion, a single ablation or multiple over-
lapping ablation sessions are performed.
The ablation process can be monitored
by ultrasound imaging; the ablated lesion
becomes hyperechoic because of outgas-
sing from heated tissues. However, the
hyperechoic area does not always cor-
respond exactly to the ablated lesion.
In addition, the hyperechoic area will
obscure the ultrasound images around
the tumor. Therefore, it is critical to
make a good plan for multiple ablation
sessions before the first ablation starts.
When the cannula is withdrawn at the
completion of ablation, the cannula track
is also ablated under ultrasound visuali-
zation to prevent bleeding and possible
tumor seeding.
If color or power Doppler imaging dem-
onstrates intratumoral blood flow prior to
ablation, Doppler imaging can be repeated
after ablation to confirm loss of blood
flow within the tumor. Contrast-enhanced
Doppler imaging is useful for this pur-
pose, but intravenous ultrasound contract
agents are commercially not available yet
in the United States. Preoperative tests or
preparation, postoperative follow-up, and
complications of RFA are described in the
following sections.
120 J. Machi

PROGNOSIS USING assess the survival of these patients,

RADIOFREQUENCY “unresectability” was classified into three
ABLATION: OUR STUDY grades, on the basis of oncological (cura-
tive or not), technical (tumor size, number,
Patients and Methods and location), and medical (surgical risk:
age, comorbidity, and liver dysfunction)
Patients undergoing RFA for unresect- factors. Three groups of patients were
able metastatic liver tumors from August defined:
1997 to July 2003 were followed. Data
were prospectively acquired, and were Grade 1: Relatively unresectable (not opti-
retrospectively analyzed. Metastases were mally resectable)
confirmed histologically by biopsy in all Patients with liver only metastases that
patients prior to RFA treatment. During were not optimally resectable or considered
this period of study, patients who had not resectable by the majority of surgeons
resectable metastatic tumors underwent due to oncological or medical reasons
surgical resection, while patients who were were included in this group although they
deemed to have “unresectable” tumors might have been technically resected by
were treated by RFA. aggressive approaches but with high risks.
Patients who were judged to be una- Most patients in this group had unfavorable
ble to tolerate RFA because of serious risk factors for recurrence: e.g., large size
comorbidity or because of limited liver tumors (> 5 cm), large number of tumors
function (e.g., underlying liver disease (> 4), bilobar tumors, high carcinoem-
such as cirrhosis, following prior hepatic bryonic antigen (CEA), short disease-free
resection) were excluded. Patients with interval from primary to metastases, node-
extensive extrahepatic metastases such as positive primary tumor, and/or moder-
advanced peritoneal carcinomatosis and ate to poorly differentiated tumor grade:
diffuse lung metastases who would not often two or more out of five clinical risk
benefit from RFA of liver metastases were scores deve-loped by Fong et al. (1999).
also excluded. On the other hand, patients Some patients in this group had high (but
who were considered likely to benefit from not prohibitive) surgical risks to perform
RFA for possible cure, possible prolon- resection, including certain comorbidity
gation of life, or symptomatic palliation (e.g., cardiovascular, respiratory, or renal
underwent RFA. Tumor number, tumor dysfunction) or limited liver function.
size, and invasion or proximity to major Grade 2: Absolutely unresectable
blood vessels were not considered for a Patients with liver only metastases that
contraindication to RFA. were absolutely unresectable due to tech-
nical, medical, or oncological reasons were
Unresectability of Tumors
included in this group. These patients freq-
As compared to patients with optimally uently had tumors invading major vascular
resectable liver metastases who underwent structures (such as main right and left
hepatic resection, patients with “unresect- portal veins, two or more major hepatic
able” metastases undergoing RFA included veins, or vena cava) and/or right and
a heterogeneous population. In order to left hepatic ducts or their confluence.
9. Unresectable Liver Metastases from Colorectal Cancer
Also included in this group were patients
who were judged not to be able to tolerate
resection. For example, some patients had
too many tumors and/or tumors too large
to resect by anatomical major hepatectomy
and/or by multiple metastasectomy, unless
two-stage hepatectomies combined with
portal vein embolization were performed
(Jaeck et al., 2004). Other patients had too
high (prohibitive) surgical risks such as
significant comorbidity or liver dysfunction
to undergo resection. Many of the patients
in this group also had multiple risk factors
for recurrence as patients in grade 1.
Grade 3: Extrahepatic metastasis
Patients who had extrahepatic metastases
in addition to liver metastases were included
in this group. Extrahepatic metastases
(usually limited lung or peritoneal metas-
tases) in these patients were not extensive
as compared to liver metastases so that RFA
would be considered to possibly provide
benefits to patients.
Preoperative and Intraoperative
Evaluation
Patients had preRFA laboratory tests includ-
ing liver function tests and CEA. Patients
had computed tomography (or magnetic
resonance imaging) and transabdominal
ultrasound preoperatively. Positron emis-
sion tomography (PET) was introduced
in October 1999; it was selectively used
mainly to diagnose or exclude extrahepatic
metastases. The timing of offering RFA to
each patient in relation to chemotherapy
for metastases was individualized, and was
determined by medical oncologists and
surgeons together.
Radiofrequency ablation approaches
(percutaneous, laparoscopic, or open sur-
gical) were determined by considering
factors including number, size, and location
121
of metastatic tumors, visualization or acces-
sibility on transabdominal ultrasound, need
for concurrent operations (i.e., synchronous
colorectal or other organ resection), need
for precise staging by laparoscopy or open
surgery, newly diagnosed tumors versus
recurrent tumors following previous resec-
tion or ablation, and patient condition and
liver function with regard to surgical and
anesthesia risks (Bowles et al., 2001; Machi
et al., 2000, 2001, 2002, 2006). The selec-
tion of RFA approach was individualized,
and informed consent was obtained from
each patient after full explanation of ben-
efits and risks of approaches.
During RFA operation, final number,
size, and location of the tumors, vascular or
bile duct invasion were determined. When
laparoscopic or open surgical RFA was
performed, high-frequency (5–10 MHz)
laparoscopic or open intraoperative ultra-
sound was routinely performed to screen
the entire liver for occult tumors. All occult
tumors were ablated when detected.
Many patients had multiple tumors of
different sizes. In order to estimate the total
(cumulative) tumor volume (sum of the
volumes of all tumors to be ablated in a
given RFA session), “total tumor volume”
was approximated by “total tumor size
(sum diameter of tumors)”, which was
calculated by adding the largest dimension
(in millimeter) of each tumor for all tumors
ablated in each treatment session.
Radiofrequency Ablation Methods
The majority of patients had general anes-
thesia, although several patients underwent
RFA under local anesthesia. Radiofrequency
ablation was performed under ultrasound
guidance using a RFA system (RITA
Medical System, Mountain View, CA) as
described in the previous section. Initially, a
122
15-gauge, 3 cm ablation cannula was used.
From December 1999, a 14-gauge, 5 cm
ablation cannula was mainly used. In July
2001, a new 7 cm ablation cannula with
saline infusion system was introduced and
used to ablate tumors 5 cm or larger. The
average target temperature at the needle
tips was set at 100–110°C. The ablation was
continued to achieve an ablation margin of
1 cm or more (Bowles et al., 2001; Machi
et al., 2000, 2001, 2002, 2006).
Postoperative Follow-Up
Patients were followed for at least 18
months or until death. Computed tomog-
raphy (or magnetic resonance imaging)
was obtained 1 week postoperatively
to determine the adequacy of RFA, and
then at 3-month intervals for 1 year,
and 3–6-months intervals thereafter.
Carcinoembryonic antigen, when elevated
preoperatively, was measured a few weeks
postoperatively and then at 2–4-month
intervals. Whenever CEA increased during
follow-up, further studies were carried out
using computed tomography routinely and
PET selectively.
Postoperatively, systemic chemotherapy
was initiated or resumed as indicated. The
nature of chemotherapy regimens received
was variable, as irinotecan and oxaliplatin-
based regimens became available during
the study follow-up period. Hepatic artery
regional chemotherapy was used only in a
few selected patients.
Local recurrence at the site of RFA was
defined as increase in size of the ablated
lesion with change in the CT contrast-
enhanced appearance and/or positive PET
activity at the site of RFA treatment. When
new tumors appeared away from the site of
RFA, they were judged to be new liver (int-
rahepatic) recurrences. Recurrences were
J. Machi
confirmed by biopsy prior to subsequent
treatment including repeated RFA. When
local recurrence or new liver recurrence
occurred, repeated RFA was considered and
performed when indicated. When patients
with recurrent tumors were not candidates
for repeated RFA (e.g., extensive recur-
rence or extrahepatic recurrence), systemic
or regional chemotherapy was considered
and performed as indicated.
Statistical Analyses
Statistical analyses were performed using
SAS®, version-8.2 for windows (SAS
Institute, Inc., Cary, North Carolina, statis-
tical software). Continuous variables were
expressed as mean ± standard error (SE) or
standard deviation (SD). The level of sig-
nificance was set at p ≤ 0.05 for all tests.
Survival was calculated using the Kaplan-
Meier method, and the median survival
between groups was first compared using
the log-rank test. Any factors that showed
a possible association with survival at p
≤ 0.15 in the univariate Kaplan-Meier
analysis were then taken forward in a mul-
tivariate Cox proportional hazards method.
The results from this analysis included
the hazard ratio and the corresponding
95% confidence interval (the detail is not
reported in this chapter).
Risk factors, which were prospectively
thought to affect long-term survival out-
come, were analyzed. These included
patient characteristics (age, gender),
liver tumor factors (synchronous versus
metachronous, CEA, number of tumors,
total tumor size), unresectability grade
(including extrahepatic metastasis), RFA
approaches, previous and subsequent treat-
ment, repeated RFA, complications, and
recurrence patterns.
9. Unresectable Liver Metastases from Colorectal Cancer
Short-Term Results
During the study period, 131 patients
with a total of 653 metastatic liver tumors
underwent 187 RFA procedures. Of these,
100 patients with colon (n = 66) or rectal
(n = 34) cancers, 57 men and 43 women,
underwent 146 RFA procedures to ablate
507 metastatic tumors. Age at the time
of the first RFA was 67.7 ± 10.1 (mean ±
SD), ranging from 46 to 88 years. Clinical
characteristics of tumors ablated in 146
procedures of 100 patients were:
Number of tumors in each operation:
1–16, mean ± SD: 3.5 ± 2.8.
Size of tumors: 3–174 mm, mean ± SD:
30 ± 16 mm.
Total tumor size: 11–418 mm, mean ±
SD: 96.4 ± 80.7 mm.
Unresectability: Grade 1: 40 (27.4%),
Grade 2: 77 (52.7%), Grade 3: 29
(19.9%).
RFA approach: Percutaneous: 61
(41.8%), Laparoscopic: 23 (15.8%),
Open: 62 (42.4%).
Eight patients had previous hepatic
resection for liver metastases. Fifty-five
patients (55%) received RFA prior to ini-
tiation of therapeutic chemotherapy for
metastatic disease (first-line RFA); the
other 45 patients (45%) received RFA
for residual or progressive disease after
therapeutic chemotherapy for metastatic
disease (second-line RFA). As this trial
accrued patients largely prior to availabil-
ity of oxaliplatin, which became available
during the follow-up period, initial chemo-
therapy consisted mostly of fluorouracil
plus leucovorin and/or irinotecan.
Eighty-five of 100 patients had subse-
quent treatment after RFA. Eighty-one Long-Term Results
had systemic chemotherapy as adjuvant The median and mean follow-up periods of
therapy of RFA and/or as treatment for the 100 patients were 24.5 months and 29.1
123
recurrence, using various agents, while
other therapies were also combined as
indicated: repeated RFA in 36 patients
(for recurrence), lung metastasectomy in
6 patients, and hepatic infusion chemo-
therapy in 4 patients.
There were no surgical complications in
121 of 146 RFA procedures (82.9%). Eight
major complications occurred in seven pro-
cedures (4.8%): liver failure (n = 1), biliary
stricture (n = 3), bile leakage (n = 2), liver
abscess (n = 2). Nineteen minor complica-
tions occurred in 18 procedures (12.3%):
ascites/pleural effusion (n = 4), wound
infection (n = 3), myoglobinuria (n = 6),
thrombocytopenia (n = 1), pneumonia
(n = 2), and heart failure/arrhythmia (n =
3). One patient with limited liver function
(Child-Pugh Class B) from cirrhosis who
had multiple metastases from a colon can-
cer had progression of liver dysfunction and
died of liver failure 2 months after RFA.
Local tumor recurrence (at the site of RFA)
was diagnosed in 34 of 507 tumors (6.7%)
following 30 of 146 procedures (20.5%).
The interval between RFA and local tumor
recurrence ranged from 1 to 23 months and
averaged 8.7 ± 0.9 months (mean ± SE).
New liver (intrahepatic) recurrence and/or
extrahepatic recurrence (lung, bone, brain,
peritoneum/retroperitoneum, lymph node)
occurred following 87 of 100 first RFA
procedures (87.0%) and following 123 of
146 procedures including repeated RFA
(84.2%). The interval between RFA and
new recurrence ranged from 1 to 40 months
and averaged 10.4 ± 0.6 months (mean ±
SE). Recurrence in the RFA cannula track
(“needle track seeding”) was not noted.
124
± 17.5 months (mean ± SD), respectively,
ranging from 2 to 84 months. At the time
of data analyses in 2005, 41 patients
(41.0%) were alive and 59 patients (59%)
were dead. Twenty-one of 41 alive patients
were recurrence-free while 20 patients had
recurrent tumors. Causes of death were
liver tumors in 28 patients (47.5%), extra-
hepatic metastases in 29 patients (49.2%),
and non-cancer in 2 patients (3.4%).
The overall median and mean survival of
the 100 patients was 28 months and 36.1
± 2.4 months (mean ± SE), and 1-, 3- and
5-year survival was 90.0%, 42.0% and
30.5%, respectively. The median and mean
of the 25 patients in the unresectability
grade 1 group were 48 months, 50.9 ± 3.8
months, and the 1-, 3- and 5-year survival
were 100%, 83.6% and 49.4%, respectively.
The recurrence-free median and mean sur-
vival of the 100 patients was 13 months and
23.5 ± 2.2 months (mean ± SE), and 1-,
3- and 5-year recurrence-free survival was
53.0%, 23.2%, and 21.7%, respectively.
Prognostic factors that were hypoth-
esized to affect overall and recurrence-free
survival of 100 patients were analyzed first
by univariate analysis. Carcinoembryonic
antigen level (200 ng/ml), total tumor size
(100 mm), RFA approach, previous thera-
peutic chemotherapy (not adjuvant but
therapy for metastases, namely first- versus
second-line chemotherapy), local recur-
rence and new recurrence were signifi-
cantly associated with both overall and
recurrence-free survival (p < 0.05), while
age, gender, synchronous versus meta-
chronous, number of tumors, and repeated
RFA were not. Unresectability grade and
complications were significantly associated
with recurrence-free survival (p < 0.05) and
appeared to be associated with overall
survival approaching statistical significance
(p = 0.0576 and p = 0.0540, respectively).
J. Machi
Subsequent treatment affected recurrence-free
survival but not overall survival. The lower
unresectability grade demonstrated better
overall and recurrence-free survival. Extra-
hepatic metastasis (grade 3) was associated
with worst survival. Fifty five patients who
underwent RFA as a first-line treatment
prior to chemotherapy for metastatic dis-
ease showed a better median survival of 48
months compared to a median survival of
22 months among 45 patients who under-
went RFA as second-line or salvage treat-
ment after chemotherapy.
Laparoscopic approach provided the best
survival, open surgical approach provided
intermediate survival, and percutaneous
approach provided the worst survival. In
order to further assess differences among
RFA approaches, tumor characteristics
stratified by approaches were evaluated and
analyzed. Tumors treated by the laparo-
scopic approach were smaller in size and
number and lower in unresectability grade,
whereas tumors treated by the percutaneous
approach were larger in size and higher in
unresectability grade. Such tumor character-
istics by RFA approach affected survival.
In the multivariate analysis, only age (70
years) and unresectability grade including
extrahepatic metastasis showed significant
association with overall and recurrence-
free survival (p < 0.05). The median
survival of patients without extrahepatic
metastasis was significantly better than
that of patients with extrahepatic metasta-
sis (30 vs., 20 months).
CURRENT ROLE OF RADIO-
FREQUENCY ABLATION
Hepatic resection is the most effective
established method for cure of colorectal
metastases to the liver. The 5-year overall
9. Unresectable Liver Metastases from Colorectal Cancer
survival rate after resection has been
reported in a range of 25–40% (Fong et al.,
1999; Nordlinger et al., 1996; Registry
of hepatic metastases, 1988; Scheele et al.,
1995): the median 5-year survival of 19
study series is 30% (Fernandez et al.,
2004). Recently, even higher 5-year sur-
vival rates of 58% were achieved in three
studies (Abdalla et al., 2004; Choti et al.,
2002; Fernandez et al., 2004). This excel-
lent outcome was attributed to improve-
ment in patient selection (e.g., use of
PET), perioperative care, multidisciplinary
treatment including newer chemotherapy,
and aggressive surgical management.
However, resection can be performed only
in 10–20% of patients with colorectal
metastases. Even with aggressive surgical
approaches such as a two-stage hepate-
ctomy (Jaeck et al., 2004), resectability
rates would still be ~ 30%.
For unresectable metastatic liver tumors,
ablation therapies including RFA were
initially used for local control, but more
recently its use was also aimed for curative
purpose. A number of reports documented
and confirmed the safety and local control
ability of RFA for metastatic liver tumors
(Berber et al., 2002; Bilchik et al., 1999;
Bowles et al., 2001; Curley et al., 1999;
Hellman et al., 2002; Iannitti et al., 2002;
Jiao et al., 1999; Machi et al., 2001;
Parikh et al., 2002; Siperstein et al.,
2000; Solbiati et al., 1997). Our study (as
described above) demonstrated the safety,
repeatability and local control efficacy of
RFA consistent with other studies. Major
complications occurred in fewer than 10%
of RFA procedures and were similar to
those noted by others including Curley
et al. (2004) who recently reported early
complications in 7.1% and late complica-
tions in 2.4% of 608 patients. The repeat-
ability is one advantage of RFA due to its
125
relative safety, less invasive nature and
patient willingness to undergo repeated
procedures, as compared to liver resec-
tion. Approximately, 30% of treatments in
our study were repeated RFA for recurrent
metastatic tumors. Local tumor recurrence
at the site of RFA occurred in 6.7% of
tumors, which is similar or superior to
other studies (~ 5–15%). The diagnosis of
local recurrence was made mostly within
1 year after RFA (mean 8.7 months). With
longer follow-up, new tumor recurrence
occurred much more frequently (higher
than 80%) in other areas of the liver or
extrahepatic organs. The diagnosis of new
tumor recurrence was made slightly later
(mean 10.4 months) than local tumor
recurrence, even as late as 40 months after
RFA.
In order to evaluate the effect of RFA as a
potentially curative treatment, a long-term
follow-up of patients after RFA is essen-
tial. However, the long-term outcome for
liver metastases treated by RFA has rarely
been documented until recently. Most of
the data on long-term survival have been
reported by investigators in Europe, par-
ticularly Italy, where RFA was introduced
in the early to mid-1990s. Initial survival
outcomes for colorectal metastases were
reported by Solbiati et al. (1997) and
by Lencioni et al. (1998). Solbiati et al.
(2003) updated their data, and Lencioni et
al. (2004) also updated data, which were
included in a long-term multicenter study
of 423 patients. Table 9.1 summarizes
studies showing long-term prognosis of
RFA. In these Italian and other European
studies with long-term follow-up (3–5-
year survival), RFA was performed by the
percutaneous approach by radiologists (De
Baere et al., 2003; De Meijer et al., 2006;
Gillams and Lees, 2004; Lencioni et al.,
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-
duced around 1996–1997. The first long-
term follow-up outcome (3-year survival)
of a United States multicenter study was
presented by Siperstein et al. (2000) at
the congress of the American College of
Surgeons in October 2002; however, the
data were not published. Curley et al.
(1999), working with Italian investigators,
have been demonstrating excellent results
of RFA for unresectable primary and met-
astatic hepatic malignancies, and their
most recent report by Abdalla et al. (2004)
described long-term outcomes of open
surgical RFA and combined RFA plus
resection (4-year survival) as compared to
resection for colorectal metastases (Table
9.1). Berber et al. (2005), using exclusively
a laparoscopic RFA approach, reported
a median survival of 28.9 months in 135
patients with colorectal metastases. In our
study (as described in the previous section),
we have reported follow-up results of 100
patients undergoing RFA for colorectal
metastatic tumors, up to 84 months, with
24.5-month median and 29-month mean
follow-up, which is the longest period ever
reported in the United States (Machi et al.,
2006). We used percutaneous, laparoscopic
and open surgical approaches to optimize
the potential benefit to each patient, reco-
gnizing the limitation of the percutaneous
approach alone (Bowles et al., 2001; Machi
et al., 2000, 2001, 2002, 2006).
The long-term follow-up in our study has
demonstrated that RFA can contribute to
encouraging survival in patients with unre-
sectable colorectal cancer liver metastases.
The survival outcome was similar or supe-
rior to those of Italian/European studies
and similar to that reported by Berber et al.
(2005) (Table 9.1). This may be due to
our aggressive performance of RFA using
laparoscopic and open approaches when
the percutaneous approach is considered
suboptimal. Additionally, the inclusion of
the unresectability grade 1 patients, who
might have otherwise undergone hepatic
resection although not optimally, may have
skewed our survival rates. However, the
9. Unresectable Liver Metastases from Colorectal Cancer
overall 5-year survival rate of 30.5% was
comparable to those historically achieved
by hepatic resection (Fong et al., 1999;
Nordlinger et al., 1996; Registry of hepatic
metastases, 1988; Scheele et al., 1995),
although lower than the 58% achieved in
recent three reports of resection (Abdalla et
al., 2004; Choti et al., 2002; Fernandez et
al., 2004).
Prognostic factors such as the clinical
risk score for recurrence after hepatic resec-
tion have been identified and are useful for
predicting post-resection survival outcome
in patients with colorectal metastatic liver
tumors (Fong et al., 1999; Nordlinger
et al., 1996). In our study, we have iden-
tified factors predictive of survival after
RFA for unresectable metastatic tumors.
Among the factors not associated with
overall and recurrent-free survivals was
the presence or absence of repeated RFA.
Repeated RFA did not result in worse sur-
vival, suggesting that RFA should be con-
sidered for recurrent metastatic tumors.
On the other hand, the factors associated
with survival were CEA, total tumor size,
unresectability grade including extrahe-
patic metastasis, RFA approach, previous
therapeutic chemotherapy, complications,
local tumor recurrence and new recur-
rence. Many of these factors are related
to one another. Complications and local
tumor recurrence are more frequent when
total tumor size (volume) to be ablated is
larger (Bowles et al., 2001; Machi et al.,
2001; Siperstein et al., 2000). Differences
in survival outcome among three RFA
approaches in our study are likely due
to selection of patients and tumors (e.g.,
unresectability grade) to be ablated: the
laparoscopic approach was superior to
other approaches because patients in lower
unresectability grade were selected.
127
Radiofrequency ablation following pre-
vious chemotherapy for metastatic dis-
ease was associated with worse survival
likely because patients who had pre-RFA
treatment generally had more extensive
tumors, chemotherapy-refractory tumors
or extrahepatic metastasis. However, the
median survival of 48 months among
patients receiving RFA prior to first-line
chemotherapy compares favorably with
survival reported from various trials of
first-line chemotherapy alone, which have
reported survival in the range of 19–22
months (Grothey et al., 2004; Hurwitz
et al., 2004). Similarly, the median survival
of 22 months in this study for RFA when
given in the second-line setting compares
favorably with survival of 11–15 months
reported from recent trials of second-line
chemotherapy alone (Giantonio et al.,
2005; Pitot et al., 2005; Rougier et al.,
2002). In order to preserve the opportu-
nity to measure secondary end points of
tumor response and response duration,
many chemotherapy trial protocols pro-
hibit patients from receiving a potentially
curative resection or ablation after the
start of chemotherapy, even though initial
response to such therapy may move many
of these initially unresectable tumors to a
potentially resectable or ablatable status.
Berber et al. (2005) have reported pre-
dictors of survival after RFA: significant
factors associated with overall survival
were CEA and tumor size, which are simi-
lar to our study. However, the presence of
limited extrahepatic disease did not appear
to affect survival significantly in their study.
This is in contradiction to patients with
extrahepatic metastasis in our study, prob-
ably due to differences in inclusion crite-
ria of patients for RFA depending on the
extent of extrahepatic metastases.
128
In our study, we have attempted to classify
patients with unresectable liver metastatic
tumors into groups in the hope of pre-
dicting prognosis or outcome after RFA
treatment. As compared to resectable meta-
static tumors for which easily applicable
and valuable patient stratification methods
such as the clinical risk score are avail-
able (Fong et al., 1999; Nordlinger et al.,
1996), unresectable liver tumors encom-
pass a wider variety of patient and tumor
characteristics, and thus stratification of
such heterogeneous patients by a straight-
forward scale is difficult. Unresectability
is generally determined based on oncolo-
gical, technical and medical factors and, in
a certain extent, the unresectability criteria
differ among surgeons. Radiofrequency
ablation can be used for the purpose
of cure, palliation or life prolongation.
Taking these factors and RFA purposes
into account, we have presented an “unre-
sectability grade” by defining three groups:
grade 1 (relatively unresectable or not
optimally resectable), grade 2 (absolutely
unresectable), and grade 3 (extrahepatic
metastasis). This unresectability grade sys-
tem has demonstrated significant survival
differences among patients with colorec-
tal metastatic liver tumors after RFA in
this study, and therefore may be a useful
predictor in future studies.
FUTURE PERSPECTIVE
The current methodology of RFA is
advancing, and a 5 cm ablation and even
a 7 cm ablation can be performed. It is
anticipated that further advances of RFA
technology will improve the ablation size
larger and the speed of ablation faster
with lower complications and lower local
J. Machi
cancer recurrence. The precise monitoring
of the size or extent of ablation intraopera-
tively by ultrasound is presently not pos-
sible, and local cancer recurrence is higher
than resection. New ultrasound or other
intraoperative technology such as 3-D
(three-dimensional) ultrasound or contrast
enhancement may improve the accuracy
of monitoring RFA with improvement of
ablation negative margin.
Hepatic resection is considered as the
gold standard for resectable colorectal
metastatic liver tumors. On the other hand,
the true role of RFA has not been determined.
Our study has demonstrated that RFA and
repeated RFA can contribute to a relatively
good survival for patients with unresectable
metastatic tumors, and has characterized
survival following RFA in relation to first- and
second-line chemotherapy (Machi et al.,
2006). However, new liver or extrahepatic
recurrence is common after RFA; new liver
tumors probably recur more frequently
after RFA than after hepatic resection, as
noted by Abdalla et al. (2004), probably
because larger liver parenchyma fields,
which are at risk for recurrence remain
after RFA compared to resection.
To prevent and control such recurrences
and to achieve a long survival, multimo-
dality therapies, including locoregional
and systemic methods, are definitively
required. Because there is a limitation in
hepatic resection as a locoregional treat-
ment method, RFA is an additionally
valuable modality, particularly for unresect-
able metastatic tumors. Our study sug-
gests greater survival benefit when RFA
is used in the first-line setting than when
used in the second-line or salvage setting.
Further investigation is needed to evaluate
the role and timing of RFA in combination
with systemic treatment, especially with
9. Unresectable Liver Metastases from Colorectal Cancer
the modern chemo-biological regimens.
The relatively longer survival provided by
RFA with chemotherapy in our study may
also suggest the value of prechemotherapy
cytoreduction by means of RFA for patients
with unresectable colorectal cancer.
We think that currently RFA cannot
completely replace hepatic resection, par-
ticularly for large tumors (e.g., > 7–8 cm)
because of high local recurrence after RFA
of large tumors. Radiofrequency ablation
can be combined with hepatic resection
for selected patients with multiple bilobar
metastases (Abdalla et al., 2004; Pawlik
et al., 2003). This resection plus ablation
method will likely increase the number of
patients who are surgically treatable more
than resection alone. When patients have
synchronous liver metastases with primary
colorectal cancer, RFA can be used more
frequently than liver resection in conjunc-
tion with colorectal resection because of
less invasiveness and less complications
with RFA. This is currently performed by
open surgery (Machi et al., 2000), but can
be performed more by laparoscopic surgery
in future because of anticipated increase in
laparoscopic colorectal cancer operations.
The true impact of RFA in comparison to
resection for optimally and not-optimally
resectable tumors cannot be determined
unless a randomized study is performed.
We believe that there may be a role of RFA
for smaller resectable metastatic tumors
and that the long-term follow-up results of
ours, especially in patients of unresectability
grade 1 showing encouraging survival, and
the results of others would support a pro-
posal for justification of such resection-
versus-RFA randomization for resectable
metastatic liver tumors. This is particularly
true in face of newer chemo-biological
systemic therapies.
129
In conclusion, the safety, repeatability
and local control efficacy of RFA have
been well documented. Furthermore, more
data have been lately reported, show-
ing the long-term survival benefits of
RFA. Radiofrequency ablation technology
and methodology will continue evolving
with provision of better ablation outcome.
However, the exact role of RFA in the treat-
ment of colorectal cancer liver metastases
needs to be defined with future investiga-
tions. Multi-disciplinary approach by more
appropriate local therapy by resection and
ablation or combination, together with
advancement of systemic therapy, will fur-
ther improve the prognosis of patients with
liver metastases.
REFERENCES
Abdalla, E.K., Vauthey, J.N., Ellis, L.M., Ellis, V.,
Pollock, R., Broglio, K.R., Hess, K., and Curley,
S.A. 2004. Recurrence and outcomes following
hepatic resection, radiofrequency ablation, and
combined resection/ablation for colorectal liver
metastases. Ann. Surg. 239: 818–827.
Berber, E., Flesher, N., and Siperstein, A.E. 2002.
Laparoscopic radiofrequency ablation of neu-
roendocrine liver metastases. World J. Surg. 26:
985–990.
Berber, E., Pelley, R., and Siperstein, A.E. 2005.
Predictors of survival after radiofrequency ther-
mal ablation of colorectal cancer metastases to
the liver: a prospective study. J. Clin. Oncol. 23:
1358–1364.
Bilchik, A.J., Rose, D.M., and Allegra, D.P. 1999.
Radiofrequency ablation: A minimally invasive
technique with multiple applications. Cancer
J. Sci. Am. 5: 356–361.
Bowles, B.J., Machi, J., Limm, W.M.L., Severino,
R., Oishi, A.J., Furumoto, N.L., Wong, L.L., and
Oishi, R.H. 2001. Safety and efficacy of radi-
ofrequency thermal ablation in advanced liver
tumors. Arch. Surg. 136: 864–869.
Choti, M.A., Sitzmann, J.V., Tiburi, M.F.,
Sumetchotimetha, W., Rangsin, R., Schulick,
R.D., Lillemoe, K.D., Yeo, C.J., and Cameron,
130
J.L. 2002. Trends in long-term survival following
liver resection for hepatic colorectal metastases.
Ann. Surg. 235: 759–766.
Curley, S.A., Izzo, F., Delrio, P., Ellis, L.M., Granchi,
J., Vallone, P., Fiore, F., Pignata, S., Daniele, B.,
and Cremona, F. 1999. Radiofrequency ablation
of unresectable primary and metastatic hepatic
malignancies. Results in 123 patients. Ann. Surg.
230:1–8.
Curley, S.A., Marra, P., Beaty, K., Ellis, L.M.,
Vauthey, J.N., Abdalla, E.K., Scaite, C., Rant,
C., Wolff, R., Choi, H., Loyer, E., Vallone, P.,
Fiore, F., Scardino, F., DeRosa, V., Orlando,
R., Pignata, S., Daniele, B., and Izzo, F. 2004.
Early and late complications after radiofre-
quency ablation of malignant liver tumors in 608
patients. Ann. Surg. 239: 450–458.
De Baere, T.J., Elias, D., Ducreux, M., Kuoch, V.,
Boige, V., Dromain, C., Lasser, P., and Tursz,
T. 2003. Percutaneous radiofrequency of liver
metastases: single center experience over a five
year period. Proc. Am. Soc. Clin. Oncol. 22: 345
(Abstract 1387).
De Meijer, V.E., Verhoef, C., Kuiper, J.W., Alwayn,
I.P.J., Kazemier, G., and Ijzermans, J.N.M. 2006.
Radiofrequency ablation in patients with pri-
mary and secondary hepatic malignancies. J.
Gastrointest. Surg. 10: 960–973.
Fernandez, F.G., Drebin, J.A., Linehan, D.C.,
Dehdashti, F., Siegel, B.A., and Strasberg,
S.M. 2004. Five-year survival after resection
of hepatic metastases from colorectal cancer in
patients screened by positron emission tomogra-
phy with F-18 fluorodeoxyglucose (FDG-PET).
Ann. Surg. 240: 438–450.
Fong, Y., Fortner, J., Sun, R.L., Brennan, M.F.,
and Blumgart, L.H. 1999. Clinical score for
predicting recurrence after hepatic resection for
metastatic colorectal cancer. Analysis of 1001
consecutive cases. Ann. Surg. 230: 309–321.
Giantonio, B.J., Catalano, P.J., and Meropol, N.J.
2005. High-dose bevacizumab improves survival
when combined with FOLFOX 4 in previously
treated advanced colorectal cancer: results from
ECOG study E3200. Proc. Am. Soc. Clin. Oncol.
(Abstract 2). 23: Part 1.
Gillams, A.R., and Lees, W.R. 2004. Radio-
frequency ablation of colorectal liver metastases
in 167 patients. Eur. Radiol. 14: 2261–2267.
J. Machi
Grothey, A., Sargent, D., Goldberg R.M., and
Schmell, H.J. 2004. Survival of patients with
advanced colorectal cancer improves with avail-
ability of fluorouracil-leucovorin, irinotecan,
and oxaliplatin in the course of treatment. J. Clin.
Oncol. 22: 1209–1214.
Hellman, P., Ladjevardi, S., Skogseid, B., Akerstrom,
G., and Elvin, A. 2002. Radiofrequency tissue
ablation using cooled tip for liver metastases
of endocrine tumors. World J. Surg. 26: 1052–
1056.
Hurwitz, H., Fehrenbacher, L., Novotny, W.,
Cartwright, T., Hainsworth, J., Heim, W., Berlin,
J., Baron, A., Griffing, S., Holmgren, E., Ferrara,
N., Fyfe, G., Rogers, B., Ross, R., and Kabbinavar,
F. 2004. Bevacizumab plus irinotecan, fluorour-
acil, and leucovorin for metastatic colorectal
cancer. N. Engl. J. Med. 350: 2335–2342.
Iannitti, D.A., Dupuy, D.E., Mayo-Smith, W.W.,
and Murphy, B. 2002. Hepatic radiofrequency
ablation. Arch. Surg. 137: 422–427.
Jaeck, D., Oussoultzoglou, E., Rosso, E., Greget,
M., Weber, J.C., and Bachellier, P. 2004. A two-
stage hepatectomy procedure combined with
portal vein embolization to achieve curative
resection for initially unresectable multiple and
bilobar colorectal liver metastases. Ann. Surg.
240: 1037–1051.
Jiao, L.R., Hansen, P.D., Havlik, R., Mitry, R.R.,
Pignatelli, M., and Habib, N. 1999. Clinical
short-term results of radiofrequency ablation in
primary and secondary liver tumors. Am. J. Surg.
177: 303–306.
Lencioni, R., Crocetti, L., Cioni, D., Dellapina, C.,
and Bartolozzi, C. 2004. Percutaneous radiofre-
quency ablation of hepatic colorectal metastases.
Technique, indications, results, and new prom-
ises. Invest. Radiol. 39: 689–697.
Lencioni, R., Goletti, O., Armillotta, N., Paolicchi,
A., Moretti, M., Cioni, D., and Donati, F. 1998.
Radio-frequency thermal ablation of liver metas-
tases with cooled-tip electrode needle: results of
a pilot clinical trial. Eur. Radiol. 8: 1205–1211.
Machi, J., Oishi, A.J., Morioka, W.K., Yu, M.,
Hundahl, S.A., Furumoto, N.L., and Oishi, R.H.
2000. Radiofrequency thermal ablation of syn-
chronous metastatic liver tumors can be performed
safely in conjunction with colorectal cancer resec-
tion. Cancer J. 6(Suppl 4): S344–S350.
9. Unresectable Liver Metastases from Colorectal Cancer
Machi, J., Uchida, S., Sumida, K., Limm, W.M.L.,
Hundahl, S.A., Oishi, A.J., Furumoto, N.L.,
and Oishi, R.H. 2001. Ultrasound-guided radi-
ofrequency thermal ablation of liver tumors:
percutaneous, laparoscopic, and open surgical
approaches. J. Gastrointest. Surg. 5:477–489.
Machi, J., Oishi, A.J., Mossing, A.J., Furumoto,
N.L., and Oishi, R.H. 2002. Hand-assisted laparo-
scopic ultrasound-guided radiofrequency thermal
ablation of liver tumors: a technical report. Surg.
Laparosc. Endosc. Percutan. Tech. 12: 160–164.
Machi, J., Oishi, A.J., Sumida, K., Sakamoto, K.,
Furumoto, N.L., Oishi, R.H., and Kylstra, J.W.
2006. Long-term outcome of radiofrequency
ablation for unresectable liver metastases from
colorectal cancer: evaluation of prognostic fac-
tors and effectiveness in first- and second-line
management. Cancer J. 12: 318–326.
Nordlinger, B., Guiguet, M., Vaillant, J.C., Balladur,
P., Boudjema, K., Bachellier, P., and Jaeck, D.
1996. Surgical resection of colorectal carcinoma
metastases to the liver. A prognostic scoring
system to improve case selection, based on 1568
patients. Cancer 77: 1254–1262.
Oshowo, A., Gillams, A., Harrison, E., Lees, W.R.,
and Taylor, I. 2003. Comparison of resection
and radiofrequency ablation for treatment of
solitary colorectal liver metastases. Br. J. Surg.
90: 1240–1243.
Parikh, A.A., Curley, S.A., Fornage, B.D., and
Ellis, L.M. 2002. Radiofrequency ablation of
hepatic metastases. Semin. Oncol. 29: 168–182.
Pawlik, T.M., Izzo, F., Cohen, D.S., Morris, J.S.,
and Curley, S.A. 2003. Combined resection and
radiofrequency ablation for advanced hepatic
malignancies: results in 172 patients. Ann. Surg.
Oncol. 10: 1059–1069.
Pitot, H.C., Rowland, K.M., Sargent, D.J., Philip,
P.A., Mitchell, E.P., Goldberg, R.M., and Alberts,
S.R. 2005. N9841: a randomized, phase III
equivalence trial of irinotecan (CPT-11) ver-
sus oxaliplatin/5-fluorouracil (FU)/leucovorin
131
(FOLFOX4) in patients with advanced colorectal
cancer previously treated with 5FU. Proc. Am.
Soc. Clin. Oncol. 24 (Abstract 3506), page 3506.
Registry of hepatic metastases. 1988. Resection of
the liver for colorectal carcinoma metastases: a
multi-institutional study of indications for resec-
tion. Surgery 103: 278–288.
Rougier, P., Lepille, D., Bennouna, J., Marre, A.,
Ducreux, M., Mignot, L., Hua, A., and Mery-
Mignard, D. 2002. Antitumour activity of three
second-line treatment combinations in patients
with metastatic colorectal cancer after optimal
5-FU regimen failure: a randomized, multicentre
phase II study. Ann. Oncol. 13: 1558–1567.
Scheele, J., Stang, R., Altendorf-Hofmann, A.,
and Paul, M. 1995. Resection of colorectal liver
metastases. World J. Surg. 19: 59–71.
Seidenfeld, J., Korn, A., and Aronson, N. 2002.
Radiofrequency ablation of unresectable liver
metastases. J. Am. Coll. Surg. 195: 378–386.
Siperstein, A., Garland, A., Engle, K., Rogers,
S., Barbar, E., Foroutani, A., String, A., Ryan,
T., and Ituarte, P. 2000. Local recurrence after
laparoscopic radiofrequency thermal ablation of
hepatic tumors. Ann. Surg. Oncol. 7: 106–113.
Solbiati, L., Goldberg, S.N., Ierace, T., Livraghi, T.,
Meloni, F., Dellanoce, M., Sironi, S., and Gazelle,
G.S. 1997. Hepatic metastases: Percutaneous
radio-frequency ablation with cooled-tip elec-
trodes. Radiology 205: 367–373.
Solbiati, L., Ierace, T., Tonolini, M., Bellobuono,
A., and Cova, L. 2003. Long-term survival of
patients treated with radiofrequency ablation
for liver colorectal metastases: improved out-
come with increasing experience. Radiology
229: 411–412.
White, T.J., Roy-Choudhury, S.H., Breen, D.J.,
Cast, J., Maraveyas, A., Smyth, E.F., Hartley,
J.E., and Monson, J.R.T. 2004. Percutaneous
radiofrequency ablation of colorectal hepatic
metastases – initial experience. Dig. Surg.
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
INTRODUCTION
Sequelae of chronic hepatitis B virus (HBV)
and hepatitis C virus (HCV) infection,
including hepatocellular carcinoma (HCC),
liver cirrhosis (LC) and their complications,
are health burdens of the world. Improving
treatment results for early stage HCC
makes the screening meaningful. Two-
staged HCC screening identifying high risk
subjects in the first stage and screening for
HCC by ultrasonography (US) in the second
stage is a practical and workable concept.
The benefit of US screening on high risk
groups has been well documented, but “who
should be included in the US screening”
depends on cost-benefits and considera-
tions of feasibility.
Epidemiology of Hepatocellular
Carcinoma and Liver Cirrhosis
There are different etiologies of HCC
around the world, such as chronic HBV and
HCV infections, which lead to advanced
liver disease and eventual development of
LC and HCC. These two viral infections
are the major risk factors for HCC.
HBV-related HCC is a common cause
of HCC in Asia and sub-Saharan Africa.
The major transmission of HBV infection
in these countries is vertical transmission
from carrier mother to newborn at the
time of birth. The majority of infected
newborns will develop chronic and sub-
clinical hepatitis, subsequently leading to
LC and HCC. This transmission route is
rare in Western countries, where sexual
transmission or intravenous drugs use are
the major causes of HBV infections during
adolescences, and which mean infected
individuals are at lower risk of develop-
ing into chronic infection. Consequently,
HBV-related HCC in Western countries is
relatively low.
On the other hand, HCV infections in
adults have high rates (~ 75%) of chronic
hepatitis and liver disease. Approximately
25% of patients with chronic HCV infec-
tion progress to LC within 20 years
(Fattovich et al., 1997). Among those
patients with cirrhosis, 1–4% develop HCC
each year. The burden of HBV-related
HCC will decrease in the future because
HBV mass vaccination programs will con-
sequently reduce the incidence of HCC.
137
138
However, HCV-related HCC is increasing
in some countries such as Taiwan (Lu et al.,
2006a), Japan (Kiyosawa et al., 2004), and
the United States (Wong et al., 2000). In
the United States, after incidents of acute
HCV infection peaked in the mid-1980s,
increasing numbers of future cases of
HCV-related HCC can be expected. Until
an HCV vaccination is developed and
become readily available, chronic HCV
infection will remain a major cause of LC
and HCC.
Benefit of HCC Screening
Individuals at risk for developing HCC
usually present with clinical symptoms
when HCC is in a far advanced stage,
without effective therapeutic options.
There is clearly a need for early diagnosis
of HCC. Thus, screening and surveillance
appear to be very appropriate. Surveillance
programs are ideal for a disease such as
HCC, because it is relatively common
in patients with chronic liver disease,
has readily identifiable risk factors, high
mortality rates, and potentially curative
therapy (Di Bisceglie, 2004). However,
there is considerable controversy regard-
ing the role of screening and surveillance
in HCC management. The value of screen-
ing and surveillance for HCC should be
assessed by its impact on detecting earlier
stage of the disease, increasing the effec-
tive application of curative treatment, and
thus reducing mortality. Although there
are inconsistent conclusions, most cohort
studies from Western and Eastern countries
have shown that surveillance of patients
with viral hepatitis or cirrhosis increases
early detection and effective application
of curative treatments (Zoli et al., 1996;
Yuen et al., 2000; Bolondi et al., 2001;
S.-N. Lu et al.
Toyoda et al., 2006; Leykum et al., 2007;
Chen et al., 2002). A community study
with a cohort of 4,843 Taiwanese at risk
for HCC has demonstrated that the morta-
lity was 41% lower in the screening group
compared to the control group (Chen et al.,
2002). To assess the value of screening,
one randomized controlled study with
18,816 Chinese at risk for HCC indicated
that biannual screening reduces HCC mor-
tality by 37%. Compared with the control
group, the screened group had a higher
number of HCC patients diagnosed at an
earlier stage, with better overall survival
rates (Zhang et al., 2004). It is obvious that
surveillance of high risk groups for HCC
has a sound scientific basis.
Alpha-fetoprotein (AFP) and/or ultra-
sonography (US) are the most widely used
tools for surveillance. For HCC screening,
AFP has a sensitivity of 17–65% and a
specificity of 80–99%, depending on the
prevalence of HCC in the screened popula-
tion as well as the cut-off level for the diag-
nosis (Daniele et al., 2004). The results for
US examination depend on the expertise of
operator and the equipment available. In a
systemic review, the pooled estimated sen-
sitivity and specificity for US in diagnosing
HCC is 60% and 97%, respectively (Colli
et al., 2006). Although neither AFP nor US
alone is an effective surveillance tool, their
combination increases the sensitivity of
screening by 5–10% (Daniele et al., 2004;
Colli et al., 2006). The optimal surveil-
lance interval for HCC has yet to be deter-
mined. However, based on the estimated
growth rate of HCC, the suggested interval
for surveillance in patients has been set at
6 months. Although there are no firm data
to support the hypothesis that 6 months
is better than 12 months, recent practice
guideline on HCC management recommend
10. Screening with Ultrasonography of Patients at High-Risk for Hepatocellular Carcinoma
US surveillance of patients at high risk of
developing HCC within an interval of 6–12
months (Bruix and Sherman, 2005).
Cost-effectiveness analyses of various
HCC screening strategies differ widely.
The cost per treatable HCC and per year
of life saved is US$1,667–17,934 and
US$25,232–284,000, respectively. Such
varied results might be due to differences
in the incidence of HCC in the screening
population, screening intervals, diagnostic
modalities, available treatment methods,
and tumor recurrence rates (Daniele et al.,
2004). New markers in identifying patients
at high risk of developing HCC, curative
and effective HCC treatment modalities
and better strategies might make HCC
surveillance more cost-effective and acce-
ptable in the future.
Surrogate Tests for Liver Cirrhosis
and Fibrosis
Many tools are used to evaluate the sever-
ity of liver fibrosis, such as liver biopsy,
image modalities, and laboratory tests.
Liver biopsy has been the gold stand-
ard for evaluating liver fibrosis, but has
been limited by interobserver and intraob-
server variability and sampling error.
Furthermore, the invasive procedure is
not suitable for community screening.
Fibroscan, which measures the stiffness
of hepatic tissue, is sensitive for stag-
ing liver fibrosis. It is a reliable tool for
detecting LC with areas under ROC curves
(AUC) of 0.97 (Kelleher and Afdhal,
2005). Although it is quick, painless, and
reproducible, the equipment is not readily
available worldwide.
On the other hand, blood samples are
easily collected in large scale community
screening. Hence, serum markers (Table 10.1)
139
are an ideal approach to detect LC in
the first stage screening. Direct fibrosis
markers include collagens, glycoproteins,
polysaccharides, metalloproteinases, tissue
inhibitors of metalloproteinases, and cyto-
kines. None of these biomarkers is liver
specific, and alterations in the levels of
these markers depend on the etiology of
underlying liver disease. Due to lack of
assay standardization, it is difficult to
draw strong conclusions regarding which
assays have better performance. Indirect
markers of liver fibrosis include aspar-
tate transaminase (AST)/ alanine transam-
inase (ALT) ratio (AAR), prothrombin
time, platelet counts, and AST to platelet
ratio index (APRI). More sophisticated
indices, including cirrhosis discriminant
score (CDS), PGA index, PGAA index,
Fibrotest, AcitTest, and Forns index have
also been developed. The cost of these pro-
prietary tests is similar to that of a biopsy,
and is consequently not suitable for large-
scale community screening. PGA index has
shown greater accuracy in detecting LC
(66–72%) in alcoholic liver disease than
viral liver disease (Kelleher and Afdhal,
2005). Fibrotest (Imbert-Bismut et al.,
2001) was used for correctly identifying
the stage of disease in 46% of all patients.
However, false-positives may be caused by
Gilbert’s syndrome, cholestasis and acute
liver inflammation. ActiTest is a modifica-
tion of Fibrotest, with the addition of ALT
to reflect both liver fibrosis and necroin-
flammatory activity. Major criticisms of
the Forns index include concerns regarding
the impact of lipid abnormality and cho-
lesterol altering medications (Afdhal and
Nunes, 2004; Kelleher and Afdhal, 2005).
AAR ≥ 1 had wide range of sensiti-
vity (31.5–77.8%) and PPV (13.5–100%),
based on differing LC prevalences.
140 S.-N. Lu et al.

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
of liver diseases, especially alcoholic liver
disease and autoimmune hepatitis. APRI is
a simple test for predicting cirrhosis with
an AUC of 0.94 (Wai et al., 2003). When
using upper (1.0) and lower (2.0) as cut-off
values, 81% of the patients with or without
cirrhosis were characterized. Its perform-
ance for evaluating cirrhosis is similar
to other, more complex indexes, such as
Forns index and Fibrotest. However, APRI
cannot be completely standardized due to
the variability of AST ranges considered
normal in different laboratories.
A blood test for platelet count is the
simplest, most inexpensive, and easily
available serum marker for predicting LC.
The possible causes of thrombocytopenia
are decreased thrombopoietin production,
antibody mediated platelet destruction,
hypersplenism, and myelotoxic effects.
Confounding factors are infectious disease
and hematological disease. The study by
Lu et al. (2006b) investigated the diagnos-
tic performance of platelet counts in pre-
dicting pathological LC in chronic HBV
and HCV patients and US LC in HCV
endemic community residents. Platelet
10. Screening with Ultrasonography of Patients at High-Risk for Hepatocellular Carcinoma
counts decrease while pathological fibro-
sis stages increase in chronic HCV, but
not in HBV patients. AUC was larger
in HCV (0.799) than HBV (0.672) and
the validity was acceptable when cutoff
was set at 150 × 109/l in cases of chronic
hepatitis B. At this cutoff level, sensitivity
and specificity were 52.5% and 78% in
HBV patients, and 68.2% and 76.4% in
HCV patients, respectively. In a small
scale HCV endemic community (Lu et al.,
2006b), the prevalence of HBsAg, anti-
HCV, both, and neither were 9.0%, 37.3%,
3.5% and 50.2%; the patients’ prevalence
of thrombocytopenia were 5.6%, 26.7%,
42.9%, and 3.0%, respectively. The mean
platelet counts decreased as the US paren-
chyma score increased. Using this cutoff
to identify US cirrhosis, sensitivity was
76.2%, specificity was 87.8%, and accu-
racy was 86.6%. For cases with chronic
HCV infection, platelet counts inversely
correlated to both histological and US
scores for hepatic fibrosis. Platelet count
is cheap, easily interpreted and not inferior
to others tests, and should be one of most
feasible surrogates. Platelet counts can be
employed not only for the diagnosis of
hepatic fibrosis and cirrhosis, but also for
identification of the population at high risk
of developing HCC.
Ultrasonographic HCC Screening on the
Thrombocytopenic Adult
Two-staged community-based HCC
screening seems reasonable. The first
stage is identification of high risk groups,
and the second stage is US screening of
the high risk group only. In a community
study in Taiwan, HBsAg, anti-HCV, AST,
ALT, AFP, and a family history of HCC
were selected as markers for the identifi-
cation of high risk groups. Twenty-nine
141
percent of all participants were identified
as belonging to the high risk group, and
were scheduled for US (Chen et al., 2002).
Although more than 95% of HCC cases
could be identified in this manner, the cost
of screening was high, and it is a daunting
task to perform US on 29% of the adult
population in a community.
Among 4,042 proven HCC cases at a
medical center in Taiwan, 2,032 (50.3%)
were HBVAg positive, 1,151 (28.5%)
were anti-HCV positive, 449 (11.1%)
were positive for both, and 410 (10.1%)
were negative for both. Using 150 × 109/l
as the cutoff, 1,926 (48%) of the HCC
cases were found to be thrombocyto-
penic. Thrombocytopenic patients made
up 42% (860/2,032) of the HBV-related
HCC cases, 63% (720/1,151) of the HCV-
related cases, 50% (225/449) of cases
positive to both, and 30% (121/410) of the
cases tested negative for both. In short,
HCV-related HCC cases had a higher rate
of thrombocytopenia than HBV-related
cases (Lu et al., 2006b).
In an HBsAg (12.4%) and HCV
(40.8%)-endemic area, 825 (17.9%) of
4,616 participants were thrombocyto-
penic. Of 586 (71%) participants in HCC
screening by US and AFP, 116 (19.8%)
were diagnosed as LC. Forty-two cases
had suspected liver nodules, and 25 (4.3%)
cases were confirmed as HCC. Only 13
of 25 HCC cases had elevated AFP. In
the low HCV prevalence area (HBsAg
12%, anti-HCV 2%), only 104 throm-
bocytopenic cases (6.1%) were found
among 1,694 participants. Forty-four
patients (42.3%) participated in US and
AFP screening, but only two (4.5%)
were diagnosed at an early stage of liver
cirrhosis and no incidence of HCC was
found (Lu et al., 2006b). Among the 25
detected HCC cases, tumor diameters
142
were < 3 cm in 13 cases, 3–5 cm in
6 cases, 5–10 cm in 3 cases, and > 10 cm
in 3 cases.
In 2004 a community-based HCC screen-
ing was conducted and 56,702 adults parti-
cipated. Subjects with thrombocytopenia
(platelet counts < 150 × 109/l) or with AFP
> 20 ng/ml were enrolled for the second
stage US screening. Three thousand two
hundred and thirty four subjects (5.7%)
were candidates for US screening, and 2,983
(92.2%) responded. Of 132 suspected cases,
123 (89.8%) complied with referral for con-
firmation, and 72 (58.5%) were confirmed
to be HCC cases. Among the confirmed
patients, only five were too advanced to be
actively treated. In addition to thrombocyto-
penia, AFP was added as a marker of high risk
identification, with a sensitivity estimated
at 70%. That is, 70% of HCC cases among
the population had either thrombocytopenia
or AFP > 20 ng/ml. Among 3,234 high risk
subjects, the 42% without chronic HBV
or HCV infections should not be at high
risk for HCC. Therefore, before entering
a long-term surveillance program, their
HBsAg and anti-HCV should be checked
(Lu et al., 2008).
In conclusion, for cases of chronic HCV
infection, platelet counts are inversely cor-
related to both histological and US scores
for hepatic fibrosis. This method is cheap,
easily interpreted, and not inferior to other
diagnostic methods, and should be consid-
ered one of the most feasible surrogates.
Platelet counts can be employed not only
for the diagnosis of hepatic fibrosis and
cirrhosis, but also for identification of the
population groups at high risk of develop-
ing HCC. Due to economic considera-
tions, this design is especially feasible in
HCV and HCC endemic areas with limited
medical resources.
S.-N. Lu et al.
REFERENCES
Afdhal, N.H., and Nunes, D. 2004. Evaluation of liver
fibrosis: a concise review. Am. J. Gastroenterol.
99: 1160–1174.
Boeker, K.H., Haberkorn, C.I., Michels, D.,
Flemming, P., Manns, M.P., and Lichtinghagen,
R. 2002. Diagnostic potential of circulating
TIMP-1 and MMP-2 as markers of liver fibrosis
in patients with chronic hepatitis C. Clin. Chim.
Acta 316: 71–81.
Bolondi, L., Sofia, S., Siringo, S., Gaiani, S., Casali,
A., Zironi, G., Piscaglia, F., Gramantieri, L.,
Zanetti, M., and Sherman, M. 2001. Surveillance
programme of cirrhotic patients for early detec-
tion and treatment of hepatocellular carcinoma:
a cost effectiveness analysis. Gut 48: 251–259.
Borroni, G., Ceriani, R., Cazzaniga, M., Tommasini,
M., Roncalli, M., Maltempo, C., Felline, C.,
and Salerno, F. 2006. Comparison of simple
tests for the non-invasive diagnosis of clinically
silent cirrhosis in chronic hepatitis C. Aliment
Pharmacol. Ther. 24: 797–804.
Bruix, J., and Sherman, M. 2005. Management of hepa-
tocellular carcinoma. Hepatology 5: 1208–1236.
Chen, T.H., Chen, C.J., Yen, M.F., Lu, S.N., Sun,
C.A., Huang, G.T., Yang, P.M., Lee, H.S., and
Duffy, S.W. 2002. Ultrasound screening and risk
of factors for death from heptocellular carci-
noma in a high risk group in Taiwan. Int. J.
Cancer 98: 257–261.
Colli, A., Fraquelli, M., Casazza, G., Massironi,
S., Colucci, A., Conte, D., and Duca, P. 2006.
Accuracy of ultrasonography, spiral CT, magnetic
resonance, and alpha-fetoprotein in diagnosing
hepatocellular carcinoma: a systemic review. Am.
J. Gastroenterol. 101: 513–523.
Daniele, B., Bencivenga, A., Megna, A.S., and
Tinessa, V. 2004. α-fetoprotein and ultrasonog-
raphy screening for hepatocellular carcinoma.
Gastroenterology 127: 108–112.
Di Bisceglie, A.M. 2004. Issues in screening
and surveillance for hepatocellular carcinoma.
Gastroenterology 127: S104–107.
Fattovich, G., Giustina, G., Degos, F., Tremolada, F.,
Diodati, G., Almasio, P., Nevens, F., Solinas, A.,
Mura, D., Brouwer, J.T., Thomas, H., Njapoum,
C., Casarin, C., Bonetti, P., Fuschi, P., Basho, J.,
Tocco, A., Bhalla, A., Galassini, R., Noventa, F.,
Schalm, S.W., and Realdi, G. 1997. Morbidity
10. Screening with Ultrasonography of Patients at High-Risk for Hepatocellular Carcinoma
and mortality in compensated cirrhosis type C:
a retrospective follow-up study of 384 patients.
Gastroenterology 112: 463–472.
Forns, X., Ampurdanes, S., Llovet, J.M., Aponte,
J., Quintó, L., Martínez-Bauer, E., Bruguera,
M., Sánchez-Tapias, J.M., and Rodés, J. 2002.
Identification of chronic hepatitis C patients
without hepatic fibrosis by a simple predictive
model. Hepatology 36: 986–992.
Giannini, E., and Testa, R. 2003. The metabolic
syndrome: all criteria are equal, but some criteria
are more equal than others. Arch. Intern. Med.
163: 2787–2788.
Guechot, J., Laudat, A., Loria, A., Serfaty, L.,
Poupon, R., and Giboudeau, J. 1996. Diagnostic
accuracy of hyaluronan and type III procollagen
amino-terminal peptide serum assays as markers of
liver fibrosis in chronic viral hepatitis C evalu-
ated by ROC curve analysis. Clin. Chem. 42:
558–563.
Imbert-Bismut, F., Ratziu, V., Pieroni, L., Charlotte,
F., Benhamou, Y., Poynard, T. and MULTIVIRC
Group. 2001. Biochemical markers of liver fibro-
sis in patients with hepatitis C virus infection: a
prospective study. Lancet 357: 1069–1075.
Kelleher, T.B., and Afdhal, N. 2005. Noninvasive
assessment of liver fibrosis. Clin. Liver Dis. 9:
667–683.
Kiyosawa, K., Umemura, T., Ichijo, T., Matsumoto,
A., Yoshizawa, K., Gad, A., and Tanaka, E.
2004. Hepatocellular carcinoma: recent trends in
Japan. Gastroenterology 127: S17–26.
Leykum, L.K., El-Serg, H.B., Cornell, J., and
Papadopoulos, K.P. 2007. Screening for hepato-
cellular carcinoma veterans with hepatitis C on
disease stage, treatment received, and survival.
Clin. Gastroenterol. Hepatol. 5: 508–512.
Lu, S.N., Su, W.W., Yang, S.S., Chang, T.T.,
Cheng, K.S., Wu, J.C., Lin, H.H., Wu, S.S.,
Lee, C.M., Changchien, C.S., Chen, C.J., Sheu,
J.C., Chen, D.S., and Chen, C.H. 2006a. Secular
trends and geographic variations of hepatitis B
virus and hepatitis C virus-associated hepatocel-
lular carcinoma in Taiwan. Int. J. Cancer 119:
1946–1952.
Lu, S.N., Wang, J.H., Liu, S.L., Hung, C.H., Chen,
C.H., Tung, H.D., Chen, T.M., Huang, W.S., Lee,
C.M., Chen, C.C., and Changchien, C.S. 2006b.
Thrombocytopenia as a surrogate for cirrhosis
and a marker for the identification of patients at
143
high-risk for hepatocellular carcinoma. Cancer
107: 2212–2222.
Lu, S.N., Wang, J.H., Chen, P.F., Tung, H.D.,
Tseng, P.L., Hung, C.H., Kee, K,M., Chen,
C.H., Chang, K.C., Lee, C.M., Changchien,
C.S., Chen, Y.D., Tsai, L.S., and Chen, T.H.
2008. Community-based mass ultrasonographic
screening of hepatocellular carcinoma among
thrombocytopenic adults. Cancer Epidemiol.
Biomarkers Prev. 17: 1813–1821.
Murawaki, Y., Ikuta, Y., Koda, M., Yamada, S., and
Kawasaki, H. 1996. Comparison of serum 7S
fragment of type IV collagen and serum central
triple-helix of type IV collagen for assessment of
liver fibrosis in patients with chronic viral liver
disease. J. Hepatol. 24: 148–154.
Rossi, E., Adams, L., Prins, A., Bulsara, M., de
Boer, B., Garas, G., MacQuillan, G., Speers, D.,
and Jeffrey, G. 2003. Validation of the FibroTest
biochemical markers score in assessing liver
fibrosis in hepatitis C patients. Clin. Chem. 49:
450–454.
Saitou, Y., Shiraki, K., Yamanaka, Y.,
Yamaguchi, Y., Kawakita, T., Yamamoto, N.,
Sugimoto, K., Murata, K., and Nakano, T.
2005. Noninvasive estimation of liver fibrosis
and response to interferon therapy by a serum
fibrogenesis marker, YKL-40, in patients
with HCV-associated liver disease. World J.
Gastroenterol. 11: 476–481.
Teare, J.P., Sherman, D., Greenfield, S.M.,
Simpson, J., Bray, G., Catterall, A.P., Murray-
Lyon, I.M., Peters, T.J., Williams, R., and
Thompson, R.P. 1993. Comparison of serum
procollagen III peptide concentrations and PGA
index for assessment of hepatic fibrosis. Lancet
342: 895–898.
Toyoda, H., Kumada, T., Kiriyama, S., Sone, Y.,
Tanikawa, M., Hisanaga, Y., Yamaguchi, A.,
Isogai, M., Kaneoka, Y., and Washizu, J. 2006.
Impact of surveillance on survival of patients
with initial hepatocellular carcinoma: a study
from Japan. Clin. Gastroenterol. Hepatol. 4:
1170–1176.
Wai, C.T., Greenson, J.K., Fontana, R.J.,
Kalbfleisch, J.D., Marrero, J.A., Conjeevaram,
H.S., and Lok, A.S. 2003. A simple noninvasive
index can predict both significant fibrosis and
cirrhosis in patients with chronic hepatitis C.
Hepatology 38: 518–526.
144
Walsh, K.M., Fletcher, A., MacSween, R.N., and
Morris, A.J. 2000. Basement membrane peptides
as markers of liver disease in chronic hepatitis C.
J. Hepatol. 32: 325–330.
Wong, J.B., McQuillan, G.M., McHutchison, J.G.,
and Poynard, T. 2000. Estimating future hepatitis
C morbidity, mortality, and costs in the United
States. Am. J. Public Health 90: 1562–1569.
Yuen, M.F., Cheng, C.C., Lauder, I.J., Lam, S.K.,
Ooi, C.G., and Lai, C.L. 2000. Early detection of
hepatocellular carcinoma increases the chance of
S.-N. Lu et al.
treatment: Hong Kong experience. Hepatology
31: 330–335.
Zhang, B.H., Yang, B.H., and Tang, Z.Y. 2004.
Randomized controlled trial of screening for
hepatocellular carcinoma. J. Cancer. Res. Clin.
Oncol. 130: 417–422.
Zoli, M., Magalotti, D., Bianchi, G., Gueli, C.,
Marchesini, G., and Pisi, E. 1996. Efficacy
of surveillance program for early detection
of hepato-cellular carcinoma. Cancer 78:
977–985.
11
Hepatocellular Carcinoma:
Contrast-Enhanced Sonography
Byung Ihn Choi and Se Hyung Kim
INTRODUCTION
In patients with chronic liver disease, sono-
graphy and measurements of serum alpha-
fetoprotein (αFP) levels are frequently used
to screen for hepatocellular carcinoma
(HCC). The sensitivity of sonography for
HCC detection, however, in the end-stage
cirrhotic liver is only 50% (Dodd et al.,
1992). Furthermore, sonographic find-
ings of HCC are variable and frequently
nonspecific. Hepatocellular carcinoma is
highly vascularized and often shows irreg-
ular neovascularization within the tumor.
Therefore, detection and characterization
of tumor vascularity are important in the
differential diagnosis, the choice of treat-
ment method, and assessment of the thera-
peutic response for HCC. Although color
and power Doppler sonography are helpful
in detecting vascularity of hepatic tumors,
it has not proved satisfactory.
Since intravenous microbubble contrast
agents including SH U 508A (Levovist;
Schering AG, Berlin, Germany) for
sonography have become available, many
researchers have reported that contrast-
enhanced sonography could be a promis-
ing technique for assessing post-treatment
response of HCC (Bartolozzi et al., 1998;
Choi, D. et al., 2000; Ding et al.,
2 0 0 1 ; Meloni et al., 2001; Solbiati
et al., 1999; Youk et al., 2003b) as well
as for making a correct diagnosis of HCC
by observing tumoral vascularity (Choi
et al., 2000a; Dill-Macky et al., 2002; Kim
et al., 1998; Wen et al., 2004). Indeed,
there is an increasing consensus that the
use of contrast agents improves the ability
of sonography to detect and characterize
focal hepatic lesions. Levovist, one of the
most popular sonographic contrast agents,
is a suspension of galactose microparticles
in sterile water. The microbubbles (2–8 μm
in diameter, with a mean of 3 μm), which
are stabilized in the microparticle suspen-
sion, can traverse the pulmonary capillary
bed and diverse into whole blood circula-
tion (Kim et al., 1998).
In this chapter, the physics of micro-
bubbles is briefly reviewed, and various
contrast-enhanced sonographic techniques
for HCC in sequence of their develop-
ments and their strengths and weaknesses
are discussed: contrast-enhanced color
Doppler sonography, contrast-enhanced
power Doppler sonography, contrast-
enhanced harmonic power Doppler sonog-
raphy, contrast-enhanced pulse-inversion
harmonic sonography, contrast-enhanced
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
differential diagnosis of a focal hepatic
lesion. Tanaka et al. (1990) have described
a color Doppler pattern of intra-lesional
flow which may be helpful in differen-
tiating HCC from other hepatic tumors.
A basket pattern, fine blood flow net-
work around the HCC nodule, and vessels
within the tumor are seen only in HCC.
This contrasts with the detour or dot pat-
tern seen in metastases or hemangioma,
respectively. However, on conventional
color Doppler sonography, the ability to
evaluate low-velocity blood flow in very
small intratumoral vessels, in particular in
small hepatic lesions and lesions located
deep within hepatic parenchyma, is less
impressive. Attempts to improve color
Doppler imaging led to the development of
ultrasound contrast agents. Consequently,
contrast-enhanced color Doppler sono-
graphy demonstrated more intratumoral
vascularity in HCCs than did conventional
color Doppler sonography (Ernst et al.,
1996; Strobel et al., 2000). Tanaka et al.
(1998) have also shown an improvement
in the detection rate of intratumoral vascu-
larity. In a study on contrast-enhanced
color Doppler sonography, the majority of
HCCs showed extensive or moderate intra-
tumoral vascularity (Strobel et al., 2000).
CONTRAST-ENHANCED
POWER DOPPLER
SONOGRAPHY
Unlike color Doppler sonography, which
emphasizes velocity, variance informa-
tion, or both, power Doppler sonography
is a technique based on the total integrated
power of the Doppler spectrum, which is
related to the number of red blood cells
that produce the Doppler shift. Findings
147
of several studies have shown that power
Doppler sonography is more sensitive than
color Doppler US in the depiction of vas-
cular flow in focal hepatic lesions, includ-
ing HCC (Choi B.I. et al., 1996; Lencioni
et al., 1996). Even using the optimized
parameters for power Doppler sonography,
however, the ability to detect low-velocity
blood flow in very small intratumoral
vessels, particularly in small HCCs and
hemangioma, is still limited.
According to the previous report by Kim
et al. (1998), intra- and peritumoral flow
signals increased in most HCCs (95%)
after administration of contrast agent and
enhanced vascular flow signals in the tumor
changed from an internal flow pattern or a
basket pattern to a prominent mixed pattern.
Contrast enhancement of the Doppler signal
in the tumor began ~ 15–20 s after injection
of contrast agent and reached a maximum
at 30–60 s. This study concluded that con-
trast-enhanced power Doppler sonography
was superior to unenhanced power Doppler
sonography in the demonstration and char-
acterization of tumor vascularity in nodular
or massive HCC.
CONTRAST-ENHANCED
HARMONIC POWER
DOPPLER SONOGRAPHY
Even though contrast-enhanced power
Doppler sonography showed a good per-
formance to depict tumor vascularity in
HCC, it still has several limitations related
to intrinsic weaknesses of power Doppler
technique. First, high susceptibility to tissue
motion limits the application of this tech-
nique in patients with poor breath-holding
and those with hepatic masses near the heart
or pulsatile great vessels. Second, color
148
blooming artifact, which can degrade image
quality, is frequently encountered immedi-
ately after bolus injection of contrast agent
(Kim et al., 1998). In one study, the size
of the vessel enhanced with contrast agent
frequently appeared larger than its actual
size, even though reduction of Doppler
gain was done until artifacts outside the
vessel nearly disappeared (Choi et al.,
2000a). Third, power Doppler artifacts pro-
duced from areas with high echogenicity
frequently mimic intratumoral vessels and
make assessing intratumoral vascularity
difficult (Kim et al., 1999). To circumvent
such limitations, the harmonic techniques
have been introduced and merged into con-
trast-enhanced power Doppler sonography.
The use of harmonic mode can effectively
reduce artifacts including blooming and
motion-related artifacts.
In our previous study, we performed a
prospective study to compare contrast-
enhanced harmonic power Doppler sonog-
raphy with conventional power Doppler
sonography in the assessment of tumor
vascularity in HCC (Choi et al., 2000b).
In most cases, the intensity of enhanced
signals in the tumor and hepatic paren-
chyma in conventional mode was stronger
than that in harmonic mode (Choi et al.,
2000b). This is not surprising when consid-
ering the relatively low intensity of second
harmonic echoes compared with funda-
mental echoes (Burns, 1996). However,
harmonic mode was superior to conven-
tional mode in terms of power Doppler
artifacts at any imaging time. In harmonic
power Doppler mode, continuous scan-
ning during a patient’s quiet breathing was
possible without producing considerable
motion-related artifacts. Overestimation of
the size of contrast-enhanced vessel was
evident on conventional power Doppler
B.I. Choi and S.H. Kim
sonography. Similarly, the sizes of intrahe-
patic and intratumoral vessels likely were
overestimated. Enhanced vessels adjacent
to a liver tumor can be seen as if they are
located within the tumor on conventional
power Doppler sonography. This can be an
important problem when a physician needs
to know whether any residual viable tumor
is present after local treatment for a malig-
nant tumor. In that case, harmonic power
Doppler sonography might be superior to
conventional power Doppler sonography
(Choi et al., 2000b).
CONTRAST-ENHANCED
PULSE-INVERSION
HARMONIC SONOGRAPHY
Like tissue harmonic imaging, pulse-
inversion harmonic imaging (PIHI) displays
the amplitude of the harmonic signals result-
ing from non-linear echoes. However, in
PIHI, two identical pulses with reversed
polarity which are transmitted down each
ray line were used. When these result-
ant returned waveforms are added, the
harmonic component gives the multiplied
harmonic level of a single waveform,
while all linear fundamental components
are canceled out more effectively than
those on tissue harmonic imaging using a
simple bandpass filter (Hohl et al., 2004;
Jang et al., 2000). This technique, instead
of using a narrow receiver filter tuned
around harmonic frequency components
which is used in a simple bandpass fil-
tering method, allows the use of broader
transmit and receive bandwidths to achieve
improved resolution. Therefore, PIHI can
overcome the bandwidth limitations of
filtering harmonic imaging (Hohl et al.,
2004; Jang et al., 2000).
11. Hepatocellular Carcinoma: Contrast-Enhanced Sonography 149

Although PIHI produces significantly bet-
ter image quality for focal hepatic lesions,
it has several limitations that should be
overcome for it to replace conventional
sonography in routine practice. Because
multiple pulses are used, the frame rate is
relatively low compared with that of con-
ventional sonography or tissue harmonic
imaging. Although we often found that PIHI
improves the conspicuousness of a HCC
barely recognizable on conventional US,
we cannot use PIHI during biopsy because
of its slow frame rate. In addition, the fact
that multiple pulses are used to form each
ray line, makes it susceptible to motion arti-
facts. Anything that moves between pulses
is not completely cancelled, so some tissue
motion artifacts are seen when a transducer
moves between frames.
In contrast to power Doppler sonography,
contrast-enhanced PIHI with interval delay
scanning can provide strong gray-scale
enhancement by microbubble contrast agents
and detect signals from microbubbles in very
slow flow without Doppler-related arti-
facts (Figure 11.1). It has been shown that
this technique can depict the enhancement
patterns of various hepatic tumors includ-
ing HCC. Kim et al. (2000) performed a
prospective study with contrast-enhanced
PIHI in various liver tumors. Hepatocellular
carcinomas with or without irregular intra-
tumoral enhancing vessels show findings
similar to those on hepatic angiography

Figure 11.1. Contrast-enhanced pulse-inversion harmonic sonography of hepatocellular carcinoma.

Gray-scale sonogram before contrast administration (left upper) shows a heterogeneous hyperechoic mass
(arrows) in the liver. Contrast-enhanced pulse-inversion harmonic sonograms obtained 33 s (right upper)
and 59 s (left lower) after injection shows heterogeneous enhancement of the tumor (arrows). A tubular
structure with intense enhancement (arrowhead) that represents an irregular intratumoral vessel is seen.
On contrast-enhanced pulse-inversion harmonic sonogram (right lower) obtained 213 s after injection
reveals mild wash-out of contrast material within the tumor (arrows). (Reproduced with permission from
Kim et al., 2000.)
150
(Figure 11.1). The pattern of enhancement
on PIHI in that study was thought to be
useful in the differential diagnosis of liver
tumors (Kim et al., 2000). As compared
with computed tomography (CT) or mag-
netic resonance imaging (MRI), there were
a few limitations of the sonographic tech-
nique used in that study. The first is that
dynamic sonographic scanning is possible
in only one scanning plane; thus, it is not
possible to characterize all lesions simul-
taneously in patients with multiple lesions.
Second, interval delay scanning in the same
area is not easy for unskilled examiners.
It may be difficult even for experienced
examiners to scan the lesion in the liver
region that is visible only during a patient’s
deep inspiration. Sufficient practice of inter-
val delay scanning at the same area, includ-
ing the tumor, must be done before injection
of the contrast agent in such cases. However,
sonography is superior to CT or MRI in
that immediate, accurate characterization
of a newly detected focal hepatic lesion on
sonographic examination is possible (Kim
et al., 2000).
CONTRAST-ENHANCED
CODED HARMONIC
SONOGRAPHY
In general, transmitted sonographic pulse
is distorted in shape and loses energy
as it travels through tissue. It is true for
contrast-enhanced signals as well as for
fundamental sonic waves. To improve the
detectability of the enhanced microbub-
ble signals, researchers have studied the
behavior of microbubbles in blood flow
and developed newer sonographic tech-
nology. Coded harmonic angio (CHA)
is one of the new imaging techniques
B.I. Choi and S.H. Kim
(Lee et al., 2002). It combines PIHI with
a coded excitation technique, which trans-
mits coded pulse sequence, decoding them
on receipt. Coded technology uses a code
and decode pulse sequence system, so-
called digitally encoded sonographic tech-
nology, to enhance the signals from the
contrast agent and, at the same time, sup-
press unwanted background, fundamental,
and harmonic signals from the tissue. This
technique can be paired with the benefits
of B-flow. B-flow optimizes direct visuali-
zation of blood cells on gray-scale imag-
ing. Because blood echoes are very weak,
it uses codes to boost weak blood echoes
and equalize non-moving tissue signal.
For that reason, B-flow technique does
not cause Doppler artifact. Ultimately,
contrast-enhanced CHA can provide
extremely high quality and detailed vas-
cular information without the artifacts
associated with contrast agent as well as
preservation of wideband resolution of
harmonic signals.
In a previous study using contrast-
enhanced CHA, the typical hemodynamic
pattern of HCCs was shown as abundant
tumor vessels supplied from the periphery
to the center of the tumor and dense paren-
chymal tumor staining with fast washout
(Figure 11.2) and the sensitivity and
specificity of those findings were 92.8%
and 92.3%, respectively (Wen et al., 2004).
Our previous study showed similar results
that contrast-enhanced CHA was an effec-
tive sonographic technique in evaluating
the vascularity of HCCs because it has the
advantages such as the depiction of excel-
lent dynamic enhancement of tumors, high-
resolution imaging of detailed intratumoral
vessels, and no sonographic artifacts (Lee et
al., 2003) (Figure 11.2). However, despite
many advantages of CHA, CHA had several
11. Hepatocellular Carcinoma: Contrast-Enhanced Sonography 151

Figure 11.2. Contrast-enhanced coded harmonic sonography of hepatocellular carcinoma. Unenhanced

transverse sonogram (upper left) shows a heterogeneous hypochoic mass (arrow) in the liver. The image
obtained at 30 s (upper right) after contrast injection reveals multiple feeding vessels and intratumoral ves-
sels (arrows) within the tumor. Strong and homogeneous enhancement of the tumor is seen on the image
obtained at 50 s (lower left) after injection of the contrast agent. These enhanced flow signals within the
tumor markedly decrease and disappear at 100 s (lower right) after injection

drawbacks. First, the stationary tissue back-

technique, is a multi-pulse technique that
ground is severely suppressed. Therefore,
provides very high-resolution images,
the quality of unenhanced CHA is inferior to
which detail the effect of microbubble
conventional gray-scale images, particularly
emission, and uses the property of bubble
in a deep-seated lesion (Lee et al., 2003).
emission to dramatically show regions
Second, interval delay scanning in the same
where the contrast agent has localized.
area is not easy for unskilled examiners.
Pulse inversion harmonic imaging uses
Sufficient practice of interval delay scan-
two alternately phased pulses and adds
ning in the area that includes the tumor must
echoes from both pulses together. On the
be done before injection of contrast agent.
contrary, ADI uses two pulses with the
same polarity and subtracts the signals
from the two pulses; consequently, only
CONTRAST-ENHANCED fundamental and harmonic contrast agent
AGENT DETECTION signals remain. Therefore, ADI may be
IMAGING referred to as contrast-only imaging. As
a result, ADI may depict signals from
Agent detection imaging (ADI), a most microbubbles better than those from PIHI.
recently developed high-MI harmonic When an intermittent imaging strategy is
152 B.I. Choi and S.H. Kim

used with ADI, it may effectively reveal
the vascularity of focal liver lesions and
may also help in lesion characterization.
Furthermore, ADI may separate the tis-
sue image from the contrast-only image
(Figure 11.3). Because the tissue image
and contrast-only image can be shown sep-
arately, the positional relationship between
the tumor and other structures can be accu-
rately assessed (Figure 11.3).
Several reports have suggested that
contrast-enhanced ADI sonography has
improved the detection and characteriza-
tion of focal liver lesions (Kim et al.,
2005b; Youk et al., 2003a). According to
our recent report, contrast-enhanced ADI
sonography can be used to characterize
focal hepatic lesions with a higher diag-
nostic confidence than baseline sonogra-
phy and furthermore, it reduced the need
for further diagnostic workups for focal
hepatic lesion characterization (Kim et al.,
2005b) (Figure 11.3). Our another recent
report showed that single-level dynamic
sonography with ADI was potentially
useful in diagnosing HCC, with excellent
visibility of intra- and peri-tumoral vascu-
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
delayed capsular enhancement (Kim et al.,
2003).
LOW MECHANICAL INDEX
IMAGING
Until recently, visualization of flow in
the microcirculation has required some
form of triggering or low-frame rate imag-
ing. To detect the bubbles, the ultrasound
had to be strong enough to destroy them.
However, the increased sensitivity pro-
vided by newer imaging techniques makes
it possible to image contrast microbubbles
relatively nondestructively in real time at
very low acoustic pressures. Therefore,
low MI scanning technique makes it
possible to perform contrast-enhanced
perfusional sonography for various hepatic
lesions (Schneider et al., 1995; Schneider,
1999; von Herbay et al., 2004).
Low MI scanning is important for two
reasons. First, at a low MI, bubble destruc-
tion is avoided. Although microbubbles
differ in their shell composition, previous
work indicates that at an MI of ~ 0.15, the
microbubbles examined are not destroyed
markedly, yet give a good harmonic con-
trast signal. The second major reason for
low MI scanning is the reduction of the
harmonic component in the tissue echoes
relative to bubble echoes. Although tissue
harmonics have benefited routine diagnos-
tic scanning, it is the background “noise”
signal above which the contrast signal must
rise. Because tissue is less nonlinear than
bubbles, it requires a higher MI than the
contrast microbubbles for a certain har-
monic response. Therefore, at a low MI, the
contrast-to-tissue ratio is higher than at a
high MI, helping to remove the tissue signal
and leaving only the contrast. The details of
153
contrast-enhanced perfusional sonography
for HCC will be introduced below.
CONTRAST-ENHANCED
SONOGRAPHY
IN ASSESSING THE
THERAPEUTIC RESPONSE
OF HEPATOCELLULAR
CARCINOMAS
Sonography or CT is most often used as
the primary guidance technique in local
ablation therapy such as percutaneous
ethanol injection (PEI) and radiofrequency
ablation (RFA) for HCC. Of these modali-
ties, gray-scale sonography is preferred
because it is more widely available, less
expensive, and allows real-time monitor-
ing, which facilitates placement of the
electrode needle. However, it is of little
help in determining the completeness of
the procedure or in long-term follow-up
of therapeutic efficacy (Lencioni et al.,
1995). Postablation gray-scale sonogra-
phy cannot help differentiate viable tumor
from necrotic tissue because of variable
echogenicity and because postablation
sonographic findings do not correlate well
with the overall shape and volume of the
necrosis (Goldberg et al., 2000). Although
several reports have shown that color
Doppler sonography is somewhat useful
in the assessment of therapeutic efficacy
after local ablation therapy for HCC, nei-
ther color nor power Doppler sonography
has proved entirely satisfactory because of
low sensitivity to microvasculature within
residual viable tumor (Choi et al., 2000b;
Rossi et al., 1996).
Because blood pool microbubble con-
trast agents have become available for use
154 B.I. Choi and S.H. Kim

with Doppler sonography, there have been
a lot of researches underway in the use-
fulness of contrast-enhanced sonography
for assessing the therapeutic responses of
HCC after treatment. Several reports have
suggested that contrast-enhanced color and
power Doppler sonography can be useful
in detecting residual or recurrent tumors
after local ablation of HCC (Goldberg
et al., 2000) (Figure 11.4). According to
a preliminary report, contrast-enhanced
power Doppler sonography can be used
as an alternative to immediate follow-up
CT for early assessment of the therapeu-
tic efficacy of RF ablation of HCC (Choi
et al., 2000b) (Figure 11.4).
More recently, contrast-enhanced gray-
scale harmonic sonography has been
developed for the evaluation of therapeu-
tic response to RF ablation therapy (Choi
et al., 2003; Kim et al., 2005b) (Figures
11.5 and 11.6). This technique allows
detection of residual tumor after ablation
without the blooming and motion artifacts
seen at contrast-enhanced color or power
Doppler sonography. Another advantage
of contrast-enhanced gray-scale harmonic
sonography over CT and MR imaging
is that it can directly show persistent
fine vascular flow signals within the
tumor during a period of effective con-
trast enhancement that lasts for more than
10 min. Consequently, contrast-enhanced
gray-scale harmonic sonography could be
helpful in detecting residual or recurrent
tumor after local ablation of HCC. These
modalities may also be helpful in identify-
ing residual tumor when performed during

Figure 11.4. Successful radiofrequency ablation of hepatocellular carcinoma. Unenhanced, transverse,

gray-scale (upper left) and power Doppler (upper right) sonograms before radiofrequency ablation (RFA)
demonstrate a large heterogeneous mass (arrows) with hypervascularity in right hepatic lobe. Transverse
gray-scale (lower left) and contrast-enhanced power Doppler (lower right) sonograms obtained 1 day after
RFA show no flow signals within the ablated area (arrows), a finding that is consistent with complete
tumor necrosis
11. Hepatocellular Carcinoma: Contrast-Enhanced Sonography 155

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
of viable tumor is needed. This real-time
confirmation of the accurate localization
of viable tumor and of correct placement
of the electrode within the tumor cannot
be expected at CT, and may reduce the
number of treatment sessions without fur-
ther use of contrast-enhanced CT or MR
imaging. Furthermore, contrast-enhanced
CT is less than ideal because errors in
contrast bolus timing cannot always be cor-
rected immediately with repeat scanning due
to the risk of toxicity. However, real-time
sonography and the ability to administer
repeated doses of sonographic contrast
material may, in some cases, allow better
detection of residual tumor.
In conclusion, many recent advances
in contrast-enhanced sonography imag-
ing have occurred, making it a strong
contender for several problem-solving
applications. Admittedly, CT and MR
imaging continue to have complimentary
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
Dill-Macky, M.J., Burns, P.N., Khalili, K., and
Wilson, S.R. 2002. Focal hepatic masses:
enhancement patterns with SH U 508A and
pulse-inversion US. Radiology 222:95–102.
Ding, H., Kudo, M., Onda, H., Suetomi, Y.,
Minami, Y., Chung, H., Kawasaki, T., and
Maekawa, K. 2001. Evaluation of posttreatment
response of hepatocellular carcinoma with con-
trast-enhanced coded phase-inversion harmonic
US: comparison with dynamic CT. Radiology
221:721–730.
Dodd, G.D. III, Miller, W.J., Baron, R.L., Skolnick,
M.L., and Campbell, W.L. 1992. Detection of
malignant tumors in end-stage cirrhotic livers:
efficacy of sonography as a screening technique.
AJR Am. J. Roentgenol. 159:727–733.
Ernst, H., Hahn, E.G., Balzer, T., Schlief, R., and
Heyder, N. 1996. Color Doppler ultrasound of
liver lesions: signal enhancement after intrave-
nous injection of the ultrasound contrast agent
Levovist. J. Clin. Ultrasound 24:31–35.
Goldberg, S.N., Gazelle, G.S., Compton, C.C.,
Mueller, P.R., and Tanabe, K.K. 2000. Treatment
of intrahepatic malignancy with radiofrequency
ablation: radiologic-pathologic correlation.
Cancer 88:2452–2463.
Gramiak, R., and Shah, P.M. 1968. Echocardiography
of the aortic root. Invest. Radiol. 3:356–366.
Hohl, C., Schmidt, T., Haage, P., Honnef, D.,
Blaum, M., Staatz, G., and Guenther, R.W. 2004.
Phase-inversion tissue harmonic imaging com-
pared with conventional B-mode ultrasound in
the evaluation of pancreatic lesions. Eur. Radiol.
14:1109–1117.
Jang, H.J., Lim, H.K., Lee, W.J., Kim, S.H., Kim,
K.A., and Kim, E.Y. 2000. Ultrasonographic
evaluation of focal hepatic lesions: comparison
of pulse inversion harmonic, tissue harmonic, and
conventional imaging techniques. J. Ultrasound
Med. 19:293–299.
Kim, A.Y., Choi, B.I., Kim, T.K., Han, J.K.,
Yun, E.J., Lee, K.Y., and Han, M.C. 1998.
Hepatocellular carcinoma: power Doppler
US with a contrast agent–preliminary results.
Radiology 209:135–140.
Kim, C.K., Choi, D., Lim, H.K., Kim, S.H., Lee,
W.J., Kim, M.J., Lee, J.Y., Jeon, Y.H., Lee, J.,
Lee, S.J., and Lim, J.H. 2005a. Therapeutic
response assessment of percutaneous radiofre-
quency ablation for hepatocellular carcinoma:
157
utility of contrast-enhanced agent detection
imaging. Eur. J. Radiol. 56:66–73.
Kim, K.W., Choi, B.I., Park, S.H., Kim, H.C.,
Lee, M.W., Kim, S.H., Lee, K.H., Park, C.H.,
Kim, J.S., Won, H.J., and Han, J.K. 2003.
Hepatocellular carcinoma: assessment of vascu-
larity with single-level dynamic ultrasonography
during the arterial phase. J. Ultrasound Med.
22:887–896.
Kim, S.H., Lee, J.M., Lee, J.Y., Han, J.K., An,
S.K., Han, C.J., Lee, K.H., Hwang, S.S., and
Choi, B.I. 2005b. Value of contrast-enhanced
sonography for the characterization of focal
hepatic lesions in patients with diffuse liver dis-
ease: receiver operating characteristic analysis.
AJR Am. J. Roentgenol. 184:1077–1084.
Kim, T.K., Han, J.K., Kim, A.Y., Park, S.J., and
Choi, B.I. 1999. Signal from hepatic hemangi-
omas on power Doppler US: real or artefactual?
Ultrasound Med. Biol. 25:1055–1061.
Kim, T.K., Choi, B.I., Han, J.K., Hong, H.S.,
Park, S.H., and Moon, S.G. 2000. Hepatic
tumors: contrast agent-enhancement patterns
with pulse-inversion harmonic US. Radiology
216:411–417.
Lee, J.Y., Choi, B.I., Han, J.K., Kim, A.Y., Shin,
S.H., and Moon, S.G. 2002. Improved sono-
graphic imaging of hepatic hemangioma with
contrast-enhanced coded harmonic angiography:
comparison with MR imaging. Ultrasound Med.
Biol. 28:287–295.
Lee, J.Y., Choi, B.I., Han, J.K., Kim, A.Y., Shin,
S.H., and Park, C.M. 2003. Assessment of
hepatocellular carcinoma vascularity: compari-
son of contrast-enhanced coded harmonic ultra-
sound with harmonic power Doppler ultrasound.
J. Med. Ultrasound 11:147–155.
Lencioni, R., Caramella, D., and Bartolozzi, C.
1995. Hepatocellular carcinoma: use of color
Doppler US to evaluate response to treatment
with percutaneous ethanol injection. Radiology
194:113–118.
Lencioni, R., Pinto, F., Armillotta, N., and
Bartolozzi, C. 1996. Assessment of tumor vascu-
larity in hepatocellular carcinoma: comparison
of power Doppler US and color Doppler US.
Radiology 201:353–358.
Meloni, M.F., Goldberg, S.N., Livraghi, T., Calliada,
F., Ricci, P., Rossi, M., Pallavicini, D., and
Campani, R. 2001. Hepatocellular carcinoma
158
treated with radiofrequency ablation: compari-
son of pulse inversion contrast-enhanced har-
monic sonography, contrast-enhanced power
Doppler sonography, and helical CT. AJR Am. J.
Roentgenol. 177:375–380.
Porter, T.R., Xie, F., Kricsfeld, D., and Armbruster,
R.W. 1996. Improved myocardial contrast with
second harmonic transient ultrasound response
imaging in humans using intravenous perfluor-
ocarbon-exposed sonicated dextrose albumin. J.
Am. Coll. Cardiol. 27:1497–1501.
Rossi, S., Di, S.M., Buscarini, E., Quaretti, P.,
Garbagnati, F., Squassante, L., Paties, C.T.,
Silverman, D.E., and Buscarini, L. 1996.
Percutaneous RF interstitial thermal ablation
in the treatment of hepatic cancer. AJR Am. J.
Roentgenol. 167:759–768.
Schneider, M. 1999. SonoVue, a new ultrasound con-
trast agent. Eur. Radiol. 9: Suppl. 3:S347–S348.
Schneider, M., Arditi, M., Barrau, M.B., Brochot,
J., Broillet, A., Ventrone, R., and Yan, F. 1995.
BR1: a new ultrasonographic contrast agent
based on sulfur hexafluoride-filled microbub-
bles. Invest. Radiol. 30:451–457.
Solbiati, L., Goldberg, S.N., Ierace, T., Dellanoce,
M., Livraghi, T., and Gazelle, G.S.1999. Radio-
frequency ablation of hepatic metastases: post-
procedural assessment with a US microbubble
contrast agent–early experience. Radiology
211:643–649.
Strobel, D., Krodel, U., Martus, P., Hahn, E.G.,
and Becker, D. 2000. Clinical evaluation of
B.I. Choi and S.H. Kim
contrast-enhanced color Doppler sonography in
the differential diagnosis of liver tumors. J. Clin.
Ultrasound 28:1–13.
Tanaka, S., Kitamra, T., Fujita, M., and Yoshioka, F.
1998. Value of contrast-enhanced color Doppler
sonography in diagnosing hepatocellular carci-
noma with special attention to the “color-filled
pattern”. J. Clin. Ultrasound 26:207–212.
Tanaka, S., Kitamura, T., Fujita, M., Nakanishi,
K., and Okuda, S. 1990. Color Doppler flow
imaging of liver tumors. AJR Am. J. Roentgenol.
154:509–514.
von Herbay, A., Vogt, C., Willers, R., and
Haussinger, D. 2004. Real-time imaging with
the sonographic contrast agent SonoVue: differ-
entiation between benign and malignant hepatic
lesions. J. Ultrasound Med. 23:1557–1568.
Wen, Y.L., Kudo, M., Zheng, R.Q., Ding, H., Zhou,
P., Minami, Y., Chung, H., Kitano, M., Kawasaki,
T., and Maekawa, K. 2004. Characterization
of hepatic tumors: value of contrast-enhanced
coded phase-inversion harmonic angio. AJR Am.
J. Roentgenol. 182:1019–1026.
Youk, J.H., Kim, C.S., and Lee, J.M. 2003a.
Contrast-enhanced agent detection imaging:
value in the characterization of focal hepatic
lesions. J. Ultrasound Med. 22:897–910.
Youk, J.H., Lee, J.M., and Kim, C.S. 2003b.
Therapeutic response evaluation of malignant
hepatic masses treated by interventional proce-
dures with contrast-enhanced agent detection
imaging. J. Ultrasound Med. 22:911–920.
12
Focal Liver Lesion: Nonlinear
Contrast-Enhanced Ultrasound Imaging
Vincenzo Migaleddu and Giuseppe Virgilio
INTRODUCTION
Ultrasound (US) is the most widespread
imaging modality for screening focal liver
lesions. Noninvasive diagnosis (detection
and characterization) of focal liver lesions
is still carried out in many centers by com-
puterized tomography (CT) and magnetic
resonance (MR) using contrast agents.
Hepatic lesions have in effect charac-
teristic perfusion and enhancement pat-
terns during the various vascular phases,
which assist characterization and in many
cases make definite diagnosis possible.
Conventional and Doppler US are often
limited in diagnosing this kind of lesion as
color Doppler vascular information of the
lesion is limited (see chapter 11).
Recently, the detection and characteri-
zation of focal hepatic lesions by US has
increased considerably due to the devel-
opment of new nonlinear (harmonic and
nonlinear fundamental) technologies that
optimize detection of contrast agents.
Furthermore, small lesions may not be
well defined by CT or MR imaging
because real-time visualization of lesion
vascularization is not possible. Because
contrast-enhanced US (CEUS) is a
dynamic analysis that allows this, small
lesions may be detected that are some-
times missed with CT or MR imaging.
Positron emission tomography (PET)
is proving to be a molecular imaging
modality with very high sensitivity, but
not competitive as the advantages do not
outweigh the high cost of the exam.
Indeed, color Doppler is a method that
can separate echoes from blood and tissue.
This is possible because of the relatively
high velocity of blood in comparison to
that of the surrounding tissue. This dis-
tinction is valid for flow in large vessels
where the Doppler signals of blood flow
may be separated from those of the tissue
clutter using a wall filter. This fails for
the parenchymal micro-vessel flow where
blood velocity is similar or slower than
that in tissue. In fact in the latter case the
Doppler shift frequency coming from the
moving tissue is comparable or higher
(up to a thousand times) than that of the
blood which perfuses it. Thus, when we
analyze together tissue and the blood in
it, wall filters eliminate both the flow
and the clutter echoes. “Flash” artifact
in color and “thump” artifact in spectral
Doppler represent overwhelming signals
from tissue movements which corrupt
image information.
159
160
Ultrasound contrast media are sus-
pensions containing air or gas micro-
bubbles that increase reflection of the
US beam used during the investigation.
Contrast agents were previously used as
echo-enhancers as they amplified blood
reflectivity and, therefore, increased the
signal-to-noise ratio for Doppler signals
amplifying blood reflectivity and increas-
ing the signal-to-noise ratio. This enhances
the echogenicity of the blood but creates
more artifacts due to blooming and tissue
motion (Powers et al., 1997). Another,
more efficient way to avoid artifacts would
be to obtain a contrast B-mode image
that suppresses Doppler artifacts without
the use of a velocity-dependent filter.
Contrast- enhanced ultrasound, defined as
nonlinear or harmonic imaging, aims to
provide such a method.
HARMONIC IMAGING:
BASIC PRINCIPLES
Harmonics are additional frequencies that
are multiples of the fundamental (transmit-
ted) frequency and are commonly found in
acoustics. Double and fourfold overtones
(harmonics) give character to the different
musical instruments. The same pheno-
menon occurs in ultrasound (Hamilton and
Blackstock, 1998). We need to distinguish
two different modes of harmonic gene-
ration: those produced by transmission
through the tissue and those produced
by interaction with microbubbles. Tissue
harmonics are derived from the nonlin-
ear distortion of the incident beam as it
passes through the tissues. The pheno-
menon occurs for US beams of a sufficiently
high power. A well-known basic principle
of US, the speed of propagation of an
V. Migaleddu and G. Virgilio
US beam, depends on the density of the
tissue. The compression part of the US
wave increases by fractions the density of
the tissue, and thus this part of the wave
travels marginally faster than the rarefac-
tion part of the wave where the density is
lower. Thus, over distance, the shape of
the waveform becomes distorted and its
angular components represent the over-
tones or harmonics. These can be filtered
and used for imaging (Ward et al., 1997).
The advantage is that they depend on a
high acoustic pressure for their formation,
and therefore are not produced as much
by the weaker unwanted side and grat-
ing lobes or by reverberations than by the
main lobe. Thus, tissue harmonic images
are cleaner with fewer side lobe and rever-
beration artifacts.
The harmonics are generated from within
the tissue and only at the center portion
of the main US beam. This results in higher
tissue contrast. The disadvantage is that
the harmonic echoes are some 20 dB less
intense than the fundamental echoes with
the risk of reducing the signal-to-noise
ratio. The disadvantage is that the harmonic
echoes are some 20 dB less intense than
fundamental echoes with a corresponding
reduction of the signal-to-noise ratio. This
is why harmonics are better handled by
newer generation systems powered with a
high dynamic range than those with older
technology-designed scanners. When it
works well, harmonic imaging is spec-
tacularly useful in sharpening up images
since it increases contrast by eliminating
noise, especially in the evaluation of the
fluid-filled structures such as the biliary
tree, gallbladder and vascular structures
(Ortega et al., 2001). In clinical practice,
for most scanners the benefit in abdomi-
nal and general work is so great that most
12. Focal Liver Lesion: Nonlinear Contrast-Enhanced Ultrasound Imaging
users now set the systems to default in a
tissue harmonic mode; however, in our
opinion, harmonic mode is more useful
in the evaluation of fluid-filled structures
positioned in the middle field because
a penetration loss occurs in the far field
with the poor echo intensity of deeper
tissue second harmonics (Migaleddu
et al., 2002).
Microbubble nonlinear (harmonic)
responses can be produced in two dif-
ferent ways: microbubble oscillations or
disruption. Microbubbles are air or gas
bodies ranging from 1 to 7 μm in diameter,
enclosed by some sort of membrane. Two
important characteristics of microbubbles,
stability and diameter, must be optimized
in order to have efficient contrast media.
Stability means that microbubbles of con-
trast medium must dwell in the blood
stream long enough to allow the comple-
tion of an US investigation with contrast
enhancement. Microbubbles must also be
of a diameter necessary to obtain contrast
medium that can pass through the pul-
monary circulation and that, once in the
systemic circulation, can act as a contrast-
enhancing medium in all organs (http://
www.smirg.org/.contrast.php). Several
cardiac and vascular US contrast agents
are available on the American, Canadian,
European or worldwide market; Echovist
and Levovist (Schering), SonoVue (Bracco)
are available in a number of European
countries. Other agents that underwent or
are still undergoing American regulatory
approval include Optison (Amersham/GE
Healthcare), Definity (BMS), Imagent
(Alliance Pharmaceutical), Sonazoid
(Amersham/GE Healthcare), Quantison
(Quadrant Healthcare), SonoVue (Bracco),
Biosphere (Ponit Biomedical), and AI-700.
Albunex and Echovist, used mainly in
161
cardiac evaluations, are effectively one-
pass-only agents and have been replaced
by new-generation agents such as Optison,
Levovist and SonoVue.
For radiology, the two most important
are Levovist (Schering) and SonoVue
(Bracco). Levovist is a first-generation
contrast medium of air containing micro-
bubbles made by shaking galactose micro-
particles with water. The galactose
micro-particles contain micro-defects,
which force the attached air microbubbles
into the required size. They are stabilized
by a monomolecular layer of a surfactant,
palmitic acid (Nanda and Carstensen,
1997). The second-generation contrast
medium SonoVue uses an inert high-
density gas (sulfur hexafluoride), which
improves the longevity of the microbub-
bles because of its high molecular weight,
which slows diffusion. Microbubble shell
is constituted by a phospholipid layer
which is similar to that of cell membranes
(Schneider et al., 1995; Morel et al.,
2000). The essence of their function
as contrast agents is that microbubbles
behave quite differently from solid or
watery tissues in that they can be comp-
ressed and expanded much more readily.
Thus, they change their diameter from
two to tenfold, depending on the power
of the incident US and, like all reac-
tive processes, have a natural resonant
frequency at which they respond most
actively. It happens that the resonant fre-
quency for 1–7 μm microbubbles lies in
the 2–10 MHz range that is used for diag-
nostic imaging. This lucky coincidence
explains the remarkable reflectivity of
microbubbles which are many more times
echogenic than comparable tissue elements
such as red blood cells (Cosgrove in
http://www.smirg.org/lectures.php).
162
PHYSICAL AND
TECHNOLOGICAL
BACKGROUND
When an US beam encounters microbub-
bles present in an insonated field, different
information is acquired depending on the
pressure exerted by the US waves, which
in turn is related to the mechanical index
(MI). The mechanical index is the tool
which expresses the acoustic pressure of
the US beam produced on insonated struc-
tures; it enables the power of the US beam
to be quantified and is directly propor-
tionate to the peak of maximum negative
pressure at the point of focalization of the
beam (P), and is inversely proportionate to
the mean frequency of the US beam (V) in
the following
P metrically to symmetrically transmitted
ratio: M.I. =
V either the form of a change in the bubbles’
This quantity is related to the mechanical
work that can be carried out on the micro-
bubble during the negative cycle of US
propagation (Apfel and Holland, 1991).
With very low mechanical indexes reflec-
tion of ultrasounds occurs. For slightly
higher mechanical indexes (up to 0.2) a
nonlinear signal is emitted in relation to
the resonance of the microbubbles that
are not destroyed, and with a mechanical
index greater than 1.2 the microbubbles
break. This creates transitory echoes in
second harmonics with a high signal inten-
sity (Chin and Burns, 1997).
First-generation contrast agents exploit
the high MI and corresponding microbub-
bles rupture by using intermittent software
imaging. Fragile microbubbles that con-
tain air and have a galactose shell stabi-
lized by palmitic acid (such as Levovist
or first-generation contrast agent) result
V. Migaleddu and G. Virgilio
in limited stability and shortened time
available for examination. They behave
like soap bubbles and the greatest amount
of high intensity information in response is
obtained the moment they break. Second-
generation contrast agents (SonoVue) are
made up of microbubbles with notable
elasticity, containing a gas (sulfur hex-
afluoride or similar), and are wrapped
in a shell of phospholipids and/or albu-
min. This enables real-time continuous-
mode software to be used, exploiting the
information from the resonance of the
microbubbles that are maintained in time,
though their duration is not unlimited and
they break spontaneously after 8–10 min.
Microbubbles like any gas particle resist
compression more than they respond to
expansion so that they may respond asym-
US waves. This asymmetry may take
shape whereby they become geometrically
distorted or of an asymmetry in the temporal
response. The latter is perhaps more usual
because the bubbles expand more quickly
than they compress. In either case the result
is that the returning wave contains differ-
ent frequencies from those of the insonant
wave and if these can be extracted, images
favoring the contribution of the microbub-
bles can be created (Greis, 2004).
A variety of algorithms that have been
marketed involve enhancement of the sig-
nal coming from the microbubbles and
cancellation of signals from tissues. They
favor a reading of the echoes with specific
frequency of contrast agent resonance in
second harmonics, and at the same time
cancel the fundamental echoes coming
from the tissues. In pulse inversion, also
known as phase inversion, two pulses in
rapid succession are emitted into the tissue;
12. Focal Liver Lesion: Nonlinear Contrast-Enhanced Ultrasound Imaging
the second pulse undergoes a 180° phase
change so to be a mirror image of the first.
The fundamental response produced in
the tissues which will be phase-inverted
compared with the fundamental pulses is
erased because the sum of two inverted
pulse is zero. In this way the echoes from
the microbubbles are enhanced so as to
differentiate them from those coming from
stationary tissues (Burns et al., 2000).
A great spatial and details resolution (capa-
bility to identify macro and micro vessels)
is achieved by intermittent imaging and spe-
cific software – Agent Detection Imaging
(ADI by Acuson Sequoia, Siemens) – with
a high MI and first-generation contrast
agent. The ADI can be used for a short
time either with static techniques or in
real-time (Migaleddu et al., 2004).
Compared with the intermittent imaging
characteristic of the reconstruction algo-
rithms used in breaking first-generation
contrast agents, the algorithms based on
the use of a low mechanical index enable
signals from the resonance of second-
generation microbubbles to be detected
in real-time; this permits enhancement
of the contrast resolution of the vascular
phenomena present in an organ or in the
lesions it contains. These methods, however,
show low spatial and details resolution,
useful for detecting peri- and intra-lesional
macro- and micro-circulation due to the
low intensity of signal coming from micro-
bubble resonance.
The observation that there is also a non-
linear response coming from the contrast
agent in the proximity of the fundamental
frequency has enabled a new algorithm
to be developed, which permits the non-
linear fundamental echo to be detected.
The advantage of detecting this nonlinear
fundamental component is an increase in
163
signal intensity that is greater than that of
the second harmonics alone of the contrast
agent. The identification of this nonlinear
response requires the use of an algorithm
capable of modulating the pulses so that
fundamental responses are elicited from
the tissue and from the contrast agent.
Such an algorithm requires perfect pulse
modulation both in phase and in ampli-
tude. It has been marketed with the name
cadence Contrast Pulse Sequencing (CPS
by Acuson Sequoia, Siemens) and it works
in real-time with continuous observation,
using second-generation contrast agents.
Previous algorithms exploit only the non-
linear harmonic response of contrast agent.
These require cancellation of the funda-
mental echoes coming from the tissue
forcing examination to be done blindly
(without an image) until the contrast agent
arrives. Contrast Pulse Sequencing software
has the immediate advantage of enabling
observation of the fundamental signals
coming from the tissue at the same time
as signals coming from the nonlinear
fundamental response of the contrast agent
(Phillips and Gardner, 2004).
NONLINEAR CONTRAST
ENHACEMENT OF FOCAL
LIVER LESIONS
(A) Characterization of Focal Liver
Lesions
Due to the dual blood supply of the liver
tissue by the hepatic artery (25–30%) and
the portal vein (70–75%), three vascular
phases can be defined in all the contrast
imaging modalities and even in CEUS.
The arterial phase starts from 10–20 s after
contrast administration into a peripheral
164
vein and persists for ∼10–15 s; followed by
the portal vein phase which usually lasts
2 min after i.v. contrast injection. The last
phase (parenchymal or sinusoidal) persists
until clearance of the contrast medium
from the hepatic parenchyma. It has two
different lengths: 4–6 min for SonoVue
and 15–20 min for Levovist, which sug-
gests several mechanisms including sinu-
soid pooling and RES/Kupffer’s cells
uptake. This differs from the extracellular
equilibrium phase observed in contrast
agent CT and MR imaging (EFSUMB
Study Group. Guidelines for the Use of
Contrast Agents in Ultrasound 2004). In
our opinion the evaluation criteria of the
US and CEUS examinations should be
specified; meaning that the terms Hypo-,
Iso- and Hyper-echogenic should be used
in the description of the conventional US
examinations; basket patterns, peripheral
or central, regular or irregular vessels for
color Doppler examinations.
Terms to be used for CEUS should be:
1. Type of enhancement: Rim, Marginal,
Nodular; Centripetal, Centrifugal filling
2. Degree of enhancement or enhancing:
Hypo-iso- and Hyper-enhancement or
enhancing – Lower, Equal or Higher
than surrounding parenchymal
3. Time of washout of contrast in the differ-
ent phases – portal and parenchymal
4. Lack of enhancement in the different
phases – arterial portal and parenchymal
5. Sustained enhancement in the different
phases – arterial portal and parenchymal
Hypo-, Iso- and Hyper-echoic terms might
be useful to describe static phenomena
underlying the conventional US imaging
but are ambiguous for explaining dynamic
phenomena underlying the CEUS imaging.
It might be appropriate to uniform terms
V. Migaleddu and G. Virgilio
related to the perfusion behavior of the
lesions which shows the evolution of lesion
vascular volume over time.
Enhancement Patterns of Benign Focal
Lesions
Hemangioma
Hemangioma is the most frequent solid
focal lesion in the liver from 1% to
20% in the general population (Ros
et al., 2001). These can be divided into
two different types: capillary and cavern-
ous hemangioma. The first are relatively
small networks of vascular spaces
(< 3 cm). Some of them can show an
early hypervascular pattern in the arterial
phase due to concentrated arterial micro-
vessels – high-flow hemangioma (Jang
et al., 1998). These are generally charac-
terized by the presence of small arterial
branches, venous lakes and fibrous areas,
and absence of capsule. The latter, cav-
ernous hemangiomas (> 3 cm) and giant
hemangiomas (> 10 cm) may show an
inhomogeneous structure composed of
venous lakes, dense connective tissue and
thrombosed areas (Unal et al., 2002).
Arterial portal micro-fistulas can be
identified with CEUS (http://www.smirg.
it/research/avi-shunt.php?sl=13). On con-
ventional US, hemangiomas are often inci-
dental findings since they are asymptomatic
except for the relatively rare giant type.
Conventional US capillary hemangiomas
have a typical homogeneous hyper-echoic
appearance in comparison with surround-
ing hepatic parenchyma with well defined
margins; in a fatty liver it may be found as
a hypo-echoic lesion. Cavernous heman-
giomas can present an inhomogeneous
aspect with hyper-echoic and hypo-echoic
areas (Moody and Wilson, 1993). In all
12. Focal Liver Lesion: Nonlinear Contrast-Enhanced Ultrasound Imaging 165

cases conventional US findings cannot be it is smaller than 5 cm (Ros et al., 2001).

considered specific particularly in onco- On conventional US iso-, hypo-, hyper-
logic and cirrhotic patients. Intralesional echoic appearances may be found. A
blood flow Doppler signals may be identi- hypo-echoic stellate central scar can be
fied in high-flow capillary hemangioma identified in 18% of cases. Color Doppler
(Tanaka et al., 1990). can depict flow signals within the central
Contrast-enhanced ultrasound can show and radiating vessels (Tanaka et al., 1990).
several appearances (Bleuzen and Tranquart, Contrast-enhanced ultrasound may show
2004; Catalano et al., 2005). In fact cont- typical and atypical features as described
rast material uptake may be fast or slow below (Migaleddu et al., 2004; Catalano
depending on intralesional circulation speed. et al., 2005).
Typical and atypical features are summa-
Typical features:
rized as follows:
Arterial phase: hyper-enhancing complete
Typical features: Portal vein phase: hyper-enhancing
Arterial phase: peripheral-nodular Parenchymal phase: iso- hyper-enhancing
enhancement, no central enhancement, rim Additional features:
enhancement Arterial phase: early “central spider” or
Portal vein phase: partial/complete cen- “spoke wheel” arteries and centrifugal fill-
tripetal filling ing, feeding artery
Parenchymal phase: complete enhance- Portal vein phase: hypo-enhanced cen-
ment tral scar
Atypical features: Parenchymal phase: hypo-enhanced cen-
Arterial phase: (small lesions) complete tral scar.
rapid centripetal enhancement; arterial–
portal fistula and arterial inflow vessel. Fatty Liver
Parenchymal phase: non enhanced cen- Hepatic steatosis is due to decreased liver
tral areas (partial thrombosis, fibrosis). clearance of fatty acids. It is caused by
hepatocellular injury and increased pro-
Focal Nodular Hyperplasia duction or mobilization of fatty acids.
Focal nodular hyperplasia is the second Alcohol, diabetes, obesity, and drugs are
most common benign solid hepatic focal the most common etiologies. Fatty changes
lesion. It occurs more commonly in young may present several morphologic and his-
women than men (two–four females to tological appearances according to diffuse
one male) in the third and fifth decade or focal involvement. A pseudo-nodular
(Volk et al., 2001), and is a hypervascular pattern might lead to a questionable diag-
hyperplasic lesion caused by preexisting nosis particularly in oncologic patients.
vascular malformation. It is character- Also spared areas can show ambiguous US
ized by abnormally arranged hepatocytes, findings (Kroncke et al., 2000).
Kupffer’s cells, bile duct elements and Focal fatty sparing:
vascularized fibrous septae. A peripheral In diffuse fatty liver focal spared areas are
pseudo capsule and a central or eccentric frequently encountered; these are typi-
fibrous scar radiating to the periphery and cally located along the liver hilum or
containing arteries can be found. Usually around the gallbladder. On conventional
166
US they appear as hypo-echoic areas with
well-defined margins. Color Doppler
does not reveal irregular vessels due to
mass effect within the lesions. Contrast-
enhanced ultrasound reveals typical vas-
cular behavior similar to the surrounding
parenchyma.
Typical features:
Arterial phase: iso-enhancing
Portal vein phase: iso-enhancing
Parenchymal phase: iso-enhancing
Focal fatty change:
Focal areas of fatty infiltration can be
found mimicking true focal lesions; on
conventional US they appear as typically
round-shaped hyper-echoic lesion. Color
Doppler and CEUS findings are similar to
those of fatty spared areas.
Typical features:
Arterial phase: iso-enhancing
Portal vein phase: iso-enhancing
Parenchymal phase: iso-enhancing
Regenerative Nodule
Regenerative nodules are found in liver
cirrhosis surrounded by fibrous septae.
The nodular dimension characterizes the
micro-macro or mixed type of cirrhosis
evolution. Large regenerative nodules from
5 to 20 mm can be detected on conven-
tional US as hypo-echoic nodular lesions;
hyper-echoic patterns due to fatty change
are rarer. Color Doppler or CEUS real-
time vascular evaluation show no apparent
vessels in lesions (Migaleddu et al., 2004;
Nicolau et al., 2006).
Typical features:
Arterial phase: iso-enhancing
Portal vein phase: iso-enhancing
Parenchymal phase: iso-enhancing
Additional features:
Arterial phase: hypo- or hyper-enhancing
V. Migaleddu and G. Virgilio
Cyst
Simple bile cysts are most common liq-
uid benign lesions (1–18%), and are
spread throughout the world (Mirilas and
Skandalakis, 2002). Hydatid cysts are
notuncommon in countries where sheep-
breeding is practiced. Conventional US
findings (typical anechoic lesions with
posterior enhancement) are specific for
diagnosis. A nonvital parasitic cyst with
solid aspect may be mistaken for tumor;
on color Doppler flow signals are absent.
Contrast-enhanced ultrasound findings are
as follows:
Typical features:
Arterial phase: non-enhancing
Portal vein phase: non-enhancing
Parenchymal phase: non-enhancing
Additional features:
Parenchymal phase: increased ultra-
sound transmission
Adenoma
Adenoma is an uncommon primary benign
tumor that has hepatocellular origin. Young
females who use hormonal medication
are the most frequently affected patients
(Karhunen, 1986). An adenoma may be
asymptomatic or may present abdominal
pain. This tumor has a marked tendency
to intra-lesional or intra-abdominal high
risk bleeding, although very rare potential
malignant transformation could be consid-
ered. Hepatic adenoma presents (80% of
cases) as a solitary well-defined or encap-
sulated tumor. This is a hypervascular
lesion composed of cords of proliferating
hepatocytes without portal venous tracts,
terminal hepatic veins, and bile ductules,
but with typical large sub-capsular tribu-
tary arteries originating from the hepatic
arterial system. In larger masses necrotic
12. Focal Liver Lesion: Nonlinear Contrast-Enhanced Ultrasound Imaging 167

and hemorrhagic changes are frequently Typical features:

observed (Ros et al., 2001, Bartolozzi Arterial phase: rim enhancement, no
et al., 2001). On conventional US, depend- enhancement
ing on lipid contents in the hepatocytes, Portal vein phase: hyper-/iso-enhancing
HA appears as iso or Hyper-echoic a rea- rim, no central enhancement
sonably large, well-circumscribed focal Parenchymal phase: hypo-enhancing
lesion. When intralesional hemorrhages or rim, no central enhancement
necrosis occur, an inhomogeneous pattern Additional features:
is prevalent (Grazioli et al., 2000). Color Arterial phase: enhanced septa, hyper-
Doppler demonstrates large peripheral sub- enhanced liver segment
capsular arteries and veins (Golli et al., Portal vein phase: hypo-enhancing rim,
1994). Contrast-enhanced ultrasound can enhanced septae
be summarized as follows (Bleuzen and Tuberculosis is an uncommon liver infec-
Tranquart, 2004; Catalano et al., 2005): tion which occurs in the miliary form. On
Typical features: conventional ultrasound single or multiple
Arterial phase: hyper-enhancing, com- hypo-echoic, well-defined nodules spread
plete throughout the liver are generally observed
Portal vein phase: hypo-enhancing (Brauner et al., 1989; Levine 1990). Contrast-
Parenchymal phase: iso-enhancing enhanced ultrasound reveals hypo-enhancing
Atypical features: well-defined nodules, especially in the later
Arterial phase: non-enhancing areas phase. Clinical findings are mandatory for
(hemorrhage) the diagnosis (Akahn and Pringot, 2002)
Portal vein phase: hyper-enhancing, non- and differentiation from hepatic metastases
enhancing areas (hemorrhage) is necessary.
Parenchymal phase: non-enhancing
areas (hemorrhage) Inflammatory Pseudo Tumors
The inflammatory pseudo tumor of the
Abscess liver is very uncommon, and is character-
Abscesses of different etiology, pyogenic ized by a well-defined mass consisting
or amoebic, may present different features of chronic inflammatory cell infiltration
due to the varying degree of liquefac- and fibrosis (Ishak et al., 1994). Prevalent
tion and location. Furthermore, vascular- necrotic nodules are also reported (Iwase
ity depends on evolution stage; usually it et al., 2002; Abbey-Toby et al., 2003).
is located at the periphery of the lesion, The etiology of this disease is unclear:
whereas liquefied areas are positioned in micro-organisms from food, cholangitis
the central zone. Conventional US can or hepatic involvement from chronic intra-
demonstrate an inhomogeneous echo- abdominal infection with ascending infec-
structure with peripheral hyper-echoic tion via the portal blood stream could
aspects and central hypo-echoic features. explain the inflammatory hepatic lesion
Color Doppler may detect vascular signals due to oblitering phlebitis (Horiuchi et al.,
at periphery or in the septa of the lesion. 1990). On conventional ultrasound large,
Contrast-enhanced ultrasound shows as inhomogeneous, hypo-echoic masses
described below (Catalano et al., 2005): with multiple thin septae are identified
168
(Nam et al., 1996). In our experience small
encapsulated nodules may be observed.
Color Doppler presents some flow sig-
nals in the periphery of the lesion in large
masses and no flow signal in the small
nodules. Contrast-enhanced ultrasound
shows marginal enhancement in the arte-
rial phase, while hypo enhancement areas
are demonstrated in the portal and paren-
chymal phases. Small nodules present
hypo or lack of enhancement especially in
the late phase. This ambiguous behavior
may mimic metastases.
Enhancement Patterns of Malignant Focal
Lesions
Dysplastic Nodule, Early and Advanced
Hepatocellular Carcinoma
Hepatocellular carcinoma (HCC) is the
most common primary hepatic malignancy.
Virus infections (B and particularly C),
genetic hemochromatosis, primary bil-
iary cirrhosis, environmental and dietary
factors such as persistent organic pol-
lutant substances, alcohol and aflotoxins
are the most important causes. Regional
distribution worldwide is related to the
different etiologies. Chronic liver diseases
correlated with C virus infection and alco-
hol abuse are becoming the most diffuse
causes of HCC. Dysplastic nodules occupy
an intermediate position between regenera-
tive nodules and well-differentiated HCC.
Dysplastic nodules are characterized by
cellular atypia without the development
of small arteries with a size ranging from
8–10 to 15 mm. Their pre-malignant nature
has been well demonstrated. Hepatic car-
cinogenesis is described as a multi-step
process where the progressive arterializa-
tion and gradual loss of portal vessels are
the principal features (Ueda et al., 1992;
Matsui et al., 1991). It is evident that the
V. Migaleddu and G. Virgilio
vascular enhancement behavior is related
to the evolution of the lesion. Thus, dur-
ing the arterial phase dysplastic nodules
or early HCC appear usually hypo- iso-
vascularized, while advanced HCCs are
hyper-vascularized. On conventional US,
findings are non-specific; usually small
HCCs (< 3 cm) are hypo-echoic and homo-
geneous but may appear hyper-echoic due
to fatty content (Kim et al., 2001; Ward and
Robinson, 2002). Color Doppler or power
Doppler may detect characteristic hyper-
vascularity; the “basket” pattern of vessels
surrounding and penetrating or “vessel-
within the tumor” pattern may be visual-
ized in HCC as has been reported (Tanaka
et al., 1990). Contrast-enhanced ultrasound
may show any selective enhancement in
the arterial phase differentiating HCCs
from regenerative nodules and dysplastic
nodules (Migaleddu et al., 2004; Nicolau
et al., 2006) (Figure 12.1a–d). The most
common features are described below:
Typical features:
Arterial phase: hyper-enhancing, com-
plete, non-enhancing areas (necrosis)
Portal vein phase: iso-, hypo-enhancing,
non-enhancing areas (necrosis)
Parenchymal phase: hypo-enhancing or
wash-out
Additional features:
Arterial phase: “chaotic” vessels, enhanc-
ing tumor thrombus in Portal Vein + HCC/
portal vein: hypo-enhancing with intral-
esional vessels
Portal vein phase: hypo-iso-enhancing
Parenchymal phase: hypo-iso-enhancing
Liver Metastases
Metastatic liver disease is the most fre-
quent malignant hepatic affliction, the liver
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)

malignant involvement. Depending on the surrounding parenchyma but with a loss of

original organ, the liver metastases may
be hyper-vascular or hypo-vascular, small
or large; the latter can present roughly
central necrotic areas. All metastases are
characterized by some degree of arte-
rial neoangiogenesis (macro-vessels in
peripheral margins and micro-vessels in
the central zone) and lack or reduced
portal supply. The total hematic volume
of the lesion is inferior to the surrounding
healthy parenchyma. Some metastases like
those from neuro-endocrine, thyroid, lym-
phoma, plasmacytoma, and kidney tumors
may have a superior arterial supply to the
portal vessels so at least the blood volume
is inferior. On conventional US, metastases
may present hypo-, iso- or hyper-echoic
patterns matching the primary tumor or
necrosis or calcification. Usually small
lesions appear as homogeneous hypo- or
iso-echoic solid nodules, while the larger
lesions frequently show an inhomogene-
ous aspect with hypo-echoic areas due
to necrosis and iso-, hyper-echoic areas
related to fibrosis and calcification. The
margins may be reasonably well-defined,
and sometimes a halo sign or target pat-
tern can be found. Cystic appearances may
170 V. Migaleddu and G. Virgilio

be related to ovarian and neuro-endocrine Hypo-vascular Metastases

primary tumors (Cosgrove, 2001; Harvey
and Albrecht, 2001). Color Doppler may
detect flow signals in the outer zone of
the lesions, whereas this finding in the
inner zone is relatively rare (Numata et al.,
1997). Depending on whether the lesion
is hyper- or hypo-vascularized CEUS
may present different aspects (Albrecht et
al., 2004; Catalano et al., 2005) (Figure
12.2a–d).
Typical features:
Arterial phase: rim or marginal enhancing
Portal vein phase: hypo-enhancing
Parenchymal phase: hypo-, lack of enhan-
cement
Additional features:
Arterial phase: complete enhancing,
non-enhancing areas (necrosis)
Portal vein phase: non-enhancing areas
(necrosis)

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

Hyper-vascular Metastases who were exposed to Thorotast vinyl

Typical features: chloride or other toxic agents (Bartolozzi
Arterial phase: hyper-enhancing, complete et al., 2001; Ros et al., 2001). Other
Portal vein phase: hypo-enhancing less common sarcomas such as leiomy-
Parenchymal phase: hypo-, lack of enhan- oma, fibrous, histiocytoma can be found.
cement Sarcomas usually appear as large size
Additional features: uncapsulated lesions at imaging detec-
Arterial phase: “chaotic” vessels tion, but very small multifocal nodules
may also be seen (Ros et al., 2001). On
Cholangiocarcinoma conventional US mixed echogenicity pre-
vails due to hemorrhagic or necrotic areas.
Cholangiocarcinoma is a malignant tumor
Color Doppler might capture minimal int-
originating from the bile duct epithelium.
ralesional flow signals. Contrast-enhanced
Peripheral cholangiocarcinoma arises from
ultrasound presents hypo-enhancement in
the small intra-hepatic bile duct. It con-
all the vascular phases.
sists of multiple nodules or a large lobu-
Lymphoma as a primary hepatic involve-
lated mass without capsule. It usually has
ment is very rare; however, it is possible to
hypo-vascularization regarding the arterial
find liver involvement during the outcome
vessels and the portal supply also is lim-
of Hodgkin’s or non Hodgkin’s systemic
ited. Arterial hyper-vascularization is less
lymphoma. Immune deficiency patients
common (Ros et al., 2001). On conven-
are mainly concerned. On conventional
tional US the small lesions are hypo- or
US hypo-echoic nodules or masses may
iso-echoic, whereas the larger lesions are
be found. The diffuse hepatic involvement
mixed-hyper-hypo-echoic. Color Doppler
appears as a diffused alteration of the
reveals the absence of intralesional blood
normal liver echo-texture. Lymphomatous
flow. During color or power Doppler evalu-
nodules present on CEUS iso- or hyper-
ation it is easy to detect bile duct dilatation
enhancement in the arterial phase but
and portal venous encasement (Harvey
hypo- or lack of enhancement in the fol-
and Albrecht, 2001). Contrast-enhanced
lowing phases (Catalano et al., 2005).
ultrasound reveals features close to those
of metastasis.
Typical features: (B) Detection of Focal Liver Lesions
Arterial phase: marginal enhancing Conventional US is the most widely dif-
Portal vein phase: hypo-, lack of enhan-fused cross-sectional imaging procedure for
cement the first approach in diagnosis of abdominal
Parenchymal phase: hypo-, lack of enhan-organs and liver diseases due to its rela-
cement tive low cost, its availability, safety, and
high patient acceptance. In the detection
Rare Malignant Neoplasm and staging of a liver lesion conventional
Angiosarcoma is the most frequent malig- US is less accurate than contrast-enhanced
nant mesenchymal liver malignancy, but spiral CT and MR because of the difficulty
proves to be a very rare tumor prevalently in detecting small size and/or iso-echoic
diagnosed in 60–80 year old male patients lesions, deep lesions especially in difficult
172
anatomical areas (sub-capsular or sub-
diaphragmatic areas). The recent advances
in nonlinear imaging and the introduction
of new contrast agents permit a large
number of limitations to be overcome, but
the role of CEUS is under debate.
Clinical investigations evaluating CEUS
have demonstrated that the accuracy in
detecting liver metastases is improved
and may be comparable to spiral CT
Additionally, others have reported that
CEUS can detect lesions not visible on
CT (Harvey et al., 2000; Albrecht et al.,
2003a). Data on diagnostic performance
have also been published using the much
simpler real-time method with low MI and
SonoVue (Albrecht et al., 2003b). Even if
further studies are required, current data
from single center and multicenter clini-
cal trials are becoming more robust. It is
not expected that US will replace CT or
MR; nevertheless, US remains useful as a
complementary procedure in some clinical
situations.
Incidental Detection of Hepatic Focal Lesion
The widespread use of a new generation of
imaging modalities has permitted a higher
frequency in detecting incidental focal
liver lesions in patients without symp-
toms or evidence of liver disease (Ros
and Davis, 1998). Benign hepatic lesions
are very frequent. In an autoptic series of
95 patients, benign lesions were detected
in 52% of cases; the most common were
bile duct tumor in 27% and hemangiomas
in 20% (Karhunen, 1986). In ∼50% of
the cases multiple lesions were found. It
has also been reported that small hepatic
lesions are frequently detected on routine
imaging evaluations. In a non-spiral con-
trast enhanced CT review of 2,978 cancer
patients, 378 (12.7%) small hepatic lesions
V. Migaleddu and G. Virgilio
measuring 10 mm or less were found,
among which 59 (15.6%) were considered
metastases; while 303 (80.2%) were clas-
sified as benign lesions, 16 (4.2%) were
not classified (Schwartz et al., 1999).
Current US, CT and MR technological
advances improve detection of incidental
liver focal lesions. Differentiation between
the benign or malignant lesions is diffi-
cult in small lesions that do not always
show enough characteristic features on
static cross-sectional imaging as CT or
MR; besides, depending on dimension
and position, it may also be difficult to
perform a biopsy. In these cases CEUS,
being a dynamic real-time imaging modal-
ity with good detail resolution, may be
used to play a crucial role (Bleuzen and
Tranquart, 2004).
Staging and Follow-Up of Oncologic Patients
For optimal management of cancer patients
with suspected metastases we need accu-
rate tumor staging. Molecular imaging
with PET and cross-sectional imaging
with CT and MR prove that they have an
important role in total body evaluation.
However, the liver is the most common
target organ for metastases of many pri-
mary tumors, especially for those arising
from abdominal organs and structures.
Conventional US being readily available
and cheap remains in many European
centers the first-line imaging modality in
detecting hepatic metastases. Ultrasound
sensitivity has been reported to be as high
as 85% (Clarke et al., 1989); especially
for rectal-colon tumor, intra-operative US
is the most sensitive imaging modality
compared to current multi-phasic contrast
CT and RM. Thus, it is usually performed
before liver resection to identify synchro-
nous metastases (Conlon et al., 2003).
12. Focal Liver Lesion: Nonlinear Contrast-Enhanced Ultrasound Imaging
Data related to the use of intra-operative
CEUS are still scarce (Siösteen and
Elvin, 2004).
In rectal-colon cancer, patients who
underwent liver resection range between
30–40% in terms of 5 years survival at
best because other metastases at the time
of resection were undetectable. To detect
metachronous metastases, a 3-monthly
conventional US and yearly CT follow-up
program may identify 88% of patients who
will develop liver metastases in an asympto-
matic stage in order to plan liver resection,
chemotherapy or percutaneous treatments
(Howell et al., 1999). Additionally, it has
been reported that patients with diffuse
metastatic diseases receiving systemic
chemotherapy at an asymptomatic stage
have higher response rates, better quality
of life, and survival, than those who begin
the chemotherapy at a symptomatic stage
(Nordic Gastrointestinal Tumor Adjuvant
Therapy Group, 1992). In these follow-up
programs CEUS results as the more cost-
effectiveness imaging procedure in the
detection of liver metastases.
Evaluation in the late parenchymal phase
with high M.I., using a first-generation
contrast agent (Levovist) has been recom-
mended in some reports (Albrecht et al.,
2003b). A long liver-specific late phase
of up to 20 min has been described due
to adherence of the microbubbles to the
hepatic sinusoid or to phagocytosis by
Kupffer’s cells (Albrecht et al., 2001).
The late phase is the most useful time
to detect metastases that usually appear
as hypo-enhanced or lack enhancement
areas. In comparison, most benign lesions
show uptake at this time and are therefore
not likely to be confused with metastases.
Second-generation gas filled microbub-
bles as in SonoVue are considered a truly
173
blood-pool contrast agent without any
interstitial equilibrium phase. The appear-
ance of metastases in the arterial phase is
variable. Hypo-vascular metastases show
up in CEUS as hypo-enhanced lesions
with and without an additional marginal
enhancement, while hyper-vascular metas-
tases appear as brightly hyper-enhanced
more or less homogeneous lesions. Hyper-
vascular metastases occur most often from
primary tumors of neuro-endocrine origin
or from renal or breast cancer.
A common pitfall is that small cysts are
sometimes detected on late phase scanning.
These can usually be distinguished from
metastases as they characteristically show
a higher US beam transmission (Bleuzen
and Tranquart, 2004). Differential diag-
nosis may be difficult for metastases with
dominant cystic component, as in ovarian
metastases. Contrast-enhanced ultrasound
could help to observe enhanced septae
or nodules that are very frequent in such
metastases. Thrombosed hemangiomas
may mimic the hypo enhancing behav-
ior of the metastases. Contrast-enhanced
ultrasound follow-up could identify the
modification of lesion vascularity due to
intralesional thrombus changes over time.
Surveillance for Hepatocellular Carcinoma
Surveillance means repeated application
of screening tests. Conventional US and
serum alpha-fetoprotein are the most
widespread screening tests for hepatocel-
lular carcinoma particularly in European
countries; a biannual interval between US
examinations is generally accepted. A ran-
domized study demonstrated the benefit of
such a program considering that in the sur-
veillance arm the HCC-related mortality
was reduced by 37% (Zhang et al., 2004).
During the Consensus Conference of
174
European Association for the study of the
liver (held in Barcelona in the year 2000),
a biannual US and AFL level evaluations
were recommended as the minimal follow up.
A spiral CT investigation is suggested for
patients who do not have nodules on US
but an increasing AFP level. For patients
who have a US finding of a nodule of
10 mm or less in diameter, 3 monthly US
is suggested taking into account that these
lesions are too small to be accurately char-
acterized and at least 50% of these lesions
will not be HCC. In patients with a nod-
ule of > 20 mm in diameter HCC can be
confirmed if the AFP level is over 400 ng/
ml and there is contrast-enhanced CT,
MR, or angiographic proof of the hyper-
vascularity of the lesion. At least two pieces
of positive imaging evidence are required
for HCC diagnosis (Bruix et al., 2001).
Biopsy is a controversial option espe-
cially for small nodules as it has a poten-
tial risk of bleeding and tumor seeding in
the needle track, whereas it is not always
useful due to the heterogeneity of vari-
ous types of hepatic nodular lesions (Brú
et al., 1989). Microbubble contrast as echo-
enhancer permits useful information to be
obtained for diagnosis if prevalent hyper-
vascularity of the lesion is demonstrated
(Lencioni et al., 1996; Fracanzani et al.,
2001). These imaging modalities are asso-
ciated with motion and blooming artifacts.
With first and second-generation con-
trast media, CEUS reveals a well defined
hyper-enhancement in arterial phase and
prevalent washout in the late phase. For
real-time evaluation of all three phases, the
use of second-generation contrast medium
and low M.I., is more suitable (Nicolau
et al., 2006).
In a series of 104 patients, using a second-
generation of contrast media SonoVue and
V. Migaleddu and G. Virgilio
low M.I., 100 primary tumors (96.2%)
demonstrated hyper-enhancement in the
arterial phase but 4 did not (3.8%) (Nicolau
et al., 2004). One study did not find arterial
enhancement in three well differentiated
HCCs out of 60 cases consequently high-
lighted the relationship between absence
of enhancement in the arterial phase and
histological differentiation (Tanaka et al.,
2001). After diagnosis of HCC, the pres-
ence and degree of cirrhosis, as well as
the tumor extension, need to be evalu-
ated for staging and management. In fact,
the presence of satellite nodules, biliary
extension, vascular, and extra-hepatic
involvement modify therapeutic planning.
Entire hepatic evaluation during the arte-
rial phase (20–30 s) might be difficult, also
because the sweeping scanning might fail
to detect some lesions. The administration
of several doses of contrast agents and
scanning of different hepatic segments
are suggested as well as the use of a high
M.I. at intervals to evaluate the destruction
and parenchyma replenishment in order to
detect hyper-enhancing foci. To date, there
are no data enabling comparison between
CEUS and spiral CT MRI in this chal-
lenging field. The extra-hepatic staging
favors total body cross-sectional imaging.
The evaluation of portal tumor invasion is
obviously easier with US examination.
(C) Monitoring of Percutaneous Ablative
Treatment
In the last 10 years numerous changes
have occurred in therapeutic approach
for primary and metastatic hepatic
malignancies. Many minimally invasive
therapies as the percutaneous ethanol
injection (PEI) and the thermal ablation
using different energy sources (radio
12. Focal Liver Lesion: Nonlinear Contrast-Enhanced Ultrasound Imaging 175

frequencies, laser or micro-wave) Livraghi et al., 2004). Comparative

are proposed (Livraghi et al., 1998;
Goldberg et al., 2000; Dodd et al., 2000;
De Baere et al., 2000; Lencioni et al.,
2003). Percutaneous ablation therapies
play a key role in the management of
patients with liver malignancies (HCC
and metastases). Numerous reports have
shown the efficacy and safety of these
ablating therapies especially in non surgi-
cal candidates. The possibility to repeat
treatment sessions in the case of recur-
rent or new lesions, the lower morbidity
and mortality as well as the low costs are
notable advantages (Solbiati et al., 2001;
data between surgical and percutaneous
treated patients show advantages for the
latter (Livraghi in http://www.smirg.org/
lectures.php).
As previously described, US, CEUS,
contrast-enhanced CT, and MRI play a
key role in the diagnosis of liver malignan-
cies in patients that are candidates for any
treatment plan. For the safety, low cost,
and absence of radiation and iodine con-
trast administration, CEUS play a crucial
role in the early and long-term follow-up
of percutaneous treated patients, usually
in the first week post treatment and after

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
1, 3, 6 months and then yearly. No reli-
able information about the outcome of
ablation treatments is provided by con-
ventional US, color Doppler, and power
Doppler. In fact, the assessment of vas-
cularization and tissue perfusion is essen-
tial to distinguish necrosis from residual
viable tumor. Triphasic helical CT or
dynamic gadolinium-enhanced MR can
predict the extent of the coagulation area
to within 2–3 mm.
The disappearance, if any was previously
visualized, of intralesional enhancement
on contrast-enhanced imaging is the most
important finding that suggests complete
ablation (Figure 12.3a–d). While persist-
ence of hyper-enhancing in the arterial
phase or clearly enhancing in the portal
phase suggest suspicious residual viable
tumor. Conatrast-enhanced ultrasound
can provide any kind of vascular related
information. Comparing pre- and post
treatment findings, the size modification
of the lesions permits a good evaluation
of the necrotic coagulation area and of
treatment results. The efficacy of treat-
ment in the hypo-enhanced lesions, as
for most liver metastases, can be identi-
fied when the lack of enhanced area is
increased determining in this manner the
sufficiency of the perilesional “safety”
margin achieved.
Contrast-enhanced ultrasound and con-
trast-enhanced CT, within the first 30
days, post-ablative evaluation may reveal
a thin and uniform enhancing rim of
hyper-vascularity along the periphery of
the necrotic area. Misinterpretation of this
perilesional hyperemic halo as residual
viable tumor can be avoided by comparing
post-ablation with pre-ablation findings
(Solbiati et al., 2004).
V. Migaleddu and G. Virgilio
CONCLUSION
Conventional US, being readily available,
economical and safe, remains the imaging
modality most widespread throughout the
world in detecting hepatic focal lesions.
In detection with CEUS, the number of
confirmed metastases increase from 3.06
to 5.42 in base-line comparison. After con-
trast medium administration the sensitivity
for detecting individual liver metastases
improves dramatically from 63% to 91%.
Furthermore, the capability to detect
very small metastases (> 10 mm) passes
with CEUS from 54% to 92% (Albrecht
et al., 2001). Comparing conventional US
and CEUS in a multicenter series of 123,
sensitivity in the detection of single metas-
tasis improved from 71% to 87%, while
the specificity increased from 60% to 88%
following contrast injection (Albrecht
et al., 2003).
Moreover, contrast-enhanced intra-
operative US (CEIOUS) also seems to
improve these results in a recent report.
In a series of 24 patients CEIOUS found
lesions missed at preoperative imaging
and at intra-operative US in four patients
and confirmed all of the new findings of
intra-operative US in four patients. In
addition CEIOUS helped to define the
tumor margins of the main lesion in 29%
of patients with rectal colon liver metas-
tases (Torzilli et al., 2005). This intra-
operative approach is suitable for the
synchronous metastases of rectal-colon
or other abdominal tumors. Contrast
sectional imaging (US, CT and RM)
conserve a central role for pre-treatment
diagnosis of other kinds of secondary
tumor. Surgeons’ and physicians’ interest
in CEUS is rapidly increasing.
12. Focal Liver Lesion: Nonlinear Contrast-Enhanced Ultrasound Imaging
Earlier we have described the guide-
lines for primary tumor management pro-
duced by the European Association for the
Study of Liver Disease (EASLD) (Bruix
et al., 2001). The American Association for
the Study of Liver Disease (AASLD) has
recently published “Practice Guidelines
for the management of HCCs”. The use
of only one contrast enhanced imaging
modality including CEUS to demonstrate
hyper-vascularization and diagnose a HCC
was suggested for a lesion larger than
20 mm (Bruix and Sherman, 2005).
The characterization of the focal liver
lesions by evaluation of hemodynamic
contrast enhancement behavior is con-
firmed by an important number of studies,
using both first or second-generation con-
trast media (Leen et al., 2002; Migaleddu
et al., 2004; Wen et al., 2004; Quaia et al.,
2004; Catalano et al., 2005).
Whereas some authors have emphasized
the importance of observing enhancement
at a late time, in our experience it is also
important to observe arterial and early
portal enhancement. Other experiences
seem to confirm this (Nicolau et al.,
2006). In particular, detection through
good details resolution of peri- and intra-
lesional macro- and micro-vessels con-
tributes to the diagnosis of focal liver
lesions, above all in the cirrhotic liver and
therefore in differential diagnosis between
Regenerative macro-nodules, Dysplastic
nodules in the transformation phase and
well-differentiated HCC; as well, early
rim enhancement can suggest diagnosis
of high flow hemangioma. The study of
the only late phase is more useful in the
diagnosis of metastases.
In order to obtain excellent details reso-
lution, especially in the arterial phase, it is
177
useful to read together fundamental echoes
from tissue as well as nonlinear fundamental
echoes from the contrast agent; the CPS
(the most recent technological approach)
permits this also with second-generation
contrast media in real-time. The possibility
of aiming at the lesion, and in particular
seeing both extra- and intra-lesional vascular
contribution in the arterial phase, consid-
erably assists early phase characteriza-
tion of HCC. Using second-generation
of contrast media, nonlinear fundamental
(CPS) and “conventional” harmonic imag-
ing findings proves to be equivalent in
the study of the late phase useful for
metastases diagnosis. This is true except
for the lesions positioned in the distal
field because there is poor intensity of the
contrast harmonic echoes in comparison
with contrast nonlinear fundamental ones
(Migaleddu et al., 2005). Excellent details
resolution especially in the arterial phase
is also valuable in outcome assessment of
ablative therapy.
REFERENCES
Abbey-Toby, A., Cazals-Hatem, D., Colombat, M.,
Belghiti, J., Vilgrain, V., and Degott, C. 2003.
Inflammatory pseudo-tumor of the liver: is pre-
operative diagnosis possible? Gastroenterol.
Clin. Biol. 27: 883–890.
Akahn, O., and Pringot, J. 2002. Imaging of abdom-
inal tubercolosis. Eur. Radiol. 12: 321–323.
Albrecht, T., Hoffmann, C.W., Schmitz, S.A.,
Schettler, S., Ovemberg, A., Germen, C.T., and
Wolf, K.J. 2001. Phase-inversione sonography
during the liver-specific late phase of contrast
enhancement: improved detection of liver metas-
tases. Am. J. Roentgenol. 176 (5): 1191–1198.
Albrecht, T., Blomley, M.J., Burns, P.N., Wilson, S.,
Harvey, C.J., Leen, E., Claudon, M., Calliada,
F., Correas, J.M., LaFortune, M., Campani,
R., Hoffmann, C.W., Cosgrove, D.O., and
LeFevre, F. 2003a. Improved detecion of hepatic
178
metastases with pulse-inversion US during the
liver-specific phase of SHU 505A: multicenter
study. Radiology 227: 361–370.
Albrecht, T., Oldenburg, A., Hohmann, J., Skrok, J.,
Hoffmann, C.W., Schettler, S., and Wolf, K.J.
2003b. Imaging of liver metastases with contrast-
specific low-MI real-time ultrasound and SonoVue.
Eur. Radiol. 13 (Suppl 3): 79–86.
Albrecht, T., Hohmann, J., Oldenburg, A., Skrok, J.,
and Wolf, K.J. 2004. Detection and characterisa-
tion of liver metastases. Eur. Radiol. 14 (Suppl 8):
25–33.
Apfel, R.E., and Holland, C.K. 1991. Gauging the
likelihood of cavitation from short-pulse, low
duty cycle diagnostic ultrasound. Ultrasound
Med. Biol. 17: 175–185.
Bartolozzi, C., Cioni, D., Donati, F., and Lencioni, R.
2001. Focal liver lesions: MR imaging-patho-
logic correlation. Eur. Radiol. 11: 1374–1388.
Bleuzen, A., and Tranquart, F. 2004. Incidental
liver lesions: diagnostic value of cadence con-
trast pulse sequencing (CPS) and SonoVue. Eur.
Radiol. 14 (Suppl 8): 53–62.
Brauner, M., Buffard, M.D., Jaentils, V., Legrand, I.,
and Gotheil, C. 1989. Sonography and computed
tomography of mocroscopic tuberculosis of the
liver. J. Clin. Ultrasound 17: 563–568.
Brú, C., Marotto, A., Bruix, J., Faus, R., Biachi, L.,
Calvet, X., Ayuso, C., Vilana, R., Gilalbert, R.,
and Rodes, J. 1989. Diagnostic accurancy of
fine-needle biopsy in patients with hepatocellular
carcinoma. Dig. Dis. Sci. 34: 1765–1769.
Bruix, J., and Sherman, M., 2005. Management
of Hepatocellular carninoma. J. Hepatol. 42:
1208–1236.
Bruix, J., Sherman, M., Llovet, J.M., Beaugrand, M.,
Lencioni, R., Burroughs, A.K., Christensen, E.,
Pagliaro, L., Colombo, M., and Rodes, J. 2001.
Clinical management of hepatocellular carci-
noma. Conclusions of the Barcelona. EASL
Conference. J. Hepatol. 35: 421–430.
Burns, P.N., Wilson, S.R., and Simpson, D.H.
2000. Pulse inversion imaging of liver blood
flow: an improved method for characterization
of focal masses with microbubble contrast.
Invest. Radiol. 35: 58–71.
Catalano, O., Nunziata, A., Lobianco, R., and Siani, A.
2005. Real-time harmonic contrast material-
specific US of focal liver lesions. RadioGraphics
25: 333–349.
V. Migaleddu and G. Virgilio
Chin C.T., and Burns P.N. 1997. Predicting the
acoustic response of a microbubble population
for contrast imaging. In: IEEE Ultrason. Symp.
22: 1557–1560.
Clarke, M.P., Kane, R.A., Steele, G. Jr., Hamilton,
E.S., Ravikumar, T.S., Onik, G., and Clouse,
M.E. 1989. Prospective comparision of pre-
operative imaging and intraoperative ultrasonog-
raphy in the detecion of liver tumours. Surgery
106: 849–855.
Conlon, R., Jacobs, M., Dasgupta, D., and Loge,
J.P. 2003. The value of intraoperative ultra-
sound during hepatic resection compared with
improved preoperative magnetic resonace imag-
ing. Eur. Ultrasound 16: 211–216.
Cosgrove, D.O. 2001. Malignant liver disease. In:
Meire, H.B., Cosgrove, D.O., Dewbury, K.C.,
and Farrant, P. (eds). Clinical Ultrasound: A
Comprehensive Text. Abdominal and General
Ultrasound, vol. 1, 2nd ed. Churchill Livingstone,
London, 211–231.
Cosgrove, D.O. 2005. Harmonic imaging: tissue
and microbubbles in http://www.smirg.org/lec-
tures.php.
De Baere, T., Elias, D., Dromain, C., Din, M.G.,
Kuoch, V., Ducreux, M., Boige, V., Lassau, N.,
Marteau, V., Lasser, P., and Roche, A. 2000.
Radiofrequency ablation of 100 hepatic metas-
tases with a mean follow-up of more than 1 year.
Am. J. Roentgenol. 175: 1619–1625.
Dodd, G.D. 3rd., Soulen, M.C., Kane, R.A.,
Livraghi, T., Lees, W.R., Yamashita, Y.,
Gillams, A.R., Karahan, O.I., and Rhim, H.
2000. Minimally invasive treatment of malig-
nant hepatic tumors: at the threshold of a major
breakthrough. Radiographics 20: 9–21.
EFSUMB Study Group Albrecht, T., Blomley, M.,
Bolondi, L., Claudon, M., Correas, J.M.,
Cosgrove, D., Greiner, L., Jager, K., Jong, N.D.,
Leen, E., Lencioni, R., Lindsell, D., Martegani,
A., Solbiati, L., Thorelius, L., Tranquart,
F., Weskott, H.P., and Whittingham, T.
2004. Guidelines for the Use of Contrast
Agents in Ultrasound. Ultraschall. Med. 25:
249–256.
Fracanzani, A.L., Burdick, L., Borzio, M., Roncalli, M.,
Bonelli, N., Borzio, F., Maraschi, A., Fiorelli,
G., and Fargion, S. 2001. Contrast-enhanced
Doppler ultrasonography in the diagnosis of
hepatocellular carcinoma and premalignant
12. Focal Liver Lesion: Nonlinear Contrast-Enhanced Ultrasound Imaging
lesions in patients with cirrhosis. Hepatology 34:
1109–1112.
Goldberg, S.N., Gazelle, C.S., Compton, C.C.,
Mueller, P.R., and Tanabe, K.K. 2000. Treatment
of intrahepatic malignancy with radiofrequency
ablation: radiologic-pathologic correlation.
Cancer 88: 2452–2463.
Golli, M., Van Nhieu, J.T., Mathieu, D., Zafrani,
E.S., Cherqui, D., Dhumeaux, D., Vasile, N., and
Rahmouni, A. 1994. Hepatocellular adenoma:
Color Doppler US and pathologic correlations.
Radiology 190: 741–744.
Grazioli, L., Federle, M.P., Ichikawa, T., Balzano, E.,
Nalesnik, M., and Madariaga, J. 2000. Liver
adenomatosis: clinical, histopathologic and
imaging findings in 15 patients. Radiology 216:
395–402.
Greis, C. 2004. Technology overview: SonoVue
(Bracco, Milan). Eur. Radiol. 14 (Suppl 8): 11–15.
Hamilton, M.F., and Blackstock, D.T. 1998.
Nonlinear Acoustics. Academic, San Diego, CA.
Harvey, C.J., and Albrecht, T. 2001. Ultrasound of
focal liver lesions. Eur. Radiol. 11: 1578–1593.
Harvey, C.J., Blomley, M.J., Eckersley, R.J.,
Cosgrove, D.O., Patel, N., Heckemann, R.A., and
Butler-Barnes, J. 2000. Hepatic malignancies:
improved detection with pulse-inversion US in
late phase of enhancement with SH U 508A-early
experience. Radiology 216: 903–908.
Horiuchi, R., Uchida, T., Kojima, T., and Shikata, T.
1990. Inflammatory pseudotumor of the liver.
Clinicalpathological study and review of litera-
ture. Cancer 65: 1583–1590.
Howell, J.D., Wotherspoon, H., Leen, E., Cooke,
T.C., and MacArlde, C.S. 1999. Evaluation of
a follow up programme after curative resec-
tion for colorectal cancer. Br. J. Cancer 79:
308–310.
Ishak, K.G., Anthony, P.P., and Sobin, L.H. 1994.
World Health Organization: International Histo-
logic Classification of Tumors, 2nd ed. Berlin,
Springer.
Iwase, K., Higaki, J., Yoon, H.E., Mikata, S.,
Miyazaki, M., Torikai, K., Shirai,Y., Awai, K.,
Imakita, M., and Kamiike, W. 2002. Solitary
necrotic nodule of the liver. J. Hepatobiliary
Pancreat. Surg. 9: 120–124.
Jang, H.J., Choi, B.I., Kim, T.K., Jun, A.J., Kim.,
K.W., Han, J.K., and Han, M.C. 1998. Atypical
small hemangiomas of the liver: “bright dot”
179
sign at two-phase spiral CT. Radiology 208:
543–548.
Karhunen, P.J. 1986. Benign hepatic tumours and
tumour-like conditions in men. Clin. Pathol. 39:
183–188.
Kim, C.K., Lim, J.H., and Lee, W.J. 2001. Detection
of hepatocellular carcinomas and dysplastic nod-
ules in cirrhotic liver: accuracy of ultrasonogra-
phy in transplant patients. J. Ultrasound Med.
20: 99–104.
Kroncke, T.J., Taupitz, M., Kivelitz, D., Scheer,
I., Daberkow, U., Rudolph, B., and Hamm, B.
2000. Multifocal nodular fatty infiltration of the
liver mimicking metastatic disease on CT: imag-
ing findings and diagnosis using MR imaging.
Eur. Radiol. 10: 1095–1100.
Leen, E., Angerson, W.J., Yarmenitis, S.,
Bongartz, G., Blomley, M., Del Maschio, A.,
Summaria, V., Maresca, G., Pezzoli, C., and
Llul, J.B. 2002. Multi-centre clinical study
evaluating the efficacy of SonoVue (BR1),
a new ultrasound contrast agent in Doppler
investigation of focal hepatic lesions. Eur. J.
Radiol. 42: 200–206.
Lencioni, R., Pinto, F., Armillota, N., and
Bartolozzi, C. 1996. Assessment of tumor vasco-
larity in hepatocarcinoma: comparison of power
Doppler and Color Doppler US. Radiology 201:
353–358.
Lencioni, R.A., Allgaier, H.P., Cioni, D.,
Olschewski, M., Deibert, P., Crocetti, L., Frings,
H., Laubenberger, J., Zuber, I., Blum, H.E.,
and Bartolozzi, C. 2003. Small hepatocellular
carcinoma in cirrhosis: randomized compari-
son of radiofrequency thermal ablation versus
percutanous ethanol injection. Radiology 228:
235–240.
Levine, C. 1990. Primary macronodular hepatic
tuberculosis: US and CT appearances. Gastro-
intest. Radiol. 15: 307–309.
Livraghi, T. 2006. Role of image-guided thera-
pies in hepatocellular carcinoma in http://www.
smirg.org/lectures.php.
Livraghi, T., Benedini, V., Lazzaron, S., Meloni, F.,
Torzilli, G., and Vettori, C. 1998. Long term
results of single session PEI in patients with large
hepatocellular carcinoma. Cancer 83: 48–57.
Livraghi, T., Meloni, F., Morabito, A., and Vettori,
C. 2004. Multimodal image-guided tailored
therapy of early and intermediate hepatocellular
180
carcinoma: long-term survival in the experi-
ence of a single radiologic referral center. Liver
Transpl. 10: 98–106.
Matsui, O., Kadoja, M., Kamejama, T., Takashima, T.,
Nakanuma, J., Unoura, M., Kobayashi, K.,
Izumi, R., and Ida, M. 1991. Benign and malig-
nant nodules in cirrhotic patients: distinction on
blood supply. Radiology 178: 493–497.
Migaleddu, V., Virgilio, G., Campisi, G., Sirigu, D.,
and Canalis, G.C. 2002. Conventional ultra-
sonography versus tissue harmonic imaging for
the assessment of the common bile duct in chole-
cystectomized patients. Radiol. Med. 103: 1–6.
Migaleddu, V., Virgilio, G., Turilli, D., Conti, M.,
Campisi, G., Canu, M., Sirigu, D., and
Vincentelli, I. 2004. Charaterization of focal
liver lesions in real-time using harmonic
imaging with high mechanical index and con-
trast agent levovist. Am. J. Roentgenol. 182:
1505–1512.
Migaleddu, V., Manca, A., Turilli, D., Luciano, P.,
Canu, N., Campisi, G., and Virgilio, G. 2005.
Fundamental nonlinear imaging versus harmonic
imaging with contrast medium sonovue in the
evaluation of the hepatic focal lesions: prelimi-
nary results. Eur. Radiol. 14 (Suppl. 1): B-126.
Mirilas, P., and Skandalakis, J.E. 2002. Benign
anatomical mistakes: incidentaloma. Am. Surg.
68: 1026–1028.
Moody, A.R., and Wilson, S.R. 1993. Atypical
hepatic hemangioma: a suggestive sonographic
morphology. Radiology 188: 413–417.
Morel, D.R., Schwieger, I., Hohn, L., Terrettaz. J.,
Llull, J.B., Cornioley, Y.A., and Schneider, M.
2000. Human pharmacokinectics and safety eval-
uation of SonoVueTM, a new contrast agent for
ultrasound imaging. Invest. Radiol. 35 (Suppl 1):
80–85.
Nam, K.J., Kang, H.K., and Lim, J.H. 1996.
Inflammatory pseudotumor of the liver: CT and
sonographic features. Am. J. Roentgenol. 167:
485–487.
Nanda, N.C., and Carstensen, E.L. 1997. Echo-
enhancing agents: safety. In: Nanda N.C.,
Schlief, R., and Goldberg, B.B. (eds). Advances
in Echo Imaging Using Contrast Enhancement,
2nd, ed.Kluwer, Dordrecht, 139–170.
Nicolau, C., Vilana, R., and Brù, C. 2004. The
use of contrast-enhanced ultrasound in the
V. Migaleddu and G. Virgilio
management of the cirrothic patient and for detec-
tion of HCC. Eur. Radiol. 14 (Suppl 8): 63–71.
Nicolau, C., Vilana, R., Catalá, V., Bianchi, L.,
Gilabert, R., Carcía, A., and Brú, C. 2006.
Importance of evaluating all vascular phases on
contrast enhanced sonography in the diferentia-
tion of benign from malignant focal lesions. Am.
J. Roentgenol. 186: 158–167.
Nordic Gastrointestinal Tumour Adjuvant Therapy
Group, 1992. Expectancy or primary chemo-
therapy in patients with advanced asymptomatic
colorectal cancer; a randomisede trial. J. Clin.
Oncol. 10: 904–911.
Numata, K., Tanaka, K., Kiba, T., Norimoto, M.,
Arata, S., Kondo, M., and Sekihara, H. 1997.
Use of hepatic tumor index on color Doppler
sonography for differentiating large hepatic
tumors. Am. J. Roentgenol. 168: 991–995.
Ortega, D., Burns, P.N., Simpson, D.H., and
Wilson, S.R. 2001. Tissue harmionic imaging:
is it a benefit for bile duct sonography? Am. J.
Roentgenol. 176: 653–659.
Phillips, P., and Gardner, E. 2004. Contrast-agent
detection and quantification. Eur. Radiol. 14
(Suppl 8): 4–10.
Powers, J.E., Burns, P.N., and Souquet, J. 1997.
Imaging instrumentation for ultrasound con-
trast agents. In: Nanda, N.C., Schlief, R., and
Goldberg, B.B. (eds). Advances in Echo Imaging
Using Contrast Enhancement, 2nd, ed. Kluwer,
Dordrecht, 139–170.
Quaia, E., Calliada, F., Bertolotto, M., Rossi, S.,
Garioni, L., Rosa, L., and Pozzi-Mucelli, R.
2004. Characterization of focal liver lesions with
contrast-specific US modes and a sulfur hexaflu-
oride-filled microbubble contrast agent: diag-
nostic performance and confidence. Radiology
232: 420–430.
Ros, P., and Davis, G. 1998. The incidental focal liver
lesion: photon, proton, or needle? Hepatology
98: 1183–1190.
Ros, P.R., Menu, Y., Vilgrain, V., Mortele, K.J., and
Terris, B. 2001. Liver neoplasms and tumor-like
conditions. Eur. Radiol. 11 (Suppl 2): 145–165.
Schneider, M., Arditi, M., Barrau, M.B., Brochot,
J., Broillet, A., Ventron, R., and Yan, F. 1995.
BR1: a new ultrasonographic contrast agent
based on sulfur hexafluoride-filled microbub-
bles. Invest. Radiol. 30: 451–457.
12. Focal Liver Lesion: Nonlinear Contrast-Enhanced Ultrasound Imaging
Schwartz, L.H., Gandras, E.J., Colangelo, S.M.,
Ercolani, M.C., and Panicek, D.M. 1999.
Prevalence and importance of small hepatic
lesions found at CT in patients with cancer.
Radiology 210: 71–74.
Siösteen, A.K., and Elvin, A. 2004. Intra-operative
uses of contrast-enhanced ultrasound. Eur.
Radiol. 14 (Suppl 8): 87–95.
Solbiati, L., Livraghi, T., Golberg, S.N., Ierace, T.,
Meloni, F., Dellanoce, M., Cova, L., Halpern,
E.F., and Gazelle, G.S. 2001. Percutaneous
radio-frequency ablation of liver metastase
from colon rectal cancer: long term risults 117
patients. Radiology 221: 159–166.
Solbiati, L., Tonolini, M., and Cova, L. 2004.
Monitoring RF ablation. Eur. Radiol. 14 (Suppl 8):
34–42.
Tanaka, S., Kitamura, T., Fuijta, M., Nakanishi, K.,
and Okuda, S. 1990. Color Doppler flow imag-
ing of liver tumors. Am. J. Roentgenol. 154:
509–514.
Tanaka, S., Ioka, T., Oshikawa, O., Hamada, Y., and
Yoshioka, F. 2001. Dynamic sonography of hepatic
tumors. Am. J. Roentgenol. 177: 799–805.
Torzilli, G., Del Fabro, D., Palmisano, A., Donadon, M.,
Bianchi, P., Roncalli, M., Balzarini, L., and
Montorsi, M. 2005. Contrast-enhanced intraop-
erative ultrasonography during hepatectomies
for Colon rectal metastaes. J. Gastrointest. Surg.
9: 1148–1154.
181
Ueda, K., Terada, T., Nakanuma, Y., and Matsui, O.
1992. Vascular supply in adenomatous hyperplasia
of the liver and hepatocellular carcinoma: a mor-
phometric study. Hum. Pathol. 23: 619–626.
Unal, O., Sakarya, M.E., Arslan, H., Tuncer, I.,
and Etlik, O. 2002. Hepatic cavernous hemangi-
omas: patterns of contrast enhancement on MR
fluoroscopy imaging. Clin. Imaging 26: 39–42.
Volk, M., Strotzer, M., Lenhart, M., Techert, J.,
Seitz, J., and Feuerbach, S. 2001. Frequency of
benign hepatic lesions incidentally detected with
contrast-enhanced section portal venous pahase
spiral CT. Acta Radiol. 42: 172–175.
Ward, B., Baker, A.C., and Humprey, V.F. 1997.
Non linear propagation applied to the improve-
ment of resolution in diagnostic medical ultra-
sound. J. Acoust. Soc. Am. 101: 143–154.
Ward, J., and Robinson, J. 2002. How to detect
hepatocellular carcinoma in chirrhosis. Eur.
Radiol. 12: 2258–2272.
Wen, Y.L., Kudo, M., Zheng, R.Q., Ding, H., Zhou, P.,
Minami, Y., Chung, H., Kitano, M., Kawasaki, T.,
and Maekawa, K. 2004. Characterizazione of
hepatic tumors: value of contrast-enhanced
coded phase-inversion harmonic angio. Am. J.
Roentgenol. 182: 1019–1026.
Zhang, B.H., Yang, B.H., and Tang, Z.Y. 2004.
Randomized controll trial of screening for hepa-
tocellular carcinoma. J. Cancer Res. Clin. Oncol.
130: 417–422.
13
Hepatocellular Carcinoma: Magnetic
Resonance Imaging
Bachir Taouli
INTRODUCTION
Hepatocellular carcinoma (HCC) is the
most common primary hepatic malig-
nancy, accounting for 90% of primary
liver neoplasms and is responsible for up
to 1 million deaths annually worldwide
(Fung and Marsh, 2002). The major risk
factor for HCC is cirrhosis. All types
of cirrhoses predispose to HCC, but the
incidence is particularly high in patients
with chronic hepatitis B (HBV), chronic
hepatitis C (HCV), and in alcoholic liver
disease. HCC develops by sequential steps
from low grade dysplastic nodules to
frank malignant HCC. The recognition of
HCC is critical because curative treatment
(including surgical resection and liver
transplantation) and prognosis depend on
early diagnosis. The 5-year survival rate
for untreated, symptomatic HCC is < 5%
(El-Serag and Mason, 1999), whereas the
5-year survival rate in patients with cirrhosis
who have a small (< 2 cm) HCC lesion and
undergo liver transplantation is 80%; there-
fore, the detection of small HCC is highly
critical to patient outcome (Mazzaferro
et al., 1996).
MAGNETIC RESONANCE
IMAGING TECHNIQUE
There is a considerable overlap in the
imaging appearances of the different types
of cirrhotic nodules. Because it provides
excellent contrast resolution and mul-
tiplanar evaluation, magnetic resonance
imaging (MRI) is the examination of
choice for detecting and characterizing
cirrhotic nodules. Compared to ultrasound
and computerized tomography (CT), MRI
is superior in detecting and characterizing
HCC in the cirrhotic liver (Yamashita
et al., 1996; Aschoff et al., 2001). The use
of intravenous contrast material with mul-
tiphasic imaging (arterial, portal venous,
and equilibrium) is essential to accurately
identify focal hepatic lesions. However,
MRI remains less accurate for lesions
< 2 cm in diameter. Recent advances in MRI
software and hardware, including the use of
parallel imaging and surface phased-array
coils, provide faster sequences that can be
acquired within a breath-hold, decreasing
motion and respiratory artifacts.
In our institution, we use breath-hold seq-
uences for routine liver MRI examination
183
184
(whole acquisition time < 30 min). The
protocol includes: T1-weighted in- and out-
of-phase gradient-recalled echo (GRE) to
evaluate for microscopic fat content, Turbo
Spin Echo T2-weighted with spectral fat
saturation or T2-fast STIR (short tau inver-
sion recovery), T2-weighted single-shot
RARE (rapid acquisition with relaxation
enhancement) and 3D T1-weighted fat
suppressed GRE before and after dynamic
injection of 15–20 ml of gadolinium
chelates using an automatic injector, with
acquisitions at the arterial, portal venous,
and equilibrium phases. We use the short-
est possible TR (repetition time) in order
to minimize acquisition time and allow
sequential acquisitions through the entire
liver during breath-holding. Fat suppres-
sion improves the conspicuity of contrast
enhancement, and eliminates chemical-
shift artifacts at fat-water interfaces.
MAGNETIC RESONANCE
IMAGING APPEARANCE
OF HEPATOCELLULAR
CARCINOMA
The macronodular architecture of the
cirrhotic liver and the morphology of
HCC are usually well demonstrated on T1-
and T2-weighted images. On T1-weighted
images, HCC can be hypo-, iso- or hyper-
intense, compared to the surrounding liver
(Aschoff et al., 2001; Honda et al., 1997).
Most well differentiated HCCs and high-
grade dysplastic nodules have high signal
intensity on T1-weighted images, and T1
hyperintensity can be related to copper,
Fe3+, glycogen deposition, or high pro-
tein or lipid content, and possibly to the
B. Taouli
degree of differentiation (Ebara et al.,
1999; Amano et al., 2003). In-phase and
opposed-phase images can demonstrate
microscopic fat components within HCC,
areas of fat accumulation showing a signal
drop on opposed-phase images.
On T2-weighted images, HCCs are
classically hyperintense, whereas regener-
ative nodules (Figure 13.1) and most dys-
plastic nodules are hypointense (Earls et al.,
1996; Matsui et al., 1989; Krinsky et al.,
2001) (Figures 13.2 and 13.4). However,
well-differentiated HCC lesions can be
iso- or rarely hypointense on T2-weighted
images (Earls et al., 1996; Hecht et al.,
2006). A focus of high signal intensity on
T2 within a hypo- or isointense nodule
(“nodule within a nodule” sign) is highly
suggestive of HCC developing within a
dysplastic nodule (Mitchell et al., 1991;
van den Bos et al., 2006). Ebara et al.
(1999) have shown that the signal inten-
sity of HCC on T1- and T2-weighted
images is related to the degree of his-
tologic differentiation and intratumoral
copper content, whereas the signal inten-
sity on T2-weighted images is related to
the degree of histologic differentiation.
However, a relationship between copper
content and signal intensity of HCC was
not confirmed by other studies (Nakakoshi
et al., 1996). Hemosiderin deposition is
common in regenerative nodules (called
siderotic nodules), producing specific imag-
ing features on MRI, such as hypointensity
on T1-GRE images and T2-weighted GRE
images (Krinsky et al., 2001). Dysplastic
nodules are usually homogeneously hyper-
intense on T1-weighted images, but small
HCC nodules may exhibit similar findings
(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
GADOLINIUM ENHANCEMENT
OF HEPATOCELLULAR
CARCINOMA
The overlapping signal intensities of regen-
erative nodules, dysplastic nodules, and
HCC on T1- and T2-weighted images
necessitate the use of contrast media, includ-
ing non-specific extracellular gadolinium
chelates or tissue-specific contrast media.
Gadolinium-DTPA is an extracellu-
lar contrast medium that produces T1
shortening and enhancement of HCC on
T1-weighted images. HCC obtains its
blood supply almost exclusively from the
hepatic artery, and most HCCs are best
seen on arterial phase images (Figures
13.1–13.3) (Kim et al., 1995; Murakami
et al., 1995; Oi et al., 1996), allow-
ing differentiation from regenerative and
dysplastic nodules, as well as from focal
fibrosis, which enhances during the portal
and equilibrium phases, with typical capsular
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
B. Taouli
retraction. In the portal venous phase,
HCC appears hyper-, iso- or hypointense
relative to the background liver (Figures
13.1, 13.2). A minority of well-differ-
entiated HCC lesions are hypovascular
(Matsui et al., 1991) and best seen at the
portal venous phase, because liver paren-
chymal enhancement peaks at this phase.
Delayed images can show late enhancement
of the fibrous capsule which is typically a
feature of HCC with a size > 2 cm, but not
of regenerative or dysplastic nodules. Both
necrosis and hemorrhage are common in
large tumors. Regenerative and dysplastic
nodules may be difficult to identify on
post-gadolinium images because these
are typically hypovascular lesions with
predominantly portal blood supply, with
an enhancement similar to that of the
adjacent parenchyma, although increased
arterial flow has been described in a small
minority of dysplastic nodules (Matsui
et al., 1991).
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

field inhomogeneities, causing a marked

MAGNETIC RESONANCE
IMAGING TISSUE-SPECIFIC
CONTRAST AGENTS
The possibility of improving the diagnosis
of HCC by the use of contrast agents spe-
cifically targeted to the liver cells has been
evaluated in multiple studies.
Tissue-specific MR contrast media
which produce T1 enhancement in hepa-
tocytes include mangafodipir (MnDPDP),
gadobenate (Gd-BOPTA), and gadoxetic
acid (Gd-EOB-DTPA). All of these agents
show uptake in well-differentiated HCC, as
well as in benign hepatocellular nodules,
so their value in discriminating between
benign and malignant lesions is limited
(Rofsky et al., 1993; Murakami et al.,
1996). Results with these agents have not
been shown to be superior to those with
gadolinium for the detection of HCC.
Superparamagnetic iron oxide (SPIO)
particles are taken up by Kupffer cells of the
reticuloendothelial system. They enhance
T2 and T2* relaxation by increasing local
decrease in the signal intensity of liver tis-
sue, particularly on T2-weighted images.
A breath-hold T2-weighted gradient-echo
sequence is generally used after SPIO
injection, because the susceptibility effects
are maximized and respiratory artifacts are
eliminated. Because of the decrease of the
signal intensity of the background liver,
SPIO particles increase the detection and
the characterization of focal liver lesions
(benign lesions will show uptake of SPIO).
Kupffer cells are rarely present in HCC,
and therefore the tumor will show little or
no uptake of SPIO, and will appear more
conspicuous after SPIO injection.
Combined Use of Superparamagnetic
Iron Oxide and Gadolinium
Using both gadolinium and ferumoxides
(an SPIO agent) MR agents, Bhartia
et al. (2003) demonstrated a 78% sen-
sitivity for detection of HCC in 31
patients transplanted between 3 and 245
days after MR imaging. The sensitivity
188
dropped to 38% for lesions smaller than
1 cm. However, these authors did not
evaluate dysplastic nodules. In a previ-
ous study (Ward et al., 2000), using
SPIO and dynamic gadolinium injection
in 27 patients (with explant correla-
tion available in 15 patients), the same
authors were able to detect significantly
more HCC nodules (particularly HCCs
< 1 cm), although few lesions were vis-
ible only on SPIO-enhanced images.
Dysplastic nodules were also character-
ized with greater confidence on the basis
of their combined enhancement charac-
teristics (no enhancement with gadolin-
ium, enhancement with SPIO). However,
this approach is more complex and adds
to the cost and duration of MRI. Another
study found that the benefit of the com-
bined use of SPIO and gadolinium is 2. Large HCC: A large HCC can have
negligible (Halavaara et al., 2002).
ATYPICAL APPEARANCES
OF HEPATOCELLULAR
CARCINOMA
1. Small HCC: A small HCC is defined
as a lesion < 2 cm in size. Most small
HCC will show intense arterial phase
enhancement. The widespread use of
arterial phase CT and MRI in cir-
rhotic patients has shown the existence
of enhancing “pseudo-lesions” (small
hepatic nodules with arterial phase
enhancement, not seen on the portal
venous phase images), that may be
wedge-shaped, geographic, oval or
round in shape (Shimizu et al., 2003).
These lesions may represent arterioportal vascular invasion appears as non-
shunts, but the underlying structure is
still not clear (Yu et al., 1997). The
differentiation between small hyper-
B. Taouli
intense pseudo-lesions and HCCs is
difficult with MRI, and follow-up is
often required. Hyperintense pseudo-
lesions should be recognized by their
lack of interval growth or frank dis-
appearance (Shimizu et al., 2003;
Jeong et al., 2002), whereas HCC will
have a tendency to grow. Jeong et al.
(2002) have calculated the doubling
time of small enhancing lesions, and
showed that all nodules with interval
growth and/or change in MR signal are
HCCs (mean doubling time of HCCs,
2 months 2 weeks). The same authors
reported positive and negative predic-
tive values of 100% and 98% respec-
tively, on the basis of nodule growth,
and suggested a close interval follow-up
in such patients.
a number of distinctive features, such
as a mosaic pattern, tumor capsule,
satellite nodules, vascular invasion,
and metastasis. The mosaic pattern is
related to confluent nodules separated
by areas of necrosis and thin septa,
best seen on T2-weighted images and
post-contrast images (Kadoya et al.,
1992). A large HCC often has a fibrous
tumor capsule (seen in 60–82% of
cases) (Kadoya et al., 1992; Choi et
al., 1993), appearing hypointense on
T1- and T2-weighted images, rarely
hyperintense on T2-weighted images,
with delayed enhancement. Vascular
invasion is often present histologically
in the HCC, affecting the portal vein
and the hepatic veins, particularly in a
large non-encapsulated HCC. On MRI,
visualization of the portal branches or
as an intraluminal mass, with enhance-
ment at the arterial phase, and a filling
13. Hepatocellular Carcinoma: Magnetic Resonance Imaging 189

defect on later phases. Extrahepatic NEW PERSPECTIVES

dissemination of HCC can occur in
advanced HCC, and has been described
in 37% of a series of 403 patients with
HCC evaluated with CT (Katyal et al.,
2000), occurring mostly as lung and
lymph node metastases.
3. Diffuse type HCC: Infiltrative or dif-
fuse type HCC represented 13% of
HCCs in a recent series, showing ill-
defined margins with extensive portal
venous tumor thrombosis, variable
enhancement, and markedly elevated
alpha-fetoprotein levels in the blood
(Kanematsu et al., 2003) (Figure 13.4).
Diffusion-weighted MRI: Diffusion-weighted
MRI represents a potential new tool for the
detection and characterization of HCC, and
differentiation of benign from malignant
lesions in the cirrhotic liver. Diffusion-
weighted MRI is noninvasive, requires no
contrast injection, and can be performed
within a breath-hold. Previous studies
have shown that benign lesions (cysts
and hemangiomas) show higher appar-
ent diffusion coefficients (ADCs) than
malignant lesions (HCCs and metastases)
(Kim et al., 1999; Taouli et al., 2003).

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
This is likely related to free water motion in
benign lesions, and restricted water motion
in the presence of tumor. We showed in
a previous study (Taouli et al., 2003) a
significant difference between ADCs of
benign and malignant liver lesions (2.45 ±
0.96 × 10−3 and 1.08 ± 0.50 × 10−3 mm2/s
for b = 0 and 500 s/mm2; respectively,
p < 0.001). The mean ± SD ADCs (x10−3
mm2/s) of the different groups of lesions
were: HCCs 1.33 ± 0.13, hemangiomas
2.95 ± 0.67 and cysts 3.63 ± 0.56 (Taouli
et al., 2003).
Perfusion-weighted MRI: Perfusion-weighted
MRI provides the ability to detect regional
and global alterations in liver blood flow,
and could be performed with CT or MRI.
Whole-liver CT imaging has the potential
to provide both high-temporal-resolution
and high-spatial-resolution imaging of the
entire liver for HCC detection; however,
the radiation dose must be considered.
Perfusion-weighted MRI represents a
promising technique for HCC surveillance
(Pandharipande et al., 2005). The closest
technique to perfusion-weighted MRI has
been use of a double-arterial phase MR
imaging technique enabled by a rapid par-
allel imaging method; one group reported
sensitivity of HCC detection of 91.7%
using double arterial phase vs. 76.3% with
a single arterial phase (Yoshioka et al.,
2002). These findings suggest the poten-
tial of high-temporal-resolution perfusion-
weighted MRI to improve detection of
HCC.
3T MRI: Advantages of imaging at 3T
include higher signal-to-noise ratio, better
image contrast, particularly in gadolinium-
enhanced applications, which translate to
higher spatial resolution images and faster
imaging. This could potentially improve
HCC detection in the future.
B. Taouli
In conclusion, although MRI with
dynamic gadolinium enhancement appears
to be superior to helical CT for detecting
HCC, its sensitivity for the detection of
small lesions is still low, and improvement
in accuracy is still needed. The use of new
contrast media, high field (3T) magnets
and new MRI techniques such as perfusion
and diffusion MRI will likely improve the
image quality and the accuracy of MRI in
the future.
REFERENCES
Amano, S., Ebara, M., Yajima, T., Fukuda, H.,
Yoshikawa, M., Sugiura, N., Kato, K., Kondo, F.,
Matsumoto, T., and Saisho, H. 2003. Assessment
of cancer cell differentiation in small hepatocel-
lular carcinoma by computed tomography and
magnetic resonance imaging. J. Gastroenterol.
Hepatol. 18:273–279.
Aschoff, A.J., Merkle, E.M., Wong, V., Zhang,
Q., Mendez, M.M., Duerk, J.L., and Lewin,
J.S. 2001. How does alteration of hepatic blood
flow affect liver perfusion and radiofrequency-
induced thermal lesion size in rabbit liver? J.
Magn. Reson. Imaging 13:57–63.
Bhartia, B., Ward, J., Guthrie, J.A., and Robinson,
P.J. 2003. Hepatocellular carcinoma in cirrhotic
livers: double-contrast thin-section MR imaging
with pathologic correlation of explanted tissue.
AJR Am. J. Roentgenol. 180:577–584.
Choi, B.I., Takayasu, K., and Han, M.C. 1993.
Small hepatocellular carcinomas and associ-
ated nodular lesions of the liver: pathology,
pathogenesis, and imaging findings. AJR Am. J.
Roentgenol. 160:1177–1187.
Earls, J.P., Theise, N.D., Weinreb, J.C., DeCorato,
D.R., Krinsky, G.A., Rofsky, N.M., Mizrachi, H.,
and Teperman, L.W. 1996. Dysplastic nodules
and hepatocellular carcinoma: thin-section MR
imaging of explanted cirrhotic livers with patho-
logic correlation. Radiology 201:207–214.
Ebara, M., Fukuda, H., Kojima, Y., Morimoto, N.,
Yoshikawa, M., Sugiura, N., Satoh, T., Kondo, F.,
Yukawa, M., Matsumoto, T., and Saisho, H. 1999.
Small hepatocellular carcinoma: relationship of
signal intensity to histopathologic findings and
13. Hepatocellular Carcinoma: Magnetic Resonance Imaging
metal content of the tumor and surrounding
hepatic parenchyma. Radiology 210:81–88.
El-Serag, H.B., and Mason, A.C. 1999. Rising inci-
dence of hepatocellular carcinoma in the United
States. N. Engl. J. Med. 340:745–750.
Fung, J., and Marsh, W. 2002. The quandary over
liver transplantation for hepatocellular carcinoma:
the greater sin? Liver Transpl. 8:775–777.
Halavaara, J., Tervahartiala, P., Isoniemi, H., and
Hockerstedt, K. 2002. Efficacy of sequential use
of superparamagnetic iron oxide and gadolinium
in liver MR imaging. Acta. Radiol. 43:180–185.
Hecht, E.M., Holland, A.E., Israel, G.M., Hahn, W.Y.,
Kim, D.C., West, A.B., Babb, J.S., Taouli, B., Lee,
V.S., and Krinsky, G.A. 2006. Hepatocellular car-
cinoma in the cirrhotic liver: gadolinium-enhanced
3D T1-weighted MR imaging as a stand-alone
sequence for diagnosis. Radiology 239:438–447.
Honda, H., Kaneko, K., Kanazawa, Y., Hayashi,
T., Fukuya, T., Matsumata, T., Maeda, T., and
Masuda, K. 1997. MR imaging of hepatocellular
carcinomas: effect of Cu and Fe contents on sig-
nal intensity. Abdom. Imaging 22:60–66.
Jeong, Y.Y., Mitchell, D.G., and Kamishima, T.
2002. Small (<20 mm) enhancing hepatic nod-
ules seen on arterial phase MR imaging of the
cirrhotic liver: clinical implications. AJR Am. J.
Roentgenol. 178:1327–1334.
Kadoya, M., Matsui, O., Takashima, T., and
Nonomura, A. 1992. Hepatocellular carci-noma:
correlation of MR imaging and histopathologic
findings. Radiology 183:819–825.
Katyal, S., Oliver, J.H. 3rd, Peterson, M.S., Ferris,
J.V., Carr, B.S., and Baron, R.L. 2000. Extrahe-
patic metastases of hepatocellular carcinoma.
Radiology 216:698–703.
Kanematsu, M., Semelka, R.C., Leonardou,
P., Mastropasqua, M., and Lee, J.K. 2003.
Hepatocellular carcinoma of diffuse type: MR
imaging findings and clinical manifestations. J.
Magn. Reson. Imaging 18:189–195.
Kim, T., Murakami, T., Oi, H., Matsushita, M.,
Kishimoto, H., Igarashi, H., Nakamura, H., and
Okamura, J. 1995. Detection of hypervascular
hepatocellular carcinoma by dynamic MRI and
dynamic spiral CT. J. Comput. Assist. Tomogr.
19:948–954.
Kim, T., Murakami, T., Takahashi, S., Hori, M.,
Tsuda, K., and Nakamura, H. 1999. Diffusion-
weighted single-shot echoplanar MR imag-
191
ing for liver disease. AJR Am. J. Roentgenol.
173:393–398.
Krinsky, G.A., and Israel, G. 2003. Nondysplastic
nodules that are hyperintense on T1-weighted
gradient-echo MR imaging: frequency in cir-
rhotic patients undergoing transplantation. AJR
Am. J. Roentgenol. 180:1023–1027.
Krinsky, G.A., Lee, V.S., Nguyen, M.T., Rofsky,
N.M., Theise, N.D., Morgan, G.R., Teperman,
L.W., and Weinreb, J.C. 2001. Siderotic nodules
in the cirrhotic liver at MR imaging with explant
correlation: no increased frequency of dys-
plastic nodules and hepatocellular carcinoma.
Radiology 218:47–53.
Matsui, O., Kadoya, M., Kameyama, T., Yoshikawa,
J., Arai, K., Gabata, T., Takashima, T., Nakan-uma,
Y., Terada, T., and Ida, M. 1989. Adenomatous
hyperplastic nodules in the cirrhotic liver: differ-
entiation from hepatocellular carcinoma with MR
imaging. Radiology 173:123–126.
Matsui, O., Kadoya, M., Kameyama, T., Yoshikawa,
J., Takashima, T., Nakanuma, Y., Unoura, M.,
Kobayash, K., Izumi, R., Ida, M., et al. 1991.
Benign and malignant nodules in cirrhotic livers:
distinction based on blood supply. Radiology
178:493–497.
Mazzaferro, V., Regalia, E., Doci, R., Andreola,
S., Pulvirenti, A., Bozzetti, F., Montalto, F.,
Ammortuna, M., Morabito, A., and Gennari, L.
1996. Liver transplantation for the treatment of
small hepatocellular carcinomas in patients
with cirrhosis. N. Engl. J. Med. 334:693–699.
Mitchell, D.G., Kim, I., Chang, T.S., Vinitski, S.,
Consigny, P.M., Saponaro, S.A., Ehrlich, S.M.,
Rifkin, M.D., and Rubin, R. 1991. Fatty liver.
Chemical shift phase-difference and suppres-
sion magnetic resonance imaging techniques in
animals, phantoms, and humans. Invest. Radiol.
26:1041–1052.
Murakami, T., Kim, T., Oi, H., Nakamura, H.,
Igarashi, H., Matsushita, M., Okamura, J., and
Kozuka, T. 1995. Detectability of hypervascu-
lar hepatocellular carcinoma by arterial phase
images of MR and spiral CT. Acta Radiol.
36:372–376.
Murakami, T., Baron, R.L., Peterson, M.S., Oliver,
J.H. 3rd, Davis, P.L., Confer, S.R., and Federle,
M.P. 1996. Hepatocellular carcinoma: MR imag-
ing with mangafodipir trisodium (Mn-DPDP).
Radiology 200:69–77.
192
Nakakoshi, T., Kajiyama, M., Fujita, N., Jong-
Hon, K., Takeichi, N., and Miyasaka, K. 1996.
Quantitative analyses of correlations of signal
intensity on T1-weighted images and T1 relaxa-
tion time with copper concentration in the rat
liver. Acad. Radiol. 3:36–39.
Oi, H., Murakami, T., Kim, T., Matsushita, M.,
Kishimoto, H., and Nakamura, H. 1996. Dynamic
MR imaging and early-phase helical CT for
detecting small intrahepatic metastases of hepa-
tocellular carcinoma. AJR Am. J. Roentgenol.
166:369–374.
Pandharipande, P.V., Krinsky, G.A., Rusinek, H.,
and Lee, V.S. 2005. Perfusion imaging of the
liver: current challenges and future goals.
Radiology 234:661–673.
Rofsky, N.M., Weinreb, J.C., Bernardino, M.E.,
Young, S.W., Lee, J.K., and Noz, M.E. 1993.
Hepatocellular tumors: characterization with
Mn-DPDP-enhanced MR imaging. Radiology
188:53–59.
Shimizu, A., Ito, K., Koike, S., Fujita, T., Shimizu,
K., and Matsunaga, N. 2003. Cirrhosis or chronic
hepatitis: evaluation of small (< or =2-cm) early-
enhancing hepatic lesions with serial contrast-
enhanced dynamic MR imaging. Radiology
226:550–555.
Taouli, B., Vilgrain, V., Dumont, E., Daire, J.L.,
Fan, B., and Menu, Y. 2003. Evaluation of liver
B. Taouli
diffusion isotropy and characterization of focal
hepatic lesions with two single-shot echo-planar
MR imaging sequences: prospective study in 66
patients. Radiology 226:71–78.
van den Bos, I.C., Hussain, S.M., Terkivatan, T.,
Zondervan, P.E., and de Man, R.A. 2006. Stepwise
carcinogenesis of hepatocellular carcinoma in the
cirrhotic liver: demonstration on serial MR imag-
ing. J. Magn. Reson. Imaging 24:1071–1080.
Ward, J., Guthrie, J.A., Scott, D.J., Atchley, J.,
Wilson, D., Davies, M.H., Wyatt, J.I., and
Robinson, P.J. 2000. Hepatocellular carcinoma
in the cirrhotic liver: double-contrast MR imag-
ing for diagnosis. Radiology 216:154–162.
Yamashita, Y., Mitsuzaki, K., Yi, T., Ogata, I.,
Nishiharu, T., Urata, J., and Takahashi, M. 1996.
Small hepatocellular carcinoma in patients with
chronic liver damage: prospective comparison of
detection with dynamic MR imaging and helical
CT of the whole liver. Radiology 200:79–84.
Yoshioka, H., Takahashi, N., Yamaguchi, M., Lou,
D., Saida, Y., and Itai, Y. 2002. Double arterial
phase dynamic MRI with sensitivity encoding
(SENSE) for hypervascular hepatocellular carci-
nomas. J. Magn. Reson. Imaging 16:259–266.
Yu, J.S., Kim, K.W., Sung, K.B., Lee, J.T., Yoo,
and H.S. 1997. Small arterial-portal venous
shunts: a cause of pseudolesions at hepatic imag-
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
INTRODUCTION
Angiogenesis is the process whereby new
blood vessels develop from the preexisting
vasculature. It takes place physiologically
during embryonic development, during the
normal growth of tissues and the wound
healing, and during the female reproduc-
tive cycle (i.e., ovulation, menstruation,
and placental development) as well as dur-
ing the pathologic growth and metastatic
spread of malignant neoplasms (Folkman,
1971). A variety of humoral agents need
to be activated to generate a neovascular
blood supply or angiogenesis in the human
body, and vascular endothelial growth fac-
tor (VEGF) is one of the most important
humoral agents to be activated to ensure
the growth of the vascular endothelium.
Various types of hepatocellular nodules
with malignant potential are known to
develop in livers with chronic liver dam-
age: they are large regenerative nodules,
low-grade dysplastic nodules, high-grade
dysplastic nodules, well-differentiated
hepatocellular carcinoma (HCC), and
moderately or poorly differentiated HCC
(International Working Party, 1995). In the
process of HCC development, the expres-
sion of proangiogenetic factors such as
VEGF appears to play an important role.
The expression of this factor in HCC tis-
sues has been considered to be associated
with tumor size or histological tumor
grade (Yamaguchi et al., 1998). Even in
the nontumorous liver parenchyma, VEGF
is expressed by sinusoidal endothelial cells
and hepatocytes (Yamane et al., 1994).
Although radiologic examinations such
as conventional angiography, computed
tomography (CT), and magnetic resonance
imaging (MRI) have been widely used as
tools for the diagnosis of HCCs, relations
between the imaging findings and biomo-
lecular angiogenetic activities in HCC and
in the surrounding liver have yet to be ascer-
tained. Investigations of this relation may
help radiologists understand radiologic
imaging findings related to molecular bio-
logical treatments using an antiangiogenic
agent such as bevacizmab, an antibody that
binds to VEGF, which has been approved
by the US Food and Drug Administration
for the treatment of metastatic colorectal
cancer. Furthermore, PTK787/ZK 222584
(PTK/ZK) is a potent, orally active and
selective inhibitor of the VEGF-receptor
tyrosine kinases VEGFR-1 (Flt-1) and
VEGFR-2 (KDR) under development. The
anti-angiogenetic effects of PTK/ZK in
193
194
several advanced cancers, including color-
ectal cancer, breast cancer, glioblastoma
multiforme, prostate, and renal cancer
where VEGF is known to play a role,
have been expected. Most interestingly,
previous results have demonstrated that
gadolinium-enhanced MRI findings may
be useful biomarkers for defining the
pharmacological response of colorectal
cancer and liver metastases and for dose
of angiogenesis inhibitors, such as PTK/
ZK (Morgan et al., 2003). However, rela-
tions between VEGF activity in tumors
and radiologic findings have yet to be
ascertained.
We previously assessed the correla-
tion between radiologic findings in ang-
iographically-assisted CT and MRI and
angiogenetic activities determined by
immunohistochemistry (Kanematsu et al.,
2004a) and Western blotting (Kanematsu
et al., 2004b, 2005) for VEGF in vari-
ous hepatocellular nodules in cirrhosis,
HCCs, and the surrounding liver. They
have described the potential usefulness
of radiologic imaging in the evaluation of
angiogenetic activity of hepatic nodules
and the surrounding liver. We summa-
rize the results and discuss the relations
between radiologic findings and VEGF
expression in this chapter.
HISTORY OF RADIOLOGIC
DIAGNOSIS
OF HEPATOCELLULAR
CARCINOMA
Hepatocellular carcinoma is a malignant
neoplasm that arises from hepatocytes
chiefly in patients with chronic liver dam-
age. Hepatocellular carcinoma is either
M. Kanematsu et al.
the first or second most common cause
of cancer death worldwide, and in 1990
the World Health Organization estimated
that there were ∼430,000 new cases of
liver cancer worldwide, and a similar
number of patients died as a result of this
disease. The most common risk factors for
HCC include hepatitis B or C viral infec-
tion, aflatoxins, iron overload, and alcohol
abuse. This disease is particularly preva-
lent in parts of Asia, Africa, and Europe.
Eighty percent of HCC patients have cir-
rhosis of the liver, and 70% have elevated
serum alpha fetoprotein levels. Treatment
options include liver transplantation, par-
tial hepatectomy, transcatheter arterial
embolization, percutaneous ablation, and
radiation therapy. Reported 5-year survival
rates range from 10% to 60%, depend-
ing on the tumor stage and the treatment
method. The number of hepatitis virus
carriers in Japan is estimated to be > 3
million (1.6% of the population), and of
these 1 million develop chronic hepati-
tis and 30,000 HCC. A national survey
(the third National Health and Nutrition
Examination Survey [NHANES III]) of the
civilian, non-institutionalized U.S. popula-
tion found that 1.8% of Americans (3.9
million) have been infected with hepatitis
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
the risk of HCV infection by ninefold.
Hepatocellular carcinoma may become
the most common malignant tumor in
North America in a decade or two, and is
the only major cancer with a substantial
increase in incidence.
A pathology-based study from 1984
reported that the common major gross
patterns were expanding, spreading, and
14. Expression of Vascular Endothelial Growth Factor in Hepatocellular Carcinoma
multifocal by means of aggregating a total
of 529 HCC cases from Japan, the United
States, and South Africa. Further modifica-
tions have divided HCC into infiltrative or
expansive, single or multinodular, and mixed
types, using observations on encapsulation
and intrahepatic venous spread. Recently,
however, with the advent of high-perform-
ance radiologic diagnosis equipment such
as sonography, CT and MRI, small HCCs,
of 10 mm in size or less, have been preoper-
atively diagnosed in greater numbers, which
has impacted upon improvement in early
detection of HCCs.
In the mid 1970s, selective hepatic
angiography was introduced as an inva-
sive diagnostic tool of importance in the
preoperative evaluation of the extent and
the vascular supply of HCCs and the liver.
Many tumors were readily diagnosed on
angiograms, but there were limitations
which included: invasiveness of the pro-
cedure, difficulty distinguishing multi-
centric HCCs from hepatic metastases
and poorly vascularized lesions from
primary cholangiocarcinoma, and inabil-
ity to detect iso- or hypovascular HCCs.
During the same period, liver scintig-
raphy using 99mTc-sulfur colloid was
introduced as a technique to detect focal
abnormalities, but it was nonspecific,
since most focal hepatic lesions from
metastases to hepatic cysts appeared as
cold defects. In the early 1980s, transab-
dominal ultrasonography was developed
as a noninvasive tool in the diagnosis of
HCCs. Ultrasound experienced difficulty
in visualization the entire liver due to
acoustic blockage by ribs or intestinal or
pulmonary air, and sound attenuation due
to thick truncal fat in obese patients, and
these represented significant limitations.
Although a variety of benign, premalignant
195
or borderline hepatocellular nodules in
cirrhosis were found on sonography, dif-
ficulty in the differentiation of HCCs
from those nodules was a significant
problem.
In the early 1980s, radiologists began
using early-generation CT for diagnos-
ing HCCs. Detection was not satisfactory
because 10-mm slices were incrementally
obtained after intravenous drip infusion
of iodinated contrast material in a lengthy
examination time, and small lesions were
obscured in the thick slice acquisition, or
missed due to slice misregistration caused
by irregular patient breath holding, or
were isodense due to poor contrast admin-
istration dynamics, which did not allow
detection of small hypervascular HCCs. In
the mid 1980s, an incremental CT acqui-
sition during conventional trans-superior
mesenteric arterial portography (CT during
arterial portography) was introduced as a
mean to examine small HCCs. This exami-
nation was very sensitive but suffered from
a variety of false-positive lesions due to
portal venous obstruction, arterioportal
shunting, non-portal splanchnic venous
return, or benign focal hepatic lesions. As
a further development in the mid 1990s, a
combination method of CT during arterial
portography and CT during conventional
hepatic arteriography was reported, and
found to have high diagnostic accuracy
(Kanematsu et al., 1997). This procedure
did not gain widespread utilization because
of its invasiveness, and technical challeng-
ing nature. We have employed combined
CT during arterial portography and CT
hepatic arteriography for the preoperative
detection and evaluation of vascularity
of HCCs and of surrounding liver in our
clinical and research practice, and have cor-
related VEGF expression and vascularity
196
of HCCs (Kanematsu et al., 2004b).
In the early 1990s, helical or spiral CTs
were introduced into clinical routine and
have undergone continued evaluation since
then. The most up-to-date 64-row multi-
detector CT scanner completes whole liver
scanning within no more than 1.8 s, and
hepatic arterial dominant-phase CT images
obtained immediately after intravenous
bolus injection of iodinated contrast mate-
rial permit detection of tiny hypervascular
HCC foci in the liver.
In the mid 1980s, MRI had been reported
as an efficient diagnostic tool for HCCs,
but only conventional spin-echo sequences
with lengthy acquisition time were avail-
able in the clinical setting. By the early
1990s, fast gradient-echo sequences that
allowed whole liver scanning during a
breath-hold, became available and gado-
linium-enhanced dynamic MRI of the
liver began to be used in daily radiologic
practice. Magnetic resonance imaging has
continued to evolve in speed of data acqui-
sition and in contrast agent design, and
it has maintained its status as the most
accurate examination in the diagnosis of
HCCs, which it initially acquired in the
mid 1990s (Semelka et al., 1992). By
allowing evaluation of T1 and T2 relaxa-
tion, protein, hemorrhage, fat deposition,
metal deposition, blood flow, perfusion,
and diffusion, MRI provides more abun-
dant information on underlying pathology
than CT for the diagnosis of diffuse and
focal hepatic diseases.
In the near future, MRI will continue to
improve in temporal and spatial resolu-
tions, develop and refine contrast agents
including liver-specific contrast agents,
develop diffusion-weighted imaging, and
allow performing perfusion and spectro-
scopic studies on a routine clinical basis.
M. Kanematsu et al.
Although CT will also continue to shorten
its acquisition time and reduce radiation
exposure, MRI will likely remain advanta-
geous because of the superior capability of
describing histopathologic components in
tumors, detection and characterization of
lesions, and safety.
DEVELOPMENT OF
HEPATOCELLULAR
CARCINOMA, ITS
VASCULATURE, AND
VASCULAR ENDOTHELIAL
GROWTH FACTOR
EXPRESSION
Large regenerative nodule is the initial
primary form of nodular hepatocellular
lesions in cirrhosis, which has blood sup-
ply from hepatic arteries and portal veins
in similar proportion to normal liver tis-
sue. Dysplastic nodule is a nodule more
progressed towards HCC than a large
regenerative nodule, and dysplastic nodule
develops clone-like populations or nuclear
hyperchromasia. Dysplastic nodules are
further categorized as low-grade and
high-grade dysplasias, with high-grade
dysplasia showing histopathologic find-
ings similar to those of well-differentiated
HCCs. High-grade dysplastic nodule dis-
plays cell density more than twice normal,
irregular nuclear contour, and invasion of
stroma or portal tracts, which are features
of HCC. The pathologic differentiation
of high-grade dysplasias from well-dif-
ferentiated HCCs is often difficult, and
as a result lesions are occasionally called
“borderline lesions” in the clinical setting.
Large dysplastic nodules are perfused by
both hepatic arteries and portal veins.
14. Expression of Vascular Endothelial Growth Factor in Hepatocellular Carcinoma 197

Researchers using combined CT during be highly promoted at this time, due to

arterial portography and CT hepatic arte- hypoxia. By this upregulation mechanism
riography to evaluate blood supply to (Figure 14.1A) (Viglietto et al., 1995),
hepatocellular nodules in cirrhosis, have early-stage HCC begins to create newly
suggested that hepatic arterial blood sup- developed arteries, referred to as “unpaired
ply was decreased in large regenerative
and dysplastic nodules compared with
that in background liver tissue (Lim et al.,
1999). Our current belief is that oxygen
supply to such nodules might be somewhat
decreased. We have found that VEGF
expression, determined by immunohisto-
chemistry, was moderately increased in
large regenerative nodules and high-grade
dysplastic nodules, which might reflect a
moderate hypoxic status of these nodules
(Kanematsu et al., 2004a).
Early-stage well-differentiated HCC
retains histologic structures similar to nor-
mal liver, they harbor portal triads as in nor-
mal liver, and are perfused by both hepatic
arteries and portal veins, although the
relative proportions of arterial and portal
venous blood supply to well-differentiated
HCCs is highly variable. In the beginning
phases, small well-differentiated HCCs, in
which cancer cells replace noncancerous
hepatocytes, are commonly hypovascu-
lar and the proportion of arterial blood Figure 14.1. Flow charts showing mechanisms of
supply is somewhat decreased. As can- VEGF upregulation and downregulation. A, Inefficient
cer cells invade into sinusoids, which is vascular supply and the resultant decrease in tissue
referred to as “stromal invasion”, portal oxygen tension lead to neovascularization to sat-
triads decrease in number to no more than isfy the needs of tissue. In human tissues, hypoxia
induces upregulation of VEGF gene expression by
one fourth of those in normal liver. In this biomolecular mechanism, i.e., HIF-1α and HIF-2α,
early stage, well-differentiated HCC lacks produced by the signal of transformation from oxy-
both hepatic arterial and portal venous hemoglobin into deoxyhemoglobin induce VEGF
blood supply, and is in a hypoxic status peptides synthesis via mRNA, and hyperoxia is
in the liver. Consequently, early-stage, brought about by way of newly developed blood
hypovascular well-differentiated HCCs vessels by angiogenesis. B, Hyperoxia eventually
brought about in previously hypoxic tissues by
appear hypovascular on both CT during angiogenesis transforms deoxyhemoglobin back to
arterial portography and CT hepatic arte- oxyhemoglobin, which suppresses VEGF peptide
riography (Hayashi et al., 1999). Vascular synthesis via HIF-1 α/2α and mRNA, which is
endothelial growth factor expression may referred to as downregulation
198
arteries”, by recruiting sinusoidal endothe-
lial cells to supply oxygen to the tumor.
Angiogenesis may be vigorous until the
lesion acquires abundant oxygen by way
of unpaired arteries. This phase occurs
when lesion size is ∼10–15 mm. However,
the lesion may at this time still not appear
hypervascular on imaging studies.
As early-stage HCC grows up to 15 mm
in size, the number of portal triads sub-
stantially decreases, and a greater number
of small unpaired arteries develop from
the residual portal triads in the tumor.
At this stage of development, sinusoid-
like vascular spaces contribute to tumor
vascularity, and they develop vigorously,
which is referred to as “capillarization”.
Once lesions exceed 15 mm in size, HCCs
often acquire increased hepatic arterial
vascularity, which is observed on conven-
tional angiograms, contrast-enhanced CT,
and MRI, as hypervascular enhancement.
When an HCC is shown as a hypervascular
lesion on radiographic imaging such as CT
or MRI, it usually no longer has appreci-
able portal venous perfusion, and they are
referred to as “overt or frank HCC”.
Once a small well-differentiated HCC
acquires ample hepatic arterial blood sup-
ply with concomittent high oxygen tension,
expression of VEGF may be suppressed by
downregulation (Figure 14.1B). The HCC
continues to grow in size owing to its bio-
logical malignant behavior and still creates
more tumor arteries, thereby maintaining
oxygen supply. Well-differentiated HCC
progressively dedifferentiates to moderately
differentiated HCC, at which stage they do
not harbor portal triads and are almost
exclusively perfused by hepatic arteries.
Generally, most HCCs 2 cm in size
or greater have macroscopic fibrous
pseudocapsules that result in a clearly
M. Kanematsu et al.
defined nodular appearance of the tumor.
In hypervascular nodular HCCs with
fibrous pseudocapsules, blood entering
the tumor through hepatic arteries circu-
late in sinusoids or capillarized vessels in
the tumor, then outflow into portal veins
or sinusoids in the surrounding noncan-
cerous liver tissue through portal venules
penetrating the fibrous pseudocapsules,
and finally drain into the hepatic veins.
This distinctive blood flow pathway in
hypervascular nodular HCCs was well-
demonstrated by Ueda et al. (1998) using
single-level cine CT during conventional
hepatic arteriography. Ring enhancement
that is seen in the late hepatic arterial phase
was termed ‘corona enhancement’, which
may represent summation of enhance-
ment of pseudocapsules with high vessel
density and enhancement of peritumoral
sinusoids. The development of ‘corona
enhancement’ can also be observed on
intravenously contrast-enhanced CT and
MRI if two serial scans are run during
hepatic arterial-dominant phase, using
high-speed CT (Murakami et al., 2001)
or MRI (Shimizu et al., 2003).
The growth of HCCs from 15 mm to
3 cm in size, is usually accompanied
by marked hypervascularity. The tumor
grade of these very hypervascular HCCs
with fibrous pseudocapsules is often
moderately-differentiated. The HCC is
well supplied with oxygen because of
the abundant hepatic arterial vascularity
and this situation may be allegorically
explained as “satisfactory status”. Such
an HCC often is in a hyperoxic status,
and therefore, does not need to further
generate tumor arteries by angiogenesis,
with the result that VEGF activity may
be suppressed due to downregulation
mechanism (Figure 14.1B).
14. Expression of Vascular Endothelial Growth Factor in Hepatocellular Carcinoma
As the HCC grows beyond 3 cm in size,
the distance between tumor vessels and
cancer cells tends to increase. As shown by
Folkman (1971), oxygen diffuses no more
than 150 μm in tissue, thus some parts of
the enlarged HCC develop hypoxia and
resultant internal necrosis. The hypoxia
induces the generation of VEGF peptides
once again. However, the generation of
tumor vessels is physically limited when
the HCC reaches 5–15 cm in size, due to
tissue pressure and cell crowding, and large
HCCs may require anaerobic metabolism
with glycolytic pathway with hexokinase
II to sustain themselves (Gwak et al.,
2005). This is observed on imaging stud-
ies as HCCs measuring 5–15 cm in size
often do not enhance well on contrast-
enhanced hepatic arterial-phase CT or MRI,
despite highly activated VEGF expression.
Yamaguchi et al. (1998) described that sar-
comatous HCCs, the final form of dediffer-
entiation with highly malignant behavior,
showed low VEGF expression in immuno-
histochemistry studies.
CORRELATION OF VASCULAR
ENDOTHELIAL GROWTH
FACTOR EXPRESSION AND
RADIOLOGIC FINDINGS
IN HEPATOCELLULAR
CARCINOMA IN PRIOR
REPORTS
Mise et al. (1996) performed Northern
blot analysis and immunohistochemistry to
semi-quantify and localize the expression
of VEGF in 20 HCCs and nine metastatic
tumors, and reported that the degree of
VEGF mRNA expression was significantly
correlated with the intensity of tumor stain-
199
ing in angiograms, and VEGF protein was
substantial in HCC cells on immunohis-
tochemical observation. In the same year,
Suzuki et al. (1996) studied vascular per-
meability factor (VPF)/VEGF transcript
expression using reverse-transcription
polymerase chain reaction (PCR) analy-
sis, and reported that 16 (69.6%) cases
showed VEGF transcript expression in
the tumor, whereas only 9 (39%) showed
it in adjacent liver parenchyma. Vascular
endothelial growth factor mRNA expres-
sion in HCCs in their study was associated
with fibrous capsule formation and septal
formation. Vascular endothelial growth
factor activity was not correlated with the
vascularity of HCCs as determined by con-
ventional angiography.
El-Assal et al. (1998) performed quantita-
tive reverse-transcription PCR and immu-
nostaining in 71 patients with HCCs to
determine microvessel density (MVD) and
VEGF expressions in HCCs and in the sur-
rounding liver. They reported that tumor vas-
cularity on angiography was not correlated
with the MVD, and neither VEGF mRNA
levels nor protein expression in HCCs was
correlated with the MVD or any histopatho-
logical features of HCCs, although they did
not directly correlate VEGF expression and
angiographic findings.
Kwak et al. (2001) attempted to deter-
mine whether VEGF is a histopathological
factor influencing contrast enhancement of
HCC by qualitatively evaluating 22 nodular
HCCs on intravenously contrast-enhanced
CT in correlation with the qualitative
extent of anti-VEGF antibody expres-
sion determined by immunohistochemical
staining. They reported that the degree of
enhancement on the hepatic arterial-phase
CT had a significant positive correlation
with VEGF expression in HCCs.
200
OUR PREVIOUS
RESEARCH WITH
IMMUNOHISTOCHEMISTRY
We evaluated 16 patients with HCC, 1
with a large regenerative nodule, and 1
with a high-grade dysplastic nodule, who
underwent surgical resection of lesions
at University of North Carolina from
December 1999 to April 2002 (Kanematsu
et al., 2004a). Magnetic resonance imaging
included a breath-hold T1-weighted in-
phase spoiled gradient-echo and a breath-
hold T2-weighted short-tau inversion
recovery echo-train spin-echo imaging.
Spoiled gradient-echo images were acquired
before and after intravenous bolus injection
of gadolinium chelate. Postcontrast early-
phase imaging was initiated at 18 s, and
late-phase imaging was at 2 min. Lesion-to-
liver contrast-to-noise ratio was measured.
Contrast enhancement index was calculated
by subtracting the contrast-to-noise ratio
on precontrast images from that on post-
contrast images. Two blinded radiologists
reviewed the images for signal intensity.
Immunohistochemical staining was per-
formed using the standard avidin-biotin
peroxidase complex technique. An expe-
rienced liver pathologist evaluated immu-
noreactivity for VEGF semi-quantitatively:
grade 0 stood for virtually no immunore-
activity, grade 1 for patchy to diffuse weak
immunoreactivity, grade 2 for patchy to
diffuse moderate immunoreactivity, grade
3 for patchy to diffuse strong immunore-
activity, and grade 4 for diffuse intense
immunoreactivity. The histologic diagnosis
of hepatic nodules was based on the crite-
ria described by the International Working
Party (1995).
All four hepatic nodules that showed
immunoreactivity for VEGF of grade 3 or
M. Kanematsu et al.
4 were moderately-differentiated HCCs.
The one large regenerative nodule and one
dysplastic nodule showed immunoreactiv-
ity for VEGF of grade 2. No significant
correlation was found between the tumor
grade and the grade of immunoreactiv-
ity for VEGF. No correlation was found
between the lesion size and the grade of
immunoreactivity for VEGF. The contrast-
to-noise ratios on T2-weighted images
were higher (P < 0.05) with immunore-
activity for VEGF of grade 4 than with
those of grades 1, 2, and 3, and were
marginally higher (P < 0.07) than those
of grade 0. No correlation was found
between contrast enhancement indexes on
postcontrast early and late-phase images
and immunoreactivity for VEGF.
A previous pathologic study by Yamaguchi
et al. (1998) reported that in HCCs of
single histologic grade, VEGF expres-
sion was the highest in well-differentiated
HCCs, followed by moderately differenti-
ated HCCs, and then poorly differentiated
HCCs. In our study, although no statisti-
cal correlation was found between tumor
grade and immunoreactivity for VEGF,
immunoreactivity for VEGF was the high-
est in moderately-differentiated HCCs:
five (83%) of six well-differentiated HCCs
showed weak to moderate immunoreactiv-
ity for VEGF and none of them showed
strong or intense immunoreactivity, but
four (67%) of six moderately differen-
tiated HCCs showed strong to intense
immunoreactivity for VEGF. The reason
for the discrepancy between our study
and that of Yamaguchi et al. (1998) is
unclear. There might be a difference in
subjective pathologic evaluation of tumor
grades or in composition of tumor grades:
Yamaguchi’s study included ten small
well-differentiated HCCs of 13.0 ± 9.8 mm
14. Expression of Vascular Endothelial Growth Factor in Hepatocellular Carcinoma
in diameter, in our study six well-differen-
tiated HCCs measured slightly larger 22 ±
10 mm. It may be that Yamaguchi’s popu-
lation included more small hypovascular
well-differentiated HCCs, which would
predictably have higher VEGF expression
than tumors of a larger size as observed
in our study. Another reason might be
epidemiologic differences associated with
underlying liver diseases between North
American and Asian populations.
Our study included a large regenerative
nodule and a high-grade dysplastic nodule,
and both expressed patchy to diffuse mod-
erate immunoreactivity for VEGF. To our
knowledge, immunoreactivity of VEGF in
large regenerative or dysplastic nodules in
cirrhosis has not been assessed in previous
studies. Although the few number of these
hepatic nodules precluded statistical power,
our results suggest that benign or border-
line hepatic nodules in cirrhosis, such as
large regenerative or dysplastic nodules,
may show an increased immunoreactivity
for VEGF, and therefore, may have angio-
genetic potential prefiguring further devel-
opment of well-differentiated HCCs.
It is believed that an HCC develops in a
background of cirrhosis frequently by means
of a multi-step dedifferentiation processes
that progress from large regenerative to dys-
plastic nodule and to HCC (International
Working Party, 1995). Regarding vascu-
larity in precancerous hepatic nodules in
cirrhosis, one report described that large
regenerative nodules in cirrhosis were mini-
mally enhancing nodules surrounded by
enhancing fibrous septa at CT hepatic
arteriography in 15 of 20 patients with
cirrhosis (Lim et al., 1999). We believe
that most of large regenerative and dys-
plastic nodules are iso- or hypovascular,
and if they have an increased VEGF
201
activity upregulated by hypovascularity or
hypoxia, there may be a future propensity
to increased vascularity and development
into a hypervascular HCC.
In two HCCs with diffuse intense immu-
noreactivity for VEGF the lesion-to-liver
contrast-to-noise ratios on T2-weighted
images were significantly higher than
those of HCC with no, mild, moderate,
and strong immunoreactivity for VEGF.
This observation suggests that the HCCs
with intense immunoreactivity for VEGF
contained a greater amount of free water
in the extracellular or interstitial spaces,
reflected by higher T2 signal, possibly due
to increased vascular permeability caused
by activated VEGF. Vascular permeability
is known to be increased by activation
of VEGF peptides (Shweiki et al., 1992;
Monacci et al., 1993).
In our study using immunohistochemis-
try, we found no correlation between arterial
vascularity observed on postcontrast early-
phase images and VEGF activity. Our belief
is that small, well-differentiated HCCs may
well have increased VEGF activity that
serves to promote vascularity and tumor
growth, but when well-differentiated HCCs
are well vascularized and have increased in
size, VEGF activity decreases, perhaps by
downregulation mechanism (Figure 14.1B).
Vascular endothelial growth factor activ-
ity may occur temporally as a precursor to
actual vessel presence, and therefore may
not correlate directly with vascularity as
reflected by intensity of early enhancement.
We thought this might explain the poor
correlation between early enhancement
on gadolinium-enhanced MRI and VEGF
activity that we observed, and determined
to perform further studies to elucidate these
questions. Signal intensities on unenhanced
T1- and T2-weighted images were more
202 M. Kanematsu et al.

closely associated with immunoreactivitytumorous liver parenchyma resected at

for VEGF than vascularity evaluated on Gifu University Hospital. Among them,
gadolinium-enhanced MRI, which may 22 patients underwent preoperative MRI
suggest that these features, reflecting of the liver within 2 weeks of surgery, and
higher fluid content, occur before vessel
20 patients underwent preoperative angi-
development. ographically assisted CT with combined
There were some limitations to our CT during arterial portography and CT
study: first, the study population was small
hepatic arteriography. Of the 28 patients, 6
reflecting that we excluded many patients,
patients had type B viral hepatitis, and 22
including those in whom pathologic diag-had type C viral hepatitis. No patient had
nosis was established by biopsy, becausea history of alcohol abuse. The underlying
VEGF activity could not be determined. liver disease documented by histopatho-
This study was retrospective, indicatinglogic study was chronic hepatitis due to
that this was a new concept to us to mild cirrhosis in 7 patients, moderate cir-
evaluate the relationship between VEGF rhosis in 16, and severe cirrhosis in 5.
activity and MRI findings of several histo-For immunoblotting techniques, the
logic types of hepatic nodules. A prospec-
specimen was sectioned through the tumor
tive, randomized, multi-institutional study
center to ensure correlation with CT and
with a larger population would be neces-MRI. Five-gram samples were dissolved in
sary. Although we evaluated the VEGF 1 ml of radioimmunoprecipitation buffer.
expression only in hepatic nodules, the Cell lysates were subjected to 10% sodium
background liver parenchyma or large dodecyl (lauryl) sulfatepolyacrylamide gel
regenerative nodules in cirrhosis also show
electrophoresis. Proteins were transferred
varying degree of VEGF activity. Although
to polyvinylidene difluoride membranes.
none of the hepatic nodules or background
After blocking membranes with tris-buff-
livers had significantly decreased signal
ered saline containing polysorbate 20 and
intensity due to iron deposition, iron depo-
5% skim milk, membranes were incubated
sition would affect the MR signal intensi-
with anti-VEGF monoclonal antibody and
ties and expected results. then with antimouse IgG coupled with
horseradish peroxidase. Detection was
performed using enhanced chemilumines-
MAGNETIC RESONANCE cence. Vascular endothelial growth factor
AND COMPUTED expression, observed as electrophoretic
bands, was quantified using image analysis
TOMOGRAPHIC DATA software, which calculated the area under
CORRELATED TO VASCULAR the histograms. Each pair of HCC sam-
ENDOTHELIAL GROWTH ples and surrounding liver was examined
FACTOR MEASURED USING using recombinant human VEGF solution
WESTERN BLOTTING (1.25 mg/ml) for calibration purpose. The
VEGF expression index (VEGFIND) was
In a second study using Western blot- calculated by dividing the area under the
ting to determine VEGF expression, we histogram for the specimen band by that
evaluated 28 HCCs with surrounding non- of the calibration band.
14. Expression of Vascular Endothelial Growth Factor in Hepatocellular Carcinoma
Magnetic resonance imaging was per-
formed in a similar fashion to our prior
study obtaining T1-weighted gradient-
echo, T2-weighted echo-train spin-echo,
and gadolinium-enhanced multiple-phase
gradient-echo images in the hepatic arte-
rial-dominant, portal venous, and equilib-
rium phases. Computed tomography was
performed using a combination of unen-
hanced CT, CT during arterial portogra-
phy, and CT hepatic arteriography. On
CT portography, scan was started 25–35 s
after the initiation of a transcatheter supe-
rior mesenteric arterial injection of 95 ml
of nonionic contrast material containing
150 mg iodine/ml at 3 ml/s, and on CT
arteriography scan was started 5 s after
the initiation of a transcatheter common
hepatic arterial injection of 30 ml of the
same contrast at 1.5 ml/s.
This study evaluated MRI data using
a quantitative approach. Our technique
was as follows: to determine MR signal
intensity values and their standard devia-
tions (SDs) in HCCs and in the surround-
ing liver, a circular region of interest was
drawn to encompass as much of the lesion
as possible, and another circular region of
interest was drawn in a region of the sur-
rounding liver. The SD of the background
noise, SDB, was measured in the phase-
encoding direction outside the anterior
abdominal wall to calculate the following:
lesion-to-liver contrast-to-noise ratio = (SI
lesion – SI liver)/SDB, where SI lesion
and SI liver are the signal intensities of
the HCC and of the surrounding liver,
respectively. As a quantitative parameter
of signal intensity heterogeneity, the sig-
nal intensity SD ratio was calculated as
follows: SD ratio = SD lesion/SDB, where
SD lesion is the signal intensity SD of
the HCC. As a quantitative parameter of
203
fat deposition in HCCs, the phase-shift
index was calculated as follows: phase-
shift index = (SI in-phase – SI opposed-
phase)/SI in-phase, where SI in-phase and
SI opposed-phase are the signal intensi-
ties of HCC on in-phase and opposed-
phase T1-weighted gradient-echo images,
respectively.
To measure CT values in HCCs and in
the surrounding liver, a circular region
of interest was drawn to encompass as
much of the HCC as possible, and another
was drawn in the surrounding liver paren-
chyma. Quantitative degree of contrast
enhancement was expressed as contrast
enhancement index, which was calculated
by subtracting the CT values on unenhanced
CT from those on contrast-enhanced CT.
Two experienced gastrointestinal radi-
ologists independently reviewed MRI in
a retrospective manner. They subjectively
evaluated the signal intensity in HCC and
in the surrounding liver using a seven-point
scale from “strong hypointensity (−3)” to
“marked hyperintensity (+3)”. Intensity of
enhancement and degree of heterogeneity
in HCCs were determined on postcontrast
images using four-point scales from “vir-
tually no enhancement (0)” to “marked
enhancement (+3)” and from “virtually no
heterogeneity (0)” to “marked heterogene-
ity (+3)”, respectively.
The two radiologists independently
reviewed CT findings in a retrospective
manner. They evaluated the contrast-
enhanced CT with reference to unen-
hanced CT to determine the degree of
contrast enhancement in HCC and in the
surrounding liver using a four-point scale
from “virtually no enhancement (0)” to
“intense enhancement (+3)”. The radiolo-
gists evaluated the degree of heterogeneity
of hepatic arterial contrast enhancement in
204
HCC on CT hepatic arteriography using
a four-point scale from “homogeneous
enhancement (0)” to “strong heterogeneity
(+3)”. Statistical correlations were deter-
mined using simple regression analysis
for continuous data and the Spearman rank
correlation test for categorical data.
The VEGFINDs of HCCs and of the sur-
rounding liver ranged from 0.46 to 9.27
(mean ± SD, 3.28 ± 2.29) and from 0.44 to
5.91 (2.86 ± 1.49), respectively. In 14 (50%)
of the 28 HCCs, the VEGFIND of the sur-
rounding liver was greater than that of the
corresponding HCC. Pathologic tumor size
showed moderate direct correlation with
the VEGFINDs of HCCs (correlation coef-
ficient: 0.40, P = 0.035). No correlation was
found between tumor sizes and VEGFINDs
of the surrounding liver. The 28 HCCs were
predominantly well-differentiated in 3,
moderately differentiated in 20, and poorly
differentiated in 5. The mean VEGFIND of
HCCs tended to be higher with poorly dif-
ferentiated HCCs than with the other histo-
logic tumor grades, although no statistical
significance was found.
The contrast-to-noise ratios correlated
inversely with the VEGFINDs of HCCs on
opposed-phase T1-weighted gradient-echo
images (−0.46, P = 0.038), and correlated
directly with the VEGFINDs of HCCs on
T2-weighted echo-train spin-echo images
(0.49, P = 0.025) (Figure 14.2). The
contrast-to-noise ratios on gadolinium-
enhanced hepatic arterial-phase gradient-
echo images were marginally and inversely
correlated with the VEGFINDs of HCCs
(−0.39, P = 0.081) (Figure 14.2). The SD
ratios of HCCs correlated directly with the
VEGFINDs of HCCs on T2-weighted echo-
train spin-echo images (0.44, P = 0.044).
No correlation was found between phase-
shift indices of HCCs and VEGFINDs.
M. Kanematsu et al.
The qualitative degree of signal inten-
sity heterogeneity of HCCs correlated
directly with the VEGFINDs of HCCs on
opposed-phase T1-weighted gradient-echo
(0.64, P = 0.016), T2-weighted echo-train
spin-echo (0.52, P = 0.038), postcontrast
hepatic arterial-phase gradient-echo (0.48,
P = 0.045), and equilibrium-phase gradi-
ent-echo (0.56, P = 0.018) images.
On CT hepatic arteriography, the con-
trast enhancement indices of HCCs
showed moderate inverse correlation with
the VEGFINDs of HCCs (−0.57, P = 0.017)
(Figures 14.2 and 14.3). The contrast
enhancement indices of the surrounding
liver showed marginal, moderate direct
correlation with the VEGFINDs of the sur-
rounding liver (0.44, P = 0.076). On CT
during arterial portography, no significant
correlation was found between the con-
trast enhancement index and VEGFIND.
Qualitatively, the degree of heterogeneity
of hepatic arterial contrast enhancement
in HCCs on CT hepatic arteriography
showed moderate direct correlation with
the VEGFINDs of HCCs (0.55, P = 0.033).
Half of the 28 HCCs in our series had
a VEGFIND lower than in the surrounding
liver (Figure 14.3), which is comparable
to other reports that described 40–59% of
HCCs have VEGF activity that is either
equal to or less than that in the surround-
ing liver (Suzuki et al., 1996, El-Assal
et al., 1998). Some of these HCCs showed
considerably greater arterial enhancement
compared to the surrounding liver (Figure
14.3). This emphasizes that presence of
VEGF does not correlate with the pres-
ence of vascularity; rather it is a precursor
to vascularity. The biological interaction
between the VEGF activity in an HCC
and in the surrounding liver has yet to be
determined.
14. Expression of Vascular Endothelial Growth Factor in Hepatocellular Carcinoma 205

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.)

Results of correlation between VEGF employed in VEGF activity evaluation

expression in HCC tissues and tumor (Yamaguchi et al., 1998; Kwak et al.,
size have varied among researchers, partly 2001; Jeng et al., 2004; Ng et al., 2001).
this may reflect different methodologies Although we found a positive correlation
206 M. Kanematsu et al.

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
tumor size, Yamaguchi et al. (1998), who
immunohistochemically investigated the
expression of VEGF in HCCs, found that
VEGF positivity gradually decreased with
an increase in tumor size. It may be
that because the population they evalu-
ated included a greater number and per-
centage of small (15 mm or less) HCCs
than other reports, that they might have
observed VEGF activity changes during
the continuous dedifferentiation in small
HCCs. Kwak et al. (2001) also observed
that tumor size correlated negatively with
immunohistochemical VEGF expression.
While, Jeng et al. (2004) found that the
level of VEGF mRNA in HCC tissues did
not significantly correlate with tumor size,
cellular differentiation, capsule, daughter
nodules, vascular permeation, necrosis and
hemorrhage of tumors. Ng et al. (2001)
similarly reported that there was no sta-
tistically significant association between
tumor VEGF mRNA levels and tumor size
or other pathologic features. Nevertheless,
our opinion is that positive correlation
between VEGF expression and tumor
size is reasonable, because intratumoral
hypoxia or necrosis induced by tumor
growth may induce VEGF protein synthe-
sis (Shweiki et al., 1992; Mukhopadhyay
et al., 1995; von Marschall et al., 2001).
14. Expression of Vascular Endothelial Growth Factor in Hepatocellular Carcinoma
Hepatocellular carcinomas are often
hypo- to isovascular on radiologic images
during the early stage of development,
typically during the stage of small well-
differentiated HCC (Hayashi et al., 1999),
and intratumoral oxygen tension at this time
is believed to be low. Such a hypoxic state
may induce neovascularization in HCCs,
which would result in an increased supply
of arterial blood and concomitant oxygen
supply to the HCC, which then leads to the
development of a hypervascular HCC that
is typically referred to as a moderately to
poorly differentiated HCC. Once a hyper-
vascular HCC has developed, VEGF gene
expression may be suppressed due to the
downregulation of VEGF (Figure 14.1B).
This might explain prior observations that
the intensity of VEGF expression in HCCs
is related to the histological tumor grade,
i.e., highest in well-differentiated HCCs,
and lowest in poorly differentiated HCCs
(Yamaguchi et al., 1998).
In our study the HCC-to-liver contrast-
to-noise ratios on T1-weighted gradient-
echo images inversely correlated with the
VEGFINDs of HCCs, and contrast-to-noise
ratios on T2-weighted echo-train spin-
echo images directly correlated with the
VEGFINDs of HCCs, which indicates that
the stronger the VEGF expression is in
HCCs, the more prolonged are the T1-
and T2-relaxation times of HCCs (Figure
14.2). T1- and T2-relaxation time prolon-
gation commonly occurs in tissues con-
taining increased amounts of free water in
extracellular or interstitial spaces, typically
as seen in malignant tumors. Moreover, it
is known that VEGF, which was previ-
ously termed vascular permeability factor
(VPF), raises the permeability of blood
vessels (Shweiki et al., 1992; Monacci
et al., 1993). We suspect that increased
207
vascular permeability regulated by VEGF
peptides results in increased free water in
the extracellular or interstitial spaces of
HCCs and thus causes T1- and T2 relaxa-
tion time prolongation in HCCs.
The contrast-to-noise ratios on gado-
linium-enhanced hepatic arterial-phase
MRI were marginally and inversely cor-
related with the VEGFINDs, and the con-
trast enhancement indices of HCCs on
CT hepatic arteriography were moderately
inversely correlated with the VEGFINDs
of HCCs, which indicated that the more
hypervascular is an HCC, the weaker is the
VEGF expression in the tumor (Figures
14.2 and 14.3). It is known that an inef-
ficient vascular supply and the resultant
reduction in tissue oxygen tension, often
leads to neovascularization in order to
maintain the nutrition of tissue. In human
HCC, it has been suggested that hypoxia
induces an upregulation of VEGF gene
expression by biomolecular mechanism
(Figure 14.1A) (Shweiki et al., 1992), i.e.,
hypoxia inducible factor (HIF)-1α and
HIF-2α are upregulated by hypoxia and
induce pro-angiogenic peptide formation
and VEGF expression (Mukhopadhyay
et al., 1995).
Heterogeneity of HCC on CT or MRI
correlated directly with the VEGFINDs of
HCCs, which indicated that the more het-
erogeneous the tumor signal or enhance-
ment, the stronger the VEGF expression
(Figures 14.2 and 14.3). Heterogeneity
of HCC on unenhanced MRI may be
explained by an uneven distribution of
extracellular free water or more com-
monly by the presence of intratumoral
necrosis. Heterogeneity of HCC on gado-
linium-enhanced MRI and on CT hepatic
arteriography may also be indicative of
unevenness of vascularity which results
208
in regions of tumor with decreased oxy-
gen tension. Previous research indicated
that the hypoxic regions of solid tumors
produced powerful and directly acting
angiogenic proteins such as VEGF, or that,
in an experimental model using rat livers,
VEGF peptides were produced by non-
parenchymal and parenchymal cells after
necrosis (Mukhopadhyay et al., 1995).
Fat deposition is observed in various
types of HCCs. Kutami et al. (2000)
reported that fatty changes in small HCCs
are closely related to tumor size, histologic
grade, and the insufficient development
of arterial tumor vessels. Previous stud-
ies have demonstrated that lipid bodies
in endothelial cells are induced during
hypoxia in any cell type (Scarfo et al.,
2001). Based on these previous results,
we suspect that VEGF expression, hepatic
arterial perfusion, hypoxia, and fat deposi-
tion in HCCs are closely related. However,
in our study, we did not observe a cor-
relation between fat deposition of HCCs
and VEGF expression, where phase shift
imaging accurately reflects intratumoral
fat deposition.
The contrast enhancement index
of the surrounding liver on CT hepatic
arteriography showed moderate direct
correlation with the VEGFINDs of the sur-
rounding liver with marginal statistical
significance, which suggested that the
greater the hepatic arterial enhancement
of the surrounding liver, the stronger is the
VEGF expression in the liver. Rosmorduc
et al. (1999) described experimental bil-
iary cirrhosis, indicating that there was
evidence that angiogenesis was stimulated
primarily by VEGF in response to hepa-
tocellular hypoxia caused by liver dam-
age. El-Assal et al. (1998) reported that
cirrhotic livers had a significantly higher
M. Kanematsu et al.
microvascular density and VEGF expres-
sion than noncirrhotic livers. The reason
why in our study the correlations between
the intensity of hepatic arterial enhance-
ment and of VEGF expression differed
between HCCs and the surrounding liver
was unclear, this discrepancy in part may
be associated with the difference in mag-
nitude between hepatic arterial perfusion
in a HCC and in the surrounding liver: the
contrast enhancement indices of HCCs
ranged from 33 to 236 Hounsfield unit
(HU) (mean, 105.3 ± 51.7 HU), whereas
the contrast enhancement indices of the
surrounding liver ranged from 9 to 68 HU
(mean, 28.9 ± 15.7 HU). The increased
oxygen supply to the HCC secondary to
neovascularization was sufficiently great
to suppress VEGF expression via VEGF
downregulation (Figure 14.1B), but the
increased oxygen supply in the surround-
ing liver might not be sufficient to cause
VEGF downregulation (Figure 14.3).
The most contentious aspect of VEGF
correlation is the relationship with the con-
comitant presence of increased vascularity.
Kwak et al. (2001) correlated tumor atten-
uation qualitatively, and determined on
intravenously contrast-enhanced CT scans
with immunoreactivity for anti-VEGF
antibody, which was qualitatively deter-
mined by immunohistochemical staining.
They concluded that the degree of VEGF
expression in HCC was directly correlated
with the degree of contrast enhancement
during the hepatic arterial phase. The
exact reasons why their results differ from
ours (Kanematsu et al., 2004b, 2005) in
terms of the correlation between the inten-
sity of hepatic arterial tumor enhancement
and VEGF expression in HCC is unclear.
Eighteen (82%) of 22 patients had type-B
hepatitis and 1 (5%) had type-C hepatitis
14. Expression of Vascular Endothelial Growth Factor in Hepatocellular Carcinoma
in their study, whereas in our study only
6 (21%) of 28 patients had type-B hepa-
titis and 22 (79%) had type-C hepatitis.
This substantial difference in the patient
populations and in the underlying hepatic
disease might affect the results. The sensi-
tivity for enhancement and timing of data
acquisition of single detector row spiral
CT, and subjective ratings of contrast
enhancement on CT images and of VEGF
expression may also affect statistical cor-
relations.
Some researchers have reported that
VEGF activity does not correlate with
the vascularity of HCCs as determined
by conventional angiography (Suzuki
et al., 1996; El-Assal et al., 1998), while
others have reported that VEGF activity
correlates directly with the intensity of
tumor stain on angiography (Mise et al.,
1996). All these reports focused on the
clinical significance of VEGF expression
in HCC and in the surrounding liver, the
evaluation of VEGF activity by means
of VEGF mRNA quantification with
Northern blotting, and VEGF localiza-
tion by immunohistochemical staining,
while the radiological assessment of the
degree of tumor staining by angiogra-
phy was performed by subjective rat-
ings using a two- or three-point scale.
We believe that our study methodology
more accurately clarified the relation-
ship between the vascularity of HCCs
and liver and VEGF activity, as contrast
enhancement was quantitatively deter-
mined using region-of-interest measure-
ments and VEGF was determined using
Western blotting technique.
Limitations of our studies have included
that the study populations were small,
reflecting single institution accrual,
and distribution of varying types of
209
histologic tumor grades was uneven.
Multi-institutional studies will be needed
to confirm our results. Also, although
we employed Western blotting to semi-
quantify VEGF peptides, this technique is
limited in terms of its ability to differenti-
ate the multiple types of VEGF peptides
that exist in cell membranes, cytoplasm,
and interstitial spaces.
SUMMARY
We have reviewed the historical devel-
opment of various diagnostic imaging
modalities for the evaluation of HCC.
We further described the research pub-
lications on the correlation of VEGF
and imaging findings. The apparently
discrepant observations of the extent of
correlation of VEGF activity and HCC,
we believe, reflects that different studies
contained patients with HCCs of different
evolution. Based on our research studies,
we have observed that VEGF appears
to correlate with increased tissue fluid,
which on MRI appears as low signal on
T1-weighted images and high signal on
T2-weighted images. It is our impres-
sion that VEGF increases as a precursor
to increased vascularity, as vascularity
increases, seen in small HCCs, VEGF
is downregulated and as tumor increases
to larger size, > 3 cm, VEGF may again
increase in response to hypoxia.
The future clinical impact of VEGF
measurements and correlation of VEGF
values with imaging results have yet to
be determined. Our hope is that VEGF
measurements will provide insight into the
angiogenic potential of tumors, and also
predict or reflect response to therapeutic
interventions.
210
REFERENCES
El-Assal, O.N., Yamanoi, A., Soda, Y., Yamaguchi, M.,
Igarashi, M., Yamamoto, A., Nabika, T., and
Nagasue, N. 1998. Clinical significance of
microvessel density and vascular endothelial
growth factor expression in hepatocellular car-
cinoma and surrounding liver: possible involve-
ment of vascular endothelial growth factor in the
angiogenesis of cirrhotic liver. Hepatology 27:
1554–1562.
Folkman, J. 1971. Tumor angiogenesis: therapeutic
implications. N. Engl. J. Med. 285: 1182–1186.
Gwak, G.Y., Yoon, J.H., Kim, K.M., Lee, H.S.,
Chung, J.W., and Gores, G.J. 2005. Hypoxia
stimulates proliferation of human hepatoma cells
through the induction of hexokinase II expres-
sion. J. Hepatol. 42: 358–364.
Hayashi, M., Matsui, O., Ueda, K., Kawamori, Y.,
Kadoya, M., Yoshikawa, J., Gabata, T.,
Takashima, T., Nonomura, A., and Nakanuma,
Y. 1999. Correlation between the blood supply
and grade of malignancy of hepatocellular nod-
ules associated with liver cirrhosis: evaluation
by CT during intraarterial injection of contrast
medium. Am. J. Roentgenol. 172: 969–976.
International Working Party. 1995. Terminology of
nodular hepatocellular lesions. Hepatology 22:
983–993.
Jeng, K.S., Sheen, I.S., Wang, Y.C., Gu, S.L., Chu,
C.M., Shih, S.C., Wang, P.C., Chang, W.H., and
Wang, H.Y. 2004. Is the vascular endothelial
growth factor messenger RNA expression in
resectable hepatocellular carcinoma of prognos-
tic value after resection? World J. Gastroenterol.
10: 676–681.
Kanematsu, M., Hoshi, H., Imaeda, T., Murakami, T.,
Inaba, Y., Yokoyama, R., and Nakamura, H.
1997. Detection and characterization of hepatic
tumors: value of combined helical CT hepatic
arteriography and CT during arterial portogra-
phy. Am. J. Roentgenol. 168: 1193–1198.
Kanematsu, M., Semelka, R.C., Leonardou, P.,
Mastropasqua, M., Armao, D., Vaidean, G.,
Firat, Z., and Woosley, J.T. 2004a. Angiogenesis
in hepatocellular nodules: correlation of MR
imaging and vascular endothelial growth factor.
J. Magn. Reson. Imaging 20: 426–434.
Kanematsu, M., Osada, S., Amaoka, N., Goshima, S.,
Kondo, H., Nishibori, H., Kato, H., Matsuo, M.,
M. Kanematsu et al.
Yokoyama, R., Hoshi, H., and Moriyama, N.
2004b. Expression of vascular endothelial
growth factor in hepatocellular carcinoma and
the surrounding liver: correlation with angi-
ographically assisted CT. Am. J. Roentgenol.
183: 1585–1593.
Kanematsu, M., Osada, S., Amaoka, N., Goshima, S.,
Kondo, H., Kato, H., Nishibori, H., Yokoyama,
R., Hoshi, H., and Moriyama, N. 2005.
Expression of vascular endothelial growth factor
in hepatocellular carcinoma and the surrounding
liver and correlation with MRI findings. Am. J.
Roentgenol. 184: 832–841.
Kutami, R., Nakashima, Y., Nakashima, O., Shiota, K.,
and Kojiro, M. 2000. Pathomorphologic study
on the mechanism of fatty change in small hepa-
tocellular carcinoma of humans. J. Hepatol. 33:
282–289.
Kwak, B.K., Shim, H.J., Park, E.S., Kim, S.A.,
Choi, D., Lim, H.K., Park, C.K., Chung, J.W.,
and Park, J.H. 2001. Hepatocellular carcinoma:
correlation between vascular endothelial growth
factor level and degree of enhancement by mul-
tiphase contrast-enhanced computed tomogra-
phy. Invest. Radiol. 36: 487–492.
Lim, J.H., Kim, E.Y., Lee, W.J., Lim, H.K., Do, Y.S.,
Choo, I.W., and Park, C.K. 1999. Regenerative
nodules in liver cirrhosis: findings at CT during
arterial portography and CT hepatic arteriogra-
phy with histopathologic correlation. Radiology
210: 451–458.
Mise, M., Arii, S., Higashituji, H., Furutani, M.,
Niwano, M., Harada, T., Ishigami, S., Toda,
Y., Nakayama, H., Fukumoto, M., Fujita, J.,
and Imamura, M. 1996. Clinical significance
of vascular endothelial growth factor and basic
fibroblast growth factor gene expression in liver
tumors. Hepatology 23: 455–464.
Monacci, W.T., Merrill, M.J., and Oldfield, E.H.
1993. Expression of vascular permeability fac-
tor/vascular endothelial factor in normal rat tis-
sues. Am. J. Physiol. 264: 995–1002.
Morgan, B., Thomas, A.L., Drevs, J., Hennig, J.,
Buchert, M., Jivan, A., Horsfield, M.A., Mross,
K., Ball, H.A., Lee, L., Mietlowski, W., Fuxuis, S.,
Unger, C., O’Byrne, K., Henry, A., Cherryman,
G.R., Laurent, D., Dugan, M., Marme, D., and
Steward, W.P. 2003. Dynamic contrast-enhanced
magnetic resonance imaging as a biomarker for
the pharmacological response of PTK787/ZK
14. Expression of Vascular Endothelial Growth Factor in Hepatocellular Carcinoma
222584, an inhibitor of the vascular endothe-
lial growth factor receptor tyrosine kinases, in
patients with advanced colorectal cancer and
liver metastases: results from two phase I stud-
ies. J. Clin. Oncol. 21: 3955–3964.
Mukhopadhyay, D., Tsiokas, L., Zhou, X.M.,
Foster, D., Brugge, J.S., and Sukhatme, V.P.
1995. Hypoxic induction of human vascular
endothelial growth factor expression through
c-Src activation. Nature 375: 577–581.
Murakami, T., Kim, T., Takamura, M., Hori, M.,
Takahashi, S., Federle, M.P., Tsuda, K., Osuga,
K., Kawata, S., Nakamura, H., and Kudo, M.
2001. Hypervascular hepatocellular carcinoma:
detection with double arterial phase multi-detector
row helical CT. Radiology 218: 763–767.
Ng, I.O., Poon, R.T., Lee, J.M., Fan, S.T., Ng, M.,
and Tso, W.K. 2001. Microvessel density, vas-
cular endothelial growth factor and its receptors
Flt-1 and Flk-1/KDR in hepatocellular carci-
noma. Am. J. Clin. Pathol. 116: 838–845.
Rosmorduc, O., Wendum, D., Corpechot, C.,
Galy, B., Sebbagh, N., Raleigh, J., Housset, C.,
and Poupon, R. 1999. Hepatocellular hypoxia-
induced vascular endothelial growth factor
expression and angiogenesis in experimental bil-
iary cirrhosis. Am. J. Pathol. 155: 1065–1073.
Scarfo, L.M., Weller, P.F., and Farber, H.W. 2001.
Induction of endothelial cell cytoplasmic lipid
bodies during hypoxia. Am. J. Physiol. Heart
Circ. Physiol. 280: 294–301.
Semelka, R.C., Shoenut, J.P., Kroeker, M.A.,
Greenberg, H.M., Simm, F.C., Minuk, G.Y.,
Kroeker, R.M., and Micflikier, A.B. 1992. Focal
liver disease: comparison of dynamic contrast-
enhanced CT and T2-weighted fat-suppressed,
FLASH, and dynamic gadolinium-enhanced MR
imaging at 1.5 T. Radiology 184: 687–694.
Shimizu, A., Ito, K., Koike, S., Fujita, T., Shimizu, K.,
and Matsunaga, N. 2003. Cirrhosis or chronic
hepatitis: evaluation of small (< or = 2-cm) early-
enhancing hepatic lesions with serial contrast-
enhanced dynamic MR imaging. Radiology 226:
550–555.
211
Shweiki, D., Itin, A., Soffer, D., and Keshet, E.
1992. Vascular endothelial growth factor induced
by hypoxia may mediate hypoxia-initiated ang-
iogenesis. Nature 359: 843–845.
Suzuki, K., Hayashi, N., Miyamoto, Y., Yamamoto, M.,
Ohkawa, K., Ito, Y., Sasaki, Y., Yamaguchi, Y.,
Nakase, H., Noda, K., Enomoto, N., Arai, K.,
Yamada, Y., Yoshihara, H., Tujimura, T.,
Kawano, K., Yoshikawa, K., and Kamada, T.
1996. Expression of vascular permeability fac-
tor/vascular endothelial growth factor in human
hepatocellular carcinoma. Cancer Res. 56:
3004–3009.
Ueda, K., Matsui, O., Kawamori, Y., Nakanuma,
Y., Kadoya, M., Yoshikawa, J., Gabata, T.,
Nonomura, A., and Takashima, T. 1998.
Hypervascular hepatocellular carcinoma: evalu-
ation of hemodynamics with dynamic CT during
hepatic arteriography. Radiology 206: 161–166.
Viglietto, G., Maglione, D., Rambaldi, M., Cerutti,
J., Romano, A., Trapasso, F., Fedele, M.,
Ippolito, P., Chiappetta, G., and Botti, G. 1995.
Upregulation of vascular endothelial growth
factor (VEGF) and downregulation of placenta
growth factor (PlGF) associated with malig-
nancy in human thyroid tumors and cell lines.
Oncogene 11: 1569–1579.
von Marschall, Z., Cramer, T., Hocker, M.,
Finkenzeller, G., Wiedenmann, B., and Rosewicz S.
2001. Dual mechanism of vascular endothelial
growth factor upregulation by hypoxia in human
hepatocellular carcinoma. Gut 48: 87–96.
Yamaguchi, R., Yano, H., Iemura, A., Ogasawara,
S., Haramaki, M., and Kojiro, M. 1998.
Expression of vascular endothelial growth factor
in human hepatocellular carcinoma. Hepatology
28: 68–77.
Yamane, A., Seetharam, L., Yamaguchi, S., Gotoh, N.,
Takahashi, T., Neufeld, G., and Shibuya, M.
1994. A new communication system between
hepatocytes and sinusoidal endothelial cells in
liver through vascular endothelial growth factor
and Flt tyrosine kinase receptor family (Flt-1
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
The liver is a common site for metastases
from colorectal carcinoma and other neo-
plasms. Most colorectal carcinoma deaths
are attributable to hepatic metastases,
and the number of these metastases is
increasing (Ballantyne and Quin, 1993).
The prognosis of patients without treat-
ment is dismal because most cases die
within a few years after discovering the
liver lesion. Hepatic resection is the sole
treatment that regularly leads to long-
term survival with a possible chance of
cure in patients with hepatic metastases,
especially those of colorectal origin.
With techniques such as intraoperative
ultrasonography (Makuuchi et al., 1991),
which improves the management of peri-
operative patients, hepatic resection has
extended the possibility of liver surgery in
patients with advanced metastatic tumors
(Kawasaki et al., 1994). Also, hepatic
resection has been proved to improve
the prognosis of patients (Penna and
Nordlinger, 2002). Therefore, accurately
detecting small lesions is indispensable
in preoperative imaging assessment when
planning hepatic resection.
SUPERPARAMAGNETIC
IRON OXIDE-ENHANCED
MAGNETIC RESONANCE
IMAGING
There are several methods such as com-
puted tomography (CT), magnetic reso-
nance imaging (MRI), and ultrasonography
for preoperative imaging assessment in
detecting liver metastases. Presently, super-
paramagnetic iron oxide (SPIO)-enhanced
MRI is being widely used to detect liver
metastases. This oxide is taken up by
phagocytosis of the reticuloendothelial
system, including Kupffer cells of the liver,
substantially shortening T2 of liver tissue
(Stark et al., 1988). The hepatic metastases
that contain negligible or a few Kupffer
cells remain largely nonenhanced, while
the normal liver is enhanced and shows
a low signal intensity on T2-weighted
images, with the result that the lesion-to-
liver signal intensity ratio is improved on
SPIO-enhanced MRI, as compared with
nonenhanced T2-weighted images.
In a multicenter trial, SPIO-enhanced MRI
revealed additional lesions not seen on
unenhanced images in 27% of cases and
additional lesions not seen by conventional
213
214
CT in 40% (Ros et al., 1995). In a mul-
tiobserver study that compared the sen-
sitivities of dual phase spiral CT with
SPIO-enhanced MRI for detecting metas-
tases, both techniques showed high sen-
sitivities for detecting lesions larger than
1 cm at 94% vs. 99%, whereas the sen-
sitivities of dual phase spiral CT and
SPIO-enhanced MRI for detecting lesions
smaller than 1 cm were 52% and 47%,
respectively (Ward et al., 1999). Computed
tomography during arterial portography
(CTAP) has been regarded as a technique
that is more sensitive at detecting small
lesions than dual phase spiral CT and
SPIO-enhanced MRI, but CTAP is inva-
sive and has a greater risk of providing
false-positive results than other techniques
(Soyer et al., 1993). CTAP is also insensi-
tive to lesions smaller than 1 cm (Young
et al., 1997). Therefore, the strategy to
improve the detection rate of small lesions
is anticipated.
SPIO-enhanced MRI increases the rate
of false-positive findings compared to
spiral-CT (Muller et al., 1998), and these
false-positive results are responsible for the
difficulty in distinguishing small lesions.
One of the common false-positive lesions is
a small hepatic cyst (Ward et al., 1999). Some
techniques are recommended for detecting
hepatic cysts, such as T1-weighted breath-
hold GRE imaging after SPIO enhance-
ment and short- and long-TE SSFSE
(Oudkerk et al., 1997; Kiryu et al., 2002).
Adding these techniques to routine SPIO-
enhanced MRI, hepatic cysts can be distin-
guished from true hepatic lesions. Another
cause for false-positives is the signals
from hepatic vessels. SPIO-enhanced MRI
increases the signal of cross-sectioned
vessels in the low signal background liver
S. Kiryu and K. Ohtomo
parenchyma; therefore, the small vessels
and lesions exhibit similar signal intensities.
Furthermore, SPIO-enhanced MRI dis-
played more peripheral vascular structures
than contrast enhanced CT (Fretz et al.,
1990). Distinguishing small lesions from
the signal intensities from hepatic vessels
is the key to detect small hepatic lesions.
DIFFUSION-WEIGHTED
IMAGING FOR SUPPRESSION
OF SIGNALS FROM HEPATIC
VESSELS
Diffusion-weighted imaging (DWI) is
sensitive to molecular diffusion, which
is the thermally induced motion of water
molecules in biological tissues called
Brownian motion. The changes in proton
self-diffusion are an early indicator of
alterations of cellular homeostasis in acute
ischemic stroke (Moseley et al., 1990);
therefore, DWI has a significant impact
in the detection of early ischemic change
and has played an important role in the
clinical setting in the field of neuroradiol-
ogy. DWI is able to suppress the signals
of vessels and has been used to improve
the sensitivity of the detection of small
hepatic lesions (Okada et al., 1998).
In the DWI pulse sequence, motion-
probing gradient (MPG) pulses are
placed before and after a 180° RF pulse,
which comprises the spin echo method
along with a 90° pulse. The protons
that have moved during or after the
first MPG randomly change their spin
phase as they move from one region
of the magnetic field to another. Thus,
the signal intensity of diffusion protons
15. Detection of Small Hepatic Lesions: Superparamagnetic Oxide-Enhanced
decreases whereas the slow diffusion
appears as a retention of signal intensity
(Mosley and Butts, 1999).
The strength of MPG is represented by
the b-value (measured in s/mm2). Using a
strong MPG with a high b-value provides
images contrasted by proton diffusion,
whereas via DWI pulse a small MPG
with a small b-value provides images
contrasted by perfusion, mainly micro-
circulation of blood. Using this feature,
Okada et al. (1998) reported the use-
fulness of MPG with small a b-value
in the suppression of signals from ves-
sels. Patients with focal liver lesions were
evaluated with DWI with a small b-value.
They found that the signals from hepatic
vessels were suppressed and the con-
spicuity of small lesions was improved.
The lesion-to-liver signal intensity ratio of
DWI images was significantly better than
that of T2-weighted fast spin echo images.
However, some problems were mentioned
and these were mainly due to the EPI tech-
nique, which was combined with MPG in
DWI and widely used in most DWI stud-
ies in clinical settings. EPI samples all the
data necessary for reconstruction of an
image, and images can be acquired within
a second. EPI utilizes gradient rephrasing,
which results in spatial distortions from
magnetic susceptibility effects, especially
at the diaphragm near the air interface. The
spatial resolution is also limited (Farzaneh
et al., 1990). These are inevitable prob-
lems associated with the EPI technique
and are disadvantageous to detect small
hepatic lesions. To overcome these prob-
lems, MPG should be combined with other
suitable pulse sequences, which provide
high spatial resolution images with less
susceptibility-induced distortion.
215
PERIODICALLY ROTATED
OVERLAPPING PARALLEL
LINES WITH ENHANCED
RECONSTRUCTION
TECHNIQUE
Fast spin echo (FSE), characterized by a
series of 180° rephasing pulses and multiple
echoes, is a rapid scanning technique com-
pared with conventional T2-weighted spin
echo imaging. Conventional T2-weighted
spin echo imaging requires quite a long
scanning time (at least 5 min), whereas data
acquisition with FSE can be completed
within a few minutes or less. Moreover,
the breathhold FSE technique shows bet-
ter lesion conspicuity than conventional
T2-weighted imaging on liver imaging
(Gaa et al., 1996). FSE can be com-
bined with MPG, and FSE DWI has far
less B0-related artifacts (e.g., from metal,
sinuses, and eddy currents) and can be
implemented with a less intensive gradient
requirement than EPI. However, FSE DWU
faces a major challenge because the signal
will not meet Carr-Purcell-Meiboom-Gill
conditions, leading to unstable echoes
(Bammer et al., 2002)
Periodically Rotated Overlapping
ParallEL Lines with Enhanced Recon-
struction (PROPELLER) is a radial scan-
ning variation based on the FSE. MRI data
are collected in a series of rotating blades
containing phase encoding lines, with each
blade sampling a common central k-space
data set, providing most of the image’s sig-
nal intensity and determining gross tissue
contrast throughout the image (Pipe, 1999)
(Figure 15.1). PROPELLER has inherent
2D navigator information in each FSE echo
train, and the basic idea of PROPELLER is
216
Figure 15.1. Illustration of PROPELLER k-space
acquisition. MRI data are collected in a series
of rotating blades containing phase encoding
lines, with each blade sampling a common central
k-space data set
that this oversampled region in the center
of the k-space can be compared between
blades to correct for inconsistencies prior
to combining the data. PROPELLER pro-
vides far greater immunity against geomet-
ric distortion than that obtained with EPI
sequences. It also mitigates signal instabil-
ity, present in DWI FSE (Bammer, 2003).
Diffusion-weighted PROPELLER has been
introduced and successful in high-resolution
diffusion-weighted tensor imaging of the
brain (Pipe et al., 2002).
SPIO-ENHANCED DWI T2 FSE
IMAGING USING PROPELLER
Based on the hypothesis that SPIO-
enhanced diffusion-weighted PROPEL-
LER MRI (SPIO DWI PROPELLER)
of the liver has advantages in terms of
detecting small hepatic metastases, the
S. Kiryu and K. Ohtomo
efficacy of SPIO DWI PROPELLER in
the detection of hepatic metastases was
evaluated in a recent study (Kiryu et
al., 2006). Fourteen patients with hepatic
metastases were examined with a 1.5 T
MR scanner. After the infusion of SPIO
(Ferumoxides), SPIO-enhanced FSE
(SPIO FSE) and SPIO DWI PROPELLER
were acquired with a respiratory trigger
using a body coil. The b-value of MPG for
SPIO DWI PROPELLER was 10 s/mm2
to suppress signals from vessels. Hepatic
resection was carried out on all patients,
and all resections were performed with an
ultrasound-guided procedure. Among the
38 metastases 1 cm or larger, 37 (97.4%)
were detected with SPIO FSE and 36
(94.7%) were detected with SPIO DWI
PROPELLER. There was no significant
difference between both pulse sequences.
Among the 30 metastases smaller than
1 cm, 16 (53.3%) were detected with SPIO
FSE and 24 (80.0%) were detected with
SPIO DWI PROPELLER. The detectability
of metastases by SPIO DWI PROPELLER
was significantly higher than that of SPIO
FSE. As expected, the small hepatic metas-
tases near to the hepatic vessels were dif-
ficult to detect by SPIO FSE. SPIO DWI
PROPELLER provided images with a
high spatial resolution and the signals
from the vessels were suppressed by MPG
(Figure 15.2).
Although a respiratory trigger was
applied, motion artifacts were noted on
SPIO FSE in some cases whereas no
motion artifacts were found on the SPIO
DWI (Figure 15.3).
Patient movement also introduces
ghosting, blurring, and the appearance
of mottled bands parallel to the phase-
encoding direction of the image (Bellon
et al., 1986). As Pipe (1999) showed that
15. Detection of Small Hepatic Lesions: Superparamagnetic Oxide-Enhanced 217

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

PROPELLER MRI was useful for cor- artifacts is an advantage of PROPELLER

recting bulk motion in head images and
respiratory motion in nongated cardiac
images, such artifacts were not found on
SPIO DWI PROPELLER. Motion artifacts
have a risk of obscuring small hepatic
lesions; therefore, the correction of motion
in detecting such small lesions.
On PROPELLER images, a streak arti-
fact occasionally appears around structures
with high signal intensity (Pipe, 1999). The
bright signal, which moved with cardiac
pulsation, produced a spatially and rapidly
218
varying phase, which could not be cor-
rected by the PROPELLER method. Large
high signal intensity structures caused a
severe streak artifact, and small lesions
near these structures may be obscured.
This is a pitfall of PROPELLER imaging.
It is concluded that SPIO-enhanced MRI
is widely used to detect hepatic metastases;
however, a problem arises in the detection
of small lesions. The signal from hepatic
vessels mimicking small lesions causes
false-positive lesions in SPIO-enhanced
MRI. MPG with a small b-value is a fea-
sible technique to suppress such signals
from hepatic vessels. The combination of
MPG and an appropriate pulse sequence,
such as PROPELLER, provides high spa-
tial resolution SPIO images and is useful
for detecting small hepatic lesions.
REFERENCES
Ballantyne, G.H., and Quin, J. 1993. Surgical treat-
ment of liver metastases in patients with colorectal
cancer. Cancer 71: 4252–4266.
Bammer, R. 2003. Basic principles of diffusion-
weighted imaging. Eur. J. Radiol. 45: 169–184.
Bammer, R., Augustin, M., Prokesch, R.W.,
Stollberger, R., and Fazekas, F. 2002. Diffusion-
weighted imaging of the spinal cord: interleaved
echo-planar imaging is superior to fast spin-
echo. J. Magn. Reson. Imaging 15: 364–373.
Bellon, E.M., Haacke, E.M., Coleman, P.E., Sacco,
D.C., Steiger, D.A., and Gangarosa, R.E. 1986.
MR artifacts: a review. Am. J. Roentgenol. 147:
1271–1281.
Farzaneh, F., Riederer, S.J., and Pelc, N.J. 1990.
Analysis of T2 limitations and off-resonance
effects on spatial resolution and artifacts in
echo-planar imaging. Magn. Reson. Med. 14:
123–139.
Fretz, C.J., Stark, D.D., Metz, C.E., Elizondo, G.,
Weissleder, R., Shen, J.H., Wittenberg, J., Simeone, J.,
and Ferrucci, J.T. 1990. Detection of hepatic
metastases: comparison of contrast-enhanced CT,
unenhanced MR imaging, and iron oxide-enhanced
MR imaging. Am. J. Roentgenol. 155: 763–770.
S. Kiryu and K. Ohtomo
Gaa, J., Hatabu, H., Jenkins, R.L., Finn, J.P., and
Edelman, R.R. 1996. Liver masses: replace-
ment of conventional T2-weighted spin-echo
MR imaging with breath-hold MR imaging.
Radiology 200: 459–464.
Kawasaki,S.,Makuuchi,M.,Kakazu, T., Miyagawa, S.,
Takayama, T., Kosuge, T., Sugihara, K., and
Moriya, Y. 1994. Resection for multiple meta-
static liver tumors after portal embolization.
Surgery 115: 674–677.
Kiryu, S., Okada, Y., and Ohtomo, K. 2002.
Differentiation between hemangiomas and cysts
of the liver with single-shot fast-spin echo image
using short and long TE. J. Comput. Assist.
Tomogr. 26: 687–690.
Kiryu, S., Watanabe, M., Kabasawa, H., Akahane, M.,
Aoki, S., and Ohtomo, K. 2006. Evaluation of
super paramagnetic iron oxide-enhanced diffu-
sion-weighted PROPELLER T2-fast spin echo
magnetic resonance imaging: preliminary expe-
rience. J. Comput. Assist. Tomogr. 30: 197–200.
Makuuchi, M., Takayama, T., Kosuge, T.,
Yamazaki, S., Yamamoto, J., Hasegawa, H., and
Takayasu, K. 1991. The value of ultrasonogra-
phy for hepatic surgery. Hepatogastroenterology
38: 64–70.
Mosley, M.E., and Butts, K. 1999. Diffusion and
perfusion. In: Stark, D. D., Bradley, W. G., Jr.,
eds. Magnetic Resonance Imaging. St. Louis,
MO: Mosby, pp1515–1538.
Moseley, M.E., Cohen, Y., Mintorovitch, J.,
Chileuitt, L., Shimizu, H., Kucharczyk, J.,
Wendland, M.F., and Weinstein, P.R. 1990.
Early detection of regional cerebral ischemia in
cats: comparison of diffusion- and T2-weighted
MRI and spectroscopy. Magn. Reson. Med. 14:
330–346.
Muller, R.D., Vogel, K., Neumann, K., Hirche, H.,
Barkhausen, J., Stoblen, F., Henrich, H., and
Langer, R. 1998. MRI with supermagnetic iron
particles versus double-spiral CT in identifi-
cation of malignant liver lesions. Rofo 168:
436–443.
Okada, Y., Ohtomo, K., Kiryu, S., and Sasaki, Y.
1998. Breath-hold T2-weighted MRI of hepatic
tumors: value of echo planar imaging with
diffusion-sensitizing gradient. J. Comput. Assist.
Tomogr. 22: 364–371.
Oudkerk, M., van den Heuvel, A.G., Wielopolski,
P.A., Schmitz, P.I., Borel Rinkes, I.H., and
15. Detection of Small Hepatic Lesions: Superparamagnetic Oxide-Enhanced
Wiggers, T. 1997. Hepatic lesions: detection
with ferumoxide-enhanced T1-weighted MR
imaging. Radiology 203: 449–456.
Penna, C., and Nordlinger, B. 2002. Colorectal
metastasis (liver and lung). Surg. Clin. North.
Am. 82: 1075–1090, x–xi.
Pipe, J.G. 1999. Motion correction with
PROPELLER MRI: application to head motion
and free-breathing cardiac imaging. Magn. Reson.
Med. 42: 963–969.
Pipe, J.G., Farthing, V.G., and Forbes, K.P.
2002. Multishot diffusion-weighted FSE using
PROPELLER MRI. Magn. Reson. Med. 47:
42–52.
Ros, P.R., Freeny, P.C., Harms, S.E., Seltzer, S.E.,
Davis, P.L., Chan, T.W., Stillman, A.E., Muroff,
L.R., Runge, V.M., Nissenbaum, M.A., et al.
1995. Hepatic MR imaging with ferumoxides:
a multicenter clinical trial of the safety and
efficacy in the detection of focal hepatic lesions.
Radiology 196: 481–488.
219
Soyer, P., Lacheheb, D., and Levesque, M. 1993.
False-positive CT portography: correlation with
pathologic findings. Am. J. Roentgenol. 160:
285–289.
Stark, D.D., Weissleder, R., Elizondo, G., Hahn, P.F.,
Saini, S., Todd, L.E., Wittenberg, J., and Ferrucci,
J.T. 1988. Superparamagnetic iron oxide: clinical
application as a contrast agent for MR imaging of
the liver. Radiology 168: 297–301.
Ward, J., Naik, K.S., Guthrie, J.A., Wilson, D.,
and Robinson, P.J. 1999. Hepatic lesion detec-
tion: comparison of MR imaging after the
administration of superparamagnetic iron oxide
with dual-phase CT by using alternative-free
response receiver operating characteristic analy-
sis. Radiology 210: 459–466.
Young, N., Sing, T., Wong, K.P., Hollands, M., and
Tait, N. 1997. Use of spiral and non-spiral com-
puted tomography arterial portography in the
detection of potentially malignant liver masses.
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
INTRODUCTION
Diagnostic imaging plays an important
role in the diagnosis of early-stage hepato-
cellular carcinoma (HCC); with contrast-
enhanced computed tomography (CT) and
magnetic resonance (MR) imaging exami-
nation being especially important for the
precise and objective evaluation of HCC.
At present, both imaging techniques are
widely used for diagnosis of HCC. For
the CT examination, a contrast-enhanced
dynamic study consisting of an arterial
phase and a portal venous phase is useful
for detecting HCC, while multidetector-
row helical CT (MDCT) scanner offers
major advantages for the evaluation of
HCC due to its faster acquisition capabil-
ity and higher spatial resolution.
For contrast-enhanced MR imaging,
a dynamic study using gadolinium-
based contrast material is also useful for
evaluating HCC. In addition, a tissue-
specific contrast material such as super-
paramagnetic iron oxide (SPIO), which
is available only for MR imaging is a
valuable aid for the diagnosis of HCC.
This chapter presents an overview of
current multidetector-row helical com-
puted tomography (MDCT) technique
including the underlying principles of
contrast injection and MR techniques,
including recent advanced technology
for the diagnosis of HCC.
MULTIDETECTOR-ROW
HELICAL COMPUTED
TOMOGRAPHY
Multidetector-row helical computed tom-
ography scanners were introduced to clinical
practice in 1998 and currently are in use in
many clinical institutes. The MDCT scanner
has an X-ray detector divided into multiple
rows along the longitudinal axis direction
of patient’s body, and can simultaneously
acquire imaging data sets from multiple
rows. The number of the rows from which
imaging data can be acquired simultane-
ously depends on the number of channels
of the scanner’s data acquisition systems.
As the number of channels increases, the
scanner can acquire imaging data sets faster
and provide images with thinner slice thick-
nesses and shorter scanning times.
221
222
MULTIDETECTOR-ROW
HELICAL COMPUTED
TOMOGRAPHY SCANNING
TECHNIQUE
The selection of scanning parameters
depends largely on the number of chan-
nels of the data acquisition system. At our
institute, two 64-channel MDCT scanners
and two 4-channel MDCT scanners were
in operation at the end of 2005. For heli-
cal scanning with a 64-channel MDCT,
a detector configuration of 0.625 mm ×
64 = 40 mm (LightSpeed VCT64; GE
Healthcare, Milwaukee, WI) or 0.5 mm ×
64 = 32 mm (Aquilion 64; Toshiba, Nasu,
Tochigi) is invariably used because the
number of channels of the data acquisition
system is equal to the number of detector
rows in the 64-channel MDCT scanner.
For the helical pitch, GE Medical Systems
scanners offer a choice of 0.6, 0.984, 1.375
or 1.75 and Toshiba scanners provide a
choice of 0.641, 0.828 and 1.484. Because
the scanning speed of the 64-channel
MDCT scanner is sufficiently fast, an
extremely large helical pitch may not be
required. The helical pitch from 0.8 to
1.5 is considered appropriate for contrast-
enhanced dynamic studies of the upper
abdomen, while a helical pitch of 1.375
(GE) or 0.828 (Toshiba) is employed for
the abdominal region. As for the rotation
speed of the X-ray tube, a lower rotation
speed of 0.5 or 0.6 s per rotation may be
used for the abdominal region, so that the
scanning period of the upper abdomen
ranges from 2 to 5 s.
For the X-ray tube current, the automatic
tube current modulation technique should
be employed in order to optimize the
radiation dose (Kalra et al., 2004). Based
on a single scout scan, the system adjusts
H. Onishi et al.
the tube current to attain an acceptable
noise level across the region of interest
(3D Auto mA; GE Healthcare, or Real
EC; Toshiba,). For the X-ray tube voltage,
120 kVp is invariably employed.
CONTRAST ADMINISTRATION
FOR THE DYNAMIC
MULTIDETECTOR-ROW
HELICAL COMPUTED
TOMOGRAPHY STUDY
Contrast enhancement is essential for the
evaluation of various diseases, especially
tumorous ones. In addition, a dynamic
study that consists of an arterial phase
and a portal venous phase after contrast
enhancement has proven to be useful for
the evaluation of tumorous diseases of
the liver, especially hypervascular tumors
such as HCC (Murakami et al., 2001). Our
basic dynamic CT scanning procedure for
the evaluation of HCC using 64-channel
MDCT scanners consists of precontrast
images, late arterial phase images and
portal venous phase images. Early arterial
and/or equilibrium phase images can be
added as an option depending on the clinical
requirements. For an effective dynamic
study, however, an optimal contrast injection
protocol is most important.
Administration Dose of Contrast Material
Heiken et al. (1995) and Brink et al. (1995)
reported that a maximum hepatic enhance-
ment of at least 50 Hounsfield Units (HU) is
needed to perform high-diagnostic-quality
hepatic CT and that 521 mgI/kg of iodine
is necessary to achieve this level of hepatic
enhancement. This dose corresponds to
~ 1.73 ml/kg of contrast material with
16. Diagnosis of Hepatocellular Carcinoma: Multidetector-Row Computed Tomography
an iodine concentration of 300 mgI/ml.
Yamashita et al. (2000) reported that for
abdominal helical CT the dose of intravenous
contrast medium should be 2.0–2.5 ml/kg
of body weight with an iodine concentration
of 300 mgI/ml.
Iodine Concentration of Contrast Material
Iodine concentration of contrast materials
for clinical CT examinations is available
at 300, 320, 350 and 370 mgI/ml. Awai
et al. (2002) reported that the high iodine
concentration of 370 mgI/ml was more
effective than the moderate iodine concen-
tration of 300 mgI/ml for the detection of
hypervascular HCC when the same injec-
tion flow rate and iodine dose were used,
both of which were adapted to the patient’s
body weight.
Injection Flow Rate and Duration of
Contrast Material
When the administered dose is adapted
to the patient’s weight, the injection flow
rate and/or injection duration will vary.
There are two mainstream approaches
to determine these parameters, the fixed
injection duration method, which involves
adjustment of the injection flow rate and
the fixed injection rate method, in which
the injection duration is adjusted. Awai
et al. (2004) reported that the fixed injec-
tion duration method is more effective
for determining the time needed to reach
peak aortic enhancement after initiation
of the contrast injection in the abdominal
aorta because this time span is equal to
the duration of the injection time (Bae,
2003). With the fixed injection duration
method, constant scan delay after contrast
arrival can be employed for each phase in
a dynamic study. When the fixed injec-
223
tion rate method is employed, however,
the scan delay after the contrast arrival
should be varied for each phase accord-
ing to the injection duration. Although
a high injection rate with short injection
duration is considered to be advantageous
for obtaining strong enhancement of the
arteries or hypervascular lesions in arterial
phase imaging, it may increase the risk of
extravasation or other complications. A
fixed injection duration of 25–30 s, result-
ing in an injection flow rate of 4–5 ml/s, is
therefore generally employed for contrast-
enhanced dynamic studies.
At our institute, contrast material at
2.0 ml/kg of 300 mgI/ml or 1.7 ml/kg of
370 mgI/ml is injected intravenously by
power injectors equipped with a double
injection drive system (Dual Shot; Nemoto
Kyorindo Co., Ltd., Tokyo, Japan) in 25 s
through a 20-gauge plastic catheter placed
in an antecuvital vein. However, the limi-
tation for the contrast amount is 150 ml
and that for the injection flow is 5 ml/s.
OPTIMAL SCANNING DELAY
There are two ways to determine scan
timing. One uses a fixed scan delay and
the other a scan delay adjusted to the
patient’s parameters by means of a test
bolus method or an automatic bolus track-
ing system. The contrast arrival time var-
ies even with normal cardiac function,
and it can be protracted to 30 s or more
in patients with heart failure. These inter-
patient differences cannot be ignored for
a good dynamic study because the images
acquired with incorrect timing can affect
the clinical diagnosis, especially in terms
of arterial phase imaging. Although the
fixed scan delay method may actually be
224
preferable for the majority of institutes,
to minimize scanning at the incorrect
enhancement time, inclusion of the test
bolus method (Murakami et al., 2001) or
the automatic bolus tracking system (Kim
et al., 2002) is recommended.
Determination of the optimal scan delay
with the test bolus method is performed
after precontrast imaging. A mini bolus
(e.g., 15 ml at 5 ml/s) of contrast mate-
rial is injected intravenously and flushing
with a saline chaser (e.g., 25 ml at 5 ml/s)
is recommended. A series of single-level
CT scans with low dose exposure (e.g.,
120 kVp, 10 mA) at the abdominal aorta
is acquired every 2 s from 8 to 40 s after
the initiation of the mini bolus injection.
A region-of-interest cursor is then placed
within the aorta and a time-enhancement
curve of the aorta is generated on the
scanner console. Contrast arrival corre-
sponds to the beginning of the upslope of
the time-enhancement curve, and contrast
arrival time (AT) to the span between the
initiation of the mini bolus injection and
the contrast arrival at the abdominal aorta.
Although the test bolus method can
be used with any MDCT scanner, the
automatic bolus tracking method can be
used only with a scanner equipped with
a system such as Real Prep (Toshiba) or
SmartPrep (GE Healthcare). Capturing the
optimal scan delay using the automatic
bolus tracking system is done as follows.
On the precontrast CT image, a region-of-
interest cursor is placed within the aorta
and the threshold value for enhancement
(e.g., 50 HU) or the CT value is input to the
scanner. The monitoring scan is started 8 s
after the initiation of the contrast injection,
and when the enhancement of the aorta
reaches the threshold on the monitoring
scan images, initiation of the diagnostic
H. Onishi et al.
CT scan is triggered automatically (Real
Prep) or manually (SmartPrep). Because
the mini bolus injection is not necessary
with the automatic bolus tracking method,
the total dose of contrast material required
is less than that for the test bolus method.
SCANNING TIMING
The early arterial phase images which are
useful for the precise evaluation of the
arteries are acquired 10–15 s after the con-
trast arrival (Murakami et al., 2006). At
our institute, the early arterial scanning is
started 10 s after the triggering (the point
that the time-enhancement curve raises to
the threshold value) using the bolus track-
ing system. The threshold is set at 50 HU
of enhancement value.
The late arterial phase images which are
useful for visualization of the hypervascu-
lar HCC or for the precise evaluation of
the intrahepatic portal veins are acquired
20–25 s (22 s at our institute) after trig-
gering. For this timing, maximal tumor-
to-liver contrast can be achieved in most
cases. The portal venous phase images
which are useful for evaluation of the por-
tal venous system including tumor throm-
bus are acquired ~ 60 s after the triggering.
The equilibrium phase images which are
useful for the detection of the hypovascu-
lar HCC are acquired 180 s after the initia-
tion of contrast injection.
IMAGE PROCESSING
Multidetector-row helical computed tom-
ography can generate not only 5 mm-slice
thickness images but also thin-slice images
with the same scanning. Usually, 5 mm-
16. Diagnosis of Hepatocellular Carcinoma: Multidetector-Row Computed Tomography
slice thickness transverse images are used
for the evaluation of HCC (Kawata et al.,
2002), while the thin-slice images are used
for detailed diagnosis or postprocessing.
Truly isotropic voxel data consisting of
submillimeter-slice thickness images can
be acquired easily when using an MDCT
scanner with ³ 16 channels. Multiplanar
reformatted (MPR) images reconstructed
from isotropic voxel data have the same
spatial resolution as the original transverse
images, and three-dimensional (3D)
CT-angiography and maximum intensity
projection (MIP) images can visualize
even small branching arteries. These imaging
techniques are thus useful for acquiring
the clinical information needed to select
the appropriate treatment course and to
plan the treatment procedure including
surgical resection or transcatheter arterial
chemoembolization of HCC. To gener-
ate effective MPR, 3D, and MIP images,
however, we need a clear understanding of
the features and management of the disease.
MAGNETIC RESONANCE
IMAGING
Magnetic resonance imaging has also been
used for the evaluation of HCC, and faster
imaging techniques have been developed,
such as the gradient echo imaging and
three-dimensional Fourier transformation
techniques. Recently, even faster imaging
was achieved with the parallel imaging
technique. In addition, a sequence dedi-
cated for dynamic studies of the abdomi-
nal region, especially the liver, was newly
developed. However, MR imaging used for
liver dynamic studies is inferior to MDCT
imaging in terms of spatial and temporal
resolution, even when these faster MR
225
imaging techniques are employed. The
main advantage of MR imaging is its
higher sensitivity for contrast, which plays
a critically important role in accurate
evaluation of HCC.
MAGNETIC RESONANCE
SCANNING TECHNIQUE
Our standard technique for evaluation
of HCC using the 1.5-Tesla unit (Signa
Excite HD; GE Healthcare) employs the
array spatial sensitivity encoding tech-
nique (ASSET; GE Healthcare), which is
one of the parallel imaging techniques.
For signal reception, an 8-channel phased
array coil system for the abdomino-pelvic
region is used. The parallel imaging tech-
nique is based on the fact that receiver sen-
sitivity generally has an encoding effect
complementary to Fourier preparation by
linear field gradients. In this technique the
spatial information related to the spatially
varying sensitivity of receiver coils is used
to reduce the number of k-space lines
needed to form an image. By using mul-
tiple receiver coils in parallel, scan time
for Fourier imaging can be considerably
reduced.
After localizer scanning, T2-weighted
single shot first spin echo (SSFSE)
images are obtained during breath-hold.
The sequence parameters are: repeti-
tion time [TR]: 900 ms, echo time [TE]:
90 ms, band width: 62.5 kHz, matrix size:
320 × 160, field of view [FOV]: 320
× 320 mm, slice thickness: 5.0–6.0 mm,
slice interval: 2.0–2.5 mm, number of
excitations [NEX]: 0.6. The slice thick-
ness and interval are increased accord-
ing to the size of the liver to include the
whole liver. After calibration scanning
226
for parallel imaging, T1-weighted fast
spoiled gradient recalled acquisition in
steady state (FSPGR) images are obtained
during breath-hold. The images consist of
in-phase and opposed-phase echoes. The
sequence parameters are, TR: 175 ms,
TE: 2.2 ms for opposed-phase imaging
and 4.5 ms for in-phase imaging, flip
angle: 90°, band width: 62.5 kHz, matrix
size: 256 × 160, FOV: 320 × 320 mm,
slice thickness: 5.0–6.0 mm, slice interval:
2.0–2.5 mm, NEX: 1. T2-weighted fast
spin echo images (TR: 3,750–8,000 ms,
TE: 90 ms, echo train length: 8, band
width: 62.5 kHz, matrix size: 512 × 160,
FOV: 320 × 320 mm, slice thickness: 5.0–
6.0 mm, slice interval: 2.0–2.5 mm, NEX: 3)
were obtained with the respiratory trigger-
ing technique, in which TR depends on
the patient’s respiratory interval. After the
precontrast scanning mentioned earlier,
contrast-enhanced imaging is performed.
CONTRAST-ENHANCED
DYNAMIC MAGNETIC
RESONANCE IMAGING
The “Liver Acquisition with Volume
Acceleration” (LAVA; GE Healthcare,)
sequence is employed for the contrast-
enhanced dynamic study. The sequence
employed is the fast gradient echo sequence
combined with the three-dimensional
Fourier transformation technique, which
uses the fat suppression sequence by
means of the improved spectral inver-
sion recovery method. Other manu-
facturers also offer fast gradient echo
sequences for the dynamic study such as
the Volumetric Interpolated Breath-hold
Examination [VIBE]; Siemens Medical
Systems, Erlangen, Germany, or the T1
H. Onishi et al.
High Resolution Interpolated Volume
Examination [THRIVE]; Philips Medical
Systems, Best, The Netherlands). The
parameters of the LAVA sequence that is
employed at our institute are, TR: 4.5 ms,
TE: 2.2 ms, flip angle: 12°, inversion
time: 7 ms, band width: 62.5 kHz, matrix
size: 320 × 192, FOV: 320 × 320 mm,
slice thickness: 4.0 mm, slice interval:
0 mm, NEX: 0.73. The dynamic study
consists of precontrast, arterial-phase,
portal venous-phase and equilibrium-
phase images.
After the precontrast imaging, 20 ml
of gadolinium chelate contrast material
is injected intravenously with a power
injector at a flow rate of 2 ml/s through
a 22-gauge plastic catheter placed in an
antecubital vein. This is followed by a
20-ml saline chaser at a flow rate of 2 ml/s.
The scan delay for the arterial phase imag-
ing is determined by means of the bolus
tracking system, and the arterial scanning
is started 10 s after the triggering. The
portal venous phase and the equilibrium
phase scanning are started 60 and 180 s,
respectively, after the initiation of the con-
trast injection.
TISSUE-SPECIFIC CONTRAST-
ENHANCED MAGNETIC
RESONANCE IMAGING
Superparamagnetic iron oxide is a liver-
specific particulate MR contrast material
that is taken up by the reticuloendothe-
lial system of the liver and improves
hepatic tumor detection by means of
increasing the focal hepatic lesion-to-
liver contrast. Two SPIO contrast materi-
als are commercially available in Japan:
ferumoxides (Feridex; Eiken and Tanabe
16. Diagnosis of Hepatocellular Carcinoma: Multidetector-Row Computed Tomography
Pharmaceutical, Osaka, Japan) and fer-
ucarbotoran (Resovist; Nihon Schering,
Osaka, Japan). While the ferumoxide is
administrated with an intravenous drip, the
ferucarbotran can be injected as a bolus.
Superparamagnetic iron oxide contrast
materials have a T2 shortening effect and
reduce the signal intensity on T2-weighted
images of the normal liver parenchyma,
where SPIO are mainly distributed. In
addition, there are other liver-specific MR
contrast materials such as gadoxetic acid
disodium (Gd-EOB-DTPA), gadobenate
dimeglumine (Gd-BOPTA) and manga-
fodipir trisodium (Mn-DPDP), which are
adsorbed into the liver cells and excreted
into the bile (Murakami et al., 1996).
The scan protocol for SPIO-enhanced
MR imaging at our institute is as follows.
After the precontrast imaging, 0.7–1.4 ml
of ferucarbotran is injected intravenously
and manually through a 22-gauge plas-
tic catheter placed in an antecubital vein,
which is followed by a 20-ml saline chaser.
Post-contrast images are obtained > 10 min
after the SPIO injection. One is a heavily
T1-weighted gradient echo with minimum
echo time (TR: 150 ms, TE: 1.4 ms, flip
angle: 90°, band width: 62.5 kHz, matrix
size: 256 × 160, FOV: 320 × 320 mm, slice
thickness: 5.0–6.0 mm, slice interval: 2.0–
2.5 mm, NEX: 1) The second is T2-weighted
fast spin echo (TR: 3,750–8,000 ms, TE:
90 ms, flip angle: 90°, echo train length: 8,
band width: 62.5 kHz, matrix size: 512 ×
224, FOV: 320 × 320 mm, slice thickness:
5.0–6.0 mm, slice interval: 2.0–2.5 mm,
NEX: 2). The third is T2*-weighted gradi-
ent echo with a long echo time (TR: 200 ms,
TE: 8.5 ms, flip angle: 60°, band width:
62.5 kHz, matrix size: 512 × 160, FOV: 320
× 320 mm, slice thickness: 5.0–6.0 mm,
slice interval: 2.0–2.5 mm, NEX: 1).
227
COMPUTED TOMOGRAPHY
IMAGING FEATURES
OF HEPATOCELLULAR
CARCINOMA
On nonenhanced CT images, HCCs can
appear as slightly less attenuated than
adjacent liver parenchyma (Figure 16.1a).
It often happens that small HCCs are
not detected on noncontrast CT images.
In HCC, however, there may be met-
amorphosis, necrosis, periotic changes
or hemorrhage; therefore, it is usually
heterogeneous attenuation, especially
in large HCCs. Fatty metamorphism is
often observed especially in the well-
differentiated type of HCC and is shown
as a low-attenuation area. Hepatocellular
carcinoma develops tumor vessels, thus
increasing the arterial blood supply of the
tumor. Conversely, in most HCCs, except
for the well-differentiated type, portal
venous blood flow in the tumor disap-
pears. Generally, therefore, hypervascular
HCC is strongly enhanced in the arterial
dominant phase while the attenuation of
HCC decreases in the portal venous domi-
nant phase as seen in contrast-enhanced
dynamic studies. Meanwhile, the liver
parenchyma that is supplied predominantly
by the portal vein is strongly enhanced in
the portal venous dominant phase.
In the early arterial phase, the aorta and
hepatic arteries are enhanced strongly and
HCCs moderately, so that the early arte-
rial images are useful for accurate evalu-
ation of the arteries, especially by means
of 3D-CT angiography or MIP images.
The early arterial phase imaging can be
skipped to reduce the radiation exposure
dose when there is no clinical need for
evaluation of the arteries by means of CT
angiography.
228 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
enhanced and tumor-to-liver contrast is
maximal in the late arterial phase, the
images obtained during this phase are
essential to detect the hypervascular HCCs
(Figure 16.1b). The intrahepatic vein
and tumor thrombus are also thoroughly
enhanced during this phase.
In the portal venous and equilibrium
phases, the contrast medium is moving
from the intravascular to the intercellular
space. The washout of contrast material
from the tumor is rapid because the HCC
contains less intercellular space than the
surrounding liver parenchyma. Although
most HCCs are not supplied by the portal
vein, the liver parenchyma is maximally
enhanced in the portal venous phase, so
that low attenuation is typically observed in
these phases (Figures 16.1c, d) (Kim et al.,
1995). Well-differentiated HCCs are usu-
ally hypovascular and are not visualized
as enhanced nodules in the arterial phase.
Computer tomography images obtained
during the portal venous and equilib-
rium phases have proven to be useful for
the detection of less vascularized tumors
such as well-differentiated or early HCCs.
Iannaccone et al. (2005) reported that
the equilibrium phase images that were
acquired 180 s after initiation of the con-
trast injection were a useful adjunct to the
biphasic images (arterial and portal venous
phase) for the detection of HCC.
COMPUTED TOMOGRAPHIC
ANGIOGRAPHY
The hepatic arteries are depicted with
CT angiography using the thin-slice data
obtained from early arterial phase imaging
16. Diagnosis of Hepatocellular Carcinoma: Multidetector-Row Computed Tomography 229

(Figure 16.1e). It has been reported that
100% of major arterial trunks and ~ 90%
or more of the small branches of the
hepatic arteries were detected on partial
MIP images and original thin-slice cross-
sectional images in the early arterial phase
(Takahashi et al., 2002). Visualization of
the anatomy of hepatic arteries is essen-
tial for planning not only surgical resec-
tion but also the transcatheter arterial
chemoembolization or reservoir infusion
system. The portal and hepatic veins can
be visualized by CT angiography from
late arterial and portal venous phases
imaging (Figure 16.1f). Hepatocellular
carcinomas often invade the portal vein
and/or hepatic vein and form the tumor
thrombus. Computed tomographic por-
tography and CT venography are thus
useful for detection of tumor invasion to
the portal or the hepatic vein. Patients
with HCC usually have cirrhosis as an
underlying disease and have collateral
vessels and varices due to portal hyper-
tension. Computed tomographic angi-
ography can also accurately show the
anatomy of these vessels.
MAGNETIC RESONANCE
IMAGING FEATURES
OF HEPATOCELLULAR
CARCINOMA
On nonenhanced MR imaging, 35% of HCC
cases show high intensity on T1-weighted
images, 25% show isointensity and 40%
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
for the evaluation of the tumor hemody-
namics (Hori et al., 2006). Hepatobiliary
contrast agents enhance normal liver
parenchyma on T1-weighted images due
to the T1 shortening effect. The tumor is
seen as a low-intensity area because the
tumor cell does not take up the agents.
With certain agents that can be injected
in bolus form, such as Gd-EOB-DTPA
and Gd-BOPTA (Kuwatsuru et al., 1999),
contrast-enhanced dynamic studies can
also be performed for the evaluation of
tumor hemodynamics.
STAGING OF
HEPATOCELLULAR
CARCINOMA
Several staging systems for HCC are cur-
rently in use. The Union Interbationale
Contre le Cancer [UICC], the General
Rules for the Clinical and Pathological
Study of Primary Liver Cancer [Liver
Cancer Study Group of Japan classifi-
cation system], and the American Joint
Committee on Cancer [AJCC] have estab-
lished staging systems that are based on
the results of pathologic analyses, which
incorporate anatomic and histologic find-
ings obtained at tumor resection. The
number and size of the primary tumors as
well as the presence or absence of portal
or hepatic vein involvement are included
in the staging of the primary lesion
because HCC is frequently associated
with portal vein or hepatic vein tumor
thrombus. Both MDCT imaging and MR
imaging are also useful for the evalua-
tion of the tumor thrombus. Lymph node
metastasis and distant metastasis which
are taken into consideration in the staging
231
systems can also be evaluated by both of
these imaging techniques.
ACCURACY FOR THE
DIAGNOSIS OF
HEPATOCELLULAR
CARCINOMA
The accuracy of tumor detection is often
evaluated by means of receiver oper-
ating characteristic (ROC) analyses or
alternative-free response ROC (AFROC)
analyses as well as in terms of sensitiv-
ity and positive predictive values. While
the conventional ROC method allows for
only one response per image, the AFROC
method allows for multiple responses per
image (Chakraborty and Winter, 1990).
Several studies have demonstrated the Az
value, which is the area under the ROC
curve, and sensitivity for the detection of
HCC. Az value and sensitivity determined
by means of contrast-enhanced dynamic
MDCT imaging ranged from 0.82 to 0.99
and from 77% to 93%, respectively (Kim
et al., 2006; Iannaccone et al., 2005).
When determined by means of gadolin-
ium-enhanced dynamic MR imaging, Az
value and sensitivity ranged from 0.68 to
0.97 and from 65% to 91%, respectively
(Mori et al., 2005; Kim et al., 2004), while
the corresponding values ranged from 0.76
to 0.98 and from 56% to 95%, respec-
tively, when SPIO-enhanced MR imaging
was used (Ward et al., 2000; Kang et al.,
2003).
Contrast-enhanced dynamic CT imag-
ing using a 16-channel MDCT scanner
and SPIO-enhanced MR imaging showed
similar diagnostic accuracy, sensitivity,
and positive predictive value for the
232
detection of HCC (Kim et al., 2006).
However, gadolinium-enhanced dynamic
MR imaging proved to be significantly more
accurate than SPIO-enhanced MR imaging
(Pauleit et al., 2002; Kim et al., 2004).
HEPATOCELLULAR
CARCINOMA AFTER
TRANSCATHERTER ARTERIAL
CHEMOEMBOLIZATION Percutaneous radiofrequency ablation
Transcatheter arterial chemoemboliza-
tion (TACE) with iodized oil is often
performed for treatment of HCC, after
which the tumor is seen as a nodule with
markedly high attenuation on nonen-
hanced CT images, since high concentra-
tions of iodized oil remain in the tumor.
Because the portion of the tumor where
iodized oil is retained at high density is
considered to be necrotic, CT is suit-
able for the evaluation of the therapeutic
efficacy of TACE with iodized oil for
HCC. However, incomplete retention of
iodized oil corresponds to a wide vari-
ety of tumor necroses, from completely
necrotic to viable cancer tissue (Choi
et al., 1992). Although contrast-enhanced
dynamic CT may be helpful to evaluate
viability for HCC with incomplete reten-
tion after the TACE treatment, the high
level of attenuation of iodized oil in the
tumor can negatively affect the evalua-
tion by dynamic CT. On the other hand,
as iodized oil does not have much effect
on the intensity observed on MR images,
MR images can clearly show intratu-
moral signal changes (Murakami et al.,
1993a, b). Contrast-enhanced dynamic
MR imaging is thus useful for evaluating
viability for HCC after TACE treatment
because it shows enhancement only in
H. Onishi et al.
the viable portion of the tumor in the
arterial phase.
HEPATOCELLULAR
CARCINOMA AFTER
PERCUTANEOUS
RADIOFREQUENCY
ABLATION THERAPY
therapy can be used for local treatment
of HCC, with coagulation necrosis gener-
ally occurring after the treatment and the
ablated area often showing a zonal pattern
on both T1- and T2-weighted images (Lee
et al., 2001). The completely ablated area
appears as hyperintense on T1-weighted
images and as hypointense on T2-weighted
images, reflecting coagulation necrosis or
hemorrhage (Figures 16.2e, f). On the
other hand, the incompletely ablated area
that is often seen around the completely
ablated area appears as hypointense on
T1-weighted images and as hyperin-
tense on T2-weighted images, reflecting
fibrous change or sinusoidal congestion
(Onishi et al., 2004). These zonal pat-
terns on non-enhanced MR imaging may
be important for the assessment of the
effect of ablation therapy on HCC soon
after completion of the therapy, since
these signal changes are seldom affected
by the arterio-portal venous shunt or por-
tal thrombus that is often associated with
percutaneous ablation therapy. In clinical
practice, contrast-enhanced dynamic CT
and contrast-enhanced dynamic MR imag-
ing are used as the standard techniques to
evaluate the therapeutic effect of ablation
therapy, with the necrotic area shown as a
nonenhanced area and the residual viable
lesion as an enhanced area. However, it is
16. Diagnosis of Hepatocellular Carcinoma: Multidetector-Row Computed Tomography 233

often difficult to distinguish whether the SUMMARY

enhanced area is a residual viable lesion or
an arterio-portal venous shunt (Dromain Multidetector-row helical computed tom-
et al., 2002). ography and MR imaging are powerful
tools for the accurate diagnosis of HCC.
Both techniques are important not only for
LIMITATIONS AND PROSPECTS detection but also for staging. Moreover,
CT angiography using MDCT is helpful
The most serious drawbacks of MDCT when performing pre-interventional radio-
imaging may be ionized radiation exposure. logical or surgical treatment for HCC. For
Although the scanning time is markedly useful and accurate clinical evaluation, it is
reduced by means of 64-channel MDCT essential to understand the theories behind
scanners, it is difficult to increase the number imaging techniques as well as the clinical
of phases for the dynamic study because of pathology of HCC and to employ optimal
the need to limit the radiation exposure scanning and contrast injection protocols.
dose. One solution for reducing the radia- Because MDCT and MR imaging have
tion exposure dose is the post-processing their respective advantages and some limi-
technique, while the CT images obtained tations, it is essential in clinical practice to
by scanning with a low exposure dose can make the best use of each modality so that
be improved by means of noise reduction they can complement each other.
filters (Kalra et al., 2003). It can be expected
that the number of detector rows in MDCT REFERENCES
scanners will continue to increase and the Awai, K., Takada, K., Onishi, H., and Hori, S.
scanning time will be reduced much further. 2002. Aortic and hepatic enhancement and
In addition, perfusion imaging for the whole tumor-to-liver contrast: analysis of the effect of
liver may become available in the near different concentrations of contrast material at
future when a wide detector covering the multi-detector row helical CT. Radiology 224:
757–763.
entire liver has been developed. Awai, K., Hiraishi, K., and Hori, S. 2004. Effect
In comparison with MDCT imaging, of contrast material injection duration and rate
the spatial and temporal resolution of on aortic peak time and peak enhancement at
MR imaging may currently be considered dynamic CT involving injection protocol with
a drawback. However, boosting of the dose tailored to patient weight. Radiology 230:
acceleration factor of the parallel imaging 142–150.
Bae, K.T. 2003. Peak contrast enhancement in CT
technique is expected to overcome this and MR angiography: when does it occur and
disadvantage and to give MR imaging an why? Pharmacokinetic study in a porcine model.
edge for the evaluation of HCC. Recently, Radiology 227: 809–816.
MR scanners equipped with an ultra-high Brink, J.A., Heiken, J.P., Forman, H.P., Sagel, S.S.,
magnetic field such as 3.0-Tesla have Molina, P.L., and Brown, P.C. 1995. Hepatic spi-
become commercially available for use ral CT: reduction of dose of intravenous contrast
material. Radiology 197: 83–88.
in the abdominal region as well. Their Chakraborty, D.P., and Winter L.H. 1990. Free-
higher signal-to-noise ratio represents response methodology: alternate analysis and a
a significant advantage for high spatial new observer-performance experiment. Radiology
resolution imaging. 174: 873–881.
234
Choi, B.I., Kim, H.C., Han, J.K., Park, J.H., Kim,
Y.I., Kim, S.T., Lee, H.S., Kim, C.Y., and Han,
M.C. 1992. Therapeutic effect of transcatheter
oily chemoembolization therapy for encapsu-
lated nodular hepatocellular carcinoma: CT and
pathologic findings. Radiology 182: 709–713.
Dromain, C., de Baere, T., Elias, D., Kuoch, V.,
Ducreux, M., Boige, V., Petrow, P., Roche, A.,
and Sigal, R. 2002. Hepatic tumors treated
with percutaneous radio-frequency ablation: CT
and MR imaging follow-up. Radiology 223:
255–262.
Heiken, J.P., Brink, J.A., McClennan, B.L., Sagel,
S.S., Crowe, T.M., and Gaines, M.V. 1995.
Dynamic incremental CT: effect of volume and
concentration of contrast material and patient
weight on hepatic enhancement. Radiology 195:
353–357.
Hori, M., Murakami, T., Kim, T, Iannaccone, R.,
Abe, H., Onishi, H., Tomoda, K., Catalano, C,
Passariello, R., and Nakamura, H. 2006.
Hemodynamic characterization of focal hepatic
lesions: role of ferucarbotran-enhanced dynamic
MR imaging using T2-weighted multishot spin-
echo echo planar sequence. J. Magn. Reson.
Imag. 23: 509–519.
Iannaccone, R., Laghi, A., Catalano, C., Rossi, P.,
Mangiapane, F., Murakami, T., Hori, M.,
Piacentini, F., Nofroni, I., and Passariello, R.
2005. Hepatocellular carcinoma: role of unen-
hanced and delayed phase multi-detector row
helical CT in patients with cirrhosis. Radiology
234: 460–467.
Imai, Y., Murakami, T., Yoshida, S., Nishikawa, M.,
Ohsawa, M., Tokunaga, K., Murata, M., Shibata, K.,
Zushi, S., Kurokawa, M., Yonezawa, T., Kawata, S.,
Takamura, M., Nagano, H., Sakon, M.,
Monden, M., Wakasa, K., and Nakamura, H. 2000.
Superparamagnetic iron oxide-enhanced magnetic
resonance images of hepatocellular carcinoma:
correlation with histological grading. Hepatology
32: 205–212.
Kadoya, M., Matsui, O., Takashima, T., and
Nonomura, A. 1992. Hepatocellular carcinoma:
correlation of MR imaging and histopathologic
findings. Radiology 183: 819–825.
Kalra, M.K., Maher, M.M., Sahani, D.V., Blake,
M.A., Hahn, P.F., Avinash, G.B., Toth, T.L.,
Halpern, E., and Saini, S. 2003. Low-dose CT of
the abdomen: evaluation of image improvement
H. Onishi et al.
with use of noise reduction filters pilot study.
Radiology 228: 251–256.
Kalra, M.K., Maher, M.M., Toth, T.L., Schmidt, B.,
Westerman, B.L., Morgan, H.T., and Saini, S.
2004. Techniques and applications of automatic
tube current modulation for CT. Radiology 233:
649–657.
Kang, B.K., Lim, J.H., Kim, S.H., Choi, D., Lim,
H.K., Lee, W.J., and Lee, S.J. 2003. Preoperative
depiction of hepatocellular carcinoma: ferumox-
ides-enhanced MR imaging versus triple-phase
helical CT. Radiology 226: 79–85.
Kawata, S., Murakami, T., Kim, T., Hori, M.,
Federle, M.P., Kumano, S., Sugihara, E.,
Makino, S., Nakamura, H., and Kudo, M. 2002.
Multidetector CT: diagnostic impact of slice
thickness on detection of hypervascular hepa-
tocellular carcinoma. AJR. Am. J. Roentgenol.
179: 61–66.
Kim, T., Murakami, T., Oi, H., Matsushita, M.,
Kishimoto, H., Igarashi, H., Nakamura, H., and
Okamura, J. 1995. Detection of hypervascular
hepatocellular carcinoma by dynamic MRI and
dynamic spiral CT. J. Comput. Assist. Tomogr.
19: 948–954.
Kim, T., Murakami, T., Hori, M., Takamura, M.,
Takahashi, S., Okada, A., Kawata, S., Cruz, M.,
Federle, M.P., and Nakamura, H. 2002. Small
hypervascular hepatocellular carcinoma revealed
by double arterial phase CT performed with
single breath-hold scanning and automatic bolus
tracking. AJR Am. J. Roentgenol. 178: 899–904.
Kim, Y.K., Kim, C.S., Kwak, H.S., and Lee, J.M.
2004. Three-dimensional dynamic liver MR
imaging using sensitivity encoding for detection
of hepatocellular carcinomas: comparison with
superparamagnetic iron oxide-enhanced MR
imaging. J. Magn. Reson. Imag. 20: 826–837.
Kim, Y.K., Kwak, H.S., Kim, C.S., Chung, G.H.,
Han, Y.M., and Lee, J.M. 2006. Hepatocellular
Carcinoma in Patients with Chronic Liver
Disease: Comparison of SPIO-enhanced MR
Imaging and 16-Detector Row CT. Radiology
238: 531–541.
Kuwatsuru, R., Kadoya, M., Ohtomo, K., Tanimoto, A.,
Hirohashi, S., Murakami, T., Tanaka, Y.,
Yoshikawa, K., and Katayama, H. 1999. Clinical
late phase II trials of MultiHance (Gd-BOPTA) for
the magnetic resonance imaging of liver tumors in
Japan. J. Comput. Assist. Tomogr. 23: 65–74.
16. Diagnosis of Hepatocellular Carcinoma: Multidetector-Row Computed Tomography
Lee, J.D., Lee, J.M., Kim, S.W., Kim, C.S., and Mun,
W.S. 2001. MR imaging-histopathologic correla-
tion of radiofrequency thermal ablation lesion in
a rabbit liver model: observation during acute and
chronic stages. Korean J. Radiol. 2: 151–158.
Mori, K., Yoshioka, H., Takahashi, N., Yamaguchi,
M., Ueno, T., Yamaki, T., and Saida, Y. 2005.
Triple arterial phase dynamic MRI with sensitiv-
ity encoding for hypervascular hepatocellular car-
cinoma: comparison of the diagnostic accuracy
among the early, middle, late, and whole triple
arterial phase imaging. AJR Am. J. Roentgenol.
184: 63–69.
Murakami, T., Nakamura, H., Hori, S., Tomoda, K.,
Mitani, T., Nakanishi, K., Hashimoto, T., Tsuda, K.,
Kozuka, T., Monden, M., and Wakasa, K. 1993a.
Detection of viable tumor cells in hepatocellular
carcinoma following transcatheter arterial chem-
oembolization with iodized oil. Pathologic cor-
relation with dynamic turbo-FLASH MR imaging
with Gd-DTPA. Acta. Radiol. 34: 399–403.
Murakami, T., Nakamura, H., Tsuda, K., Nakanishi, K.,
Hori, S., Tomoda, K., Mitani, T., Kozuka, T.,
Monden, M., and Wakasa, K. 1993b. Treatment of
hepatocellular carcinoma by chemoembolization:
evaluation with 3DFT MR imaging. AJR Am. J.
Roentgenol. 160: 295–299.
Murakami, T., Baron, R.L., Peterson, M.S., Oliver,
J.H. 3rd, Davis, P.L., Confer, S.R., and Federle,
M.P. 1996. Hepatocellular carcinoma: MR imag-
ing with mangafodipir trisodium (Mn-DPDP).
Radiology 200: 69–77.
Murakami, T., Kim, T., Takamura, M., Hori, M.,
Takahashi, S., Federle, M.P., Tsuda, K., Osuga, K.,
Kawata, S., Nakamura, H., and Kudo, M. 2001.
Hypervascular hepatocellular carcinoma: detec-
tion with double arterial phase multi-detector row
helical CT. Radiology 218: 763–767.
235
Murakami, T., Onishi, H., Mikami, K., Iannaccone, R.,
Federle, M.P., Kim, T., Hori, M., Okada, A.,
Marukawa, T., Higashihara, H., Passariello, R.,
and Nakamura, H. 2006. Determining the optimal
timing for early arterial phase hepatic CT imaging
by measuring abdominal aortic enhancement in
variable contrast injection protocols. J. Comput.
Assist. Tomogr. 30: 206–211.
Onishi, H., Matsushita, M., Murakami, T., Tono, T.,
Okamoto, S., Aoki, Y., Iannaccone, R., Hori,
M., Kim, T., Osuga, K., Tomoda, K., Passariello,
R., and Nakamura, H. 2004. MR appearances
of radiofrequency thermal ablation region: his-
topathologic correlation with dog liver models
and an autopsy case. Acad. Radiol. 11: 1180–
1189.
Pauleit, D., Textor, J., Bachmann, R., Conrad, R.,
Flacke, S., Layer, G., Kreft, B., and Schild, H.
2002. Hepatocellular carcinoma: detection with
gadolinium- and ferumoxides-enhanced MR
imaging of the liver. Radiology 222: 73–80.
Takahashi, S., Murakami, T., Takamura, M., Kim, T.,
Hori, M., Narumi, Y., Nakamura, H., and Kudo, M.
2002. Multi-detector row helical CT angiog-
raphy of hepatic vessels: depiction with dual-
arterial phase acquisition during single breath
hold. Radiology 222: 81–88.
Ward, J., Guthrie, J.A., Scott, D.J., Atchley, J.,
Wilson, D., Davies, M.H., Wyatt, J.I., and
Robinson, P.J. 2000. Hepatocellular carcinoma
in the cirrhotic liver: double-contrast MR imag-
ing for diagnosis. Radiology 216: 154–162.
Yamashita, Y., Komohara, Y., Takahashi, M.,
Uchida, M., Hayabuchi, N., Shimizu, T., and
Narabayashi, I. 2000. Abdominal helical CT:
evaluation of optimal doses of intravenous con-
trast material -a prospective randomized study.
Radiology 216: 718–723.
17
Hepatocellular Carcinoma: Effect
of Injection Rate/Injection Duration
of Contrast Material on Computed
Tomography
Tomoaki Ichikawa and Tsutomu Araki
INTRODUCTION
Multiphasic, contrast-enhanced computed
tomography (CT) scanning has improved
the detection and characterization of focal
hepatic lesions, such as hepatocellular
carcinomas (HCC) (Ohashi et al., 1993;
Hollett et al., 1995; Baron et al., 1996).
It has been well established that hepatic
arterial-dominant phase (HAP) scanning
is essential to detect hypervascular HCC.
Hepatic arterial-dominant phase images
can detect a larger number of HCC (67/81,
83%) than any other phases including
unenhanced, portal venous phase (PVP),
and delayed phase (DP) images (Hwang
et al., 1997). In addition, some hypervas-
cular HCC (26–32%) are only seen or are
most conspicuous on HAP images (Hwang
et al., 1997; Mitsuzaki et al., 1996).
The injection rate of the contrast mate-
rial is one of the most important technique-
related factors for detecting hypervascular
HCC on HAP images. The enhancement
of the aorta (including hepatic artery) on
HAP images significantly increases at
higher injection rates of contrast material
(Tublin et al., 1999; Kim et al., 1998).
Based on these results, it was suggested
in recent HAP imaging studies of the liver
that higher injection rates of contrast mate-
rial can improve hypervascular hepatic
lesion detection (Kim et al., 1995).
There has been another controversy that the
detection rate of hypervascular false-positive
lesions, for which small arterial-portal venous
shunts (A-P shunts) may play a major role,
may increase on HAP images in cirrhotic
liver (Yu et al., 1997). Frequent visualization
of A-P shunts on HAP images may dimin-
ish lesion specificity, which in turn, may
diminish overall diagnostic performance for
detecting hypervascular HCC underly-
ing cirrhotic liver. Considering the possible
increase of A-P shunts as well as hypervas-
cular HCC on HAP images obtained with
a higher injection rate of contrast material,
little agreement has been found with regard
to the use of higher injection rates of contrast
material for the evaluation of hypervascular
HCC underlying liver cirrhosis.
FIXED INJECTION RATE
AND INJECTION DURATION
OF CONTRAST MATERIAL
There are two methods to administer
patients’ body-weight-tailored dose of
contrast material for individual patients:
237
238
(1) the administration of contrast material
with a fixed injection rate, which pres-
ently is the common method, and (2) with
a fixed injection duration. In one injection
protocol, when the alternative of injection
rate or injection duration is fixed, the other
factor should be variable by depending on
the patients’ body-weight-tailored dose
of contrast material. However, it should
be recognized that the injection duration
should be fixed in an injection protocol
rather than the injection rate to establish an
optimal multiphasic contrast-enhanced CT
protocol of the liver because the injection
duration is the most important and the only
factor of all technical factors to predict
the scan timing for each phase (Ichikawa
et al., 2006a).
Fixed Injection Rate of Contrast Material
The injection rate of the contrast material
is the most important technical factor for
determining the amplitude of contrast-
enhancement in hypervascular HCC on
HAP images. The enhancement of the
aorta (including hepatic artery) on HAP
images significantly increases at higher
injection rates of contrast material (Bae
et al., 1998; Awai and Hori, 2003). Because
hypervascular HCC is always contrast-
enhanced by contrast material supplied
from the hepatic artery, higher injection
rates of contrast material may increase the
maximum enhancement value of hyper-
vascular HCC. Such a hypothesis has been
supported by a previous report that 4 ml/s
injection rate significantly improved the
detection of small hypervascular HCC
(< 3 cm) compared with 2 ml/s (Mitsuzaki
et al., 1996). However, the determina-
tion of the best injection rate of contrast
material underlying cirrhotic liver is still
T. Ichikawa and T. Araki
controversial, considering the possible
increase of detection rate of hypervascular
false-positive lesions represented by A-P
shunts with a higher injection rate of con-
trast material (Yu et al., 1997).
In our study with a comparison of 3
and 5 ml/s of injection rates for detecting
hypervascular HCC underlying cirrhotic
liver, the sensitivity values for detecting
hypervascular HCC tended to be higher
with 5 ml/s (mean, 36/41 [88%]) than
with 3 ml/s of contrast material (mean,
28/35 [80%]), although there was no
significant difference (Ichikawa et al.,
2006b). While our study also showed
that the detection rate of the false-posi-
tive lesions on HAP images significantly
increased with 5 ml/s than with 3 ml/s
of contrast material, the lesion specifi-
city was significantly lower with 5 ml/s
than with 3 ml/s. In conclusion, it can
be stated that a high injection rate of
contrast material, such as 5 ml/s, is not
always better than medium injection rate
of contrast material, such as 3 ml/s, for
the purpose of evaluating hypervascu-
lar HCC in patients with cirrhotic liver.
Finally, we recommend that one should
use a medium injection rate of contrast
material for hepatic CT imaging in cir-
rhotic patients because the specificity
and the positive predictive values in the
detection of hypervascular HCC are sig-
nificantly higher with a medium injec-
tion rate (3 ml/s) of contrast material
(mean, 236/250 [95%] and 28/42 [67%],
respectively) than with a high injection
rate (5 ml/s) of contrast material (mean,
227/265 [86%] and 36/73 [49%], respec-
tively). Our recommendation seems to be
reliable, especially for general institutes
involving non-expert radiologists for the
liver into hepatic CT image interpretation
17. Hepatocellular Carcinoma: Effect of Injection Rate/Injection Duration 239

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
dose tailored to patient weight. Radiology 230:
142–150.
Bae, K.T. 2003. Peak contrast enhancement in CT
and MR angiography: when does it occur and
why? Pharmacokinetic study in a porcine model.
Radiology 227: 809–816.
Bae, K.T., Heiken, J.P., and Brink, J.A. 1998. Aortic
and hepatic contrast medium enhancement at
CT. Part II. Effect of reduced cardiac output in a
porcine model. Radiology 207: 657–662.
Baron, R.L., Oliver, J.H., Dodd, G.D., Nalesnik,
M., Holbert, B.L., and Carr, B. 1996.
Hepatocellular carcinoma: Evaluation with
biphasic, contrast-enhanced, helical CT.
Radiology 199: 505–511.
Hollett, M.D., Jeffrey, R.B., Nino-Murcia, M.,
Jorgensen, M.J., and Harris, D.P. 1995.
Dual-phase helical CT of the liver: value of arterial
phase scans in the detection of small (<1.5 cm)
malignant hepatic neoplasms. Am. J. Radiol.
164: 879–884.
Hwang, G.J., Kim, M.J., Yoo, H.S., and Lee,
J.T. 1997. Nodular hepatocellular carcinomas:
detection with arterial-, portal-, and delayed-
phase images at spiral CT. Radiology 202:
383–388.
Ichikawa, T., Erturk, S.M., and Araki, T. 2006a.
Multiphasic contrast-enhanced multidetector-
row CT of liver: contrast-enhancement theory
and practical scan protocol with a combination
of fixed injection duration and patients’ body-
T. Ichikawa and T. Araki
weight-tailored dose of contrast material. Eur. J.
Radiol. 58: 165–176.
Ichikawa, T., Nakajima, H., Nanbu, A., Hori, M.,
and Araki, T. 2006b. Effect of injection rate of
contrast material on CT of hepatocellular carci-
noma. Am. J. Radiol. 186: 1413–1418.
Kim, T., Murakami, T., and Oi, H. 1995. Detection
of hypervascular hepatocellular carcinoma by
dynamic MRI and dynamic spiral CT. J. Comput.
Assist. Tomogr. 19: 948–954.
Kim, T., Murakami, T., and Takahashi, S. 1998. Effects
of injection rates of contrast material on arterial
phase hepatic CT. Am. J. Radiol. 171: 429–432.
Mitsuzaki, K., Yamashita, Y., Ogata, I., Nishiharu,
T., Urata, J., and Takahashi, M. 1996. Multiple-
phase helical CT of the liver for detecting small
hepatomas in patients with liver cirrhosis: con-
trast-injection protocol and optimal timing. Am.
J. Radiol. 167: 753–757.
Ohashi, I., Hanafusa, K., and Yoshida, T. 1993.
Small hepatocellular carcinomas: two-phase
dynamic incremental CT in detection and evalu-
ation. Radiology 189: 851–855.
Tublin, M.E., Tessler, F.N., Cheng, S.L., Peters,
T.L., and McGovern, P.C. 1999. Effect of injec-
tion rate of contrast medium on pancreatic and
hepatic helical CT. Radiology 210: 97–101.
Yu, J.S., Kim, K.W., Sung, K.B., Lee, J.T., and
Yoo, H.S. 1997. Small arterial-portal venous
shunts: a cause of pseudolesions at hepatic imag-
ing. Radiology 203: 737–742.
18
Detection of Combined Hepatocellular
and Cholangiocarcinomas: Enhanced
Computed Tomography
Akihiro Nishie and Kengo Yoshimitsu
INTRODUCTION
Most primary liver cancers are grossly
divided into two histologic categories:
hepatocellular carcinoma (HCC), derived
from hepatocytes, and cholangiocellular
carcinoma (CC), arising from the intra-
hepatic bile duct epithelium. Combined
hepatocellular and cholangiocarcinoma
(cHCC-CC) is a rare form of primary liver
cancer presenting histologic evidence of
both hepatocellular and biliary epithelial
differentiation, and is seen in 1.0–3.3%
of surgically treated patients with pri-
mary liver tumors (The liver cancer study
group of Japan, 1990; Aoki et al., 1993).
Comparing the clinicopathological fea-
tures of patients with cHCC-CC with
those of HCC or CC patients has yielded
somewhat conflicting results in several
studies. In many studies, cHCC-CC was
found to be more prevalent in males (Lee
et al., 2002). The incidence of positive
hepatitis B or C serological test results
in the cHCC-CC group is reported to be
intermediate between that in the HCC
group and in the CC group (Liu et al.,
2003; Koh et al., 2005), or to be similar to
that in the HCC group (Yano et al., 2003).
Furthermore, the incidence of underlying
chronic liver disease in the cHCC-CC
group is also intermediate between that in
the HCC group and in the CC group (Yano
et al., 2003; Liu et al., 2003), or similar to
that in the HCC group (The liver cancer
study group of Japan, 1990). On the other
hand, Jarnagin et al. (2001) reported that
the incidence of positive hepatitis B or
C serology and liver cirrhosis in patients
with cHCC-CC was similar to that in CC
patients. The reasons for the difference
in this clinicopathological feature may
include the difference between Western
and Asian societies in the cause of liver
cancer or the difference in the selection
criteria of patients for hepatic resection
(Liu et al., 2003). The survival rate in
patients with cHCC-CC has been reported
to be worse than that in patients with HCC
or CC (Jarnagin et al., 2001; Yano et al.,
2003; Liu et al., 2003), while Koh et al.
(2005) concluded that the cHCC-CC group
had a prognosis that was better than the
CC group but worse than the HCC group.
We have not come to a conclusion about
clinicopathological features of cHCC-CC
at the present time, and more cases must
be analyzed to elucidate them.
In our daily practice we tend to be
distressed at the difficulty of making a
241
242
preoperative diagnosis of this type of
tumor because it is rarely encountered. In
an examination performed by The liver
cancer study group of Japan (1990), the
frequency of lymph node metastases in
patients with cHCC-CC was 76.2% and
was as high as that in patients with CC,
while that in patients with HCC is 30.3%.
Some of the literature suggests that there
was a need for a lymph node resection dur-
ing a hepatectomy for cHCC-CC (Uenishi
et al., 2000; Sasaki et al., 2001; Eguchi
et al., 2002). Therefore, it is important to
make a diagnosis of cHCC-CC preopera-
tively.
Fine needle aspiration biopsy (FNAB)
is one of the possible methods of diagnos-
ing this entity. However, making a reliable
diagnosis of cHCC-CC based on cytologic
preparations alone is very difficult because
HCC and CC have a wide spectrum of
morphologic features (Dusenbery, 1997).
Wee and Nilsson (1999) mentioned that, at
the very least, there should be a high index
of suspicion, especially when the clinical
history is highly suggestive of HCC, and
FNAB reveals only an adenocarcinoma or
tumor cells which could be interpreted as
intermediate cells. They also suggested that
the FNAB diagnosis of cHCC-CC may be
possible if the clinical, cytohistologic, and
immunohistochemical findings support
the presence of two cellular components,
HCC and adenocarcinoma. However, the
usefulness of FNAB in diagnosing cHCC-
CC has been previously discussed in only
a few studies, and presently we have not
come to a conclusion about it.
On the other hand, sonography, computed
tomography (CT), and magnetic resonance
imaging (MRI) are now routinely used
to detect liver tumors and evaluate their
characteristics. With the advent of these
A. Nishie and K. Yoshimitsu
modalities, small HCC and CC can now be
diagnosed with relative ease (Honda et al.,
1993a, b), while a prospective diagnosis of
cHCC-CC on enhanced CT remains diffi-
cult, although several reports describe CT
findings of this entity (Aoki et al., 1993;
Fukukura et al., 1997; Nishie et al., 2005).
We herein describe enhanced CT findings
of cHCC-CC and discuss the possibility of
making a diagnosis of cHCC-CC based on
enhanced CT.
PATHOGENESIS
In general, the histopathological definition
of cHCC-CC is based on criteria proposed
by the World Health Organization: a tumor
that contains unequivocal elements of both
HCC and CC. If bile production, intercel-
lular bile canaliculi, trabecular growth
pattern or all three are recognized, the
epithelium is defined as a hepatocellular
element; if mucin production, definite
gland formation, or both are noted, the
epithelium is defined as a cholangiocel-
lular component (Gibson, 1978). Allen
and Lisa (1949) first defined three types
of cHCC-CC: (1) HCC and CC originating
from different sites in the liver, but consist-
ing of a uniform cell type (double cancer);
(2) contiguous HCC and CC originating
independently, but intermingling as they
grow (combined type); and (3) completely
intermingled HCC and CC (mixed type).
On the other hand, Goodman et al.
(1985) also defined criteria for a patho-
morphological diagnosis of cHCC-CC
that were different from those proposed
by Allen and Lisa (1949). According
to this classification, type I, “collision
tumors”, contain separate and apparently
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;
however, no transitional forms between
HCC and CC are observed; type II, “tran-
sitional tumors”, exhibit hepatocellular
differentiation with a transition area to
CC or to a mixed hepatocellular glan-
dular tumor; and type III, “fibrolamellar
tumors”, resemble a fibrolamellar vari-
ant of HCC but contain mucin-producing
glands. Goodman’s type I tumor is equiva-
lent to Allen and Lisa’s double cancer and
combined type, and Goodman’s type II is
similar to Allen and Lisa’s mixed type.
However, Goodman’s type III tumor does
not correspond to any categories in Allen
and Lisa’s classification. Furthermore,
another type of cHCC-CC is also present.
This type is composed of intermediate
tumor cells that show morphological fea-
tures intermediate between HCC and CC
(Taguchi et al., 1996; Kim et al., 2004). It
is supposed that the latter type of tumor
may originate from hepatic progenitor
cells. Another pathological characteris-
tic of this tumor is a broad desmoplastic
stroma encasing the strands and trabecu-
lae of the tumor cells (Kim et al., 2004).
It is very important to be familiar with the
known histopathological characteristics
when radiological imaging of cHCC-CC
is analyzed.
ENHANCED COMPUTED
TOMOGRAPHIC FINDINGS
Investigations into the imaging character-
istics of cHCC-CC appear to be somewhat
limited to case reports and small series using
imaging techniques that are not up to current
244
standards because of the rarity of this entity.
Aoki et al. (1993) reported on the CT find-
ings of 14 lesions in 12 cases with cHCC-
CC. Enhanced CT findings were divided into
three types with reference to differences in
the enhancement pattern. Type A (n = 9) was
defined as a mass in which only the periph-
eral portion with a wide to narrow width was
variously enhanced as a lesion that remained
isodense or as a lesion in which the central
portion was of low density or isodense on
the early phase. On the late phase the central
portion showed high density or isodensity.
We can guess that viable cancer cells are
located in the periphery of a mass, and are
enhanced on the early phase but fade on
the late phase; the central portion of a mass,
which mainly comprises fibrous tissue and
scant tumor cells, is enhanced only on the
late phase. These findings are consistent with
those of liver metastasis and CC (Muramatsu
et al., 1986; Honda et al., 1993b). In fact,
there were viable cancer cells consisting of
HCC-predominant components in the mar-
ginal portion, and the central portion of the
tumor was composed of abundant fibrous
stroma containing CC-predominant compo-
nents. In the midzone between the peripheral
and central areas, HCC and CC components
intermingled. Type B (n = 4) was defined as
a mass in which the entire portion was well
enhanced on the early phase and faded to
low density on the late phase. This finding is
identical to that for HCC. These masses were
mainly composed of HCC with CC compo-
nents scattered in the mass or of a main mass
of HCC with a small CC area attached to it.
Type C (n = 1) showed low density on the late
phase as well as on the early phase, and this
finding is one of the characteristic CT find-
ings in CC (Honda et al., 1993b). The main
lesion was histopathologically composed of
CC and an attached small HCC component.
A. Nishie and K. Yoshimitsu
That is to say, these results suggest that
cHCC-CCs tend to show CT findings charac-
terized by a predominant component identi-
fied in the mass and that it is difficult to make
a diagnosis of cHCC-CC based on enhanced
CT alone. However, Aoki et al. (1993) evalu-
ated the values of alpha-fetoprotein (AFP)
and carcinoembryonic antigen (CEA), which
are tumor markers characteristic to HCC
and CC, respectively. The AFP level was
positive in 65% of patients with cHCC-CC.
Furthermore, 47% of patients had abnormal
CEA values. They concluded that cHCC-CC
might be suspected when CT findings and
AFP or CEA values are at variance.
Fukukura et al. (1997) also analyzed the
CT findings of 15 cases with cHCC-CC.
They compared the CT findings on the
basis of the gross appearance of resected
specimens. The tumors were classified
into two groups: HCC type which is green-
ish or yellow in color, has hemorrhagic
necrosis in varying degrees and is sharply
demarcated with a fibrous capsule; and CC
type which is whitish in color and shows
an infiltrative growth with an irregular mar-
gin. In the HCC type (n = 6), most parts of
the masses comprised HCC components,
and CC components were observed only
sporadically or focally within the mass.
On the other hand, in the CC type (n = 9),
four masses had a large volume of HCC
components. In the two masses the ratio
of HCC to CC components was almost
equal, and three masses consisted of pre-
dominant CC components. All HCC type
masses were entirely enhanced at the early
phase and faded into low density at the late
phase. This result is quite consistent with
that reported by Aoki et al. (1993). Two
of the masses also had a ring enhancement
representing a fibrous capsule, which is
also one of the characteristics of HCC
18. Detection of Combined Hepatocellular and Cholangiocarcinomas
(Itai et al., 1986). However, eight of nine
CC type masses were enhanced only at the
peripheral portion at the early phase and
changed to low density areas or had only
central portions enhanced at the delayed
phase; in other words, they showed CT
findings compatible with those of CC,
although a sufficient volume of HCC com-
ponents was present in at least six cases.
This was because these HCC components
presented remarkable fibrosis along the
blood space, which is also seen in scle-
rosing HCC (Yamashita et al., 1993).
Therefore, these HCC components were
considered to show CT findings similar to
those seen in CC. The volume of fibrous
stroma in a mass is one of the factors that
can change enhanced CT findings, and we
are sometimes puzzled when diagnosing
a liver mass. Fukukura et al. (1997) con-
cluded that this CC type can be diagnosed
as cHCC-CC by evaluating virus markers
and serum AFP levels, which may suggest
the presence of HCC components.
Choi et al. (1994) reported iodized-oil
CT findings in six cases with cHCC-CC.
On iodized-oil CT scans, a partial reten-
tion of iodized oil was observed in the
mass in all six cases. HCC portions of
pathologic specimens showed compact
retention of iodized oil, whereas CC por-
tions did not show compact retention.
However, it is usually impossible to make
a diagnosis of cHCC-CC based only on
an iodized-oil CT result because scleros-
ing HCC or necrotic portions of HCC also
show no retention of iodized oil.
How frequently can we diagnose cHCC-
CC based on enhanced CT alone? We
evaluated the performance of enhanced
CT in making a diagnosis of cHCC-CC
by closely comparing CT findings with
histologic findings (Nishie et al., 2005).
245
Twenty-seven patients with cHCC-CCs his-
topathologically proven by surgical resec-
tion were enrolled in our study. Before the
evaluation of enhanced CT findings, the
pathologist identified all areas of HCC or
CC cell type and subclassified cHCC-CC
according to the criteria of Allen and Lisa
(1949). Then, we focused on the enhanc-
ing pattern of a part or component of the
lesions. Based on the previous reports, we
defined the components showing high den-
sity on the early phase and low density on
the late phase (high–low pattern) as HCC-
pattern component (Honda et al., 1993a),
and we defined the components showing
low density on the early phase and low,
iso, or high density on the delayed phase
(low–low, low-iso, or low–high patterns)
as CC-pattern components (Honda et al.,
1993b). We correlated the radiologically
determined components to the histological
areas of each cell type. When a complete
concordance between the component
detected radiologically and the histological
finding was obtained in terms of site, size,
and enhancing pattern, we determined this
to be a correct concordance (Figure 18.1).
The diagnostic rate (DR) was defined as the
ratio of patients in whom CT (HCC- and
CC-pattern components) correctly depicted
HCC/CC to the total number of patients.
According to the criteria of Allen and
Lisa (1949), histological evaluation deter-
mined the presence of double cancer,
combined type, and mixed type in 2, 9 and
16 patients, respectively. In the two cases
of double cancer, two distinct masses were
described as HCC-pattern and CC-pattern
components, respectively. The DR was
100% (2/2) in double cancer.
On the other hand, the DR was 44.4%
(4/9) in the combined type. Why could we
not diagnose at a high rate? The reasons
246
for discordance included the different
enhancing pattern from the typical pat-
terns as defined earlier and the small size
of each component. The CC in one patient
showed a high–low (HCC) pattern on CT
and was composed of poorly differentiated
adenocarcinoma cells with a small amount
of desmoplastic stroma. Yoshida et al.
(1999) also reported two cases of small int-
rahepatic CCs with marked hypervascular-
ity, composed of a large number of tumor
cells and little interstitial fibrous tissue.
Honda et al. (1993b) reported that only 5%
of intrahepatic peripheral cholangiocarci-
noma showed the high–low pattern. The
HCC in another patient showed a low–low
(CC) pattern on CT and was composed
of well-differentiated hepatocellular car-
cinoma cells. Fujita et al. (1996) reported
that 17.6% of well-differentiated HCC
showed a low–low pattern. That is to say,
these results suggest that it becomes dif-
ficult to diagnose cHCC-CC on CT when
one component of cHCC-CC shows an
atypical enhancing pattern. Furthermore,
another reason for the discordance is that
either component was very small (<1 cm
in diameter) and could not be detected on
CT. This discordance was observed in the
other three patients.
The DR in the mixed type was 18.8%
(3/16), and even lower than that for the
combined type. Three masses composed
of intermediate cells were included in this
group. The main reasons for discordance
included the different enhancing pattern
from the typical patterns as defined ear-
lier, and the small size of each component.
In one patient the CC-pattern component
on CT was histologically proven to be
HCC. This HCC showed a low-iso (CC)
pattern on CT and was composed of
moderately differentiated hepatocellular
A. Nishie and K. Yoshimitsu
carcinoma with abundant fibrous stroma.
Either component was very small (< 1 cm
in diameter) in nine patients and could not
be detected on CT. Another reason for the
discordance was the presence of tumors
composed of intermediate cells as men-
tioned earlier (Taguchi et al., 1996; Kim
et al., 2004). Two different components
are not reasonably detected in this type
of tumor, which is constructed of single
tumor cells. In our series, three patients
had mixed type tumors that were shown
to be totally composed of intermediate
cells, with one showing an HCC pattern
and the other two showing a CC pattern
on CT. The two intermediate cell tumors
showing a CC pattern had a large amount
of interstitial fibrosis, which is one of the
characteristics of this type (Kim et al.,
2004) and is histopathologically similar
to CC. However, the third tumor showing
the HCC pattern was composed of malig-
nant cells proliferating in a trabecular,
solid, and tubular pattern. Some of these
cells reacted with CA19-9, and only a
small amount of fibrous tissue was seen.
This result suggests that all tumors com-
posed of intermediate cells do not show
the CC enhancement pattern. As a result,
the total DR on enhanced CT was 33.3%
(9/27), which was not very high. Variable
factors such as an atypical enhancement
pattern, the small size of the HCC or CC
components, and the presence of a tumor
composed of intermediate cells led to a
misreading interpretation.
The main differential diagnosis of
cHCC-CCs on enhanced CT is consid-
ered to be HCC or CC containing various
kinds of differentiation. An early advanced
HCC, which is a representative of such
tumors, may be differentiated from cHCC-
CCs by delayed enhancement of a low
18. Detection of Combined Hepatocellular and Cholangiocarcinomas
attenuation component seen on the early
phase, because most CCs had increased
CT values on the delayed phase compared
with the early phase (Honda et al., 1993b).
However, it may be very difficult to dif-
ferentiate such tumors from cHCC-CCs
on enhanced CT alone when HCC or CC
components with atypical enhancing pat-
terns may be contained in the tumor.
In conclusion, currently it is not easy
to make a diagnosis of cHCC-CC cor-
rectly on the basis of enhanced CT alone.
However, this type of tumor may be diag-
nosed by referring to clinical data such as
tumor markers. Cases with cHCC-CC had
to be recruited over 10 years in most of
the previous reports because of the rarity
of this entity. Therefore, most of the CT
images were scanned by the incremental
dynamic technique. Multi-detector row
technology, which presents high spatial
resolution, may be able to improve the
detection of small components that are
< 1 cm in diameter, and may lead to an
improvement in the diagnostic rate.
REFERENCES
Allen, R. A., and Lisa, J. R. 1949. Combined liver
cell and bile duct carcinoma. Am. J. Pathol.
25:647–655.
Aoki, K., Takayasu, K., Kawano, T., Muramatsu,
Y., Moriyama, N., Wakao, F., Yamamoto,
J., Shimada, K., Takayama, T., Kosuge, T.,
Yamasaki, S., Sakamoto, M., and Hirohashi,
S. 1993. Combined hepatocellular carcinoma
and cholangiocarcinoma: clinical features and
computed tomographic findings. Hepatology
18:1090–1095.
Choi, B. I., Han, J. K., Kim, Y. I., Kim, H. C.,
Park, J. H., Kim, C. W., and Han, M. C. 1994.
Combined hepatocellular and cholangiocarci-
noma of the liver: sonography, CT, angiography,
and iodized-oil CT with pathologic correlation.
Abdom. Imaging 19:43–46.
247
Dusenbery, D. 1997. Combined hepatocellular-
cholangiocarcinoma. Cytologic findings in four
cases. Acta Cytol. 41:903–909.
Eguchi, H., Nagano, H., Sakon, M., Miyamoto, A.,
Kondo, M., Arai, I., Morimoto, O., Dono, K.,
Umeshima, K., Nakamori, S., Wasaka, K., and
Monden, M. 2002. A successful resection and
long-term survival of a patient with intrahepatic
recurrences of combined hepatocellular-cholan-
giocarcinoma: report of a case. Surg. Today
32:742–746.
Fujita, M., Kuroda, C., Inoue, E., Kumatani, T.,
Hosomi, N., Narumi, Y., Kuriyama, K., Kadota,
T., Kasugai, H., Sasaki, Y., and Ishiguro, S.
1996. Usefulness of spiral CT in the detection
of hypovascular, well-differentiated hepatocel-
lular carcinoma [in Japanese]. Nippon Igaku
Hoshasen Gakkai Zasshi 56:155–159.
Fukukura, Y., Taguchi, J., Nakashima, O., Wada,
Y., and Kojiro, M. 1997. Combined hepatocellu-
lar and cholangiocarcinoma: correlation between
CT findings and clinicopathological features. J.
Comput. Assist. Tomogr. 21:52–58.
Gibson, J. 1978. Histological typing of tumors of
the liver, biliary tract, and pancreas. Geneva,
Switzerland: World Health Organization.
Goodman, Z. D., Ishak, K. D., Langloss, J. M.,
Sesterhenn, I. A., and Rabin, L. 1985. Combined
hepatocellular-cholangiocarcinoma. A histo-
logic and immunohistochemical study. Cancer
55:124–135.
Honda, H., Ochiai, K., Adachi, E., Yasumori, K.,
Hayashi, T., Kawashima, A., Fukuya, T., Gibo,
M., Matsumata T., Tsuneyoshi M., and Masuda,
K. 1993a. Hepatocellular carcinoma: correlation
of CT, angiographic, and histopathologic find-
ings. Radiology 189:857–862.
Honda, H., Onitsuka, H., Adachi, E., Ochiai,
K., Gibo, M., Yasumori, K., Matsumata, T.
Sugimachi, K., and Masuda, K. 1993b.
Intrahepatic peripheral cholangiocarcinoma:
two-phased dynamic incremental CT and patho-
logic correlation. J. Comput. Assist. Tomogr.
17:397–402.
Itai, Y., Ohtomo, K., Kokubo, T., Yamaushi, T.,
Minami, M., Yashiro, N., and Araki, T. 1986.
CT of hepatic masses: significance of prolonged
and delayed enhancement. Am. J. Roentgenol.
146:729–733.
248
Jarnagin, W. R., Weber, S., Tickoo, S. K., Koea,
J. B., Obiekwe, S., Fong, Y., DeMatteo, R.
P., Blumgart, L. H., and Klimstra, D. 2001.
Combined hepatocellular and cholangiocarcinoma:
demographic, clinical, and prognostic factors.
Cancer 94:2040–2046.
Kim, H., Park, C., Han, K., Choi, J., Kim, Y. B.,
Kim, J. K., and Park, Y. N. 2004. Primary liver
carcinoma of intermediate (hepatocyte-cholangi-
ocyte) phenotype. J. Hepatol. 40:298–304.
Koh, K. C., Lee, H., Choi, M. S., Lee, J. H., Paik,
S. W., Yoo, B. C., Rhee, J. C., Cho, J. W., Park,
C. K., and Kim, H. J. 2005. Clinicopathological
features and prognosis of combined hepatocellular
cholangiocarcinoma. Am. J. Surg. 189:120–125.
Lee, C. C., Wu, C. Y., Chen, J. T., and Chen,
G. H. 2002. Comparing combined hepatoc-
cellular-cholangiocarcinoma and cholan-
giocarcinoma: a clinicopathological study.
Hepatogastroenterology 49:1487–1490.
Liu, C. L., Fan, S., Lo, C., Ng, I. O., Lam, C., Poon,
R. T., and Wong J. 2003. Hepatic resection for
combined hepatocellular and cholangiocarci-
noma. Arch. Surg. 138:86–90.
Muramatsu, Y., Takayasu, K., Moriyama, N.,
Shima, Y., Goto, H., Ushio, K., Yamada T.,
Hasegawa, H., Koyama, Y., and Hirohashi, S.
1986. Peripheral low-density area of hepatic
tumors: CT-pathologic correlation. Radiology
160:49–52.
Nishie, A., Yoshimitsu, K., Asayama, Y., Irie, H.,
Aibe, H., Tajima, T., Shinozaki, K., Nakayama,
T., Kakihara, D., Shimada, M., Aishima, S.,
Yoshida, K., and Honda, H. 2005. Detection of
combined hepatocellular and cholangiocarcino-
mas on enhanced CT: comparison with histologic
findings. Am. J. Roentgenol. 184:1157–1162.
Sasaki, A., Kawano, K., Aramaki, M., Ohno, T.,
Tahara, K., Takeuchi, Y., Yoshida, T., and Kitano,
S. 2001. Clinicopathological study of mixed
A. Nishie and K. Yoshimitsu
hepatocellular and cholangiocellular carcinoma:
modes of spreading and choice of surgical treat-
ment by reference to macroscopic type. J. Surg.
Oncol. 76:37–46.
Taguchi, J., Nakashima, O., Tanaka, M., Hisaka, T.,
Takazawa, T., and Kojiro, M. 1996. A clinico-
pathological study on combined hepatocellu-
lar and cholangiocarcinoma. J. Gastroenterol.
Hepatol. 11:758–764.
The liver cancer study group of Japan. 1990.
Primary liver cancer in Japan: clinicopatho-
logical features and results of surgical treatment.
Ann. Surg. 221:277–287.
Uenishi, T., Hirohashi, K., Shuto, T., Yamamoto, T.,
Kubo, S., Tanaka, H., Ikebe, T., and
Kinoshita, H. 2000. Surgery for mixed hepa-
tocellular and cholangiocellular carcinoma.
Hepatogastroenterology 47:832–834.
Wee, A., and Nilsson, B. 1999. Combined hepa-
tocellular-cholangiocarcinoma: diagnostic chal-
lenge in hepatic fine needle aspiration biopsy.
Acta Cytol. 43:131–138.
Yamashita, Y., Fan, Z. M., Yamamoto, H.,
Matsukawa, T., Arakawa, A., Miyazaki, T.,
Harada, M., and Takahashi, M. 1993. Sclerosing
hepatocellular carcinoma: radiologic findings.
Abdom. Imaging 18:347–351.
Yano, Y., Yamamoto, J., Kosuge T., Sakamoto, Y.,
Yamasaki, S., Shimada, K., Ojima, H.,
Sakamoto, M., Takayama, T., and Makuuchi,
M. 2003. Combined hepatocellular and
cholangiocarcinoma: a clinicopathologic
study of 26 resected cases. Jpn. J. Clin. Oncol
33:283–287.
Yoshida, Y., Imai, Y., Murakami, T., Nishikawa,
M., Kurokawa, M., Yonezawa, T., Tokunaga, K.,
Fukushima, Y., Wakasa, K., Kim, T., Nakamura,
H., Sakon, M., and Monden, M. 1999. Intrahepatic
cholangiocarcinoma with marked hypervascular-
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
INTRODUCTION
Hepatocellular carcinoma (HCC) is the
fifth most common cancer in the world and
responsible for 500,000 deaths globally every
year (Parkin et al., 2001). Approximately,
80% of the HCC is related to hepatitis B
and/or C virus infection. The prognosis
remains poor because of advanced stage of
cancer and associated hepatic impairment
at diagnosis and high recurrence rate of ∼
80% after surgery (Imamura et al., 2003) or
percutaneous ethanol injection (PEI) (Koda
et al., 2000) in a different segment of the
liver. Therefore, early diagnosis is impera-
tive to enable patients to undergo curable
therapies. According to the nation-wide
biannual report of the Liver Cancer Study
Group of Japan, HCC accounted for 95%
of 19,920 patients with primary liver cancer
(Ikai et al., 2005).
With the advancement of imaging
modalities, such as helical computed tom-
ography (CT) (Murakami et al., 2001),
and high-quality ultrasonography and
magnetic resonance (MR) imaging, small
hepatic nodular lesions as well as small
HCC (< 3 cm) have been frequently found
in chronic hepatic disease. They include
adenomatous hyperplasia (AH) corre-
sponding to low grade dysplastic nod-
ule (DN) proposed by the International
Working Party (1995), atypical AH to high
grade dysplastic nodule, and early HCC
(Sakamoto et al., 1991).
CLASSIFICATION
OF NODULAR
HEPATOCELLULAR LESIONS
To differentiate various small nodular
lesions (< 2 cm) developed in chronic injured
liver, histological criteria were proposed
by the Liver Cancer Study Group of Japan
in 1992, 2003, for early HCC, AH, atypical
AH, and large regenerative nodule. A new
classification system of nodular hepato-
cellular lesions was introduced by the
International Working Party in 1994. It
consists of low grade DN, high grade
DN, and HCC. Subsequently, a new guide
249
250
book was published by the World Health
Organization (Hirohashi et al., 2000),
which mentioned that the morphological
criteria for differential diagnosis among
AH, atypical AH and early HCC was still
under discussion mainly due to the lack of
objective phenotypic or genotypic markers.
Recently, the discrepancy between
Japanese (Sakamoto et al., 1991) and non-
Japanese (International Working Party,
1995) investigators regarding dysplastic
nodule and early HCC was reported to
reflect differences in interpretation and
applications of nomenclature and not due
to different biological pathways (Hirohashi
et al., 2000; Theise et al., 2002):
Pathology of adenomatous hyperplasia,
atypical adenomatous hyperplasia, early
hepatocellular carcinoma, and nodule-in-
nodule hepatocellular carcinoma.
Macroscopically, the first three entities
such as AH (Figure 19.1c), atypical AH,
and early HCC (Figure 19.2c) are difficult
to differentiate. However, mean diameters
of AH, atypical AH and early HCC were
0.8, 1.2 and 1.4 cm, respectively, with a sig-
nificant difference between AH and atypi-
cal AH and between AH and early HCC
(Takayasu et al., 1989). Nodule-in-nodule
type HCC, composed of internal develop-
ment of dedifferentiated HCC within early
HCC, can be differentiated from the other
three entities when the internal nodule
is large enough to be visualized grossly.
The boundary of early HCC as well as
nodule-in-nodule type HCC is equivocal
to recognize on both intraoperative and
extracorporeal ultrasonography. Thus, care-
ful attention is needed not to leave positive
margins especially at surgical resection
and/or local ablation therapy.
Microscopically, early HCC presents
focal structural abnormalities of increased
cell density which are more than twice
K. Takayasu
that of surrounding hepatic parenchyma
(Sakamoto et al., 1991), irregular thin
trabecular pattern, acinar and/or pseu-
doglandular pattern, fatty change and/or
clear cell changes of cancer cells (Kojiro,
1997), and stromal invasion (Kondo et al.,
1994). Early HCC is deemed to be a well-
differentiated HCC, namely cancer in
situ. Meanwhile, AH shows a moderate
increase in cell density compared with
the surrounding tissue but no structural
abnormalities. The nuclear cytoplasmic
ratio is slightly increased due to a slight
decrease in the amount of cytoplasm (The
liver cancer study group of Japan, 1992,
2003). Atypical AH presents with inter-
mediate findings between AH and early
HCC. Nodule-in-nodule HCC consists of
progressed HCC within an early HCC;
emergence of less differentiated HCC
within a very well differentiated one.
Immunohistochemical examination with
heat-shock protein 70 was reported to be a
sensitive molecular marker to differentiate
early HCC from precancerous lesion or
noncancerous liver (Chuma et al., 2003).
The frequency of vascular invasion and/
or intrahepatic metastases observed micro-
scopically in early HCC, nodule-in-nodule
type HCC, and progressed HCC (Figure
19.1c) was 5%, 26%, and 35%, respec-
tively (Watanabe et al., 1990). In our
hospital, early HCC and nodule-in-nodule
HCC accounted for 14% and 5.1%, respec-
tively, of surgically resected 980 HCCs in
664 patients (Oikawa et al., 2005).
MULTISTEP PROGRESSION OF
HEPATOCARCINOGENESIS
Clinical and pathological studies have
elucidated multistep progression of hepa-
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

Adenomatous hyperplasia emerges within differentiated HCC. As another pathway,

chronic injured liver, which progresses to
early HCC. Subsequently, less differenti-
ated HCC develops within an early HCC,
namely nodule-in-nodule type HCC, which
finally develops to overt HCC which is
usually occupied by moderately to poorly
de novo hepatocarcinogenesis is also pre-
sumed to occur.
The size of early HCCs is usually < 2 cm
in diameter; however, large-size early
HCC and nodule-in-nodule HCC measur-
ing 5 cm have been rarely encountered.
252 K. Takayasu

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)

NEEDLE BIOPSY ficiently accurate imaging technique to

distinguish HCC from other benign or
Histopathologic diagnosis is still the gold malignant conditions. However, the diag-
standard, especially for nonenhanced nod- nosis of the small amount of specimen
ular lesions by dynamic CT. According obtained by needle biopsy is not always
to the Barcelona-2000 EASL conference easy due to sampling error and intratu-
(Bruix et al., 2001), needle biopsy is moral heterogeneity. One study reported
recommended for the patient whose hepatic that 9.7% of 99 histologically proven
lesion does not exceed 2 cm due to insuf- regenerative nodules obtained by needle
19. Hepatocellular Carcinoma and Adenomatous Hyperplasia
aspiration biopsy showed malignant trans-
formation at 3 years (Kobayashi et al.,
2006). This might evoke a critical problem
of whether it suggests limitation of needle
biopsy or whether regenerative nodule
actually develops to carcinoma. On the
other hand, needle tract seeding and intra-
peritoneal hemorrhage are still possible
risks of needle biopsy.
METHODS OF
MULTIDETECTOR
COMPUTED TOMOGRAPHY
Helical CT with two (arterial and delayed
phases) or three phases (arterial, portal,
and delayed phases) is widely used by
injecting 100–120 ml of nonionic contrast
medium into antecubital veins at a speed of
3 ml/s. Computed tomography is started at
35–40 s (arterial phase), 70 s (portal phase),
and 3 min (delayed phase) after the start of
the injection with contrast medium. Entire
livers are scanned within one breath-hold
for ∼ 8–20 s, which depends on liver size.
Scanning parameters are as follows: axial
single- or 4-slice mode, 5–10 mm beam
collimation, 0.5–1.0 s/rotation, 0.7–3.0 of
helical pitch (0.7–1.0 of pitch factor),
120–150 kVp, and 200–250 mAs. Image
reconstructions are 5–7 mm in thickness.
METHODS OF A
COMBINATION OF
COMPUTED TOMOGRAPHY
AND ANGIOGRAPHY
As a further examination of hemodynamics
of lesions, a combination study of helical
CT and angiography was performed. One
was CT arterial portography (CTAP) and
253
the other was CT hepatic arteriography
(CTHA) in which the catheter tip is
advanced into the superior mesenteric
artery or the proper hepatic artery, respec-
tively. The former was started 20 s after
the injection of 90 ml of Ioversol (Optiray;
350 mgI/ml, Yamanouchi, Tokyo) diluted
with saline (1:3 ratio; iodine, 87.5 mgI/
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
was commenced 10 s after the injection of
60 ml of the same diluted contrast medium
into the proper, right or left hepatic artery
at a speed of 1.3–2 ml/s based on the
variation of arterial anatomy. Both the
beam collimation and image reconstruc-
tion was 7 mm. Usually CTAP was first
performed to obtain unequivocal contrast
image, followed by celiac arteriography
to know the variation of artery, and next
CTHA, and superselective arteriography,
if necessary.
CT IMAGES OF ADVANCED
HEPATOCELLULAR
CARCINOMA
In general, advanced HCC is demon-
strated as a hyper-attenuating lesion in
arterial phase and hypo-attenuating in
portal and/or delayed phases of dynamic
CT (Takayasu et al., 1990). The washout
of contrast medium from a lesion is more
clearly seen in delayed phase (3–5 min
after contrast injection). The small HCC
(< 3 cm) is homogeneously enhanced by
contrast medium (Takayasu et al., 1995a),
and faint ring enhancement is recognized
if it is encompassed by fibrous capsule.
Large HCC (∼ 10 cm) is inhomogeneously
enhanced mainly due to associated necro-
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

To understand the serial change of atten- 33.3% on CTHA, respectively (Takayasu

uation of the nodular lesions on dynamic et al., 1989). Even CTAP and CTHA often
CT, both pathologic (Ueda et al., 1992) failed to detect AH (Figure 19.1A, B).
and hemodynamic studies with CTAP and When the lesions of AH and/or atypical
CTHA (Hayashi et al., 1999) have been AH were detected on CT and/or CTHA,
performed by focusing on multistep pro- they were usually shown as hypoattenu-
gression of hepatocarcinogenesis. At the ated ones which is similar to early HCC.
initial stage of multistep hepatocarcino- AH is more difficult to detect on various
genesis, AH receives almost similar blood images due to its smaller size as compared
supply to nontumorous parenchyma from to atypical AH and early HCC (Takayasu
both arterial and portal blood flow; subse- et al., 1989).
quently a gradual decline of blood supply
from both vessels occurs mainly due to the
decrease in the number of the portal tracts COMPUTED TOMOGRAPHY
such as in atypical AH and/or early HCC. IMAGES OF NODULE-IN-
However, arterial blood flow only continu- NODULE HEPATOCELLULAR
ously increases following the temporary CARCINOMA
decrease secondary to the development of
unpaired artery which might occur during When a dedifferentiated component (mod-
the process from early HCC to nodule-in- erately or poorly differentiated HCC)
nodule HCC. Finally, arterial blood flow emerges within a well-differentiated HCC,
alone nourishes the lesion of advanced the lesion presents with a nodule-in-nodule
HCC. The portal blood supply shows one- appearance. The arterial phase CT dis-
way decrease during the multistep pro- closes the central area hyperattenuated
gression, resulting in complete depletion and the peripheral area remains iso- or
in advanced HCC. The arterial blood flow hypoattenuated. In the delayed phase, the
into the lesions draws the inverted para- entire lesion becomes hypoattenuated
bolic curve during the process from AH because the central enhanced area changes
to advanced HCC which might explain the to hypoattenuation (Winter et al., 1994). In
unstable state of attenuation of lesions onother words, the CT attenuation behavior
dynamic CT and CTHA. of the central and peripheral portions of
nodule-in-nodule type HCC is similar to
that of isolated advanced HCC and iso-
COMPUTED TOMOGRAPHY lated early HCC, respectively. CT arterial
IMAGES OF ADENOMATOUS portography and CT hepatic arteriography
HYPERPLASIA AND could show more precise hemodynamics
ATYPICAL ADENOMATOUS of this entity (Winter et al., 1994).
HYPERPLASIA A lesion requires differential diagnosis
from early HCC, complicated cyst (hemor-
The detection rate of 14 AHs and 12 atypi- rhagic cyst), atypical hemangioma, which
cal AHs surgically resected was 0% and is hardly enhanced due to entirely hyali-
50% on US, 0% and 9.1% on dynamic nized degeneration, necrotic regenerative
CT, 50% and 33.3% on CTAP, and 0% and nodule and hypovascular metastasis.
256
NATURAL OUTCOME OF
HYPO-ATTENUATING
NODULAR LESIONS
It is important to elucidate the natural out-
come of hypo-attenuating nodules mainly
corresponding to early HCC as well as
atypical AH in order to know the bio-
logical behavior, to determine the optical
treatment, and to predict risk factors of
malignant transformation. Therefore, we
studied the natural outcome of 60 hypo-
attenuating nodular lesions in chronic liver
disease using follow-up CT. As a result, 36
(60%) lesions showed attenuation conver-
sion from hypo- to hyper-attenuation sug-
gesting malignant transformation occurred.
Twenty-one hypo-attenuating lesions were
unchanged in attenuation, and the remain-
ing three disappeared spontaneously
(Takayasu et al., 2006). Overall cumulative
attenuation conversion rates from hypo- to
hyper-attenuation were 15.8% at 1-year,
44.3% at 2-years, and 58.7% at 3-years.
Thirty-six lesions, converted to hyper-
attenuation, consisted of two subtypes;
one was hyper-in-hypoattenuating lesion
suggesting nodule-in-nodule type HCC
in 25 lesions and the other was entirely
hyperattenuating nodule corresponding to
overt HCC in 11 lesions. Between the two
subtype groups, the attenuation conversion
rate from hypo- to hyper-in-hypo or from
hypo- to entirely hyperattenuation was not
significantly different.
Takayama et al. (1990) reported that
histologically proven AH developed to
advanced HCC with a malignant transfor-
mation of 22% (4 of 18 lesions) at 1-year,
50% (5 of 10) at 2-years, and 80% (4 of
5) at 3-years. Sakamoto and Hirohashi
K. Takayasu
(1998) studied the natural outcome of
hypovascular nodules on angiography and
well-differentiated HCC histologically
with ultrasonography; as a result, 13 (72%)
and 1 (6%) of 18 AH nodules developed to
early HCC and overt HCC, respectively;
while, 4 (33%) of 12 early HCCs pro-
gressed to overt HCC without observation
of the stage of early HCC on ultrasonogra-
phy. Borzio et al. (2003) reported that 28
(31%) ultrasound-detected macronodules
(including large regenerative nodule, low
grade DN and high grade DN) in cirrhosis
transformed into HCC at 3-years. Recently,
Kobayashi et al. (2006) reported that the
malignant transformation was recognized
in 29 of 154 (18.8%) nodular lesions. The
biopsy showed high or low grade DN and
regenerative nodules which were initially
found by ultrasonography. In addition,
3-year malignant transformation rate of
high grade DN was 61.5% that was very
similar to ours of 58.7% studied by follow-
up dynamic CT (Takayasu et al., 2006). The
annual rate of malignant transformation
(Kobayashi et al., 2006) and conversion
attenuation (Takayasu et al., 2006) were
20% and 15.8%, respectively. Their data
were higher than those of occurrence
of HCC ranging 4–7% in patients with
chronic viral hepatic disease (Ikeda et al.,
1993; Tsai et al., 1997).
A hyper-in-hypoattenuating nodular
lesion suggesting nodule-in-nodule HCC
showed rapid progression to an entirely
hyper-attenuating one (overt HCC)
(Takayasu et al., 2006). Sadek et al. (1995)
and Hayashi et al. (2002) also reported
similar findings in their follow-up with
MRI and a combination study of CT and
angiography, respectively.
19. Hepatocellular Carcinoma and Adenomatous Hyperplasia 257

CRITICAL CONSIDERATION TO in order to reveal the intratumoral vas-

TREAT HYPOATTENUATING culature by separately evaluating portal
LESIONS and arterial blood supply (Takayasu et al.,
1995b; Hayashi et al., 1999). However, it
Consensus has not been reached regard- is invasive and periodical follow-up is not
ing the management of hypo-attenuating adequate. Instead, dynamic MR imaging or
nodular lesion in virus-related chronic multidetector CT is an alternative. Because
liver disease. Until now, several treat- radiation hazard of CT is problematic,
ments have been performed. However, contrast enhanced ultrasonography should
they have some limitations; surgery is replace dynamic CT (Ding et al., 2001).
invasive (Takayama et al., 1998), transar- In conclusion, early HCC derived from
terial chemoembolization (TACE) is not chronic injured liver was mainly demon-
effective (Takayasu et al., 1993), and local strated as hypo- or iso-attenuating nodule
ablation therapy sometimes has severe in both arterial and delayed phases of
complications such as needle tract implan- dynamic CT. A hyper-attenuating early
tation in PEI (Ishii et al., 1998), and RFA HCC was only 5% on arterial phase of
related death (de Baere et al., 2003). dynamic CT and 15% on CT hepatic
There is no evidence for early treatment arteriography. Sixty percent of 60 hypo-
of hypo-attenuating nodular lesions to attenuating lesions developed to hyper-
provide benefit to patients with chronic in-hypo (nodule-in-nodule type HCC) or
injured liver. Some researchers do not entirely enhanced lesion (overt HCC) with
agree with early treatment of such lesions cumulative attenuation conversion rates
because good survival could be obtained of 15.8% at 1-year, 44.3% at 2-years,
mainly due to lead-time bias. There is a and 58.7% at 3 years. Thus, careful and
belief that the real critical factor that deter- periodic follow-up helical CT is needed to
mines the prognosis is not a hypo-vascular survey early detection of attenuation con-
lesion but a hyper-vascular advanced HCC version or malignant transformation.
which abruptly appears as a multicentric
fashion of occurrence in the different seg- REFERENCES
ment of the liver during the follow-up of
hypo-vascular lesion. The uncertainness Borzio, M., Fargion, S., Borzio, F., Fracanzani,
A.L., Croce, A.M., Stroffolini, T., Oldani, S.,
of natural history of hypovascular lesion is Cotichini, R., and Roncalli, M. 2003. Impact
another problem; spontaneous disappear- of large regenerative, low grade and high grade
ance was reported with a wide range from dysplastic nodules in hepatocellular carcinoma
5% (Takayasu, in press) to 40% (Tochio development. J. Hepatol. 39: 208–214.
et al., 1994) when studied by CT and ultra- Bruix, J., Sherman, M., Llovet, J.M., Beaugrand, M.,
sonography, respectively. Lencioni, R., Burroughs, A.K., Christensen, E.,
Pagliaro, L., Colombo, M., and Rodes, J. 2001.
The combination of CT hepatic arteri- Clinical management of hepatocellular carcinoma.
ography and CT arterial portography is Conclusions of the Barcelona-2000 EASL confer-
a precise examination to determine the ence. European Association for the Study of the
malignancy of the hypo-vascular lesion Liver. J. Hepatol. 35: 421–430.
258
Chuma, M., Sakamoto, M., Yamazaki, K., Ohta,
T., Ohki, M., Asaka, M., and Hirohashi, S.
2003. Expression profiling in multistage hepa-
tocarcinogenesis: identification of HSP70 as a
molecular marker of early hepatocellular carci-
noma. Hepatology 37: 198–207.
de Baere, T., Risse, O., Kuoch, V., Dromain, C.,
Sengel, C., Smayra, T., Gamal El Din, M.,
Letoublon, C., and Elias, D. 2003. Adverse events
during radiofrequency treatment of 582 hepatic
tumors. AJR Am. J. Roentgenol. 181: 695–700.
Ding, H., Kudo, M., Onda, H., Suetomi, Y., Minami, Y.,
and Maekawa, K. 2001. Hepatocellular carci-
noma: depiction of tumor parenchymal flow
with intermittent harmonic power Doppler US
during the early arterial phase in dual-display
mode. Radiology 220: 349–356.
Hayashi, M., Matsui, O., Ueda, K., Kawamori, Y.,
Kadoya, M., Yoshikawa, J., Gabata, T.,
Takashima, T., Nonomura, A., and Nakanuma, Y.
1999. Correlation between the blood supply and
grade of malignancy of hepatocellular nodules
associated with liver cirrhosis: evaluation by CT
during intraarterial injection of contrast medium.
AJR Am. J. Roentgenol. 172: 969–976.
Hayashi, M., Matsui, O., Ueda, K., Kawamori, Y.,
Gabata, T., and Kadoya, M. 2002. Progression to
hypervascular hepatocellular carcinoma: corre-
lation with intranodular blood supply evaluated
with CT during intraarterial injection of contrast
material. Radiology 225: 143–149.
Hirohashi, S., Ishak, K., Kojiro, M., Wanless, I.,
Theise, N., Tsukuma, H., Blum, H., Deugnier, Y.,
Laurent Puig, P., Fischer, H., and Sakamoto, M.
2000. Hepatocellular carcinoma. In: Hamilton, S.,
Aaltonen, L., editors. World Health Organization
Classification of Tumours: Pathology and
Genetics of Tumours of the Digestive System.
Lyon: International Agency for Research on
Cancer Press. pp. 159–172.
Ikai, I., Arii, S., Ichida, T., Okita, K., Omata, M.,
Kojiro, M., Takayasu, K., Nakanuma, Y.,
Makuuchi, M., Matsuyama, Y., and Yamaoka, Y.
2005. Report of the 16th follow-up survey of pri-
mary liver cancer. Hepatol. Res. 32: 163–172.
Ikeda, K., Saitoh, S., Koida, I., Arase, Y., Tsubota, A.,
Chayama, K., Kumada, H., and Kawanishi, M.
1993. A multivariate analysis of risk factors
for hepatocellular carcinogenesis: a prospective
K. Takayasu
observation of 795 patients with viral and alco-
holic cirrhosis. Hepatology 18: 47–53.
Imamura, H., Matsuyama, Y., Tanaka, E., Ohkubo, T.,
Hasegawa, K., Miyagawa, S., Sugawara, Y.,
Minagawa, M., Takayama, T., Kawasaki, S., and
Makuuchi, M. 2003. Risk factors contributing to
early and late phase intrahepatic recurrence of
hepatocellular carcinoma after hepatectomy.
J. Hepatol. 38: 200–207.
International Working Party. 1995. Terminology of
nodular hepatocellular lesions. Hepatology 22:
983–993.
Ishii, H., Okada, S., Okusaka, T., Yoshimori, M.,
Nakasuka, H., Shimada, K., Yamasaki, S.,
Nakanishi, Y., and Sakamoto, M. 1998. Needle
tract implantation of hepatocellular carcinoma
after percutaneous ethanol injection. Cancer 82:
1638–1642.
Kobayashi, M., Ikeda, K., Hosaka, T., Sezaki, H.,
Someya, T., Akuta, N., Suzuki, F., Suzuki, Y.,
Saitoh, S., Arase, Y., and Kumada, H. 2006.
Dysplastic nodules frequently develop into hepa-
tocellular carcinoma in patients with chronic viral
hepatitis and cirrhosis. Cancer 106: 636–647.
Koda, M., Murawaki, Y., Mitsuda, A., Ohyama, K.,
Horie, Y., Suou, T., Kawasaki, H., and Ikawa, S.
2000. Predictive factors for intrahepatic recur-
rence after percutaneous ethanol injection ther-
apy for small hepatocellular carcinoma. Cancer
88: 529–537.
Kojiro, M. 1997. Pathology of hepatocellular
carcinoma. In: Okuda, K., Tabor, E., editors.
Liver Cancer. New York: Churchill Livingstone.
pp. 165–187.
Kondo, F., Kondo, Y., Nagato, Y., Tomizawa, M.,
and Wada, K. 1994. Interstitial tumour cell
invasion in small hepatocellular carcinoma.
Evaluation in microscopic and low magnification
views. J. Gastroenterol. Hepatol. 9: 604–612.
Lim, J.H., Kim, C.K., Lee, W.J., Park, C.K., Koh,
K.C., Paik, S.W., and Joh, J.W. 2000. Detection
of hepatocellular carcinomas and dysplastic nod-
ules in cirrhotic livers: accuracy of helical CT in
transplant patients. AJR Am. J. Roentgenol. 175:
693–698.
Murakami, T., Kim, T., Takamura, M., Hori, M.,
Takahashi, S., Federle, M.P., Tsuda, K., Osuga,
K., Kawata, S., Nakamura, H., and Kudo, M.
2001. Hypervascular hepatocellular carcinoma:
19. Hepatocellular Carcinoma and Adenomatous Hyperplasia
detection with double arterial phase multi-detec-
tor row helical CT. Radiology 218: 763–767.
Oikawa, T., Ojima, H., Yamasaki, S., Takayama, T.,
Hirohashi, S., and Sakamoto, M. 2005. Multistep
and multicentric development of hepatocellular
carcinoma: histological analysis of 980 resected
nodules. J. Hepatol. 42: 225–229.
Parkin, D.M., Bray, F., Ferlay, J., and Pisani, P.
2001. Estimating the world cancer burden:
Globocan 2000. Int. J. Cancer 94: 153–156.
Sadek, A.G., Mitchell, D.G., Siegelman, E.S.,
Outwater, E.K., Matteucci, T., and Hann, H.W.
1995. Early hepatocellular carcinoma that devel-
ops within macroregenerative nodules: growth
rate depicted at serial MR imaging. Radiology
195: 753–756.
Sakamoto, M., and Hirohashi, S. 1998. Natural
history and prognosis of adenomatous hyperpla-
sia and early hepatocellular carcinoma: multi-
institutional analysis of 53 nodules followed up
for more than 6 months and 141 patients with
single early hepatocellular carcinoma treated by
surgical resection or percutaneous ethanol injec-
tion. Jpn. J. Clin. Oncol. 28: 604–608.
Sakamoto, M., Hirohashi, S., and Shimosato, Y.
1991. Early stages of multistep hepatocarcino-
genesis: adenomatous hyperplasia and early
hepatocellular carcinoma. Hum. Pathol. 22:
172–178.
Takayama, T., Makuuchi, M., Hirohashi, S.,
Sakamoto, M., Okazaki, N., Takayasu, K.,
Kosuge, T., Motoo, Y., Yamazaki, S., and
Hasegawa, H. 1990. Malignant transformation
of adenomatous hyperplasia to hepatocellular
carcinoma. Lancet 336: 1150–1153.
Takayama, T., Makuuchi, M., Hirohashi, S.,
Sakamoto, M., Yamamoto, J., Shimada, K.,
Kosuge, T., Okada, S., Takayasu, K., and
Yamasaki, S. 1998. Early hepatocellular carci-
noma as an entity with a high rate of surgical
cure. Hepatology 28: 1241–1246.
Takayasu, K., Makuuchi, M., Hirohashi, S.,
Okazaki, N., Muramatsu, Y., Moriyama, N.,
Takayama, T., and Hasegawa, H. 1989. Imaging
of adenomatous hyperplastic lesions containing
and not containing hepatocellular carcinoma in
the liver. Jpn. J. Gastroenterol. 86: 2404–2412.
Takayasu, K., Moriyama, N., Muramatsu, Y.,
Makuuchi, M., Hasegawa, H., Okazaki, N.,
259
and Hirohashi, S. 1990. The diagnosis of small
hepatocellular carcinomas: efficacy of various
imaging procedures in 100 patients. AJR Am. J.
Roentgenol. 155: 49–54.
Takayasu, K., Wakao, F., Moriyama, N., Muramatsu, Y.,
Sakamoto, M., Hirohashi, S., Makuuchi, M.,
Kosuge, T., Takayama, T., and Yamazaki, S.
1993. Response of early-stage hepatocellular
carcinoma and borderline lesions to therapeutic
arterial emboli-zation. AJR Am. J. Roentgenol.
160: 301–306.
Takayasu, K., Furukawa, H., Wakao, F., Muramatsu, Y.,
Abe, H., Terauchi, T., Winter, T.C., 3rd,
Sakamoto, M., and Hirohashi, S. 1995a. CT
diagnosis of early hepatocellular carcinoma:
sensitivity, findings, and CT-pathologic correla-
tion. AJR Am. J. Roentgenol. 164: 885–890.
Takayasu, K., Muramatsu, Y., Furukawa, H., Wakao, F.,
Moriyama, N., Takayama, T., Yamasaki, S.,
Sakamoto, M., and Hirohashi, S. 1995b. Early
hepatocellular carcinoma: appearance at CT during
arterial portography and CT arteriography with
pathologic correlation. Radiology 194: 101–105.
Takayasu, K.M., Mizuguchi, Y., Okusaka, T., Shimada,
K., Takayama, T., and Sakamoto, M. 2006. CT
evaluation of the progression of hypoattenuating
nodular lesions in virus related chronic liver dis-
ease. AJR Am. J. Roentgenol. 187: 454–463.
Theise, N.D., Park, Y.N., and Kojiro, M. 2002.
Dysplastic nodules and hepatocarcinogenesis.
Clin. Liver Dis. 6: 497–512.
The liver cancer study group of Japan. 1992. The
General Rules for the Clinical and Pathological
Study of Primary Liver Cancer, 3rd ed. Tokyo:
Kanehara.
The liver cancer study group of Japan. 2003.
General Rules for the Clinical and Pathological
Study of Primary Liver Cancer, 2nd English ed.
Tokyo: Kanehara.
Tochio, H., Tomita, S., Kudo, M., Okabe, Y.,
Kashida, H., Hamada, M., Yamamoto, K., Terai,
Y., Itani, T., Mimura, J., Hirasa, M., Ibuki, Y.,
Komori, H., Orino, A., and Todo, A. 1994.
Clinical tracing of small hypovascular hepatic
nodules associated with chronic liver disease.
Acta Hepatol. Jpn. 35: 333–346.
Tsai, J.F., Jeng, J.E., Ho, M.S., Chang, W.Y.,
Hsieh, M.Y., Lin, Z.Y., and Tsai, J.H. 1997.
Effect of hepatitis C and B virus infection on risk
260
of hepatocellular carcinoma: a prospective study.
Br. J. Cancer 76: 968–974.
Ueda, K., Terada, T., Nakanuma, Y., and Matsui,
O. 1992. Vascular supply in adenomatous hyper-
plasia of the liver and hepatocellular carci-
noma: a morphometric study. Hum. Pathol. 23:
619–626.
Watanabe, H.H., Sakamoto, M., Hasegawa, H.,
Yamazaki, S., Makuuchi, M., Moriyama, N., and
Takayasu, K. 1990. New macroscopic subtype of
K. Takayasu
small hepatocellular carcinoma -common type in
early type- and histological consideration of the
multistep progression of HCC. Acta Hepatol.
Jpn. 31: 1267–1273.
Winter, T.C., 3rd, Takayasu, K., Muramatsu, Y.,
Furukawa, H., Wakao, F., Koga, H., Sakamoto,
M., Hirohashi, S., and Freeny, P.C. 1994. Early
advanced hepatocellular carcinoma: evaluation
of CT and MR appearance with pathologic cor-
relation. Radiology 192: 379–387
20
Hepatocellular Cancer in Cirrhotic
Patients: Radiological Imaging
Francesca Lodato, N. Davies, D. Yu, and Andrew K. Burroughs
INTRODUCTION
Hepatocellular carcinoma (HCC) is the
major cause of death in cirrhotic patients.
It is the fifth most common neoplasm
and the third most common cause of
death related to cancer in the world
(Parkin et al., 2001; Llovet et al., 2003).
Moreover, data on the epidemiology and
natural history of chronic hepatitis C
virus (HCV) infection, suggest that its
frequency has been increasing over the
past 20–30 years in the U.S (El Serag and
Mason, 1999) and in Europe (Deuffic et
al., 1998), and it is the leading cause of
death in patients with cirrhosis in Europe
(Fattovich et al., 1997).
There are several potential therapies for
HCC including resection, liver transplan-
tation and several percutaneous as well
as transarterial ablative techniques, but
their applicability is limited by several
anatomical factors, such as the size and
number of nodules, their location and also
the patients liver reserve. Unfortunately
only a minority of cirrhotic patients diag-
nosed with HCC have tumor amenable to
a potential curative therapy, but all thera-
pies have best results with single nodules
£ 3 cm (Bruix et al., 2001).
This clinical scenario has important
implications for strategies for early diag-
nosis of HCC. However, there are no
randomized controlled trials concerning
surveillance for HCC in cirrhosis, but
only cohort studies. Two studies show a
higher chance of receiving treatment at
an earlier stage (Sangiovanni et al., 2004)
and strongly support current practice of
surveillance in most centres, despite the
lack of formal evidence. The knowledge
that all treatments yield better results
when used for smaller and unifocal HCC
(Bruix et al., 2001) is the linchpin to cur-
rent practice. Therefore, it is important
to be accurate in both the diagnosis and
staging of HCC as this has prognos-
tic significance and helps programming
potential treatments. For both orthotopic
liver transplantation (OLT) and resection
the survival is higher for patients with a
single nodule smaller than 5 cm or with
three nodules each 3 cm in diameter or
less (Adam and Del Gaudio, 2003).
261
262
THE EASL CONSENSUS
STATEMENTS AND AASLD
GUIDELINES
An expert panel on HCC from the European
Association for the Study of the Liver
(EASL) formulated a consensus state-
ment in 2001 to standardize the approach
on HCC diagnosis and treatment (Figure
20.1) (Bruix et al., 2001). These have
evolved, and recent practice guidelines
from the American Association for the
Study of Liver Diseases (AASLD), have
included data from later studies and tech-
nical advances (Figure 20.2) (Bruix and
Sherman, 2005). From the diagnostic point
of view, EASL recommendations define
histological and radiological criteria for the
identification of HCC. Firstly, surveillance
is recommended for cirrhotics in whom
potentially effective treatment for HCC
can be offered if tumors can be detected
early. The available data on tumor growth
suggest that the interval from undetectable
lesions to 2 cm diameter is about 4–12
months (Ebara et al., 2005). Thus, with the
aim of detecting HCC smaller than 3 cm,
the suggested interval for surveillance is 6
Figure 20.1. European association for the study of
the liver consensus algorithm for the imaging path-
way in cirrhotic patients to detect HCC (2001)
F. Lodato et al.
months. Surveillance is recommended also
by AASLD guidelines.
Secondly, the EASL consensus state-
ments categoriz cirrhotic patients on the
basis of the size of the single nodule. For
tumors > 2 cm, diagnosis can be made
using two noninvasive radiological tech-
niques demonstrating arterial hypervas-
cularity (Lim et al., 2000; Nino-Murcia
et al., 2000). When alfa-fetoprotein is
> 400 ng/ml, only one imaging technique
is considered sufficient to confirm the
diagnosis. It should be remembered that
a biopsy for a nodule > 2 cm in a cirrhotic
liver which does not confirm malignancy
cannot be considered definitive; using
only radiological diagnostic techniques
and clinical findings, the positive predictive
value is higher than 95% (Torzilli et al.,
1999; Levy et al., 2001). However there is
no mention of discordant vascular patterns
in the radiological imaging. The American
Association for the Study of Liver Diseases
Guidelines revised these statements, giv-
ing diagnostic importance to venous/portal
wash-out together with arterial hypervas-
cularization, for the diagnosis of HCC.
Moreover, it is suggested that contrast-en-
hanced ultrasonography (US) can be used
as a non-invasive additional diagnostic
technique to spiral Computed Tomography
(CT), Magnetic Resonance Imaging (MRI)
and angiography. However, standard arte-
riography and CT arteriography are not
recommended for routine diagnosis. In
cases of an atypical vascular pattern, a
liver biopsy is recommended but no men-
tion is made of the diagnostic accuracy
of histology for these nodules or the
applicability of biopsy, e.g., difficult loca-
tion, poor coagulation. Based on AASLD
Guidelines with nodules > 2 cm, only one
dynamic technique showing a typical
20. Hepatocellular Cancer in Cirrhotic Patients: Radiological Imaging 263

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-
tern with alfa-fetoprotein > 200 ng/ml is
sufficient to make a diagnosis of HCC.
When the nodule is less than 2 cm in
diameter, diagnosis is more difficult and
may require needle liver biopsy. For nod-
ules smaller than 1 cm, the EASL consensus
recommends serial US examinations every
3 months until the nodule size exceeds
1 cm, and the same approach is suggested
by AASLD guidelines. When the nodule
is between 1 and 2 cm, EASL consensus
suggests proceeding to needle liver biopsy
to confirm the diagnosis, irrespective of the
vascular profile at imaging. The American
Association for the Study of Liver Diseases
Guidelines have refined this, suggesting
to perform two imaging techniques, and
to proceed to liver biopsy only in cases of
discordant vascular pattern or atypical vas-
cular pattern between the two techniques.
However, no mention is made of the prob-
lems of biopsy, as detailed above, with the
added consideration of operator error or
difficulty of being accurate in placing the
biopsy needle in a lesion < 2 cm.
SURVEILLANCE FOR
HCC: RADIOLOGICAL
TECHNIQUES
Radiological imaging of cirrhotic patients is
not easy and requires particular training and
264
expertise of the operator. Hepatocellular
carcinoma can exhibit different echo-
graphic features reflecting the variability of
the macroscopic and microscopic findings.
Cirrhosis itself is a disease that produces
pathological changes, such as regenerative
and dysplastic nodules (DN), so that some-
times it is impossible to discriminate these
from a small HCC. During the evolution
from benign regenerative nodule, to low
and high grade dysplastic nodule and finally
to HCC, the portal venous supply progres-
sively disappears so that new arterial ves-
sels constitute the main blood supply for
the HCC. Therefore, differences in focal
lesion perfusion can help to characterize
the nodules in a cirrhotic liver (Lencioni
et al., 1996). Following the detection of
a suspicious nodule ultrasonographically,
both EASL statements and AASLD guide-
lines require confirmation using dynamic
techniques as contrast-enhanced US, spiral
CT or dynamic magnetic resonance (MR).
However the single best technique fol-
lowing ultrasound is not well established
(Fung et al., 2004).
ULTRASONOGRAPHY
(US), DOPPLER-
ULTRASONOGRAPHY,
POWER DOPPLER-
ULTRASONOGRAPHY AND
CONTRAST-ENHANCED
ULTRASONOGRAPHY
Ultrasonography is a rapid and non inva-
sive evaluation although it may be hin-
dered by factors such as the patients body
mass or intestinal interposition.
Hepatocellular carcinoma surveillance
studies based on the use of US showed
F. Lodato et al.
variable sensitivity from 35% to 84%
depending mostly on the expertise of
the operator and the equipment avail-
able (Peterson and Baron, 2001). When
the gold standard for diagnosis was the
explanted liver, the sensitivity of US was
quite disappointing ranging from 33% to
72% (Kim et al., 2001; Rode et al., 2001;
Gambarin-Gelwan et al., 2000; Teefey
et al., 2003) in part because the livers
were explanted months after diagnosis.
Only one article (Teefey et al., 2003)
provides information on the size of the
false negative nodules, showing that all
lesions which were missed were smaller
than 2.5 cm. Ultrasonography specificity
ranged from 92% to 100%.
At US examination, HCC may appear as
round or oval lesions, small tumors usually
have smooth and sharp margins. Nodular-
type HCC with extranodular growth and
multinodular-type HCC may present with
irregular margins. Small HCC may be
hypoechoic, hyperechoic or even isoechoic
with the surrounding liver. Approximately
75% of small HCC (< 2 cm) are hypoe-
choic. The hyperechoic appearance usu-
ally reflects fatty changes of the tumor and
its differential diagnosis may be difficult
with respect to a hemangioma. As HCC
nodules grow they tend to develop hyp-
oechoic peripheral rims and become more
heterogeneous.
The use of Doppler US or power Doppler
may help in the diagnosis. Hepatocellular
carcinoma usually has a florid arterial
vascularization, and this is seen as int-
ralesional flow signal with an arterial
spectrum. Large HCC may have a basket
pattern with a fine blood network sur-
rounding the nodule and flowing into it.
Doppler examination shows the presence
of pulsatile flow with arterial waveform
20. Hepatocellular Cancer in Cirrhotic Patients: Radiological Imaging
and high resistive and pulsality indexes.
Large regenerative nodules and DN do not
have arterial intralesional vessels, so these
findings support the diagnosis of HCC.
The problem is that in small HCC the
sensitivity of Doppler techniques is low
and a typical pulsatile flow arterial aspect
can be shown in less than 50% of nodules
(Lencioni et al., 1996).
The newest microbubble contrast agents
and the development of specific contrast
techniques have opened a new era in liver
ultrasonography. The AASLD Guidelines
consider contrast enhanced US as a con-
firmatory technique when a suspicion of
HCC is made from standard US. Images
are based on non linear acoustic effects
of bubbles enhancing grey scale, with
high contrast and spatial resolution. This
technique improves the accuracy of US
in the diagnosis of focal liver lesions
(Lencioni et al., 2002; Wen et al., 2004).
Contrast agents are safe and well tolerated.
The advent of second generation contrast
agents together with low mechanical index
scanning have led the European Federation
for Ultrasound in Medicine and Biology, to
define the guidelines for its use (Albrecht
et al., 2004). Contrast-enhanced US is rec-
ommended to characterize any suspected
lesion detected at baseline US in cirrho-
sis. At contrast enhanced-US examination,
HCC typically shows a strong enhancement
in the arterial phase and a rapid washout
during the portal and late phases in which
the nodule remains iso or hypoechoic.
Regenerative nodules and DN usually do
not show any uptake during the arterial
phase and their appearance resembles the
surrounding liver. However, as for non
contrast-enhanced US, the sensitivity of
using ultrasound contrast agents decreases
with smaller lesions. Defining spiral CT as
265
the gold standard to identify hypervascular
lesions, the sensitivity of contrast enhanced
US was 97% for nodules > 3 cm in diam-
eter, 92% for nodules between 2 and 3 cm,
and 67% for nodules < 1 cm (Gaiani et al.,
2004). Another disadvantage is that only
one lesion or lesions close together can be
examined at one time. Therefore, a cost-
effective algorithm, of which technique to
use following routine US still needs to be
determined (Fung et al., 2004).
SPIRAL COMPUTED
TOMOGRAPHY
The introduction of spiral CT in the early
1990s has dramatically improved the accu-
racy of CT in the diagnosis of HCC.
Computed Tomography data acquisition
speed is greatly increased so that it is pos-
sible to scan the whole liver during a single
breath hold, allowing a satisfactory evalu-
ation of different contrast enhancement
phases. Moreover, 3-D reconstructions are
available with faster data acquisition. The
standard spiral CT evaluation for HCC
should include unenhanced and contrast
enhanced phase of the arterial phase (25–
30 s after the contrast injection), of the
portal venous phase (70–80 s after contrast
injection), and a late phase (180–210 s
after contrast injection). As HCC is a
vascular tumor that receives almost all
its blood supply from arterial vessels, it
is well seen during the arterial phase; the
rest of the liver receives the majority of
blood supply from the portal vein, and will
not enhance until the contrast reaches the
portal system. Fast scans during the dif-
ferent phases of contrast flow through the
liver help in increasing the differentiation
between the tumor and the liver.
266
The technique can be improved by
using multidetector-row scan technology
which due to the rapid acquisition time,
allows visualization of two different arte-
rial phases, early and late. The first arte-
rial phase is a CT arteriography that may
be of some help depending on the aim of
the examination. Some studies show that a
double arterial phase did not improve the
detection rate of HCC compared to a single
arterial phase (Laghi et al., 2003; Francis
et al., 2003), whereas it did improve with
the addition of a third arterial phase as the
middle phase is claimed to be the most
sensitive (Murakami et al., 2003).
However, despite the tremendous tech-
nical progress in this area, CT is still
insensitive for detecting small HCC. When
considering studies on explanted livers,
the sensitivity of spiral CT ranged between
54% and 79% (Lim et al., 2000; Rode et al.,
2001; Burrel et al., 2003; Valls et al.,
2004) mostly due to the lack of detection
of smaller lesions (< 2 cm).
Another important issue is the specificity
of this technique. It is well recognized that
a number of benign lesions can simulate
HCC on CT. The advent of arterial phase
imaging brought about a series of new
“HCC pseudolesions” that could be mis-
taken for enhancing tumors. Specificity
of spiral CT ranges between 93% and
97% in explant studies (Rode et al., 2001;
Lim et al., 2000; Burrel et al., 2003).
Among the hypervascular lesions that can
simulate HCC are pseudoaneurysms, arte-
riovenous malformations, hemangiomas
and dysplastic nodules; while hypovascu-
lar lesions simulating HCC are hemangi-
omas, confluent fibrosis and regenerative
nodules. Moreover, cirrhotics may have
transient focal enhancement of the liver
during the arterial phase, mostly due to
F. Lodato et al.
arterial-portal shunting. However these
images are often peripheral, wedge shaped
and not round and sometimes disappear
in follow up examination. These features
help to exclude HCC. The portal phase
acquisition is particularly important as
HCCs have a rapid wash out during this
phase, whereas hemangiomas or vascular
malformation show contrast enhancement.
The most problematic lesions mimicking
HCC are dysplastic nodules that may show
arterial phase enhancement, a problem that
cannot be resolved even when using the
most modern imaging technologies.
Another technique that could improve
CT performance is the injection of contrast
through the hepatic artery while perform-
ing CT. This method has shown 66% more
tumors in patients with HCC compared to
triphasic helical contrast CT (Oliver et al.,
1997), but its use is limited by cost and
the mortality and morbidity related to the
angiographic procedure. However, it can
be easily performed during chemoemboli-
zation or embolization therapy when the
catheter in the hepatic artery is needed to
perform the therapeutic procedure.
Lipiodol CT is another technique that
requires the administration of contrast
media through the hepatic artery. It is lim-
ited by the need to perform angiography
and requires the patient to be rescanned
at least 14 days after the injection. This
allows the normal liver to wash-out the
contrast while the HCC retains it for a
longer period. However retention of lipi-
odol is not specific for HCC alone, with
only one study reporting that lipiodol-
CT is more specific than spiral CT in
detecting small, less differentiated nodules
(Nakayama et al., 2001). Moreover, it can
hinder assessment of residual vascularity
following loco-regional therapies.
20. Hepatocellular Cancer in Cirrhotic Patients: Radiological Imaging
MAGNETIC RESONANCE
IMAGING
The use of MRI in the diagnosis of HCC
has expanded greatly over the last few years
due to the development of new hardware
and software that have allowed the rapid
acquisition of very detailed images, reduc-
ing the artefact secondary to respiratory
activity. Volumetric sequences allow the
reconstruction of 3-D images with very high
spatial and temporal resolution. Moreover,
the introduction of liver-specific contrast
agents into clinical practice, has enabled a
perfusional investigation of liver nodules.
The standard examination for the detection
of HCC should include:
 T1-weighted fat spoiled gradient-echo
sequences with fat suppression
 Respiration triggered or breath hold T2
weighted fat spin echo sequences with
fat suppression
 Serial dynamic T1-weighted fast spoiled
gradient-echo sequences after injection
of a gadolinium chelate (Lencioni et al.,
2004)
Once again, as HCC is a very heterogeneous
tumor, these characteristics are reflected in
an MRI examination. For example, HCC
may have variable stromal and intracellu-
a b
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
267
lar components, such as fat, glycogen, or
ions which may cause different appearance
in T1-weighted and T2-weighted images.
The signal may appear hypo, iso or hyper-
intense in T1-weighted images, and iso or
hyperintense in T2-weighted images.
Gadolinium chelates are distributed in
extracellular space and their enhancing
effect depends on the blood supply dur-
ing the early dynamic phase and on the
interstitial matrix during the delayed phase.
The new fast MRI techniques allow the
acquisition of multiple intravenous contrast
enhanced dynamic images of the liver,
including the arterial phase, the portal peak
phase and an delayed phase. As already
mentioned, HCC is usually a hypervascular
tumor, so that most HCC are characterized
by an enhancement in the arterial phase,
and rapid washout during the portal phase
(Figures 20.3a–c). Regenerative nodules
and DNs are usually not hypervascular.
As described for spiral CT, some benign
lesions, especially high grade DN, may
have increased arterial blood supply that
may render diagnosis difficult, and in cir-
rhotic livers there are often arterioportal
shunts that may produce false positive
interpretations (Ito et al., 2004).
Dynamic MRI can also be performed
by using other tissue specific agents, for
c
268
example, superparamagnetic iron oxide
(SPIO) particles are taken up by Kuppfer
cells of the reticuloendothelial system
(RES-targeted contrast agents) which are
absent in HCC. They produce a decrease
in the intensity of the liver tissue espe-
cially on T2-weighted images as HCC
nodules show little or no uptake of the
contrast. Gadolinium and SPIO agents
could be coupled (Ward and Robinson,
2002), but the usefulness of this technique
is debated (Halavaara et al., 2002).
Despite substantial progress in MRI still 1995), although foci of dysplasia may
fails in detecting small HCC; the sensitiv-
ity of dynamic MRI ranges from 33% to
77% in explant studies (Rode et al., 2001;
Burrel et al., 2003; Krinsky et al., 2002;
Teefey et al., 2003) and 78% when using
dynamic imaging plus RES-targeted agents
(Bhartia et al., 2003). When considering
only HCC smaller than 1 cm, sensitivity
drops to 4–71%. Nevertheless, a study
showed that dynamic MRI is superior to
helical CT in the detection of nodules
between 1 and 2 cm (Burrel et al., 2003).
In this study, 50 patient with HCC under-
going liver transplantation were analyzed.
Twenty-nine of them had HCC; the patho-
logic examination was the gold standard.
On a per nodule basis, sensitivity of MRI
was superior to triphasic helical CT (76%
vs. 61%, p = 0.001). Sensitivity for detec-
tion of additional nodules decreased with
size and was superior to CT for nodules 10
to 20 mm (84% vs. 47%, p = 0.016).
THE PROBLEM OF SMALL
NODULES
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
F. Lodato et al.
the nature of these small nodules and,
most importantly, the relationship between
these nodules and the development of
HCC is uncertain. It is well accepted that
cirrhosis itself represents a preneoplastic
condition, but in carcinogenesis the exact
point in time between a lesion that still
has some growth regulation and one that
has lost growth regulation is difficult to
define. The only morphological change
that is considered as intermediate to HCC
is dysplasia (International Working Party,
take place within benign regenerative nod-
ules. Moreover, it is not clear whether all
dysplastic nodules will inevitably evolve
to HCC. Some reports suggest that high
grade dysplastic nodules (HGDN) should
be considered as HCC precursors in most
cases (Borzio et al., 2003). In clinical
practice a histological diagnosis of HGDN
poses several difficulties in the interpre-
tation of this finding, as a conservative
therapeutic approach could affect the out-
come. More difficulties arise considering
that in the context of a dysplastic nodule,
foci of HCC can exist but biopsy may
miss the sampling of the malignant por-
tion of the lesion. In the setting of cir-
rhosis the picture is also complicated by
vascular lesions that are not malignant
e.g., arteroportal shunts or other perfusion
abnormalities.
The introduction of radiological tech-
niques that study the perfusional pattern
of nodules has enormously improved the
accuracy of diagnosis, and the vascular
study of lesion has become the main
parameter used for the radiological diag-
nosis of HCC. Consequently the use of
20. Hepatocellular Cancer in Cirrhotic Patients: Radiological Imaging
considering that liver biopsy can also be
related to a 0–5% of seeding (Stigliano
and Burroughs, 2005) even if associ-
ated with percutaneous ablative tech-
niques (Stigliano and Burroughs, 2005;
Ebara et al., 2005; Shiina et al., 2005).
Moreover, when a biopsy is negative for
HCC its presence cannot be excluded,
as there is a 10% risk of false negativity
especially in the presence of small nodules
(< 2 cm) (Caturelli et al., 2004).
Bolondi et al. (2005) showed that 78.5%
of nodules between 1 and 3 cm diameter
with only one positive imaging technique
in terms of hypervascularity suggestive for
HCC were found to be true HCC at biopsy.
When looking at the rate of hypervascular-
ity demonstrated by two separate radio-
logical techniques leading to the diagnosis
of HCC, this rate was significantly lower
in the group of nodules between 1 and
2 cm than in the group of nodules between
2 and 3 cm (44% vs. 84%, p = 0.001).
Amongst the group of nodules between
1 and 2 cm with discordant detection of
hypervascularity by two techniques, a defi-
nite diagnosis of HCC by histology was
made in 48%. The American Association
for the Study of Liver Diseases Guidelines
recommend performing liver biopsy in
cases of discordant imaging technique for
nodules > 1 and < 2 cm. Based on the data
by Bolondi et al. (2005) this would be
the case for almost half of these nodules
as 44% of small nodules had discordant
imaging hypervascularity. The problem
of hypovascular nodules (20% of cases
in this study) is another important one;
all cases occurred in the group of nodules
between 1 and 2 cm, and 36% were later
proven to be HCC. These data support the
approach suggested by the most recent
AASLD guidelines, which recommend
269
performing biopsy in all lesions > 1 cm and
< 2 cm without a typical vascular pattern.
However, these recommendations for
biopsy open up a problem of the clinical
applicability to perform biopsy for such
small nodules, as three factors affect this:
the reliability of obtaining tissue from the
right place, the uncertainty of the diagno-
sis histologically and the risk of seeding.
The consequences of seeding can be dev-
astating. Surgery may no longer be fea-
sible because of the spread of neoplastic
disease, which has been reported or seed-
ing may occur after potentially definitive
therapy including liver transplantation. Its
occurrence is reported in 0–5.1% of cases
following liver biopsy but may also occur
following percutaneous treatment for
HCC. Percutaneous ethanol injection (PEI)
has a risk of 0.13–6.8%. Radiofrequency
Ablation (RFA) has a reported risk of
0–3%, that increases to 12.5% in the
case of superficial nodules (Stigliano and
Burroughs, 2005). Prospective studies of
loco-regional therapy that might be con-
sidered for reducing the risk of seeding
show a seeding rate of 1.96% with PEI
(Ebara et al., 2005; Shiina et al., 2005) and
2.55% with RFA (Shiina et al., 2005).
In the setting of liver transplantation, the
need of a precise diagnosis can be dictated
by the allocation system adopted. In the
USA and in some European liver trans-
plantation centres, the Model for End-
Stage Liver Disease-HCC scoring system
is used, in which a single nodule ≥ 2 cm
or multiple nodules increases the priority
on the waiting list (Sharma et al., 2004;
Hayashi et al., 2004). Here the issue is
the need of a precise diagnosis to justify
increased priority, regardless of baseline
liver function. It is interesting to note that
explant studies in the USA have shown the
270
lack of histological confirmation in one
third of patients diagnosed with HCC pre-
transplant (Wiesner et al., 2004), which
appears higher than European studies.
If the diagnosis will not change the
priority for liver transplantation, nodules
between 1 and 2 cm without confirmatory
hypervascularity by two separate tech-
niques can be treated by locoregional
treatment even without “a priori” histolog-
ical diagnosis, associated with performing
biopsy at the time of the procedure. It is
not clear if the seeding risk is increased by
performing both biopsy and locoregional
therapy at the same time but this could be
possible. For single nodules smaller than
3 cm among various techniques PEI may
be preferable, with fewer major compli-
cations, and possibly less risk of seeding
(Ebara et al., 2005; Shiina et al., 2005);
PEI is also cheaper than radiofrequency
ablation (RFA) and the size of the nee-
dle is smaller. However, beyond 1 year a
recent randomized trial has shown higher
survival rates with RFA when compared to
PEI (Shiina et al., 2005).
Each centre should follow an established
protocol and audit their decision making
as there is insufficient evidence to select
one approach over another.
CONCLUSIONS
Current imaging techniques have sub-
stantially improved the diagnostic
accuracy for HCC in cirrhotic patients,
reducing the need of liver biopsy and its
attendant immediate complications, as
well as the risk of seeding. This is mostly
due to the development of dynamic tech-
niques that study the vascular pattern of
nodules. However, the performance of
F. Lodato et al.
these techniques is poor when nodules
are smaller than 2 cm, which may also
be hypovascular.
In the era of liver transplantation, the
need for a correct diagnosis has become
imperative in most cases particularly for
prioritisation of patients for transplanta-
tion. Although liver biopsy may not be
avoided there is a risk of seeding and this
must be taken into account during the
diagnostic pathway. Diagnostic algorithms
need to be worked out considering that
single nodules < 2 cm have in general a far
better prognosis and virtually no risk of
recurrence after liver transplantation.
In cirrhotic patients nodules suspected
of being malignant that are smaller than
1 cm should be followed up at short term
intervals of no more than 3 months, as
the use of dynamic techniques is only
sufficiently accurate when the nodules
are larger. With nodules between 1 and
2 cm, if two dynamic techniques are not
concordant in demonstrating hypervascu-
larity, liver biopsy should be performed
when a precise diagnosis would change
the immediate therapeutic option for the
patient, but could be delayed or avoided
depending on the timing of potential ther-
apy such as liver transplantation (e.g., list
and reimage). However, another therapeu-
tic option is to treat the nodule as if it is
HCC, even in the absence of a histological
diagnosis, as the risk of seeding may not
be increased by adding ablative therapy
to the biopsy (Stigliano and Burroughs,
2005) at least for PEI. For nodules larger
than 2 cm, a single dynamic technique is
sufficient if the hypervascular pattern is
typical, to diagnose HCC. In the pres-
ence of an atypical vascular pattern the
presence of alfa-fetoprotein > 200 ng/ml is
diagnostic.
20. Hepatocellular Cancer in Cirrhotic Patients: Radiological Imaging
REFERENCES
Albrecht, T., Blomley, M., Bolondi, L., Claudon, M.,
Correas, J.M., Cosgrove, D., Greiner, L., Jager K.,
Jong, N.D., Leen, E., Lencioni, R., Lindsell, D.,
Martegani, A., Solbiati, L., Thorelius, L.,
Tranquart, F., Weskott, H.P., Whittingham, T.,
and EFSUMB Study Group. 2004. Guidelines
for the use of contrast agents in ultrasounds.
Ultrashall. Med. 25: 249–256.
Adam, R., and Del Gaudio, M. 2003. Evolution
of liver transplantation for hepatocellular carci-
noma. J. Hepatol. 39: 888–895.
Bhartia, B., Ward, J., Guthrie, J.A., and Robinson,
P.J. 2003. Hepatocellular carcinoma in cirrhotic
livers: double-contrast thin section – MR imag-
ing with pathologic correlation of explanted tis-
sue. AJR Am. J. Roentgenol. 180: 577–584.
Bolondi, L., Gaiani, S., Celli, N., Golfieri, R.,
Grigioni, W.F., Leoni, S., Venturi, A.M., and
Piscaglia, F. 2005. Characterization of small
nodules in cirrhosis by assessment of vascular-
ity: the problem of hypovascular hepatocellular
carcinoma. Hepatology 42: 27–34.
Borzio, M., Fargion, S., Borzio, F., Fracanzani,
A.L., Croce, A.M., Stroffolini, T., Oldani, S.,
Cotichini, R., and Roncalli, M. 2003. Impact
of large regenerative, low grade and high grade
dysplastic nodules in hepatocellular carcinoma
development. J. Hepatol. 39: 209–214.
Bruix, J., Sherman, M., Llovet, J.M., Beaugrand, M.,
Lencioni, R., Burroughs, A.K., Christensen, E.,
Pagliaro, L., Colombo, M., Rodes J, and EASL
Panel of Experts on HCC. 2001. Clinical manage-
ment of hepatocellular carcinoma. Conclusion
of the Barcelona – 2000 EASL Conference. J.
Hepatol. 35: 421–430.
Bruix, J., and Sherman, M. 2005. Management
of hepatocellular carcinoma. Hepatology 42:
1208–1236.
Burrel, M., Llovet, J.M., Ayuso, C., Iglesias, C.,
Sala, M., Miquel, R., Caralt, T., Ayuso, J.R.,
Sole, M., Sanchez, M., Bru, C., Bruix, J., and
Barcelona Clinic Liver Cancer Group. 2003.
MRI angiography is superior to helical CT for
detection of HCC prior to liver transplanta-
tion: an explant correlation. Hepatology 38:
1034–1042.
Caturelli, E., Solmi, L., Anti, M., Fusilli, S., Roselli,
P., Andriulli, A., Fornari, F., Del Vecchio Blanco,
271
C., and De Sio, I. 2004. Ultrasound guided fine
needle biopsy of early hepatocellular carcinoma
complicating liver cirrhosis: a multicentre study.
Gut 53: 1356–1362.
Deuffic, S., Poynard, T., Buffat, L., and Valleron,
A.J. 1998. Trends in primary liver cancer. Lancet
17;351: 214–215.
Ebara, M., Okabe, S., Kita, K., Sugiura, N., Fukuda,
H., Yoshikawa, M., Kondo, F., and Saisho, H.
2005. Percutaneous ethanol injection for small
hepatocellular carcinoma: therapeutic efficacy
based on 20-year observation. J. Hepatol. 43:
377–380.
El Serag, H.B., and Mason, A.C. 1999. Rising inci-
dence of hepatocellular carcinoma in the United
States. N. Engl. J. Med. 340: 745–750.
Fattovich, G., Giustina, G., Degos, F., Tremolada, F.,
Diodati, G., Almasio, P., Nevens, F., Solinas, A.,
Mura, D., Brouwer, J.T., Thomas, H., Njapoum, C.,
Casarin, C., Bonetti, P., Fuschi, P., Basho, J.,
Tocco, A., Bhalla, A., Galassini, R., Noventa, F.,
Schalm, S.W., and Realdi, G. 1997. Morbidity
and mortality in compensated cirrhosis type C:
a retrospective follow-up study of 384 patients.
Gastroenterology 112: 463–472.
Francis, I.R., Cohan, R.H., McNulty, N.J., Platt, J.F.,
Korobkin, M., Gebremariam, A., and Ragupathi,
K.I. 2003. Multidetector CT of the liver and
hepatic neoplasm: effect of multiphasic imaging
on tumor conspicuity and vascular enhancement.
AJR Am. J. Roentgenol. 180: 1217–1224.
Fung, K.T., Li, F.T., Raimondo, M.L., Maudgil, D.,
Mancuso, A., Tibballs, J.M., Watkinson, A.A.,
Patch, D., and Burroughs, A.K. 2004. Systematic
review of radiological imaging for hepatocellu-
lar carcinoma in cirrhotic patients. Br. J. Radiol.
77: 633–640.
Gaiani, S., Celli, N., Piscaglia, F., Cecilioni, L.,
Losinno, F., Giangregorio, F., Mancini, M., Pini,
P., Fornari, F., and Bolondi, L. 2004. Usefulness
of contrast enhanced computed sonography in
the assessment of hepatocellular carcinoma
hypervascular at spiral computed tomography. J.
Hepatol. 41: 421–426.
Gambarin-Gelwan, M., Wolf, D.C., Shapiro, R.,
Schwartz, M.E., and Min, A.D. 2000. Sensitivity
of commonly available screening tests in detect-
ing hepatocellular carcinoma in cirrhotic
patients undergoing liver transplantation. Am.
J. Gastroenterol. 95: 1535–1538.
272
Halavaara, J., Tervahartiala, P., Isoniemi, H., and
Hockerstedt, K. 2002. Efficacy of sequential use
of superparamgnetic iron oxide and gadolinium
in liver MR imaging. Acta. Radiol. 43: 180–185.
Hayashi, P. H., Trotter, J.F., Forman, L., Kugelmas,
M., Steinberg, T., Russ, P., Wachs, M., Bak, T.,
Kam, I., and Everson, G.T. 2004. Impact of pre-
transplant diagnosis of hepatocellular carcinoma
on cadveric liver allocation in the era of MELD.
Liver Transplant. 10: 42–48.
International Working Party. 1995. Terminology of
nodular hepatocellular lesions. Hepatology 22:
983–989.
Ito, K., Fujita, T., Shimizu, A., Koike, S., Sasaki, K.,
Matsunaga, N., Hibino S., and Yuhara, M. 2004.
Multiarterial phase dynamic MRI of small early
enhancing hepatic lesions in cirrhosis or chronic
hepatitis: differentiating between hypervascular
hepatocellular carcinomas and pseudolesions.
AJR Am. J. Roentgenol. 183: 699–705.
Kim, C.K., Lim, J.H., and Lee, W.J. 2001. Detection
of hepatocellular carcinoma and dysplastic nod-
ules in cirrhotic liver: accuracy of ultrasonogra-
phy in transplant patients. J. Ultrasound. Med.
20: 99–104.
Krinsky, G.A., Lee, V.S., Theise, N.D., Weinreb,
J.C., Morgan, G.R., Diflo, T., John, D.,
Teperman, L.W., and Goldenberg, A.S. 2002.
Transplantation for hepatocellular carcinoma
and cirrhosis: sensitivity of magnetic resonance
imaging. Liver Transplant. 8: 1156–1164.
Laghi, A., Iannaccone, R., Rossi, P., Carbone, I.,
Ferrari, R., Mangiapane, F., Nofroni, I., and
Passariello, R. 2003. Hepatocellular carcinoma:
detection with triple-phase multidetector row
helical CT in patients with chronic hepatitis.
Radiology 226: 543–549.
Lencioni, R., Cioni, D., Crocetti, L., Donati, F.,
Franchini, C., Giusti, S., and Bartolozzi, C.
2002. Ultrasound imaging of focal liver lesion
with a second generation contrast agent. Acad.
Radiol. 9: S 371–374.
Lencioni, R., Cioni, D., Crocetti, L., Della Pina, C., and
Bartolozzi, C. 2004. Magnetic resonance imag-
ing of liver tumors. J. Hepatol. 40: 162–171.
Lencioni, R., Pinto, F., Armillotta, N., and
Bartolozzi, C. 1996. Assessment of tumor vas-
cularity in hepatocellular carcinoma: compari-
son of power-doppler US and color-doppler US.
Radiology 201: 353–358.
F. Lodato et al.
Levy, I., Greig, P.D., Gallinger, S., Langer, B., and
Sherman, M. 2001. Resection of hepatocellular
carcinoma without preoperative tumor biopsy.
Ann. Surg. 234: 206–209.
Lim, J.H., Kim, C.K., Lee, W.J., Park, C.K., Koh,
K.C., Paik, S.W., and Joh, J.W. 2000. Detection
of hepatocellular carcinomas and dysplastic nod-
ules in cirrhotic livers: accuracy of helical CT in
transplant patients. AJR Am. J. Roentgenol. 175:
693–698.
Llovet, J.M., Burroughs, A, and Bruix, J. 2003.
Hepatocellular carcinoma. Lancet 363: 1907–1917.
Murakami, T., Kim, T., Kawata, S., Kanematsu, M.,
Federle, M.P., Hori, M., Okada, A., Kumano, S.,
Sugihara, E., Tomoda, K., and Nakamura, H.
2003. Evaluation of optimal timing of arterial
phase imaging for the detection of hypervas-
cular hepatocellular carcinoma by using triple
arterial phase imaging with multidetector-row
helical computed tomography. Invest. Radiol.
38: 497–503.
Nakayama, A., Imamura, H., Matsuyama, Y., Kitamura,
H., Miwa, S., Kobayashi, A., Miyagawa, S., and
Kawasaki, S. 2001. Value of lipiodol computed
tomography and digital subtraction angiography
in the era of helical biphasic computed tomogra-
phy as preoperative assessment of hepatocellular
carcinoma. Ann. Surg. 234: 56–62.
Nino-Murcia, M., Olcott, E.W., Jeffrey, R.B.,
Lamm, R.L., Beaulieu, C.F., and Jain, K.A.
2000. Focal liver lesions: pattern-based classifi-
cation scheme for enhancement at arterial phase
CT. Radiology 215: 746–751.
Oliver, J.H., Baron, L.L., and Carr, B.E. 1997.
CT imaging of hepatocellular carcinoma:
CT-arteriography versus triphasic helical con-
trast CT. Radiology 205: 144 A.
Parkin, D.M., Bray, F., Ferlay, J., and Pisani, P. 2001.
Estimating the world cancer burden: Globocan
2000. Int. J. Cancer 15;94: 153–156.
Peterson, M.S., and Baron, R.L. 2001. Radiologic
diagnosis of hepatocellular carcinoma. Clin.
Liver Dis. 5: 123–144.
Rode, A., Bancel, B., Douek, P., Chevallier, M.,
Vilgrain, V., Picaud, G., Henry, L., Berger, F.,
Bizollon, T., Gaudin, J.L., and Ducerf, C. 2001.
Small nodule detection in cirrhotic livers: evalu-
ation with US, spiral CT, and MRI and correla-
tion with pathologic examination of explanted
liver. J. Comput. Assist. Tomogr. 25: 327–336.
20. Hepatocellular Cancer in Cirrhotic Patients: Radiological Imaging
Sangiovanni, A., Del Ninno, E., Fasani, P., De
Fazio, C., Ronchi, G., Romeo, R., Morabito,
A., De Franchis, R., and Colombo, M. 2004.
Increased survival of cirrhotic patients with a
hepatocellular carcinoma detected during sur-
veillance. Gastroenterology 126: 1005–1014.
Sharma, P., Balan, V., Hernandez, J.L., Harper,
A.M., Edwards, E.B., Rodriguez-Luna, H.,
Byrne, T., Vargas, H.E., Mulligan, D., Rakela, J.,
and Wiesner, R.H. 2004. Liver transplantation
for hepatocellular carcinoma: the MELD impact.
Liver Transpl. 10: 36–41.
Shiina, S., Teratani, T., Obi, S., Sato, S., Tateishi, R.,
Fujishima, T., Ishikawa, T., Koike, Y., Yoshida, H.,
Kawabe, T., and Omata, M. 2005. A randomized
controlled trial of radiofrequency ablation with
ethanol injection for small hepatocellular carci-
noma. Gastroenterology 129: 122–130.
Stigliano, R., and Burroughs, A.K. 2005. Should
we biopsy each liver mass suspicious for HCC
before liver transplantation? – No, please don’t.
J. Hepatol. 43: 563–568.
Teefey, S.A., Hildeboldt, C.C., Dehdashti, F.,
Siegel, B.A., Peters, M.G., Heiken, J.P.,
Brown, J.J., McFarland, E.G., Middleton, W.D.,
Balfe, D.M., and Ritter, J.H. 2003. Detection of
primary hepatic malignancyn in liver transplant
273
candidates: Prospective comparison of CT, MR
imaging, US and PET. Radiology 226: 533–542.
Torzilli, G., Minagawa, M., Takayama, T., Inoue, K.,
Hui, A.M., Kubota, K., Ohtomo, K., and
Makuuchi, M. 1999. Accurate preoperative eval-
uation of liver mass lesions without fine-needle
biopsy. Hepatology 30: 889–893.
Valls, C., Cos, M., Figueras, J., Andia, E., Ramos, E.,
Sanchez, A., Serrano, T., and Torras, J. 2004.
Pretransplantation diagnosis and staging of
hepatocellular carcinoma in patients with cir-
rhosis: value of dual phase helical CT. AJR Am.
J. Roentgenol. 182: 1011–1017.
Ward, J., and Robinson, P.J. 2002. How to detect
hepatocellular carcinoma in cirrhosis. Eur
Radiol. 12: 2258–2272.
Wen, Y.L., Kudo, M., Zheng, R.Q., Ding, H.,
Zhou, P., Minami, Y., Chung, H., Kitano,
M., Kawasaki, T., and Maekawa, K. 2004.
Charaterzation of hepatic tumors: value of
contrast enhanced coded phase-inversion har-
monic angio. AJR Am. J. Roentgenol. 182:
1019–1026.
Wiesner, R.H., Freeman, R.B., and Mulligan, D.C.
2004. Liver transplantation for hepatocellular
cancer: the impact of the MELD allocation
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
Hepatocellular carcinoma (HCC) is typically a
hypervascular tumor and anti-angiogenesis
therapy may be effective for the treatment
of HCC. Thalidomide has been shown to
inhibit angiogenesis induced by various
proangiogenic factors and may produce
objective tumor response in various can-
cers. In HCC, thalidomide may produce
complete or partial remission in ~ 4–7%
of patients and disease stabilization in
10–20% patients with advanced disease.
We have performed a prospective study
using power Doppler ultrasound to evalu-
ate the vascular response to thalidomide
treatment in patients with advanced HCC.
Patients who achieved objective tumor
response to thalidomide tended to have
a higher pretreatment vascularity index,
defined as the number of colored (vas-
cular) pixels within a well-demarcated
tumor area divided by the number of total
pixels in that area, than patients who did
not. The pretreatment vascularity index
also correlated with levels of circulating
angiogenic factors. The value of power
Doppler ultrasound in future clinical trials
of anti-angiogenesis therapy of HCC is
discussed.
INTRODUCTION
Hepatocellular carcinoma (HCC) is the
fifth most common cancer in the world,
with an estimated incidence of 1 million
new cases per year (Parkin, 2001). Most
of the patients with HCC are diagnosed at
advanced stage and not eligible for cura-
tive therapy. However, there is no standard
systemic therapy that has proven efficacy
in patients with advanced HCC. HCC
patients usually tolerate conventional cyto-
toxic chemotherapy poorly because of the
underlying cirrhosis and the accompanying
hypersplenism and peripheral cytopenia
(Hsu et al., 2004). Therefore, novel thera-
peutic agents must be developed to treat
this difficult disease.
Molecular targeted therapy, which aims
at specific molecular derangements in
cancer cells or their microenvironment,
represents breakthrough over conven-
tional cytotoxic therapy (Gschwind et
al., 2004). A rapidly expanding list of
cancers, including non-small cell lung
cancer, breast cancer, gastrointestinal stro-
mal tumor, colorectal cancer, and chronic
myelogeneous leukemia, are now treatable
by agents targeting at specific signaling
pathways (Krause and Van Etten, 2005).
277
278
In addition to targeting the cancer cells per
se, anti-angiogenic/antivascular therapy
is another promising approach to treat
cancers that are usually refractory to con-
ventional cytotoxic therapy. Development
of novel molecular targeting agents has
thus become a key endeavor for the global
pharmacological industry and numerous
agents are already in the development
pipeline.
HCC is typically a hypervascular tumor.
HCC cells and their surrounding stroma
cells have been found to express various
proangiogenic factors, and these proangio-
genic factors may be closely associated
with high histological grade, portal venous
thrombosis, and tumor capsular invasion
(Torimura et al., 1998; Chow et al., 1997,
1998; Harada et al., 1998). It appears that
a proangiogenic milieu exists to facilitate
tumor proliferation and invasion in HCC.
It is thus reasonable to hypothesize that
anti-angiogenesis therapy may be able to
control this tumor.
The anti-angiogenic activity of thalidomide
was first demonstrated in the 1990s. It
can inhibit in vivo angiogenesis induced
by vascular endothelial growth factor
(VEGF) and basic fibroblast growth factor
(D’Amato et al., 1994; Kenyon et al.,
1997). Thalidomide has also been shown to
Table 21.1. Clinical trials of thalidomide for patients with advanced HCC.
No. of Daily dose of
Author patients thalidomide
Hsu et al. 68 200 mg
Wang et al. 99 150–300 mg
Lin et al. 26 200 mg
Patt et al. 37 400 mg
Chen et al. 42 200 mg
Schwartz et al. 38 200 mg
Disease control rate: complete response + partial response + stable disease > 2 months
CR, complete response; PR, partial response
C. Hsu et al.
decrease the transcription of proangiogenic
factors such as, VEGF, and insulin-like
growth factor (Stephens et al., 2000; Vacca
et al., 2005). Its potential therapeutic
efficacy for neoplastic disorders has been
reported in patients with refractory multiple
myeloma and human immunodeficiency
virus-associated Kaposi’s sarcoma (Singhal
et al., 1999; Little et al., 2000).
Several single-arm clinical trials to test
the antitumor efficacy of thalidomide for
advanced HCC have been published, and
their results are summarized in Table 21.1
(Hsu et al., 2003; Wang et al., 2004a, b;
Lin et al., 2005; Patt et al., 2005; Chen
et al., 2005; Schwartz et al., 2005). The
results are similar among different studies,
with objective response rate of ~ 4–7% and
overall survival of 4–7 months. Dose esca-
lation of thalidomide was planned in most
studies, but the patients usually cannot
tolerate a daily dose of > 300 mg because
of treatment-related toxicity, including
fatigue, constipation, and neurotoxicity.
The short survival and poor tolerance
mostly reflect the advanced disease status
and compromised liver function reserves
in this group of patients.
Despite its unsatisfactory overall efficacy
in the treatment of advanced HCC, thali-
domide does produce tumor stabilization
Objective Disease Overall survival
response rate stabilization rate (months)
6.3% (1 CR, 3 PR) 15.9% 4.3
7.0% (6 PR) NA NA
3.9% (1 PR) 11.5% 4.0
5% (1 PR) NA 6.8
7% (3 PR) 23.8% 4.0
5% (1 CR, 1 PR) 18% 5.5
21. Treatment of Hepatocellular Carcinoma with Thalidomide
in ~ 10–25% of HCC patients. In our
study, the ten patients who had either an
objective tumor response (four patients)
or a > 50% decrease in α-fetoprotein
levels (six patients) had a median time
to disease progression of 23 weeks (95%
CI 19.5–26.5) and a median overall sur-
vival of 62.4 weeks (95% CI 31.2–93.6)
(Hsu et al., 2003). Thalidomide and other
biological agents are most likely to exert
cytostatic rather than cytotoxic effect
on established tumors. Therefore, tumor
shrinkage may not be a good surrogate
marker of efficacy for these agents. In
addition, the optimal biological dose of
these agents may be significantly lower
than the maximum tolerated dose defined
by conventional phase I trials. Therefore,
new endpoints are needed to provide
proof-of-principle evidence for the biolo-
gical efficacy of molecular targeted therapy
(Korn et al., 2001).
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.,
2002). Among these methods, Doppler
ultrasound has the advantage of non-
invasiveness and ease to use for serial
follow-up. Tumor vascularity measured
by Doppler ultrasound has been shown to
provide important prognostic information
for patients with colon, gastric, ovarian,
and cervical cancers and has shown good
correlation with vessel density determined
by immunohistochemical staining (Wu
et al., 1994; Cheng et al., 1999, 2000, 2002).
The potential of using Doppler ultrasound
279
to measure the biological efficacy of antian-
giogenesis therapy has also been suggested
in preclinical studies (Gee et al., 2001;
Goertz et al., 2002).
Power (amplitude mode) Doppler ultra-
sound, in which the color map displays the
integrated power of the Doppler signals,
has been shown to be superior to conven-
tional color Doppler ultrasound, which
estimates the mean Doppler frequency
shift (Choi et al., 1996; Lencioni et al.,
1996). Power Doppler ultrasound has the
advantages of low noise variance, rela-
tive angular independence, and increased
dynamic signal range, which make it more
sensitive and specific than the conven-
tional Doppler ultrasound (Rubin et al.,
1994). Therefore, it may provide a conve-
nient, real-time method for observing the
biological effects of novel agents targeted
at tumor vasculature.
We have performed a prospective study of
thalidomide for the treatment of advanced
HCC and power Doppler ultrasound was
used to evaluate the vascular response to
thalidomide treatment (Hsu et al., 2005).
Patients with advanced HCC that were not
suitable for curative surgery or other local
treatment were enrolled to receive thali-
domide treatment. The starting dose was
200 mg per day and the dose was escalated
by 100-mg increments every 2–3 weeks
if no grade 2 or greater treatment-related
toxicity developed. Clinical response to
280
thalidomide treatment was evaluated based
on computed tomography (CT) scan find-
ings and the clinical response. The patients
were considered to have response to thali-
domide treatment if complete or partial
response by WHO criteria (Miller et al.,
1981) or a clinical benefit response was doc-
umented. The clinical benefit response was
achieved when all the following three criteria
were met: (1) stable disease on CT imaging,
(2) a more than 50% decrease of alpha-
fetoprotein level for more than 4 weeks, and
(3) improved performance status.
Sonographic examination was performed
every 1–2 weeks during thalidomide
treatment. The HDI5000 power Doppler
ultrasound unit (Advanced Technology
Laboratories, Bothell, WA) was used. A
2–5 MHz curved array was used for evalu-
ation of liver tumors and a 5–10 MHz
broad-band linear array transducer for
superficial tumors, such as lymph nodes
metastasis. For examination of the liver
tumors, routine abdominal ultrasound was
first performed to identify the index lesion,
followed by power Doppler evaluation.
For examination of the superficial tumors,
power Doppler ultrasound was directly
used. The power Doppler ultrasound set-
ting was standardized using a medium
wall filter, a color gain of 79%, a pulse
repetition frequency of 1,000 Hz with
moderate-to-long persistence, and a slow
sweep technique to achieve the highest
sensitivity without apparent background
noise. Focusing depth was set between 1
and 12 cm. Only one representative lesion
was chosen for follow-up power Doppler
examination.
The tumor was scanned carefully in
all directions, and the tumor section with
maximal color signals was captured and
C. Hsu et al.
stored for later analysis. Each tumor was
scanned three times. The stored images
were then retrieved and the examiner
contoured the tumor margin by using
a cursor. Quantification of the vascular
color signals within the demarcated area
was performed by software (Encomate;
Electronic Business Machine Co., Ltd.,
Taipei, Taiwan). The vascularity index (VI)
was defined as the number of colored
(vascular) pixels within a well-demarcated
tumor area divided by the number of total
pixels in that area. For each tumor the
mean VI of the three representative tumor
sections, one from each scan, was used for
statistical analysis.
From April 2000 to March 2003, 144
patients with advanced HCC were screened
for enrollment into this study. Tumor vas-
cularity was evaluable by power Doppler
ultrasound in only 56 of these patients.
Reasons for nonevaluability included (1)
huge tumor size that could not be encom-
passed within the field of the sonographic
examination; (2) indiscrete tumor margin
resulting from diffuse tumor infiltration,
severe cirrhosis, or prior local therapy;
(3) location of the index tumor that could
not be approached by sonographic exami-
nation, such as lung or intra-abdominal
lymph nodes. Of the 56 patients with
evaluable tumor vascularity, 47 (38 men,
9 women, median age 62.2 years) were
enrolled into this study. The index lesion
was a liver tumor for 44 patients, a meta-
static neck lymph node for 2 patients,
and a metastatic skin tumor for 1 patient.
Patients who completed at least 1 month
of thalidomide treatment were considered
evaluable for response. Forty-four of these
patients were evaluable for response to
thalidomide treatment and all 47 patients
21. Treatment of Hepatocellular Carcinoma with Thalidomide
were evaluable for toxicity. Three of the
47 enrolled patients did not complete 1
month of treatment because of subjective
intolerance of side effects, mainly fatigue.
During thalidomide treatment, the pati-
ents were followed regularly with physical
examination, hematology, serum biochem-
istry and alpha-fetoprotein tests. Computed
tomography scan was done before the start
and after 1 month of thalidomide therapy
and then every 2 months. Sonographic
examination was performed before and
every 1–2 weeks during thalidomide treat-
ment. Five patients were considered to have
response to thalidomide treatment, includ-
ing one complete and one partial response
based on CT scan findings. Another three
patients fulfilled the criteria of clinical
benefit response described above. The total
response rate was 11.4% (95% C.I.: 1.6–
21.2%). Fifteen patients had stable disease
and 24 patients had progressive disease.
The median pretreatment value of VI
was 2.73 (range 0–25.36). Pretreatment VI
was significantly higher in patients with an
objective tumor response to thalidomide
(median 7.42 vs. 2.15, p = 0.03). After tha-
lidomide treatment, four of the five patients
with response had a decrease in VI, but 20
of the 36 patients without response had
also a decrease of VI. The tumors of three
patients became non-evaluable for tumor
vascularity during follow-up because of
progressive tumor infiltration and blurring
of the tumor margins. All of them had pro-
gressive disease.
Representative images of power Doppler
ultrasound in our patients are shown in
Figure 21.1. Figure 21.1a and b were
from a 51-year-old man who presented
with multiple liver and lung tumors.
Hepatocellular carcinoma was diagnosed
281
by liver biopsy. He had received chemo-
therapy with doxorubicin before the start
of thalidomide treatment but progressive
disease after chemotherapy was docu-
mented. The patient took thalidomide up
to 600 mg per day and had stable disease
for 2 months. Figure 21.1a was from the
baseline examination and Figure 21.1b
from examination done at the documenta-
tion of progressive disease to thalidomide
treatment. An increase in both tumor area
and vascularity index was noted. Figure
21.1c and d were from a 72-year-old
woman with hepatitis C-related cirrhosis.
Hepatocellular carcinoma was diagnosed
by compatible imaging findings and high
alpha-fetoprotein levels. She had received
multiple sessions of trans-arterial chemo-
embolization before enrollment into this
study. The patient received thalidomide
200 mg per day and stable disease was
achieved for 23 weeks. Figure 21.1c was
from the baseline examination and Figure
21.1d from examination done at the docu-
mentation of stable disease to thalidomide
treatment. The tumor area remained sta-
tionary while a decrease of vascularity
index was noted.
In addition to power Doppler ultra-
sound, we also checked the levels of cir-
culating proangiogenic factors, including
vascular endothelial growth factor, basic
fibroblast growth factor, and placental
growth factor. The pretreatment VI was
significantly correlated with pretreatment
levels of basic fibroblast growth factor and
placental growth factor, suggesting that VI
may reflect the activity of tumor angiogen-
esis. However, none of the proangiogenic
factors we measured differed significantly
between responders and non-responders to
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

FUTURE PERSPECTIVES OF tumor vascularity in clinical trials of anti-

IMAGING STUDIES FOR
EVALUATION OF ANTI-
ANGIOGENESIS THERAPY
Although conventional Doppler ultrasound
has been extensively studied for evaluation
of tumor vascularity, the low flow rate in
tumor microvessels makes this method
not sensitive enough to detect changes of
angiogenic therapy. Our study indicates
that power Doppler ultrasound may pro-
vide a convenient, real-time method for
observing the biological effects of agents
targeted at tumor vasculature. New tech-
niques are being developed to improve the
sensitivity and resolution of conventional
ultrasound. Of these techniques, micro-
bubble-based, contrast-enhanced ultrasound
21. Treatment of Hepatocellular Carcinoma with Thalidomide
is the most extensively studied. The intra-
venously injected microbubbles can signi-
ficantly enhance the Doppler signals of
blood flow and help differential diagnosis
of liver tumors (Kim et al., 1998; Vilana
et al., 2003; Quaia et al., 2004). The use-
fulness of contrast-enhanced ultrasound
in evaluation of tumor blood flow after
thalidomide treatment has been suggested
in a pilot study for patients with advanced
HCC (Bertolotto et al., 2006). New imag-
ing techniques, such as pulse inversion
harmonic ultrasound and agent detection
imaging, may further improve the image
resolution and reduce artifacts (Choi et al.,
2002; Lee et al., 2006). The potential of
these new techniques in the monitoring of
anti-angiogenesis therapy for HCC should
be further investigated.
Despite these technical advances, there
are still limitations in establishing the
role of ultrasound for evaluation of anti-
angiogenesis therapy in HCC. The first
limitation is the patient population. Most of
the HCC patients have underlying cirrhosis
and had received various local treatments,
such as trans-arterial embolization or per-
cutaneous ethanol injection, before they
started empirical antiangiogenesis therapy.
The margins of their tumors are often
indiscrete and calculation of VI is there-
fore difficult. Besides, patients with large
or deep-seated tumors in the liver may not
be assessable by ultrasound. In our study
only ~ 40% of patients with advanced HCC
who were otherwise eligible for thalido-
mide therapy were considered evaluable
by power Doppler ultrasound. The second
limitation is the difficulty in correlating
the vascular response with other clinical
outcomes, such as tumor shrinkage and
patient survival. The response rate of curr-
ently available anti-angiogenesis therapy or
283
other molecular targeted therapy for HCC
is ~ 5% (Philip et al., 2005; Schwartz et al.,
2006; Abou-Alfa et al., 2006). Therefore, it
is difficult to find a clear-cut threshold VI
value that may predict the response to anti-
angiogenesis therapy. Besides, in our study
decrease in VI after thalidomide treatment
was also found in patients with progressive
disease, suggesting that tumors in these
patients have surpassed the angiogenesis
regulatory mechanisms that may be inhib-
ited by thalidomide. The third limitation
is the inherent dependence of ultrasound
on individual examiner’s technique. Repro-
ducibility of these new sonography tech-
niques must be validated before they can
be widely used in the clinical settings.
Other functional imaging modalities
have also been extensively studied for
their use in evaluation of anti-angiogenesis
therapy. Dynamic contrast-enhanced mag-
netic resonance imaging (DCE-MRI) has
been widely used to evaluate the vascular
response, including change in vascular
perfusion and permeability, after antiang-
iogenic therapy (Padhani and Leach, 2005;
Liu et al., 2005). Guidelines for its use
in early-phase clinical trials have been
proposed (Leach et al., 2005; Evelhoch
et al., 2005). DCE-MRI is relatively non-
invasive and can be objectively measured
serially. However, detailed definition of
the examination protocols is needed to
improve reproducibility. The potential
anti-angiogenesis effects of thalidomide
in HCC have been evaluated by DCE-
MRI. Patients with stable disease after
thalidomide treatment tended to have
greater reduction of perfusion in their
tumors, as indicated by a greater decrease
in peak and maximal signal enhancement,
than patients with progressive disease (Wang
et al., 2004b). Other endpoints of vascular
284
response, such as the transfer constant
(Ktrans) and the initial area under the gado-
linium concentration time curve, should
be evaluated for their validity in future
clinical trials of anti-angiogenesis therapy
for HCC.
In conclusion, evaluation of tumor vas-
cularity by power Doppler ultrasound
is feasible in a subset of patients with
advanced HCC. The usefulness of power
Doppler ultrasound and other functional
imaging modalities should be explored in
future clinical trials of anti-angiogenesis
therapy for HCC.
REFERENCES
Abou-Alfa, G.K., Schwartz, L., Ricci, S., Amadori, D.,
Santoro, A., Figer, A., De Greve, J., Douillard, J.Y.,
Lathia, C., Schwartz, B., Taylor, I., Moscovici, M.,
and Saltz, L.B. 2006. Phase II study of sorafenib
in patients with advanced hepatocellular carcinoma.
J. Clin. Oncol. 24: 4293–4300.
Bertolotto, M., Pozzato, G., Croce, L.S., Nascimben, F.,
Gasparini, C., Cova, M.A., and Tiribelli, C. 2006.
Blood flow changes in hepatocellular carcinoma
after the administration of thalidomide assessed
by reperfusion kinetics during microbubble infu-
sion: preliminary results. Invest. Radiol. 41:
15–21.
Chen, C.N., Cheng, Y.M., Liang, J.T., Lee, P.H.,
Hsieh, F.J., Yuan, R.H., Wang, S.M., Chang,
M.F., and Chang, K.J. 2000. Color Doppler
vascular index can predict distant metastasis and
survival in colon cancer patients. Cancer Res.
60: 2892–2897.
Chen, C.N., Cheng, Y.M., Lin, M.T., Hsieh, F.J.,
Lee, P.H., and Chang, K.J. 2002. Association of
color Doppler vascularity index and microves-
sel density with survival in patients with gastric
cancer. Ann. Surg. 235: 512–518.
Chen, L.T., Liu, T.W., Chao, Y., Shiah, H.S.,
Chang, J.Y., Juang, S.H., Chen, S.C., Chuang,
T.R., Chin, Y.H., and Whang-Peng, J. 2005.
α-fetoprotein response predicts survival benefits
of thalidomide in advanced hepatocellular carci-
noma. Aliment. Pharmacol. Ther. 22: 217–226.
C. Hsu et al.
Cheng, W.F., Lee, C.N., Chu, J.S., Chen, C.A.,
Chen, T.M., Shau, W.Y., Hsieh, C.Y., and Hsieh,
F.J. 1999. Vascular index as a novel parameter
for the in vivo assessment of angiogenesis in
patients with cervical carcinoma. Cancer 85:
651–657.
Choi, B.I., Kim, T.K., Han, J.K., Chung, J.W.,
Park, J.H., and Han, M.C. 1996. Power versus
conventional color Doppler sonography: com-
parison in the depiction of vasculature in liver
tumors. Radiology 200: 55–58.
Choi, B.I., Kim, A.Y., Lee, J.Y., Kim, K.W.,
Lee, K.H., Kim, T.K., and Han, J.K. 2002.
Hepatocellular carcinoma: contrast enhancement
with Levovist. J. Ultrasound Med. 21: 77–84.
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-
noma: an immunohistochemical study. Human
Pathol. 28: 698–703.
Chow, N.H., Cheng, K.S., Lin, P.W., Chan, S.H.,
Su, W.C., Sun, Y.N., and Lin, X.Z. 1998.
Expression of fibroblast growth factor-1 and
fibroblast growth factor-2 in normal liver and
hepatocellular carcinoma. Digest. Dis. Sci. 10:
2261–2266.
Choyke, P.L., Knopp, M.V., and Libutti, S.K. 2002.
Special techniques for imaging blood flow to
tumors. Cancer J. 8: 109–118.
D’Amato, R.J., Loughnan, M.S., Flynn, E., and
Folkman, J. 1994. Thalidomide is an inhibitor of
angiogenesis. Proc. Natl. Acad. Sci. U.S.A. 91:
4082–4085.
Evelhoch, J., Garwood, M., Vigneron, D., Knopp, M.,
Sullivan, D., Menkens, A., Clarke, L., and
Liu, G. 2005. Expanding the use of magnetic
resonance in the assessment of tumor response
to therapy: workshop report. Cancer Res. 65:
7041–7044
Gee, M.S., Saunders, H.M., Lee, J.C., Sanzo,
J.F., Jenkins, W.T., Evans, S.M., Trinchieri,
G., Sehgal, C.M., Feldman, M.D., and Lee,
W.M. 2001. Doppler ultrasound imaging detects
changes in tumor perfusion during antivascu-
lar therapy associated with vascular anatomic
alterations. Cancer Res. 61: 2974–2982.
Goertz, D.E., Yu, J.L., Kerbel, R.S., Burns, P.N.,
and Foster, F.S. 2002. High-frequency Doppler
ultrasound monitors the effects of antivascular
21. Treatment of Hepatocellular Carcinoma with Thalidomide
therapy on tumor blood flow. Cancer Res. 62:
6371–6375.
Gschwind, A., Fischer, O.M., and Ullrich, A. 2004.
The discovery of receptor tyrosine kinases:
targets for cancer therapy. Nat. Rev. Cancer 4:
361–370.
Harada, T., Arii, S., Mise, M., Imamura, T.,
Higashitsuji, H., Furutani, M., Niwano, M.,
Ishigami, S., Fukumoto, M., Seiki, M., Sato, H.,
and Imamura, M. 1998. Membrane-type matrix
metalloproteinase-1 (MT1-MMP) gene is over-
expressed in highly invasive hepatocellular
carcinomas. J. Hepatol. 28: 231–239.
Hsu, C., Chen, C.N., Chen, L.T., Wu, C.Y., Yang,
P.M., Lai, M.Y., Lee, P.H., and Cheng, A.L.
2003. Low-dose thalidomide treatment for
advanced hepatocellular carcinoma. Oncology
65: 242–249.
Hsu, C., Cheng, J.C., and Cheng, A.L. 2004.
Recent advances in non-surgical treatment for
advanced hepatocellular carcinoma. J. Formos.
Med. Assoc. 103: 483–495.
Hsu, C., Chen, C.N., Chen, L.T., Wu, C.Y., Hsieh,
F.J., and Cheng, A.L. 2005. Effect of thalidomide
in hepatocellular carcinoma: assessment with
power Doppler US and analysis of circulating
angiogenic factors. Radiology 235: 509–516.
Kenyon, B.M., Browne, F., and D’Amato, R. 1997.
Effects of thalidomide and related metabolites
in a mouse corneal model of neovascularization.
Exp. Eye Res. 64: 971–978.
Kim, A.Y., Choi, B.I., Kim, T.K., Han, J.K.,
Yun, E.J., Lee, K.Y., and Han, M.C. 1998.
Hepatocellular carcinoma: power Doppler
US with a contrast agent: preliminary results.
Radiology 209: 135–140.
Korn, E.L., Arbuck, S.G., Pluda, J.M., Simon, R.,
Kaplan, R.S., and Christian, M.C. 2001.
Clinical trial designs for cytostatic agents: are
new approaches needed? J. Clin. Oncol. 19:
265–272.
Krause, D.S., and Van Etten, R.A. 2005. Tyrosine
kinases as targets for cancer therapy. N. Engl. J.
Med. 353: 172–187.
Leach, M.O., Brindle, K.M., Evelhoch, J.L.,
Griffiths, J.R., Horsman, M.R., Jackson, A.,
Jayson G.C., Judson, I.R., Knopp, M.V., Maxwell,
R.J., McIntyre, D., Padhani, A.R., Price, P.,
Rathbone, R., Rustin, G.J., Tofts, P.S., Tozer,
G.M., Vennart, W., Waterton, J.C., Williams, S.R.,
285
Workman, P., and Pharmacodynamic/Pharma-
cokinetic Technologies Advisory Committee,
Drug Development Office, Cancer Research
UK. 2005. The assessment of antiangiogenic
and antivascular therapies in early-stage clini-
cal trials using magnetic resonance imaging:
issues and recommendations. Brit. J. Cancer 92:
1599–1610.
Lee, J.Y., Choi, B.I., Han, J.K., Lee, J.M., and Kim,
S.H. 2006. State-of-the-art ultrasonography of
hepatocellular carcinoma. Eur. J. Radiol. 58:
177–185
Lencioni, R., Pinto, F., Armillotta, N., and
Bartolozzi, C. 1996. Assessment of tumor vascu-
larity in hepatocellular carcinoma: comparison
of power Doppler US and color Doppler US.
Radiology 201: 353–358.
Lin, A.Y., Brophy, N., Fisher, G.A., So, S., Biggs, C.,
Yock, T.I., and Levitt, L. 2005. Phase II
study of thalidomide in patients with unre-
sectable hepatocellular carcinoma. Cancer 103:
119–125.
Little, R.F., Wyvill, K.M., Pluda, J.M., Welles, L.,
Marshall, V., Figg, W.D., Newcomb, F.M.,
Tosato, G., Feigal, E., Steinberg, S.M.,
Whitby, D., Goedert, J.J., and Yachoan, R. 2000.
Activity of thalidomide in AIDS-related Kaposi’s
sarcoma. J. Clin. Oncol. 18: 2593–2602.
Liu, G., Rugo, H.S., Wilding, G., McShane, T.M.,
Evelhoch, J.L., Ng, C., Jackson, E., Kelcz, F.,
Yeh, B.M., Lee, F.T. Jr., Charnsangavej, C.,
Park, J.W., Ashton, E.A., Steinfeldt, H.M.,
Pithavala, Y.K., Reich, S.D., and Herbst, R.S.
2005. Dynamic contrast-enhanced magnetic
resonance imaging as a pharmacodynamic measure
of response after acute dosing of AG-013736,
an oral angiogenesis inhibitor, in patients with
advanced solid tumors: results from a phase I
study. J. Clin. Oncol. 23: 5464–5473.
Miller, A.B., Hoogstraten, B., Staquet, M., and
Winkler, A. 1981. Reporting results of cancer
treatment. Cancer 47: 207–214.
Padhani, A.R., and Leach, M.O. 2005. Antivascular
cancer treatments: functional assessments by
dynamic contrast-enhanced magnetic resonance
imaging. Abd. Imaging 30: 324–341
Parkin, D.M. 2001. Global cancer statistics in the
year 2000. Lancet Oncol. 2: 533–543.
Patt, Y.Z., Hassan, M.M., Lozano, R.D., Nooka,
A.K., Schnirer, I.I., Zeldis, J.B., Abbruzzese, J.L.,
286
and Brown, T.D. 2005. Thalidomide in the treat-
ment of patients with hepatocellular carcinoma:
a phase II trial. Cancer 103: 749–755.
Philip, P.A., Mahoney, M.R., Allmer, C., Thomas, J.,
Pitot, H.C., Kim, G., Donehower, R.C., Fitch, T.,
Picus, J., and Erlichman, C. 2005. Phase II
study of erlotinib (OSI-774) in patients with
advanced hepatocellular cancer. J. Clin. Oncol.
23: 6657–6663.
Quaia, E., Calliada, F., Bertolotto, M., Rossi, S.,
Garioni, L., Rosa, L., and Pozzi-Mucelli, R. 2004.
Characterization of focal liver lesions with
contrast-specific US modes and a sulfur
hexafluoride-filled microbubble contrast agent:
diagnostic performance and confidence. Radiology
232:420–430.
Rubin, J.M., Bude, R.O., Carson, P.L., Bree,
R.L., and Adler, R.S. 1994. Power Doppler
US: a potentially useful alternative to mean-
frequency-based color Doppler US. Radiology
190: 853–856.
Schwartz, J.D., Sung, M., Schwartz, M., Lehrer, D.,
Mandeli, J., Liebes, L., Goldenberg, A., and
Volm, M. 2005. Thalidomide in advanced hepa-
tocellular carcinoma with optional low-dose
interferon-α2a upon progression. Oncologist 10:
718–727.
Schwartz, J.D., Schwartz, M., Lehrer, D., Cohen, E.,
Sung, M., Kinkhabwala, M., Siegel, A.,
Holloway, S., Ocean, A., and Wadler, S. 2006.
Bevacizumab in unresectable hepatocellular
carcinoma (HCC) for patients without metastasis
and without invasion of the portal vein. Proc.
Am. Soc. Clin. Oncol. 24: abstr 4144.
Singhal, S., Mehta, J., Desikan, R., Ayers, D.,
Roberson, P., Eddlemon, P., Munshi, N., Anaissie,
E., Wilson, C., Dhodapkar, M., Zeddis, J.,
and Barlogie, B. 1999. Antitumor activity of
thalidomide in refractory multiple myeloma. N.
Engl. J. Med. 341: 1565–1571
C. Hsu et al.
Stephens, T.D., Bunde, C.J., and Fillmore, B.J.
2000. Mechanism of action in thalidomide tera-
togenesis. Biochem. Pharmacol. 59: 1489–1499.
Torimura, T., Sata, M., Ueno, T., Kin, M., Tsuji, R.,
Suzaku, K., Hashimoto, O., Sugawara, H., and
Tanikawa, K. 1998. Increased expression of
vascular endothelial growth factor is associated
with tumor progression in hepatocellular carci-
noma. Human Pathol. 29: 986–991.
Vacca, A., Scavelli, C., Montefusco, V., Di Pietro, G.,
Neri, A., Mattioli, M., Bicciato, S., Nico, B.,
Ribatti, D., Dammacco, F., and Corradini, P. 2005.
Thalidomide downregulates angiogenic genes in
bone marrow endothelial cells of patients with
active multiple myeloma. J. Clin. Oncol. 23:
5334–5346.
Vilana, R., Llovet, J.M., Bianchi, L., Sanchez, M.,
Pages, M., Sala. M., Gilabert, R., Nicolau, C.,
Garcia, A., Ayuso, C., Bruix, J., and Bru, C.
2003. Contrast-enhanced power Doppler sonog-
raphy and helical computed tomography for
assessment of vascularity of small hepatocellular
carcinomas before and after percutaneous abla-
tion. J. Clin. Ultrasound 31: 119–128.
Wang, J., Chen, L.T., Tsang, Y.M., Liu, T.W., and
Shih, T.T. 2004a. Dynamic contrast-enhanced
MRI analysis of perfusion changes in advanced
hepatocellular carcinoma treated with an antian-
giogenic agent: a preliminary study. Am. J.
Roentgenol. 183: 713–719.
Wang, T.E., Kao, C.R., Lin, S.C., Chang, W.H.,
Chu, C.H., Lin, J., and Hsieh, R.K. 2004b.
Salvage therapy for hepatocellular carcinoma
with thalidomide. World J. Gastroenterol. 10:
649–653.
Wu, C.C., Lee, C.N., Chen, T.M., Shyu, M.K.,
Hsieh, C.Y., Chen, H.Y., and Hsieh, F.J. 1994.
Incremental angiogenesis assessed by color
Doppler ultrasound in the tumorigenesis of ovarian
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

INTRODUCTION the use of different imaging modalities,

especially single photon emission com-
During the recent years it has been shown puted tomography (SPECT) and SPECT
that transarterial applications of antican- with integrated computed tomography
cer agents are promising approaches in (SPECT-CT), for planning and control of
the treatment of hepatic malignancies. treatment will be discussed.
While new techniques led to a reappraisal
of hepatic arterial infusion chemotherapy
(HAIC) over implanted port systems, CURRENT STATUS OF HEPATIC
there has also been a growing interest in
ARTERIAL CHEMOTHERAPY
the use of transarterial radioembolization
(RE). The concept of transarterial regional AND RADIOEMBOLIZATION
oncologic treatment of hepatic malignan-
Because liver tumors are supplied prima-
cies presupposes that the entire liver can
rily by the hepatic artery, direct intraarte-
be exclusively treated over the supplying
rial infusion of chemotherapeutic agents
arteries. Several imaging modalities are
can expose the tumor to a much higher
in use for prediction and control of per-
drug concentration than intravenous therapy.
fusion territories of intraarterial catheters.
For transarterial RE, a semiselectivity of
Besides digital subtraction angiography
drug delivery is given by the low (∼ 30% or
(DSA), computed tomography (CT) and
less) arterial blood supply of normal liver
magnetic resonance imaging (MRI), scin-
tissue with the portal vein as the main sup-
tigraphic imaging is being employed to
plying vessel.
visualize the intrahepatic perfusion pat-
tern, to exclude extrahepatic perfusion,
and to assess an intrahepatic arteriovenous Intraarterial Chemotherapy
shunting to the lung. In the following, the The role of HAIC in patients with cancers
current status of HAIC and RE in cancer confined to the liver has been controver-
therapy, the application techniques and sially discussed for many years, whereby

287
288
most experiences with this approach has
been gained with the treatment of “liver
only”-metastases of colorectal can-
cer (Cohen and Kemeny, 2003; MAGC,
1996). In addition, limited data are avail-
able for other liver neoplasms such as
hepatocellular (Tanaka et al., 2005) or
cholangiocellular carcinoma (Cantore
et al., 2005). Recently, the introduction of
minimal invasively implanted port systems
led to a reappraisal of this approach (Ricke
et al., 2004).
For colorectal cancer patients with unre-
sectable liver metastases or after hepatic
metastasectomy, more than a dozen of ran-
domized trials have compared HAIC using
fluoropyrimidines floxuridine (FUDR) or
5-flourouracil (5-FU) with intravenous
drug application (Cohen and Kemeny,
2003; Nelson and Freels, 2004). Summing
up those results, HAIC proved to improve
response rates in patients with unresect-
able disease and time to hepatic progres-
sion for both indications, but results were
counterbalanced by the lack of a survival
benefit in most of the studies.
The reasons preventing a clear-cut sur-
vival benefit for HAIC have been dis-
cussed in detail (Cohen and Kemeny,
2003; MAGC, 1996; Nelson and Freels,
2004). Particularly in the earlier trials,
small patient numbers and inappropriate
control groups may have led to mislead-
ing interpretations. In addition, exces-
sively high rates of extrahepatic failure
were observed in studies on intraarte-
rial FUDR. Adjacent trials completed in
North America suggested that systemic
control rates of intraarterial FUDR can be
improved by combinations with intrave-
nous 5-FU (Kemeny et al., 1999). In addi-
tion, European and Asian investigators
aimed to optimize results by taking the
T. Denecke et al.
advantage of systemic drug concentrations
provided by the intraarterial application of
5-FU (Lorenz and Muller, 2000; Sadahiro
et al., 2004).
Unlike hepatic arterial FUDR, which
is usually delivered through surgically
implanted infusion pumps, regional appli-
cations of 5-FU demand the use of external
pumps, due to the maximal drug concen-
tration of 25 mg/ml. Therefore, hepatic
arterial 5-FU can only be delivered by
intermittent percutaneous access, or
through port systems. There is extensive
experience with the use of port catheters in
a number of specialized treatment centers
worldwide, but this technique is associ-
ated with a considerable proportion of
primary failure which has been reported to
approximate one third of patients sched-
uled for 5-FU-based HAIC in recent mul-
ticenter trials (Lorenz and Muller, 2000;
Kerr et al., 2003). Once HAIC has been
initiated, port catheters are associated with
higher complication rates than surgically
implanted pumps, whereas secondary fail-
ure rates have been reported to be similar
(Heinrich et al., 2003). Presently, it is
accepted that the low performance level of
surgically implanted port catheters repre-
sents one major reason why the theoretical
advantage of 5-FU-based HAIC could
not be transferred to convincing clinical
improvements.
In this context, interventionally
implanted port catheter systems (IIPCS)
have evolved a promising alternative to
surgically implanted devices. The use of
IIPCS enables initiation of HAIC with-
out laparotomy, and available data sug-
gest favorable complication and failure
rates with this approach. However, studies
published so far mainly focused on either
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
dissected gastroduodenal artery (GDA).
The perfusion territory is controlled ini-
tially by injection of e.g., methylene blue.
The surgical placement of intraarterial port
systems is definite. In case of system dys-
function, e.g., catheter or artery occlusion,
a revision is difficult. Thus, dysfunction of
surgical ports usually leads to a replace-
ment of HAIC by systemic treatment
(Heinrich et al., 2003).
Employing novel catheterization tech-
niques, minimally invasive interventional
implantations of intraarterial port systems
became possible (Ricke et al., 2004). For
this alternative approach, port catheters are
usually positioned through a femoral artery
access under DSA guidance to position the
catheter tip in the hepatic artery. The subcu-
taneous port reservoir is placed just inferior
to the groin. A subclavicular approach is also
practicable. In principle, minimal invasively
implanted port systems appear to be superior
to surgically implanted port catheters with
respect to primary and secondary failure and
the possibility to easily repair or replace the
systems in case of dysfunction (Ricke et al.,
2004; Sadahiro et al., 2004). An additional
advantage is the minimal risk of the proce-
dure as compared with the mortality rate of
surgical port placement (Kerr et al., 2003).
However, the correct position of intraar-
terial catheters regarding hepatic and
extrahepatic distribution of therapeuti-
cally applied drugs, unlike in surgically
implanted catheters, cannot be confirmed
by intra-operative application of e.g.,
methylene blue, and dislocations of the
catheter after initial placement and dur-
ing the therapy course have to be consid-
ered (Ricke et al., 2004). Therefore, valid
imaging procedures are mandatory to pre-
serve the advantages of minimal-invasive
catheter implantation in clinical practice.
T. Denecke et al.
Radioembolization
Radioembolization (RE) is a one-time
procedure and is usually not thought to be
repeated. The therapeutic agent consists
of microspheres, 20–40 nm in diameter,
which contain a radionuclide emitting
beta-radiation (e.g., Yttrium90) (Gray
et al., 2001). The microspheres are deliv-
ered into the hepatic artery and embolize
the microvessels and capillaries of tissues.
Again, the dominance of arterial supply
of hepatic malignancies in contrast to the
normal liver is used for a semiselectivity
of drug delivery. The typical vascular
architecture of liver metastases leads to a
predominant seeding of the microspheres
into the outer regions of the tumors, which
are hypervascularized and thus hyperoxy-
genated, making radiation in these areas
more effective (Campbell et al., 2001).
The transarterial access is similar to that
described for interventional radiological
port placement. Alternatively, the injection
can be performed over an intraarterial port
system supplying the liver. For the procedure
of drug application, radiation protection
is mandatory. Therefore, specific devices
for application were developed in order
to avoid direct and prolonged contact of
the radiologist to the therapeutic agent. To
preserve a hepatic reserve, a fractionized
application selectively to one and sub-
sequently to the other liver lobe is under
evaluation.
Avoidance of extrahepatic implantation of
microspheres, potentially causing gastritis,
ulcers, pancreatitis and other gastrointestinal
morbidities, is even more decisive than in
HAIC (Murthy et al., 2005). Therefore,
a sufficient and careful planning of the
procedure is mandatory. The interven-
tion planning includes an arterial selec-
tive catheterization with application of
22. Perfusion Scintigraphy with Integrated Single Photon Emission 291

Tc99m-MAA and subsequent planar scin-

extrahepatic perfusion can be best
tigraphy to predict the distribution of
diagnosed by port scintigraphy with
microspheres, subsequently released at the
Tc-99m labeled macroaggregated albu-
same intraarterial location, with respect
min (MAA) employing planar imaging
to shunting to the lung and gastrointestinal
and SPECT (Cohen and Kemeny, 2003;
accumulations (Murthy et al., 2005). The
Lehner et al., 1987; Lubin et al., 1987).
most exact detection of extrahepatic accu-
Further techniques such as perfusion CT,
mulation is desirable in order to avoid
and slow-infusion magnetic resonance
unnecessary treatment complications. This
arteriography (MRA) have been used to
and the control of microsphere distribution
evaluate the intrahepatic arterial flow in
by Bremsstrahlung scintigraphy after RE
patients with hepatic intraarterial ports
are valuable indications for SPECT-CT.
(Morimoto et al., 1999; Seki et al., 2003),
The issues concerning the avoidance of
but did not gain broad clinical accept-
inadvertent deposition of therapeutic agent
ance. In the routine surveillance during
into non-target arteries are the same as
treatment, both scintigraphy with Tc-99m
discussed for HAIC. The quantification of
labeled MAA and DSA are suitable to
pulmonary shunting is more important for
check for patency of catheters and hepatic
RE than for HAIC, because the complica-
arteries as well as system leaking, but may
tions after RE with pneumonitis and pul-
lack valid information about completeness
monary insufficiency are life threatening,
and homogeneity of hepatic perfusion. A
while pulmonary complications in HAIC
combination of perfusion imaging with
are controversially discussed. However,
sufficient anatomic information may be
planar scintigraphy is sufficient for dosing
useful to ensure homogenous perfusion of
microspheres. the entire liver in patients treated with HAI
via intra-arterial port systems. This espe-
cially holds true for those patients with
VISUALIZATION aberrant hepatic arteries and/or who previ-
OF PERFUSION TERRITORIES ously underwent hepatic resection (Ricke
OF HEPATIC INTRAARTERIAL et al., 2004; Denecke et al., 2005).
CATHETERS IN PLANNING
AND CONTROL Single Photon Emission Computed
OF TRANSARTERIAL Tomography with Integrated Computed
TREATMENT Tomography for Port Perfusion
Scintigraphy
Imaging Techniques
A novel tool of nuclear medicine imaging,
There are several issues to be addressed by a hybrid SPECT-CT camera, is potentially
imaging techniques employed for the rou- able to show both parenchymal perfusion
tine control of intraarterial port systems over the port system and anatomic rela-
before HAIC. Before initiation of treat- tion for correct assignment of intra- and
ment, leaking of the port system and alter- extrahepatic tracer accumulation (Denecke
ations of supplying arteries and catheters et al., 2005). This device combines a
may be easily detected by DSA, whereas dual head gamma camera with a rotating
292
low dose X-ray tube enabling both X-ray
based attenuation correction for enhanced
image quality of SPECT and anatomic
mapping on inherently fused SPECT-CT
images with sufficient spatial resolution
to delineate the liver silhouette (Bocher
et al., 2000). Other comparable devices,
which became available more recently,
consist of a dual head gamma camera and
an integrated diagnostic multirow detector
CT scanner.
A recommendable scintigraphy proto-
col for a low dose SPECT-CT includes
planar scintigraphy covering pelvis, abdo-
men, and thorax, as well as tomographic
imaging of the relevant organs in the
upper abdomen with image fusion for ana-
tomical mapping. The patient is prepared
with a thyroid blockage (e.g., perchlorate,
potassium iodide). This is important to
avoid active Tc-99m uptake of the gastric
mucosa to enable the identification of
accidental embolization of Tc-99m-MAA
into the stomach wall. After testing the
port system for patency with a bolus of
isotonic saline (0.9%), 200 MBq Tc-99m-
MAA in 2 ml saline (the syringe must
be shook before injection to optimize
the distribution) is injected into the port
capsule followed by another saline flush
(∼ 5–10 ml) under dynamic anterior pla-
nar image acquisition of the abdomen (1
image/5 s over 1 min) with a low energy
high resolution (LEHR) collimator to rule
out tracer extravasation. The injection has
to be performed carefully and slowly.
This is important because of two major
problems: (1) the viscosity of MAA needs
a relatively high injection pressure (espe-
cially when thin supraselective catheters
are used as port lines), which may cause
rupture of the port system, and (2) a high
injection rate can cause a reflux of the
T. Denecke et al.
tracer retrogradely into arteries originating
proximally from the tip of the port line,
suggesting a catheter dislocation. Just after
sequential imaging, static planar acquisi-
tion is started. Anterior, right oblique and
right lateral planar image acquisitions of
the upper abdomen (each up to 800,000
counts) as well as anterior projections of
the lower abdomen to rule out leaking
and of the thorax to assess shunting to the
lung are obtained. In patients with two
port systems, the activity can be divided
by two and is injected into the port sys-
tems, each injection followed by dynamic
and static planar imaging as described
above. Then (∼ 20 min after tracer injec-
tion) a SPECT-CT of the upper abdomen
is performed using a dual-head gamma
camera (LEHR collimators; FOV, 540 ×
400 mm; matrix, 128 × 128; 120 frames;
3° steps; 15 s/frame) with an integrated
rotating X-ray tube (fixed tube current,
2.5 mA; 140 kV; slice thickness, 10 mm;
matrix, 128 × 128; rotation time, 13.5 s;
step and shoot technique) for attenuation
correction (“Hawkeye”Millenium VG, GE
Medical Systems). With this device, total
acquisition time is ∼ 27 min including the
initial transmission scan using the low-
dose X-ray tube (9 min for 40 cm scan
length) and the subsequent SPECT result-
ing in a delay of SPECT acquisition from
tracer injection of ∼30 min. After iterative
reconstruction (OSEM algorithm; number
of iterations, 2; number of subsets, 10;
prefiltering, Butterworth (frequency, 0.25;
order, 10); postfiltering, Butterworth (fre-
quency, 0.5; order, 10) and attenuation
correction (attenuation map derived from
the low dose CT), the reconstructed data
are visualized in sagital, coronal, and axial
slices, with the SPECT, the CT and the
fusion images side by side. Inherent image
22. Perfusion Scintigraphy with Integrated Single Photon Emission
fusions are generated from the coregis-
tered SPECT and low-dose CT images.
With a SPECT-CT unit consisting of
a diagnostic dedicated CT scanner, the
imaging protocol would be similar. In
this case, an unenhanced scan of the area
covered by SPECT would be acquired in
low dose technique (∼ 40 mAs) which is
enough for anatomic mapping. Regular
CT scans should only be performed if
restaging is necessary. The patient is asked
to hold the breath in mild expiration for a
better congruency of the diaphragm posi-
tion in SPECT and CT. This is not required
when a low-dose device is used as the
X-ray tube takes ∼ 13 s for one rotation
covering 1 cm scan length, resulting in an
averaged image of the diaphragm move-
ment comparable to that of SPECT.
Image Analysis
Image interpretation includes several issues:
(1) the patency of the port line, (2) the
assessment of intrahepatic arteriovenous
shunting to the lung, (3) tracer extravasation
indicative of leakage of the port system, (4)
the presence and location of intraabdominal
extrahepatic tracer embolization, and (5)
the intrahepatic perfusion pattern especially
regarding the perfusion intensity of tumor
baring segments (Ziessman et al., 1985;
Denecke et al., 2005).
Patency of the port system is assessed by
the performance of injection and describes
the line as blocked or smooth running. A
leakage of the system is rarely present and
usually occurs as a disconnection at the
junction of the catheter to the reservoir as
a weak point. This is indicated by a tracer
retention at the groin without activity reach-
ing the liver or upper abdomen on initial
dynamic images obtained during injection.
293
While calculation of an intrahepatic
shunting to the lung can be assessed
visually and semiquantitatively on planar
images, the intrahepatic perfusion terri-
tory of the port can be best evaluated on
SPECT images with anatomic mapping.
For readings of port perfusion scintigraphy
obtained with integrated SPECT-CT, the
physician should be aware of the patients
clinical history. Knowledge of previous
hepatic surgery, or other interventional
procedures (e.g., radiofrequency ablation)
and other conditions (e.g., portal vein
thrombosis, bile duct obstruction, tumor
burden) potentially affecting the liver sil-
houette is important. Previously performed
imaging procedures such as MRI or CT
are helpful to gather information about
the liver configuration and localization of
hepatic metastases. Additionally, it is nec-
essary to know the arterial anatomy in the
upper abdomen and the (desired) location
of the port line tip.
If the tracer is injected slowly, extra-
hepatic accumulation can occur only in
perfusion territories of arteries originating
distally from the port line tip. The desired
position of the catheter tip in the case of
a regular arterial anatomy is the proper
hepatic artery, distally of origins of arte-
rial branches supplying organs other then
the liver. However, there are numerous
variants of both the hepatic arteries and
the arterial branches to the gastrointesti-
nal tract. These normal variants may be
relevant for transarterial treatments and
the interpretation of planning and control
imaging. They have recently been exten-
sively reviewed by Liu et al., (2005).
A few arteries, which are potentially
included in the perfusion territory of
intraarterial liver catheters, surgically and
minimal invasively implanted systems
294 T. Denecke et al.

alike, are to be mentioned. The cystic artery
arises usually from the right hepatic artery
(71%). This implies the risk of therapy
related cholecystitis. If the gall bladder has
not been resected previously, intraarterial
drug delivery should be performed distally
from the origin of the cystic artery. The
right gastric artery is a minor contributor
to the gastric blood supply. Variation of
the right gastric artery is common. It may
arise from the proper hepatic artery (51%),
the left (23%) or the right (3%) hepatic
artery, the origin of the GDA (3%) or the
common hepatic artery (9%). The iden-
tification of this vessel is imperative for
regional treatment, as inflammation with
gastric necrosis, mucosal ulceration, and
perforation are potential complications
of inadvertent delivery of RE similarly
as in intraarterial infusion chemotherapy.
Here, insufficient coiling of the right gas-
tric artery during minimally invasive port
implantation, the incidence of endoscopi-
cally confirmed mucosal lesions is 36%,
compared to 3% in cases with sufficient
coiling. Thus, it is important to exclude
insufficient coiling with recanalization of
the vessel (4% of cases), subsequent open-
ing collateral vessels of the right gastric
artery before application of therapeutic
agents proximally to its origin. The GDA
and its branches may be responsible for
duodenal ulceration or pancreatitis dur-
ing regional therapy, when coiled insuffi-
ciently or subsequently opening collaterals
were initially not visible during coiling
and MAA test injection (Liu et al., 2005).
If extrahepatic perfusion of non-target ves-
sels is detected before therapy application
(Figure 22.1), a reintervention or a change
of the therapeutic management becomes
necessary (Denecke et al., 2005).
Tc-99m-MAA deposition to non-target
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
is accumulation in the spleen, indicating
reflux or dislocation of port lines into the
celiac trunk. Accumulation in the gastric
wall in relevant amounts can be depicted
on planar images, if a focal or longitudinal
tracer retention occurs clearly separated
from the liver silhouette. Discrete accumu-
lations in the right gastric wall close to the
liver, however, can be obscured by the inten-
sive uptake of the adjacent hepatic paren-
chyma. These subtle findings can usually
be depicted on axial and coronal SPECT
images. This, as well as accumulation in
the gall bladder or the GDA bed, is best
diagnosed on SPECT-CT fusion images
(Figure 22.1). The clear outline of the liver
on the anatomic images help to discriminate
extrahepatic foci from hepatic uptake. This
especially holds true in the cases with an
untypical configuration of the liver or a het-
erogeneous intrahepatic perfusion pattern.
For an effective regional chemotherapy
of liver metastases, the intrahepatic tran-
sarterial supply with chemotherapy agents
should address the entire liver or at least
reach a sufficient level in those segments
with tumor burden. Heterogeneous per-
fusion detected on Tc-99m-MAA SPECT
can be due to several reasons, such as
hyper- or hypoperfusion of tumors or
vascular alterations. Heterogeneous int-
rahepatic perfusion patterns that do not
represent malsupply of tumor baring seg-
ments and do not require an alteration of
the therapeutic management, can poten-
tially be misinterpreted on scintigraphic
images because of the lack of anatomical
background information. These are for
example, necrotic areas within hepatic
tumors and areas previously treated with
chemoembolization or interstitial tumor
ablation (e.g., radiofrequency ablation).
Other potential causes of misinterpretation
295
are change of the liver silhouette due to
surgical treatment (e.g., hemihepatectomy,
enucleation) or tumor and treatment related
dystrophy. To distinguish sufficient supply
with heterogeneous accumulation (that
are not affecting therapy efficacy) from
true hypoperfusion or perfusion defects
on scintigraphic images, it is necessary
to have anatomic images for correlation.
Ideally, image fusion helps to compare
the scintigraphically visualized perfusion
territory and the liver outlines, which is
important in cases with atypical liver con-
figuration. Additionally, heterogeneous
MAA accumulations on SPECT images
can be correlated to the size, shape, and
distribution of hepatic tumor lesions on
CT or MRI to clearly identify unsup-
plied liver segments, especially those with
tumor involvement.
A lack of supply can be caused by sev-
eral problems. A (subtotal) port line occlu-
sion is diagnosed easily by unsuccessful
saline injection and requires recanaliza-
tion. System leakage and disconnection is
discussed above. Occlusion of the arterial
vessels supplied by the port catheter can
also occur. This can be due to intima irri-
tation by the catheter or the chemotherapy
agent, resulting in dissection and/or throm-
bosis. Tracer reflux to proximal non-target
vessels is seen in the case of a total occlu-
sion of the hepatic bed. This is sometimes
accompanied by a prominent focal tracer
retention in the region of the catheter tip,
which, however, is not a pathognomonic
sign for arterial occlusion. A partial lack
of hepatic supply can be observed in the
case of occlusion of the right or left liver
artery only. Just as perfusion of non-
target vessels, these are important findings
and require recanalization efforts before
resumption of therapy infusion.
296
Another reason for hypoperfusion or
perfusion defects within the liver are aber-
rant hepatic arteries. There exist numer-
ous variants of the anatomy of hepatic
arteries. The relevant variants for minimal
invasive implantation and scintigraphic
control of intraarterial liver ports are listed
in the following. An aberrant or accessory
right hepatic artery occurs in 10–31%
of cases, with 96% of these originating
from the superior mesenteric artery (Liu
et al., 2005). The caliber of accessory
right hepatic arteries varies as well as the
supplied volume of liver. An accessory or
replaced left hepatic artery is present in
12–21% of cases. It usually arises from the
left gastric artery and thus communicates
with the gastric and esophageal bed. Other
relevant variants are double hepatic arter-
ies with separate origin from the aorta or
an accessory middle hepatic artery (Liu
et al., 2005).
Regarding the application of therapeutic
agents for HAIC as a continuous proce-
dure or for RE as a one-time procedure,
specific considerations have to be made
for both intervention and scintigraphic
control. For a one-time procedure, it is
possible to divide the amount of therapeutic
agents into two or more portions and to
infuse them over an angiographic cath-
eter selectively into the supplying arteries
of the liver or single liver lobes. This is
advantageous not only because this tech-
nique is independent from anatomic vari-
ants with separate hepatic arteries but also
for fractionation of the therapy into two
single lobe treatment sessions in order to
preserve a hepatic reserve. In contrast, con-
tinuing infusions of chemotherapy require
a single fix location of the port line, flood-
ing the arterial bed of the entire liver. In
patients with regular arterial anatomy, this
T. Denecke et al.
is possible by placing the catheter tip into
the proper hepatic artery. In patients with
aberrant hepatic arteries, however, a tactic
is needed, which allows for perfusion of the
entire liver without changing the location
of agent release. There are different strate-
gies in use: (1) catheter implantation into
the major feeder and neglect of the minor
hepatic arteries (Figure 22.2); (2) catheter
implantation into the dominant vessel and
proximal coil embolization of the minor
hepatic artery; (3) Catheter implantation
into the splenic artery with the therapeutic
agent entering the liver over the portal
vein; or (4) double port implantation into
hepatic arteries of separate origin with a
port reservoir in each groin are not to be
favored as the supply of hepatic malignan-
cies is predominantly arterial not portal,
and a double port system poses a doubled
risk of dysfunction or complication (Ricke
et al., 2004).
To ignore a minor accessory vessel results
in a parallel blood flow. Macroaggregated
albumine (MAA) as well as therapeutic
agents injected in one hepatic artery reaches
only this territory, while the neighboring
territory is fed by the ignored vessel. This
is represented by a signal defect on per-
fusion SPECT (Denecke et al., 2005). To
interpret this defect correctly, the outline
of the liver (visualized on SPECT-CT)
as well as the arterial anatomy and the
desired location of the port line are needed
(Figure 22.2).
To avoid hypoperfusion in tumor bearing
segments which are supplied by accessory
vessels, a different strategy is needed. It is
proposed that proximal coil embolization
of one liver artery induces intrahepatic
collateralization from the remaining feeder
by lowering the filling pressure in the
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)

of accessory arteries causes some col- interpretation of scintigraphic findings,

lateralization from the remaining artery,
and also increased portal vein supply.
Even though the arterial supply and thus
the supply with therapeutic agents in the
embolized territory is reduced, it is still
the best achievable result. On perfusion
SPECT, this imposes a hypoperfused area
(Denecke et al., 2005). Again, for correct
knowledge about the arterial anatomy and
the shape of the liver are required.
Therapeutic Consequences
A diagnostic modality has to be judged
by the potential impact on the therapeutic
management. More recently published
298 T. Denecke et al.

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
(Seki et al., 2003). In contrast to scintigra-
phy, both techniques provide a high level
of anatomic information, but depend on
dynamic contrast media infusion, whereas
Tc-99m-MAA enables static imaging of
the perfusion pattern. In addition, contrast
revealed by scintigraphic imaging in dif-
ferentiating perfused from non-perfused
tissue is higher compared with CT and
MRI. Thus, a combination of morphologi-
cal and scintigraphically acquired func-
tional information is warranted.
Correlation of SPECT images with sep-
arately generated CT images to assign
perfusion patterns to anatomic structures
using a simple side by side analysis is
not promising, especially when inhomo-
geneous tracer distributions are present
(Amthauer et al., 2005). For sufficient
retrospective fusion, one would need pro-
spective planning of both examinations for
optimal image fusion, despite the problems
of breathing motion of the liver which are
hard to overcome. Additionally, the lack of
anatomic information in the SPECT data
set is problematic for fusion algorithms
based on mutual information (Amthauer
et al., 2005).
To overcome these problems, a prom-
ising approach is the combination of
scintigraphic perfusion imaging with
morphologic correlation in one exami-
nation and inherent fusion of data sets.
These requirements are ideally fulfilled
by hybrid SPECT-CT systems. Breathing
excursions of the liver are averaging dur-
ing the relatively long acquisition times
of both SPECT (30 min) and low dose CT
(9 min) (Amthauer et al., 2005; Denecke
et al., 2005). The anatomic resolution of
the integrated low dose CT is sufficient
for delineation of liver silhouette and
neighboring organs (Denecke et al., 2005).
299
Therefore, it is possible to accurately
assess the perfusion territory of intra-
arterial port systems with SPECT-CT and
Tc99m-MAA. This enables not only exclu-
sion of extrahepatic accumulation but also
evaluation regarding catheter dislocation
and occlusion of hepatic arteries. Indeed,
the impact on the therapeutic management
of 24 patients observed in a recent study
was considered relevant (33%) (Denecke
et al., 2005). This was largely due to
subsequently dislocated port catheters
and extrahepatic perfusion which can be
treated by reintervention before initiating
the next cycle of chemotherapy. In another
patient in whom chemotherapy application
was stopped because of abdominal pain,
extrahepatic perfusion could be excluded
with SPECT-CT.
In addition, mapping of intrahepatic
tracer distribution with visualization of
lobes and single segments is possible on
SPECT-CT images. When correlated to
the initial CT scan it can be estimated if
metastases are located in sufficiently sup-
plied regions, or if reduced chemotherapy
supply has to be suggested. The latter was
observed predominantly in patients with
abnormal anatomy of the hepatic arteries
and consecutive problematic port place-
ment (Denecke et al., 2005). Whether a
reduced supply of segments with metas-
tases can be tolerated with the assump-
tion of increasing intrahepatic arterial
collateralization over time is not clear to
date. Further follow up and evaluation of
response rates are required.
Regarding RE, in our opinion the
employment of SPECT-CT for perfusion
scintigraphy is decisive. The risk of gas-
trointestinal morbidity is more severe
than in HAIC. Once implanted, the pro-
gression of embolic and radiation effects
300
on the tissues hardly be stopped. For safety
reasons, the best possible modality for
exclusion of inadvertent gastrointestinal
implantation of diagnostic and therapeutic
agents should be used, which is currently
perfusion scintigraphy and Bremsstrahlung
scintigraphy with SPECT-CT.
In conclusion, SPECT-CT improves
identification and localization of areas
with reduced chemotherapy agent supply
and eventual extrahepatic perfusion in
patients receiving HAIC. This has a rele-
vant impact on the therapeutic strategy and
enables treatment optimization. In patients
scheduled for RE with Y90-microspheres,
SPECT-CT is a very valuable tool to detect
or exclude inadvertent deposition of tracer
into gastrointestinal non-target vessels,
and thus helps to increase the safety of this
procedure.
REFERENCES
Amthauer, H., Denecke, T., Rohlfing, T., Ruf, J.,
Bohmig, M., Gutberlet, M., Plockinger, U.,
Felix, R., and Lemke, A.J. 2005. Value of image
fusion using single photon emission computed
tomography with integrated low dose computed
tomography in comparison with a retrospective
voxel-based method in neuroendocrine tumours.
Eur. Radiol. 15: 1456–1462.
Andrews, J.C., Walker, S.C., Ackermann, R.J.,
Cotton, L.A., Ensminger, W.D., and Shapiro, B.
1994. Hepatic radioembolization with yttrium-90
containing glass microspheres: preliminary
results and clinical follow-up. J. Nucl. Med. 35:
1637–1644.
Bocher, M., Balan, A., Krausz, Y., Shrem, Y.,
Lonn, A., Wilk, M., and Chisin, R. 2000.
Gamma camera-mounted anatomical X-ray
tomography: technology, system characteris-
tics and first images. Eur. J. Nucl. Med. 27:
619–627.
Campbell, A.M., Bailey, I.H., and Burton, M.A.
2001. Tumour dosimetry in human liver fol-
lowing hepatic yttrium-90 microsphere therapy.
Phys. Med. Biol. 46: 487–498.
T. Denecke et al.
Cantore, M., Mambrini, A., Fiorentini, G., Rabbi, C.,
Zamagni, D., Caudana, R., Pennucci, C.,
Sanguinetti, F., Lombardi, M., and Nicoli, N.
2005. Phase II study of hepatic intraarterial
epirubicin and cisplatin, with systemic 5-fluor-
ouracil in patients with unresectable biliary tract
tumors. Cancer 103: 1402–1407.
Cohen, A.D., and Kemeny, N.E. 2003. An update
on hepatic arterial infusion chemotherapy for
colorectal cancer. Oncologist 8: 553–566.
Dancey, J.E., Shepherd, F.A., Paul, K., Sniderman,
K.W., Houle, S., Gabrys, J., Hendler, A.L.,
and Goin, J.E. 2000. Treatment of nonresect-
able hepatocellular carcinoma with intra-
hepatic 90Y-microspheres. J. Nucl. Med. 41:
1673–1681.
Denecke, T., Hildebrandt, B., Lehmkuhl, L., Peters, N.,
Nicolaou, A., Pech, M., Riess, H., Ricke, J.,
Felix, R., and Amthauer, H. 2005. Fusion imag-
ing using a hybrid SPECT-CT camera improves
port perfusion scintigraphy for control of hepatic
arterial infusion of chemotherapy in colorectal
cancer patients. Eur. J. Nucl. Med. Mol. Imaging
32(9): 1003–1010.
Gray, B., Van Hazel, G., Hope, M., Burton, M.,
Moroz, P., Anderson, J., and Gebski, V. 2001.
Randomised trial of SIR-Spheres plus chemo-
therapy vs. chemotherapy alone for treating
patients with liver metastases from primary large
bowel cancer. Ann. Oncol. 12: 1711–1720.
Heinrich, S., Petrowsky, H., Schwinnen, I., Staib-
Sebler, E., Gog, C., El-Ganainy, A., Gutt, C.,
Muller, H.H., and Lorenz, M. 2003. Technical
complications of continuous intra-arterial chem-
otherapy with 5-fluorodeoxyuridine and 5-fluor-
ouracil for colorectal liver metastases. Surgery
133: 40–48.
Kemeny, N., Huang, Y., Cohen, A.M., Shi, W.,
Conti, J.A., Brennan, M.F., Bertino, J.R.,
Turnbull, A.D., Sullivan, D., Stockman, J.,
Blumgart, L.H., and Fong, Y. 1999. Hepatic
arterial infusion of chemotherapy after resection
of hepatic metastases from colorectal cancer. N.
Engl. J. Med. 341: 2039–2048.
Kerr, D.J., McArdle, C.S., Ledermann, J., Taylor, I.,
Sherlock, D.J., Schlag, P.M., Buckels, J., Mayer,
D., Cain, D., and Stephens, R.J. 2003. Medical
Research Council’s colorectal cancer study group;
European Organisation for Research and Treatment
of Cancer colorectal cancer study group.
22. Perfusion Scintigraphy with Integrated Single Photon Emission
Intrahepatic arterial versus intravenous fluor-
ouracil and folinic acid for colorectal cancer
liver metastases: a multicentre randomised trial.
Lancet 361: 368–373.
Lehner, K., Reiser, M., Gebhardt, U., Heuck, A.,
and Schaff, J. 1987. DSA – control of implanted
devices for arterial hepatic perfusion. Cardiovasc.
Intervent. Radiol. 10: 71–74.
Liu, D.M., Salem, R., Bui, J.T., Courtney, A.,
Barakat, O., Sergie, Z., Atassi, B., Barrett, K.,
Gowland, P., Oman, B., Lewandowski, R.J.,
Gates, V.L., Thurston, K.G., and Wong, C.Y.
2005. Angiographic considerations in patients
undergoing liver-directed therapy. J. Vasc. Interv.
Radiol. 16: 911–935.
Lorenz, M., and Muller, H.H. 2000. Randomized,
multicenter trial of fluorouracil plus leucov-
orin administered either via hepatic arterial or
intravenous infusion versus fluorodeoxyurid-
ine administered via hepatic arterial infusion
in patients with nonresectable liver metastases
from colorectal carcinoma. J. Clin. Oncol. 18:
243–254.
Lubin, E., Cyjon, A., and Neuman, M. 1987.
Technetium-99m macroaggregates for the study
of the distribution of arterial infusion chemother-
apy in the liver. Clin. Nucl. Med. 12: 385–388.
Meta-Analysis Group in Cancer (MAGC). 1996.
Reappraisal of hepatic arterial infusion in the
treatment of nonresectable liver metastases from
colorectal cancer. J. Natl. Cancer Inst. 88:
252–258.
Morimoto, M., Satake, M., Sekiguchi, R., Haruno, M.,
and Moriyama, N. 1999.
Optimal injection protocol for CT evaluation during
hepatic arterial infusion chemotherapy. Invest.
Radiol. 34: 744–750.
Murthy, R., Nunez, R., Szklaruk, J., Erwin, W.,
Madoff, D.C., Gupta, S., Ahrar, K., Wallace,
M.J., Cohen, A., Coldwell, D.M., Kennedy, A.S.,
and Hicks, M.E. 2005. Yttrium-90 microsphere
therapy for hepatic malignancy: devices, indica-
tions, technical considerations, and potential
complications. Radiographics 25(1): S41–S55.
Nelson, R.L., and Freels, S. 2004. A systematic
review of hepatic artery chemotherapy after
hepatic resection of colorectal cancer metastatic
to the liver. Dis. Colon Rectum. 47: 739–745.
Pelosi, E., Bar, F., Battista, S., Bello, M., Bucchi,
M.C., Alabiso, O., Molino, G., and Bisi, G.
301
1999. Hepatic arterial infusion chemotherapy
for unresectable confined liver metastases: pre-
diction of systemic toxicity with the applica-
tion of a scintigraphic and pharmacokinetic
approach. Cancer Chemother. Pharmacol. 44:
505–510.
Ricke, J., Hildebrandt, B., Miersch, A., Nicolaou, A.,
Warschewske, G., Teichgraber, U., Lopez
Hanninen, E., Riess, H., and Felix, R. 2004.
Hepatic arterial port systems for treatment of
liver metastases: factors affecting patency and
adverse events. J. Vasc. Interv. Radiol. 15:
825–833.
Sadahiro, S., Suzuki, T., Ishikawa, K., Yasuda, S.,
Tajima, T., Makuuchi, H., Saitoh, T., and
Murayama, C. 2004. Prophylactic hepatic arte-
rial infusion chemotherapy for the prevention
of liver metastasis in patients with colon carci-
noma: a randomized control trial. Cancer 100:
590–597.
Seki, H., Ozaki, T., Takaki, S., Ooi, H., Oda, J.,
and Shiina, M. 2003. Using slow-infusion MR
arteriography and an implantable port system
to assess drug distribution at hepatic arterial
infusion chemotherapy. AJR Am. J. Roentgenol.
180: 681–686.
Suzuki, Y., Kobayashi, S., and Yasuda, S. 1991.
Application of single-photon emission com-
puted tomography (SPECT) with 99mTc-MAA
in evaluation of perfusion patterns during hepatic
infusion chemotherapy. Ann. Nucl. Med. 5:
123–126.
Tanaka, S., Shimada, M., Shirabe, K., Maehara, S.,
Harimoto, N., Tsujita, E., Sugimachi, K.,
and Maehara, Y. 2005. A novel intrahepatic
arterial chemotherapy after radical resec-
tion for advanced hepatocellular carcinoma.
Hepatogastroenterology 52: 862–865.
Van Hazel, G., Blackwell, A., Anderson, J., Price, D.,
Moroz, P., Bower, G., Cardaci, G., and Gray, B.
2004 Randomised phase 2 trial of SIR-Spheres
plus fluorouracil/leucovorin chemotherapy ver-
sus fluorouracil/leucovorin chemotherapy alone
in advanced colorectal cancer. J. Surg. Oncol.
88: 78–85.
Ziessman, H.A., Gyves, J.W., Juni, J.E., Wahl,
F.L., Thrall, J.H., Ensminger, W.D., Goldstein,
H.A., and Dubiansky, V. 1985. Atlas of hepatic
arterial perfusion scintigraphy. Clin. Nucl. Med.
10: 675–681.
23
Postoperative Interferon Alpha
Treatment of Patients with Hepatocellular
Carcinoma: Expression of p48 Using
Tissue Microarray
Hui-Chuan Sun
INTRODUCTION
Hepatocellular carcinoma (HCC) is a major
health problem in China and the third most
common cause of cancer death in the world
(Parkin et al., 2005). Surgical resection
can have a curative outcome, but > 50%
of patients will have recurrence in the
first 5 years after resection (Poon, 2007),
and the recurrence usually results in death
(Regimbeau et al., 2004). No standard post-
operative adjuvant therapy has yet been
established (Sun and Tang, 2003).
In a previous study, we demonstrated
that interferon-α (IFN-α) delays tumor
growth and inhibits metastasis in a nude
mouse model (Wang et al., 2000). Further
studies revealed that both antiangio-
genesis and antiproliferative effects are
involved in the antitumor effect in the
HCC model (Wang et al., 2003; Wu et al.,
2004, 2005). In addition to the direct
antitumor effect, the antivirus effect may
also be involved in the prevention of
recurrence of HCC. Some reports have
indicated that long-term IFN-α reduces
the incidence of HCC in patients with
chronic hepatitis B virus (HBV) or hepatitis
C virus (HCV) infection (Lin et al., 2007;
Yu et al., 2006).
The consensus is that recurrence of HCC
after resection may originate from intra-
hepatic metastasis of the primary tumor,
or from de novo carcinogenesis in the
remnant liver (Sun et al., 2005). A number
of randomized controlled studies have
demonstrated the effect of postoperative
IFN-α therapy on the prevention of tumor
recurrence. Ikeda et al. (2000) and Lin
et al. (2004) concluded that the preventive
effect of IFN-α therapy can be attributed
to the clearance of HBV or HCV, result-
ing in inhibition of hepatocarcinogenesis.
However, other investigators concluded
that the direct antitumor effect of IFN-α
plays an important role in the prevention of
tumor recurrence (Kubo et al., 2001; Sun
et al., 2006). Furthermore, our study did
not find a relationship between conversion
of hepatitis B e antigen (HBeAg) and out-
come of IFN-α therapy (Sun et al., 2006).
In our previous randomized control-
led trial, 28% of IFN-α-treated patients
experienced recurrence within 1 year after
resection (Sun et al., 2006). Clearly, IFN-α
therapy was not effective in every patient.
303
304
On the other hand, IFN-α therapy pro-
duced some side effects in patients who
did not respond therapeutically to IFN-α.
Therefore, a predictive marker of response
to IFN-α is needed to avoid unnecessary
treatment costs and side effects.
A wealth of information is available
on the molecular processes underlying
IFN-α-induced signaling. Binding of
IFN-α to the type I receptor brings the
receptor-associated tyrosine kinases Tyk2
and Jak1 in close proximity, which allows
the kinases to phosphorylate, leading to
their activation and the generation of recep-
tor docking sites for the latent cytoplasmic
protein STAT (Wu et al., 2004). Tyrosine-
phosphorylated STAT1 and STAT2 also
associate as heterodimers to form IFN-
α-stimulated gene factor 3 (ISGF3α) and
translocate to the nucleus, where they asso-
ciate with the DNA-binding protein ISGF3γ
(P48) to form ISGF3, which then binds
to the IFN-stimulated response element
to induce transcription of several IFN-
inducible genes (Ghislain et al., 2001). Our
previous study showed that transduction
of P48 in resistant HCC cells can restore
sensitivity to IFN-α, a finding that further
confirms the importance of Jak/STAT sig-
naling in the antiproliferative effects of
IFN-α in HCC cells (Wu et al., 2005). We
believe that P48 expression in tumor cells
is related to the effect of IFN-α. Therefore,
we hypothesized that P48 expression would
predict outcome in the patients receiving
IFN-α therapy who were enrolled in the
randomized controlled trial presented here.
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
H.-C. Sun
surgical team at the Liver Cancer Institute
(Fudan University, Shanghai, China) were
enrolled to test the effect of postoperative
IFN-α therapy on the prevention of recur-
rence. All patients were randomized into
either the IFN-α therapy group (5 MU three
times a week for 18 months or until recur-
rence was diagnosed) or the control group
(no antitumor treatment, until recurrence
was diagnosed). The details of this trial
have been reported elsewhere (Sun et al.,
2006).
In this study, 80 samples from the IFN-α
group (group 1) and 75 samples from the
control group (group 2) were retrieved from
the Department of Pathology, Zhongshan
Hospital. All samples were from patients
who followed the treatment protocol.
Tissue Microarray and
Immunohistochemistry
After careful screening of hematoxylin-
eosin-stained slides for optimal tumor con-
tent, we constructed tissue microarray slides
in collaboration with Shanghai Biochip
Co. Briefly, two 0.8-mm-diameter punch
cores were taken from each formalin-fixed,
paraffin-embedded HCC sample. Samples
were obtained from nonnecrotic tumor foci.
Immunohistochemistry for P48 was per-
formed using the avidin-biotin complex
(Vector Laboratories, Burlingame, CA),
including heat-induced antigen retrieval
procedures. Incubation with polyclonal anti-
bodies against P48 (P48:C20, 1:100; Santa
Cruz Biotechnology, Santa Cruz, CA) was
carried out at 4°C for 18 h (Clifford et al.,
2000). Negative controls were treated identi-
cally but with the primary antibody omitted.
23. Postoperative Interferon Alpha Treatment of Patients with Hepatocellular Carcinoma
P48 immunoreactivity. The percentage of
positive tumor cells was determined by each
observer, and the average of three scores
was calculated. We randomly selected 10
high-power fields (400 ×, 100 cells per
high-power field); 1,000 cells were counted
in each core. In this study, P48 expression
was graded as follows: tumors with posi-
tive cytoplasm staining in < 20% of tumor
cells were graded as negative, and tumors
with positive cytoplasm staining in ≥ 20%
of tumor cells were graded as positive.
Statistical Analysis
Continuous variables were expressed as
mean ± standard deviation and were com-
pared between groups using the Student t
test. Categorical variables were compared
using the chi-square test. Survival curves
were computed using the Kaplan-Meier
method and were compared between groups
using the log-rank test. To elucidate factors
that could have contributed to improved sur-
vival results, we performed an analysis to
identify the prognostic factors for disease-
free survival (DFS) and overall survival.
Univariate and then multivariate analyses
using the Cox proportional hazards model
were performed on 13 factors with poten-
a b
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×)
305
tial prognostic significance, which were
all categorized as binary variables. These
included four clinical factors (age < or >
60 years, gender, HBeAg status, and serum
α-fetoprotein level ≤ or > 20 ng/ml) and 9
pathological factors (cirrhotic or noncir-
rhotic liver, tumor size ≤ or > 5 cm, single
or multiple tumor nodules, degree of tumor
encapsulation, Edmondson’s grade I/II or
III/IV, presence or absence of microvascu-
lar invasion, pTNM stage I/II or III, CLIP
0 or 1/2 (Yan and Yan, 2003), and P48
staining). Receiver operating characteristic
(ROC) analysis was used to assess the pre-
dictive accuracy of prognostic factors; we
used the free software developed by Metz at
the University of Chicago (http://xray.bsd.
uchicago.edu/krl/roc_soft.htm). All statis-
tical analyses were performed using statistical
software (SPSS Inc., Chicago, IL). Statistical
significance was defined as P < 0.05.
RESULTS
Clinicopathological Data
Positive P48 expression was found in 81
patients, and negative P48 expression was
found in 74 patients (Figure 23.1A, B).
c
306
P48 was stained mainly in the cytoplasm
of HCC cells but occasionally in the
nucleus of HCC cells and in the bile duct
(Figure 23.1C). There was no statistically
significant difference in clinicopatholog-
ical data, except gender, between patients
with positive and those with negative P48
expression. There was no statistically
significant difference from patients who
received IFN-α (group 1) and those who
did not (group 2) between patients with
positive and those with negative P48
expression.
Survival
In group 1 (IFN-α therapy group), DFS
was significantly better in patients with
positive P48 expression compared with
those with negative P48 expression (haz-
ard ratio, 1.8; 95% CI, 1.0–3.2; P = 0.036).
The cumulative 1-, 3-, and 5-year DFS
rates were 84%, 55%, and 36% among
patients with positive P48 expression and
53%, 34%, and 29% among those with
negative P48 expression, respectively. In
the control group, however, there was no
significant difference in DFS between
patients with positive and those with nega-
tive staining (hazard ratio, 1.4; 95% CI,
0.8–2.6; P = 0.208).
In group 1, overall survival was signifi-
cantly better in patients with positive P48
expression than in those with negative P48
expression (hazard ratio, 2.7; 95% CI,
1.2–6.2; P = 0.014). Again, in group 2,
there was no significant difference in over-
all survival between patients with positive
and those with negative staining (hazard
ratio, 1.2; 95% CI, 0.6–2.3; P = 0.545).
When we analyzed data from patients
with positive P48 expression, we found that
overall survival in group 1 was better than in
H.-C. Sun
group 2 (hazard ratio, 3.9; 95% CI, 1.7–9.2;
P = 0.001). The mean overall survival was
57.3 months (95% CI, 53.0–61.5 months)
in group 1 and 41.5 months (95% CI,
34.0–49.0 months) in group 2, but the differ-
ence was not statistically significant (hazard
ratio, 1.5; 95% CI, 0.8–2.9; P = 0.246).
Prognostic Factors for Disease-Free
Survival and Overall Survival
in Group 1
Univariate analysis of prognostic factors for
DFS revealed that of all evaluated factors,
only liver cirrhosis (P = 0.000) and P48
expression (P = 0.011) had a significant prog-
nostic influence. After multivariate analysis,
only liver cirrhosis (present vs. absent: risk
ratio [RR] 4.9; 95% CI, 1.4–16.9; P = 0.012)
and pattern of P48 staining (negative vs. pos-
itive: RR, 2.0; 95% CI, 1.1–3.8; P = 0.028)
remained significant prognostic factors for
DFS . Receiver operating characteristic anal-
ysis showed that these two prognostic factors
have good predictive accuracy with regard
to 1-year DFS (area under the curve [AUC]
= 0.649; 95% CI, 0.5–0.8; P < 0.01) and
acceptable accuracy in terms of 2-year DFS
(AUC = 0.604; 95% CI, 0.5–0.7; P = 0.05).
Univariate analysis of prognostic fac-
tors for overall survival revealed that of
all evaluated factors, liver cirrhosis (P =
0.032), HBeAg positivity (P = 0.009),
and P48 expression (P = 0.014) had a sig-
nificant prognostic influence. After multi-
variate analysis, only Edmondson’s stage
(high grade [III/IV] vs. low grade [I/II]:
RR, 5.6; 95% CI, 2.0–15.7; P = 0.001),
and pattern of P48 staining (negative vs.
positive: RR, 3.8; 95% CI, 1.4–10.3; P =
0.008) were significant prognostic factors
for overall survival. ROC analysis showed
that these two prognostic factors have a
23. Postoperative Interferon Alpha Treatment of Patients with Hepatocellular Carcinoma
impressive predictive accuracy for overall
survival in the first 4 years (AUCs for 1-,
2-, 3-, and 4-year survival were 0.883 [P <
0.01], 0.737 [P < 0.01], 0.646 [P = 0.01],
and 0.634 [P = 0.03], respectively).
DISCUSSION
A gene expression profile and IFN-α/type
2 interferon receptor (IFNAR2) have been
reported to predict response of HCC to the
combination of IFN-α and 5-fluorouracil
(Kurokawa et al., 2004; Ota et al., 2005).
In melanoma and transitional cell carci-
noma, a decrease in P48 expression was
associated with poor response to IFN-α
(Matin et al., 2001). Other molecular
markers, such as SOCS-1 and SOCS-3,
acted as negative feedback regulators of
Jak/STAT and determined sensitivity to
IFN (Takayama et al., 2000). However,
until now there have been no reports on a
predictive marker for outcome after post-
operative IFN-α therapy in patients with
HBV-related HCC.
In the present tissue microarray analy-
sis, we found no significant differences
in DFS and in overall survival between
control subjects (patients not treated with
postoperative IFN-α) with positive and
those with negative P48 expression; how-
ever, significant differences were found
between treated patients with positive
and those with negative P48 expression.
Furthermore, overall survival in patients
with positive P48 expression who were
treated with IFN-α was better than over-
all survival in patients with positive P48
expression who were not treated with IFN-
α. These data clearly demonstrate that P48
expression is associated with outcome after
postoperative IFN-α therapy. Our results
307
also show that patients with poorly dif-
ferentiated HCC and/or liver cirrhosis may
have a poorer outcome. We noticed that the
predictive accuracy of P48 expression in
terms of DFS and overall survival dimin-
ished over time. The reason is that IFN-α
was administered for 18 months to patients
in the treatment group, and P48 expression
was used to predict the effect of IFN-α
therapy, so the predictive accuracy of P48
expression was good in the first several
years and diminished after IFN-α therapy
was stopped. As we found in our clinical
trial (Sun et al., 2006), 18 months of IFN-α
therapy delayed tumor recurrence but did
not reduce the incidence of recurrence over
5 years; a number of patients developed
tumor recurrence after IFN-α therapy was
stopped. However, overall survival within
the 5-year period was improved by IFN-α
therapy, because of delayed recurrence in
the treatment group.
Tissue microarray analysis was first
introduced by Kononen et al. in 1998. It
is a high-throughput approach for histo-
logic examination of tissue for the pur-
pose of determining protein expression.
However, high-throughput platforms are
compromised by a loss of completeness of
data. Therefore, several important issues
need to be considered in conducting tissue
microarrays.
First, adequate sample size is important.
Tumor heterogeneity is of major concern in
interpreting results from tissue microarrays
(Rubin et al., 2002). In well-established
methods of immunohistochemistry, about
0.05% of tissue in a routine section rep-
resents 1 cc of tumor tissue (Camp et al.,
2000), whereas a tissue microarray section
represents a greatly reduced amount of tis-
sue. Obtaining more tissue microarray cores
should yield a more accurate result with
308
less variability, but this is not always pos-
sible, especially when the archived block
has been repeatedly used or when a fine-
needle biopsy sample is used. Furthermore,
obtaining too many cores reduces the value
of tissue microarray analysis. Rubin et
al. (2002) analyzed the results of a tis-
sue microarray test for Ki67 expression
in prostate cancer and found that three
0.8-mm-diameter cores are necessary and
enough to obtain accurate results, but other
researchers have reported that one or two
cores are enough (Camp et al., 2000; Lugli
et al., 2004). Currently, there is no hard-
and-fast guideline for determining optimal
sample size (Hewitt, 2006). In this study,
two 0.8-mm-diameter cores were taken
from each specimen.
Second, nonrepresentative samples should
be avoided. The location of sampling sites
may play an important role. In this study, the
location of each core was decided in col-
laboration with an experienced pathologist.
The principle is to take cores from the non-
necrotic portion of the sample and to collect
as many intact tumor cells as possible.
The third challenge is to establish a
scoring system for evaluation of P48
expression. In the present study, three
pathologists reviewed every slide and
noted a mean of 348, 338, and 361 positive
tumor cells per patient. Standard devia-
tions were 355, 348 and 359, and standard
deviation did not increase as the corre-
sponding mean increased.
An implication of this study is that tissue
microarray analysis could be used to deter-
mine the significance of gene expression
in a larger sample size. Although we found
P48 expression to be associated with sen-
sitivity of liver cancer cells to IFN-α, we
are not sure about the significance of P48
expression in patients who receive IFN-α.
H.-C. Sun
Tissue microarray analysis is the best tool
for evaluating the clinical significance
of a specific gene expression in a larger
number of specimens with different sub-
types of tumor.
It is also notable that approximately
16% of patients with positive P48 expres-
sion received a diagnosis of recurrence
within 1 year. One explanation for this
may be that expression of P48 is not suf-
ficient for antiproliferation in Jak/STAT
signaling. A previous study found that lack
of STAT1 expression also led to resistance
to growth inhibitory effects of IFN-α in
a liver cancer cell line (Murphy et al.,
2001). In addition to antiproliferation,
IFN-α has other antitumor mechanisms,
such as immunomodulation and antian-
giogenesis (Jonasch and Haluska, 2001).
Our previous study showed that the antian-
giogenesis mechanism involved another
signal transduction pathway, despite P48
deficiency in the HCC cell line (Wu et
al., 2005). Meanwhile, the natural limita-
tions of tissue microarray analysis, such as
tumor heterogeneity, nonuniform tissue-
fixation processes, or even interpretation
of immunoactivity, may also compromise
predictive accuracy.
As pointed out by Kononen et al. (1998),
the tissue microarray test can be used with
a large number of samples to profile more
general trends related to the tissue type
being studied; the test is not geared toward
determining the exact protein expression
of an individual case. Tissue microarrays
are designed to analyze a large number
of samples from archived paraffin blocks
but are not suitable for routine clinical
analysis. For the routine clinical access of
P48 expression as a screening tool, a pro-
spective study with a standard section and
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
combined therapy with subcutaneous interferon-
alpha and intra-arterial 5-fluorouracil; role of
type 1 interferon receptor expression. Br. J.
Cancer 93: 557–564.
Parkin, D.M., Bray, F., Ferlay, J., and Pisani, P.
2005. Global cancer statistics. 2002. CA Cancer
J. Clin. 55: 74–108.
Poon, R.T. 2007. Optimal initial treatment for
early hepatocellular carcinoma in patients
with preserved liver function: transplantation or
resection? Ann. Surg. Oncol. 14: 541–547.
Regimbeau, J.M., Abdalla, E.K., Vauthey, J.N.,
Lauwers, G.Y., Durand, F., Nagorney, D.M., Ikai, I.,
Yamaoka, Y., and Belghiti, J. 2004. Risk factors
for early death due to recurrence after liver resec-
tion for hepatocellular carcinoma: results of a
multicenter study. J. Surg. Oncol. 85: 36–41.
Rubin, M.A., Dunn, R., Strawderman, M., and
Pienta, K.J. 2002. Tissue microarray sampling
strategy for prostate cancer biomarker analysis.
Am. J. Surg. Pathol. 26: 312–319.
Sun, H.C., and Tang, Z.Y. 2003. Preventive treat-
ments for recurrence after curative resection of
hepatocellular carcinoma – a literature review of
randomized control trials. World J. Gastroenterol.
9: 635–640.
Sun, H.C., Tang, Z.Y., Ma, Z.C., Qin, L.X., Wang, L.,
Ye, Q.H., Fan, J., Wu, Z.Q., and Zhou, X.D.
2005. The prognostic factor for outcome follow-
ing second resection for intrahepatic recurrence
of hepatocellular carcinoma with a hepatitis B
virus infection background. J. Cancer Res. Clin.
Oncol. 131: 284–288.
Sun, H.C., Tang, Z.Y., Wang, L., Qin, L.X., Ma,
Z.C., Ye, Q.H., Zhang, B.H., Qian, Y.B., Wu,
Z.Q., Fan, J., Zhou, X.D., Zhou, J., Qiu, S.J., and
Shen, Y.F. 2006. Postoperative interferon alpha
treatment postponed recurrence and improved
overall survival in patients after curative resec-
tion of HBV-related hepatocellular carcinoma:
a randomized clinical trial. J. Cancer Res. Clin.
Oncol. 132: 458–465.
H.-C. Sun
Takayama, T., Sekine, T., Makuuchi, M., Yamasaki,
S., Kosuge, T., Yamamoto, J., Shimada, K.,
Sakamoto, M., Hirohashi, S., Ohashi, Y., and
Kakizoe, T. 2000. Adoptive immunotherapy to
lower postsurgical recurrence rates of hepatocel-
lular carcinoma: a randomised trial. Lancet 356:
802–807.
Wang, L., Tang, Z.Y., Qin, L.X., Wu, X.F., Sun,
H.C., Xue, Q., and Ye, S.L. 2000. High-dose and
long-term therapy with interferon-alfa inhibits
tumor growth and recurrence in nude mice bear-
ing human hepatocellular carcinoma xenografts
with high metastatic potential. Hepatology 32:
43–48.
Wang, L., Wu, W.Z., Sun, H.C., Wu, X.F., Qin,
L.X., Liu, Y.K., Liu, K.D., and Tang, Z.Y. 2003.
Mechanism of interferon alpha on inhibition of
metastasis and angiogenesis of hepatocellular
carcinoma after curative resection in nude mice.
J. Gastrointest. Surg. 7: 587–594.
Wu, W.Z., Sun, H.C., Gao, Y.Q., Li, Y., Wang, L.,
Zhou, K., Liu, K.D., Iliakis, G., and Tang, Z.Y.
2004. Reduction in p48-ISGFgamma levels con-
fers resistance to interferon-alpha2a in MHCC97
cells. Oncology 67: 428–440.
Wu, W.Z., Sun, H.C., Shen, Y.F., Chen, J., Wang, L.,
Tang, Z.Y., Iliakis, G., and Liu, K.D. 2005.
Interferon alpha 2a down-regulates VEGF
expression through PI3 kinase and MAP kinase
signaling pathways. J. Cancer Res. Clin. Oncol.
131: 169–178.
Yan, P., and Yan, L.N. 2003. Staging of hepatocel-
lular carcinoma. Hepatobiliary Pancreat. Dis.
Int. 2: 491–495.
Yu, M.L., Lin, S.M., Chuang, W.L., Dai, C.Y.,
Wang, J.H., Lu, S.N., Sheen, I.S., Chang, W.Y.,
Lee, C.M., and Liaw, Y.F. 2006. A sustained
virological response to interferon or interferon/
ribavirin reduces hepatocellular carcinoma and
improves survival in chronic hepatitis C: a
nationwide, multicentre study in Taiwan. Antivir.
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
INTRODUCTION
Hepatocellular carcinoma (HCC) which
occurs in 80% of all primary liver is the
fifth most common cancer and the third
most common cause of death from cancer
in the world (Parkin et al., 2001; Befeler
and Di Bisceglie, 2002). The incident
rate of HCC, which is highest in Asia and
Africa, is increasing in Western countries.
In the United States, 19,160 new cancers
of the liver and intrahepatic bile duct
were estimated in 2007, with an estimated
16,780 deaths (Jemal et al., 2007). Chronic
hepatitis B infection is the most common
cause of HCC in China and Southeast Asia
(Muller, 2006), while hepatitis C infection
is of high importance in Europe, Japan
and North America (Bosch et al., 1999).
Other common risk factors for hepatocar-
cinogenesis include alcohol, aflatoxin β1,
cirrhosis, α-1-antitrypsin deficiency and
hereditary hemochromatosis. Currently,
there is no standard treatment for advanced
HCC. Surgical treatments, including hepa-
tectomy and liver transplantation, are
major curative options for HCC. However,
long-term survival rate remains unsat-
isfactory due to tumor recurrence and
metastasis (Poon and Fan, 2004; Lo and
Fan, 2004; Llovet et al., 2005). For bet-
ter stratification and management of HCC
patients, identification of potential mark-
ers for accurate prognosis of HCC patients
in terms of malignancy and survival rate is
indispensable.
Molecular studies of hepatocarcinogen-
esis have identified a number of molecular
biomarkers conferring prognostic signifi-
cance on prognosis and survival prediction
of HCC patients. One of the approaches is
to identify genes specifically or excessively
expressed in HCC tissue as potential tumor
markers. For instance, overexpression of
proto-oncogene c-met in HCC patients is
significantly associated with intrahepatic
metastasis and a short 5-year survival rate
(Ueki et al., 1997). Survivin has been found
to be associated with disease-free survival
and a higher rate of tumor recurrence
(Fields et al., 2004; Ikeguchi et al., 2002).
Specific overexpression of Stat5b is sig-
nificantly correlated with poor disease-free
and overall survival of HCC patients (Lee
et al., 2006). However, HCC is heterogenous
and different etiological factors may lead
to diverse molecular pathways involved
in hepatocarcinogenesis. Current knowl-
edge on prediction of survival rate of
HCC patients is still insufficient so that
313
314
continuous investigation of novel potential
prognostic markers is urgently needed.
Vertebrate Six1 gene is a homolog of the
Drosophila sine oculis (so) gene, which
belongs to a subfamily of the Six class
of homeodomain-containing transcription
factors characterized with a lysine resi-
due within the DNA-binding helix of the
homeodomain (Oliver et al., 1995). The
functions of vertebrate Six1 are involved
in diverse early organ developments of
the brain, ear, eye, muscle and kidney
(Relaix and Buckingham, 1999; Laclef
et al., 2003; Ozaki et al., 2004; Zheng et al.,
2003; Xu et al., 2003). Interestingly, Six1
gene has been found to be deregulated in
human cancers. Overexpression of Six1
occurs in a large percentage of primary
breast cancers and strongly correlates with
metastatic breast lesions (Ford et al., 1998).
Moreover, upregulation of Six1 in human
rhabdomyosarcoma (RMS) is significantly
correlated with progression and metastatic
potential (Yu et al., 2004). In this study,
we provided the first evidence of Six1
expression in HCC patients after liver
resection to investigate its correlation with
clinicopathological significance and sur-
vival rate of HCC patients. We use reverse
transcription-polymerase chain reaction
(RT-PCR) and Western blot analysis to
study both the mRNA and protein level of
Six1 gene in HCC clinical cases aiming
to identify a novel prognostic marker for
predicting survival of HCC patients.
MATERIALS AND METHODS
Cell Lines Reverse Transcription-Polymerase
Nonmetastatic human HCC cell lines Chain Reaction
Hep3B, Huh7 and PLC were purchased Total RNA from liver tissues and cell
from the American Type Culture Collection lines were extracted by TRIZol regent
K. Tak-Pan Ng and K. Man
(Manassas, VA), and metastatic human
HCC cell lines MHCC97L and MHCC97H
were obtained from the Liver Cancer
Institute and Zhongshan Hospital of Fudan
University, Shanghai, the People’s of
Republic of China (Li et al., 2001). All
cells were cultured in DMEM high glucose
medium with 10% FBS (Gibco, Grand
Island, NY) and 1% penicilium and strep-
tomycin at humidified 37°C incubator sup-
plied with 5% CO2.
Clinical Samples
One hundred and three HCC patients who
underwent liver resection between March
1999 and June 2004 were recruited from
Department of Surgery, Queen Marry
Hospital, The University of Hong Kong
contributed 103 pairs of protein extracts
and 72 pairs of RNA extracts from tumor
tissues and adjacent nontumor tissues in
the study. There were 87 men and 16
women. The age of the patients ranged
from 33 to 75 years, with a median age of
55 years. Eighty-six HCC patients (83.5%)
were positive for hepatitis B surface anti-
gen, whereas only three (2.9%) were posi-
tive for hepatitis C virus antibody. Twenty
normal liver tissues were recruited from
living donors at the same hospital. All
donors were examined to ensure that they
were free of liver diseases and hepatitis B
infection. The follow-up duration for the
HCC patients ranged from July 1999 to
November 2005. The study was approved
by the Ethics Committee of the University
of Hong Kong.
24. Hepatocellular Carcinoma: Overexpression of Homeoprotein Six1
according to the manufacturer’s instruc-
tions (Invitrogen, Carlsbad, CA). Each
complementary DNA was synthesized
from 1 μg of RNA extract using the High
capacity cDNA Kit (Applied Biosystems,
Foster City, CA) under the condition of
25°C for 5 min followed by 37°C for 2 h.
For clinical samples, PCR reaction for
Six1 was performed using the Taq PCR
Kit (Promega, Madison, WI) under the
following PCR cycles: 95°C for 5 min,
40 cycles at 95°C for 1 min, 57°C for
1 min and 72°C for 1 min. For internal
control 18S, 30 PCR cycles were used.
For cell line samples, 30-cycle multiplex
PCR reaction was performed by combin-
ing Six1 and 18S primers. PCR products
were visualized by 2% agarose gel electro-
phoresis stained with ethidium bromide.
Primer sets used were as follows: for Six1,
sense 5′ AAG GAG AAG TCG AGG
GGT GT-3′, antisense 5′-TGC TTG TTG
GAG GAG GAG TT-3′; for 18S ribosomal
RNA, sense 5′-CTC TTA GCT GAG TGT
CCC GC-3′, antisense 5′-CTG ATC GTC
TTC GAA CCT CC-3′.
Western Blot
Proteins from clinical specimens and HCC
cell lines were prepared by using Urea
buffer (8M Urea, pH 8.0). The amount
of protein lysates used for Western blot
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
to PDMF membrane (Millipore, Billerica,
MA) according to the standard proto-
col. After blocking with 5% nonfat milk
at room temperature for 1 h, Six1 anti-
body with 1:1,000 dilution was hybridized
with the membrane at 4°C overnight. The
membrane was washed three times with
315
TBS/T for 10 min each and incubated
with secondary antibody at room tempera-
ture for 1 h. Protein signal was detected by
ECL Plus system (Amersham Biosciences,
Piscataway, NJ). Antibodies Six1 and
β-Actin were purchased from Santa Cruz
Biotechnology (Santa Cruz, CA).
Statistical Analysis
Statistical analysis was carried out using
SPSS 11.5 for Windows (SPSS Inc.,
Chicago, IL). The association of Six1
protein expression and clinicopathological
parameters including sex, age, pTNM
staging, venous infiltration, encapsulation,
tumor size, alpha-fetoprotein level, and
hepatitis B surface antigen was analyzed
by Chi-square test. The prognostic value
of Six1 protein for predicting the overall
survival of HCC patients after hepatic
resection was calculated by Kaplan-Meier
analysis with the log-rank test. Cox pro-
portional hazard regression model was
performed with univariable and multi-
variable analyses to test factors that were
significantly associated with the overall
survival of the HCC patients. Logistic
regression analysis was also performed
to compare those factors for predicting
1- and 5-year overall survival of the HCC
patients. A P value of < 0.05 was consid-
ered to be statistically significant.
RESULTS
Six1 Expression in HCC Cell Lines
Three nonmetastatic HCC cell lines
(Hep3B, Huh7 and PLC) and two meta-
static HCC cell lines (MHCC97L and
MHCC97H) were checked for Six1 gene
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

MHCC97H with the highest metastatic Six1 protein overexpression (p = 0.000, r

potential expressed the highest level of = 0.438). All tumor tissues with positive
Six1 mRNA. Western blot analysis showed Six1 protein expression could match with
that Six1 protein was only expressed in the their positive mRNA expression except
metastatic cells and expressed the highest for one case.
level in MHCC97H. Reverse transcription-polymerase chain
reaction and Western blot were also used
Six1 Expression in HCC Patients to analyze the Six1 mRNA and protein
and Normal Donors expression patterns in normal liver tissues
(Figure 24.1C, Table 24.1). Most of the
Seventy-two pairs of tumor and non-
normal liver tissues did not show Six1
tumor liver tissues were subjected for
mRNA expression (90%), whereas only a
detection of Six1 mRNA by RT-PCR
few of them (2 of 20) showed weak posi-
(Figure 24.1B). A single-pair primer
tive signal. Moreover, Six1 protein was not
with 40-PCR-cycle amplification was
expressed in normal liver tissues.
employed because Six1 mRNA in clinical
samples could not be detected by using
multiplex PCR condition. Overexpression Six1 Protein Expression Correlated with
of Six1 mRNA was found in ∼ 85% (61 Advanced Tumor Stage
of 72) of tumor tissues compared with Six1 protein expression pattern in HCC
nontumor tissues (Table 24.1). Most of tumor tissues was compared with their
the nontumor liver tissues (91.7%) did clinicopathological features (Table 24.2).
not show Six1 mRNA expression. A few Overexpression of Six1 protein was sig-
nontumor tissues (6 of 72) expressed Six1 nificantly correlated with pTNM stage
mRNA, but all the expression levels were (p = 0.002, r = 0.312) and venous infiltra-
lower than those in the matched tumor tion (p = 0.004, r = 0.282). Six1 protein
tissues. Western blot analysis showed expression was detected in 71% (45 of 63)
that near 60% (61 of 103) of tumor tis- of advanced stage HCC patients and in
sues expressed Six1 protein while no 71% (42 of 59) HCC patients with venous
nontumor tissue expressed Six1 protein infiltration. No significant association was
(Figure 24.1B, Table 24.1). Using Chi- found between Six1 protein and sex, age,
squared test with Fisher’s exact test, Six1 cirrhosis, encapsulation, tumor size, alpha-
mRNA overexpression in HCC patients fetoprotein level of hepatitis B surface
was significantly correlated with their antigen.

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.

Six1 Protein Expression Correlated

with Poor Survival
Kaplan-Meier analysis was employed to
analyze the overall survival rate of HCC
patients in correspondence to the Six1 protein
expression pattern. The result showed that
HCC patients who overexpressed Six1
protein were significantly associated with
poor overall survival (Log rank = 4.12,
P = 0.0423, Figure 24.2). Kaplan-Meier
analysis of pTNM staging, venous infiltra-
tion and alpha-fetoprotein level were also
significantly associated with the overall
survival of HCC patients (data not shown). Figure 24.2. Kaplan-Meier overall survival curve
To further examine which factors were of HCC patients correlated with Six1 protein
the independent predictors, univariate expression
24. Hepatocellular Carcinoma: Overexpression of Homeoprotein Six1 319

and multivariate Cox proportional hazard
regression analyses of these factors
corresponding to the overall survival of
HCC patients were performed. Univariate
Cox proportional hazard regression analy-
sis showed that Six1 protein (HR = 1.956,
95% CI 1.011–3.784, p = 0.046) and other
factors were significantly associated with
the overall survival of HCC patients after
hepatic resection (Table 24.3). However,
multivariate analysis indicated that pTNM
staging was the only independent factor
for predicting the overall survival of HCC
patients (HR = 7.698, 95% CI 1.891–31.33,
p = 0.004, Table 24.3). Logistic regression
analysis of these factors associated with
the 1- and 5-year overall survival showed
that Six1 protein was the most influential
factor for predicting the 1-year overall sur-
vival (OR = 5.405, 95% CI 1.427–20.474,
p = 0.013) while pTNM staging was the
most important factor for predicting the
5-year overall survival (OR = 12.152, 95%
CI 1.652–89.378, p = 0.004, Table 24.4).
DISCUSSION
Homeobox transcription factor Six1 located
at 14q23 of the chromosome is involved in
the early development of many organs
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
kidney (Relaix and Buckingham, 1999;
Laclef et al., 2003; Ozaki et al., 2004;
Zheng et al., 2003; Xu et al., 2003).
Alteration of Six1 expression takes place
in human breast and Wilms’ cancer as well
as rhabdomyosarcoma (RMS), indicating
its possible contribution in the tumorigenic-
ity of different cancers (Ford et al., 1998;
Yu et al., 2004; Li et al., 2002). Six1 is
overexpressed in 44% of primary tumors
and 90% of metastatic tumors in breast
cancer (Ford et al., 1998). Moreover, ele-
vation of Six1 in human RMS patients is
significantly associated with a later tumor
stage (Yu et al., 2004). To investigate
whether Six1 is also deregulated in HCC
patients, both mRNA and protein levels
of Six1 were examined in this study. In
agreement with other studies, we found
that overexpression of Six1 frequently
occurred in tumor tissues of human HCC
patients at ~ 85% in mRNA level and 60%
in protein level (Table 24.1). Although
overexpression of Six1 mRNA in tumor
tissues of HCC patients was significantly
correlated with its positive Six1 protein
expression, the percentage of Six1 protein
overexpression in tumor tissues was lower
than that of Six1 mRNA overexpression.
This phenomenon was also observed in
HCC cell lines, in which Hep3B expressed
Six1 mRNA but not Six1 protein (Figure
24.1A). Furthermore, a small portion of
HCC nontumor liver tissues (8.3%) and
normal liver tissues (10%) showed posi-
tive Six1 mRNA expression rather than
protein expression (Table 24.1). All these
data suggested that the post-transcriptional
regulation of Six1 among HCCs may be
different, and Six1 protein level, rather
than mRNA level, is more likely to reflect
its real expression status. In addition,
expression of Six1 protein in liver is
K. Tak-Pan Ng and K. Man
expected to be tumor-specific as indicated
by the evidence that no Six1 protein was
expressed in both normal liver tissues and
HCC nontumor liver tissues (Table 24.1).
The tumor-specific characteristic of Six1 in
HCC may thus provide a useful implication
for the therapeutic application of Six1 in
HCC by suppression strategy.
Overexpression of Six1 is potentially
related to metastatic status of breast cancer
and RMS (Ford et al., 1998; Yu et al., 2004).
MHCC97H and MHCC97L are human
HCC cell lines with the same genetic back-
ground but with different local and distant
metastatic potentials (Li et al., 2001).
Orthotopic implantation of the MHCC97H
cell into nude mice results in 100% of lung
metastasis, whereas MHCC97L exhibits
40% of lung metastasis (Li et al., 2001).
Our data showed that Six1 protein was
only expressed in metastatic HCC cell
lines (MHCC97L and MHCC97H) but not
in nonmetastatic HCC cell lines (Hep3B,
Huh7 and PLC), suggesting that Six1
may be a metastasis-associated oncogene.
Statistical analysis of the clinicopatho-
logical features of HCC patients revealed
that overexpression of Six1 protein was
significantly correlated with pTNM stag-
ing (p = 0.002) and venous infiltration
(p = 0.004). Approximately, 71% of HCC
patients recognized at the advanced pTNM
stage (45 of 63) or with venous infiltra-
tion (42 of 59) were found to overexpress
Six1 protein, suggesting that Six1 may
play an important role in HCC progres-
sion and invasion. Although Six1 protein
did not show significant association with
tumor size (p = 0.268), > 60% (48 of 77)
of HCC patients with tumors larger than
5 cm showed positive Six1 protein expres-
sion (Table 24.2), indicating its probable
function in tumor progression.
24. Hepatocellular Carcinoma: Overexpression of Homeoprotein Six1
Hepatic resection is one of the major
curative options for HCC patients. Different
centers have reported achieving different
survival rates (Llovet et al., 2005). Identifi-
cation of novel risk markers after hepa-
tectomy is beneficial for the manage-
ment of HCC patients and improvement of
their survival. Our results showed that
overexpression of Six1 protein in HCC
patients was significantly associated with
poor overall survival after hepatectomy
(p = 0.0423, Figure 24.2), suggesting that
this protein may be a novel marker related
to poor overall survival of HCC patients
after hepatectomy.
Although multivariate Cox proportional
hazard regression analysis showed that
Six1 protein was not an independent factor
for predicting the overall survival, logistic
regression analysis indicated that Six1 pro-
tein was the most influential factor asso-
ciated with poor 1-year overall survival
(p = 0.013, Table 24.4) rather than 5-year
overall survival (Table 24.4), proposing
that Six1 may be regarded as a predictor
for short-term overall survival of HCC
patients after hepatectomy. Hepatocellular
carcinoma patients with a higher rate of
early-year death after hepatectomy may be
caused by higher malignancy of HCC that
pathologically is, in part, determined to be
at the advanced pTNM stage or the presence
of venous infiltration. Overexpression of
Six1 protein in HCC patients was deter-
mined to be significantly associated with
these malignant factors indicating the pos-
sible role of Six1 on promoting HCC
malignancy. Even though pTNM staging
was determined to be the most important
factor in predicting long-term overall sur-
vival of HCC patients after hepatectomy
(Tables 24.3 and 24.4), assessment of Six1
protein level of resected tumor tissues of
321
HCC patients may still provide a valuable
index for effective management of HCC
patients after hepatectomy especially during
early year.
Up to now, the understanding of the
mechanism of Six1 in pathogenesis of can-
cers is still limited. Overexpression of Six1
in breast cancer cells can promote the can-
cer cells to escape from the G2 cell cycle
checkpoint after X-ray irradiation (Ford
et al., 1998). The cell cycle regulatory
activity of Six1 in breast cancer is regu-
lated by casein kinase II which inactivates
Six1 through phosphorylation (Ford et al.,
2000). Cyclin A1 is a downstream effector
for Six1 in breast cancer where over-
expression of Six1 promotes cyclin A1
expression and subsequently increases
cell proliferation and progression (Coletta
et al., 2004). Gene amplification of
Six1 is a probable mechanism which
contributes to tumorigenesis in breast
cancer (Reichenberger et al., 2005).
Overexpression of Six1 in RMS cells
can boost their pulmonary metastasis
potential, whereas downregulation of
Six1 suppresses their metastatic abil-
ity (Yu et al., 2004). Yu et al. (2006)
also demonstrated that Six1 potentially
activates several oncogenes including
cyclin D1, c-Myc and Ezrin. Moreover, the
ability of Six1 to promote metastasis of
RMS requires the function of Ezrin. The
above research in breast cancer and RMS
provides important clues to clarify the
functional roles of Six1 in HCC because
different cancers may have different regu-
latory mechanisms. In fact, Cyclin A is
overexpressed in HCC and its overexpres-
sion is associated with poor survival of
HCC patients (Chao et al., 1998; Ohashi
et al., 2001). Casein kinase II shows
an important involvement in transforming
322
growth factor-beta1-induced HCC (Gavin
et al., 2003). The relationship between Six1
and Cyclins as well as casein kinase II in
regulating the tumorigenesis and metastasis
of HCC thus needs further clarification.
The specificity of Six1 protein expression
in HCC metastatic cells may also provide a
good bridge to study the functions of Six1
involved in metastasis of HCC through
suppression strategy such as using antisense
or RNA interference means.
In conclusion, our data indicated that
Six1 protein was specifically and fre-
quently expressed in HCC tumor tissues.
Overexpression of Six1 protein in HCC
patients was significantly associated with
advanced pTNM stage and venous infiltra-
tion. In addition, Six1 protein could be a
novel marker for predicting the short-term
overall survival of HCC patients after
hepatic resection. Further functional stud-
ies are worthwhile to explore the precise
mechanism and eventually develop potential
therapies targeting Six1 in liver cancer
recurrence and metastases.
REFERENCES
Befeler, A.S., and Di Bisceglie, A.M. 2002.
Hepatocellular carcinoma: diagnosis and treat-
ment. Gastroenterology 122: 1609–1619.
Bosch, F.X., Ribes, J., and Borras, J. 1999.
Epidemiology of primary liver cancer. Semin.
Liver Dis. 19: 271–285.
Chao, Y., Shih, Y.L., Chiu, J.H., Chau, G.Y., Lui,
W.Y., Yang, W.K., Lee, S.D., and Huang, T.S.
1998. Overexpression of cyclin A but not Skp 2
correlates with the tumor relapse of human hepa-
tocellular carcinoma. Cancer Res. 58: 985–990.
Coletta, R.D., Christensen, K., Reichenberger,
K.J., Lamb, J., Micomonaco, D., Huang, L.,
Wolf, D.M., Muller-Tidow, C., Golub, T.R.,
Kawakami, K., and Ford, H.L. 2004. The Six1
homeoprotein stimulates tumorigenesis by reac-
tivation of cyclin A1. Proc. Natl. Acad. Sci. USA
101: 6478–6483.
K. Tak-Pan Ng and K. Man
Fields, A.C., Cotsonis, G., Sexton, D., Santoianni, R.,
and Cohen, C. 2004. Survivin expression in
hepatocellular carcinoma: correlation with pro-
liferation, prognostic parameters, and outcome.
Mod. Pathol. 17: 1378–1385.
Ford, H.L., Kabingu, E.N., Bump, E.A., Mutter,
G.L., and Pardee, A.B. 1998. Abrogation of the
G2 cell cycle checkpoint associated with over-
expression of HSIX1: a possible mechanism
of breast carcinogenesis. Proc. Natl. Acad. Sci.
USA 95: 12608–12613.
Ford, H.L., Landesman-Bollag, E., Dacwag, C.S.,
Stukenberg, P.T., Pardee, A.B., and Seldin, D.C.
2000. Cell cycle-regulated phosphorylation of
the human SIX1 homeodomain protein. J. Biol.
Chem. 275: 22245–22254.
Gavin, L.G., Romieu-Mourez, R., Panta, G.R., Sun,
J., Factor, V.M., Thorgeirsson, S.S., Sonenshein,
G.E., and Arsura, M. 2003. Inhibition of CK2
activity by TGF-beta1 promotes IkappaB-alpha
protein stabilization and apoptosis of immortal-
ized hepatocytes. Hepatology 38: 1540–1551.
Ikeguchi, M., Ueda, T., Sakatani, T., Hirooka, Y.,
and Kaibara, N. 2002. Expression of survivin
messenger RNA correlates with poor progno-
sis in patients with hepatocellular carcinoma.
Diagn. Mol. Pathol. 11: 33–40.
Jemal, A., Siegel, R., Ward, E., Murray, T., Xu, J.,
and Thun, M.J. 2007. Cancer statistics 2007. CA.
Cancer J. Clin. 57: 43–66.
Laclef, C., Hamard, G., Demignon, J., Souil, E.,
Houbron, C., and Maire, P. 2003. Altered myo-
genesis in Six1-deficient mice. Development
130: 2239–2252.
Lee, T.K., Man, K., Poon, R.T., Lo, C.M., Yuen,
A.P., Ng, I.O., Ng, K.T., Leonard, W., and Fan,
S.T. 2006. Signal transducers and activators of
transcription 5b activation enhances hepatocellu-
lar carcinoma aggressiveness through induction
of epithelial-mesenchymal transition. Cancer
Res. 66: 9948–9956.
Li, C.M., Guo, M., Borczuk, A., Powell, C.A.,
Wei, M., Thaker, H.M., Friedman, R., Klein, U.,
Tycko, B. 2002. Gene expression in Wilms’
tumor mimics the earliest committed stage in the
metanephric mesenchymal-epithelial transition.
Am. J. Pathol. 160: 2181–2190.
Li, Y., Tang, Z.Y., Ye, S.L., Liu, Y.K., Chen, J., Xue,
Q., Chen, J., Gao, D.M., and Bao, W.H. 2001.
Establishment of cell clones with different
24. Hepatocellular Carcinoma: Overexpression of Homeoprotein Six1
potential from the hepatocellular carcinoma
cell line MHCC97. World J. Gastroenterol. 7:
630–636.
Llovet, J.M., Schwartz, M., and Mazzaferro, V. 2005.
Resection and liver transplantation for hepatocel-
lular carcinoma. Semin. Liver Dis. 25: 181–200.
Lo, C.M., and Fan, S.T. 2004. Liver transplanta-
tion for hepatocellular carcinoma. Br. J. Surg.
91: 131–133.
Muller, C. 2006. Hepatocellular carcinoma – rising
incidence, changing therapeutic strategies. Wien.
Med. Wochenshr. 156: 404–409.
Ohashi, R., Gao, C., Miyazaki, M., Hamazaki, K.,
Tsuji, T., Inoue, Y., Uenura, T., Hirai, R.,
Shimizu, N., and Namba, M. 2001. Enhanced
expression of cyclin E and cyclin A in human
hepatocellular carcinomas. Anticancer Res. 21:
657–662.
Oliver, G., Wehr, R., Jenkins, N.A., Copeland,
N.G., Cheyette, B.N., Hartenstein, V., Zipursky,
S.L., and Gruss, P. 1995. Homeobox genes and
connective tissue patterning. Development 121:
693–705.
Ozaki, H., Nakamura, K., Funahashi, J., Ikeda,
K., Yamada, G., Tokano, H., Okamura, H.O.,
Kitamura, K., Muto, S., Kotaki, H., Sudo, K.,
Horai, R., Iwakura, Y., and Kawakami, K. 2004.
Six1 controls patterning of the mouse otic vesi-
cle. Development 131: 551–562.
Parkin, D.M., Bray, F., Ferlay, J., and Pisani, P. 2001.
Estimating the world cancer burden: Globocan
2000. Int. J. Cancer 94: 153–156.
Poon, R.T., and Fan, S.T. 2004. Hepatectomy
for hepatocellular carcinoma: patient selection
323
and postoperative outcome. Liver Transpl. 10:
S39–S45.
Reichenberger, K.J., Coletta, R.D., Schulte, A.P.,
Varella-Garcia, M., and Ford, H.L. 2005. Gene
amplification is a mechanism of Six 1over-
expression in breast cancer. Cancer Res. 65:
2668–2675.
Relaix, F., and Buckingham, M. 1999. From insect
eye to vertebrate muscle: redeployment of a
regulatory network. Genes Dev. 13: 3171–3178.
Ueki, T., Fujimoto, J., Suzuki, T., Yamamoto, H.,
and Okamoto, E. 1997. Expression of hepatocyte
growth factor and its receptor c-met proto-onco-
gene in hepatocellular carcinoma. Hepatology
25: 619–623.
Xu, P.X., Zheng, W., Huang, L., Maire, P., Laclef, C.,
and Silvius, D. 2003. Six1 is required for
the early organogenesis of mammalian kidney.
Development 130: 3085–3094.
Yu, Y., Davicioni, E., Triche, T.J., and Merlino, G.
2006. The homeoprotein six1 transcriptionally
activates multiple protumorigenic genes but
requires ezrin to promote metastasis. Cancer
Res. 66: 1982–1989.
Yu, Y.L., Khan, J., Khanna, C., Helman, L.,
Meltzer, P.S., and Merlino, G. 2004. Expression
profiling identifies the cytoskeletal organ-
izer ezrin and the developmental homeopro-
tein Six-1 as key regulators. Nat. Med. 10:
175–181.
Zheng, W., Huang, L., Wei, Z.B., Silvius, D.,
Tang, B., and Xu, P.X. 2003. The role of Six1
in mammalian auditory system development.
Development 130: 3989–4000.
25
Hepatocellular Carcinoma:
KiSS-1 Overexpression
as a Prognostic Factor
Katharina Schmid, Isabella Mosberger, and Fritz Wrba
INTRODUCTION
Knowing about reliable prognostic markers
is of great clinical importance to predict
relapse and dissemination of hepatocellular
carcinoma (HCC), one of the most common
malignant diseases worldwide. It is well
known that surgery serves as the only
curative therapy; by the end of the last
century, orthotopic liver transplantation
had become an optional treatment for a
subgroup of patients with low tumor stage,
resulting in better prognosis with regard to
tumor recurrence (Island et al., 2005).
Lee et al. (1996) reported for the first
time that the KiSS-1 gene, mapped to
chromosome 1q32–q41, functions as a
metastasis suppressor gene. This working
group found that KiSS-1 mRNA expres-
sion occurred only in nonmetastatic, but
not in metastatic melanoma. Lee and Welch
(1997) also observed that the expression of
KiSS-1 reduced the metastatic potential
of human breast carcinoma cells by 95%,
while tumorigenicity was not affected.
The study groups of Kotani et al. (2001),
Muir et al. (2001), and Ohtaki et al. (2001)
reported that the product of KiSS-1 gene,
termed as KiSS-1 (metastin, kisspeptin-54),
a predominately hydrophilic, 164 amino
acid protein, serves as the natural ligand
of an orphan G-protein-coupled receptor,
termed as AXOR12 (hOT7T175, human
GPR54), a 398-amino acid protein encoded
by the AXOR12 gene. In healthy individuals
the highest expression levels of KiSS-1
mRNA were observed in placenta, and in
testis, whereas moderate levels of KiSS-1
mRNA were found in pancreas, liver, and
small intestine. For AXOR12 transcripts,
the highest expression levels were seen
in placenta and pancreas, with moderate
expression in spleen, peripheral blood
leukocyte, testis, and lymph nodes.
During the following years, only a few
clinical data on KiSS-1 and AXOR12 gene
expression in human malignancies were
published. The majority reported that down-
regulation of KiSS-1 mRNA correlated with
tumor progression: Shirasaki et al. (2001)
found that KiSS-1 mRNA expression
was lost in advanced stage melanomas.
Sanchez-Carbayo et al. (2003) described
lower transcript levels of KiSS-1 in invasive
bladder carcinoma compared with superfi-
cial tumors and normal urothelium. Masui
et al. (2004) reported that pancreatic can-
cer showed significantly lower expression
325
326 K. Schmid et al.

levels of KiSS-1 mRNA than normal pan-
creas tissue. Ikeguchi et al. (2004) observed
that the loss of KiSS-1 gene expression in
esophageal squamous carcinoma was a sig-
nificant predictor of lymph node metastasis.
According to Dhar et al. (2004), down-
regulation of KiSS-1 expression was the
strongest independent prognostic factor for
gastric carcinoma in multivariate analysis.
Kostadima et al. (2007) found that the
KiSS-1 gene was silent in the vast majority
of resected node-positive breast carcino-
mas, strongly supporting the antimetastatic
role of the gene. Conversely, Ikeguchi et al.
(2003) reported that overexpression of
KiSS-1 correlated with tumor progression
in HCC. Also, Martin et al. (2005) found
that the expression of KiSS-1 increased in
patients with aggressive breast cancer.
In order to investigate a possible correla-
tion between KiSS-1 expression in HCC
and patient prognosis, we performed an
immunohistochemical study: expression
levels of the KiSS-1 gene product and its
cognate receptor AXOR12 were analysed
in HCC specimens on the protein level,
and results were correlated with clinical
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

MATERIALS 13. Levamisole (Sigma-Aldrich, St. Louis,

MO, USA)
1. SuperFrost Plus slides (silane coated) 14. Mayer’s Hematoxylin
2. Xylene 15. For mounting sections: Aquatex (Merck
3. Ethanol: anhydrous denaturated, 100%, Pharma, Darmstadt, Germany)
96%, 80%, and 70% 16. Placenta tissue slides for positive control
4. Distilled water (dH2O), physiologi-
cal saline (Merck Pharma, Darmstadt,
Germany)
5. Wash buffer: Tris-HCl-buffer stemsolu- METHODS
tion (TBS): Add 60.57 g Tris (Biomol,
Hamburg, Germany) to 500 ml dH2O This report presents a step-by-step proto-
and 400 ml HCl, adjust pH to 7.5, bring col for immunohistochemical staining of
to 1,000 ml with dH2O. Working solu- KiSS-1 and AXOR12 in HCC. Paying atten-
tion: add 900 ml physiological saline tion to correct tissue processing, including
to 100 ml stemsolution a reliable method and adequate duration of
6. Antigen retrieval citrate buffer: fixation, is essential for a good final result.
stemsolution A (10 mM citric acid): To choose the right tissue block containing
bring 21.06 g citric acid-1-hydrate to sufficient vital tumor tissue is indispensable
1,000 ml dH2O; stemsolution B: bring to obtaining reliable results. By incubating
29.41 g sodium citrate-dihydrate to the tissue sections with primary antibody
1,000 ml dH2O (both Merck Pharma, at 4°C for 20 h before storing them at room
Darmstadt, Germany). Working solu- temperature for another hour, a strong reac-
tion: 9 ml stemsolution A, and 41 ml tion of KiSS-1 in tumor cells will be found,
stemsolution B are brought to 500 ml and possible background staining is neg-
with dH2O ligible. Antibody binding was visualized
7. Blocking solution: Avidin/Biotin- with Supersensitive Alkaline Phosphatase
Blocking Kit (Vector, Burlingame, Complex. For visualizing the antibody bind-
CA, USA) ing the New Fuchsin Kit is applied, adding
8. Antibodies: KISS-1 and AXOR12, levamisole to block endogenous alkaline
both polyclonal rabbit antibodies phosphatase. Negative and positive controls
(Phoenix Pharmaceuticals, Belmont, should be used to ensure correct results:
CA, USA) for negative controls we recommend sub-
9. Antibody dilution: Albumine from stituting primary antibodies with irrelevant
Bovine Serum (BSA) (Sigma-Aldrich, rabbit serum, and for positive controls, pla-
St. Louis, MO, USA) centa tissue is of proper value.
10. Biotinylated secondary antibody: Goat
Anti-Rabbit IgG (Biocare Medical,
Tissue Preparation
San Diego, CA, USA)
11. Supersensitive Alkaline Phosphatase To obtain representative paraffin-embed-
Complex (Biocare Medical, San Diego, ded tissue blocks, tumor tissue pieces of
CA, USA) no more than 1 cm in diameter should be
12. New Fuchsin Kit (Nichirei Corporation, fixed in 7.5% formalin for 24 h before
Tokyo, Japan) embedding in paraffin.
328
Immunohistochemistry
1. Cut 4 μm thick sections of paraffin-
embedded tumor tissue specimens,
transfer the sections onto silane coated
SuperFrost Plus slides, and dry the
sections at room temperature.
2. Before deparaffinizing, heat the sections
for 30 min at 55°C to prevent detachment
of tissue from slides during the staining
procedure. Deparaffinize and rehydrate
the sections by submerging the slides in
xylene (2 × 5 min), then in 100% ethanol
for 5 min twice, followed by rinsing the
tissue sections in 96%, 80%, 70% ethanol
and in dH2O.
3. For antigen retrieval, apply autoclave
pretreatment in sodium citrate buffer
(10 mM, pH 6.0) at 1 bar for 20 min;
afterwards cool slides for 40 min in the
autoclave.
4. Wash sections three times in TBS for
5 min each.
5. To avoid unspecific staining, block
each section with 100–200 μl blocking
solution: first apply avidin solution for
15 min, rinse briefly with TBS, fol-
lowed by incubation with biotin solu-
tion for another 15 min as outlined by
the manufacturer’s guideline.
6. Remove blocking solution, and after
washing in TBS, apply 100–200 μl
diluted primary antibody to each
section (dilute KiSS-1 1:150 and
AXOR12 1:400 in 1% BSA in TBS),
and incubate for 20 h (overnight) at
4°C, followed by an additional hour
at room temperature.
7. Remove antibody solution and wash sec-
tions in TBS three times for 5 min each.
8. Apply 100–200 μl biotinylated goat
anti-rabbit IgG as secondary antibody
to each section, and incubate for 30 min
at room temperature.
K. Schmid et al.
9. Remove secondary antibody solution
and wash sections three times in TBS
for 5 min each.
10. Add 100–200 μl Supersensitive Alkaline
Phosphatase Complex (diluted 1:3 in
TBS) to each section and incubate for
30 min at room temperature.
11. Remove the complex and wash sections
three times in TBS for 5 min each.
12. Visualize the antibody binding using
the New Fuchsin Kit according to the
manufacturer’s instructions: incubate
for at least 10 min and at a maximum
of 30 min to obtain an optimal staining
result (check under the microscope).
13. Wash sections three times in TBS for
5 min each.
14. Counterstain the sections in Mayer’s
hematoxylin for 3 min.
15. Wash sections in tap water (10–20
times).
16. Mount sections using Aquatex.
Evaluation of Immunohistochemical
Result
Count immunopositive tumor cells in at
least ten consecutive high-power fields
(× 400), and give number of positive cells
as a percentage (use a system of 10%
steps, range 0–100%). Only samples that
exhibited 50% or more intensely stained
tumor cells should be classified as posi-
tive (weakly stained samples and samples
showing less than 50% intensely stained
tumor cells are considered negative).
RESULTS AND DISCUSSION
Patients and Tissue Samples
This immunohistochemical study was
undertaken to elucidate a possible link
between KiSS-1 and AXOR12 expression
25. Hepatocellular Carcinoma: KiSS-1 Overexpression as a Prognostic Factor
levels in HCC and patient prognosis (Schmid
et al., 2007). All patients included in the study
underwent orthotopic liver transplantation as
a curative therapeutic approach. Underlying
liver diseases were mostly alcoholic and
hepatitis C cirrhosis, less frequently hepatitis
B cirrhosis, cryptogenic cirrhosis and other
rare causes of cirrhosis (hemochromatosis,
primary biliary cirrhosis, primary sclerosing
cholangitis) as well as few cases of non-
cirrhotic liver. Before surgery was carried
out, a few patients received percutaneous,
endotumoral alcohol instillation, or were
treated by transarterial chemoembolization
of the tumors. After liver transplantation,
adjuvant chemotherapy (doxorubicin) was
administered to a subgroup of the patients
(Pokorny et al., 2005).
The results of our study revealed a gen-
erally diffuse, cytoplasmatic staining pat-
tern of both, KiSS-1 and AXOR12 antigen
showing a heterogenic reaction from focal
up to > 80% of tumors cells. Nonneoplastic
heptocytes in cirrhotic liver tissue also
expressed KiSS-1 and AXOR12 strongly
in a focal pattern; reactivity could be
observed mainly at the margin of regenera-
tive nodules (Figure 25.1). The epithelium
of bile ducts and hepatocytes of normal
liver tissue showed only weak KiSS-1 but
no AXOR12 expression. No correlation
with underlying liver disease was seen.
Immunoreactivity of KiSS-1 and
AXOR12 in HCC showed statistically
significant correlation, which confirms
previous studies reporting a coincidence
expression of KiSS-1 receptor and its
cognate ligand in various tissues (Lee et
al., 1996; Muir et al., 2001; Ohtaki et al.,
2001). Ikeguchi et al. (2003) investigated
KiSS-1 and AXOR12 mRNA expression
in HCC, and found a higher percentage of
mRNA transcripts in HCC as we did on
329
protein level by immunohistochemistry.
This difference might reflect an additional
regulation of KiSS-1/AXOR12 expression
at the translational level or by different
cleavage variants of the KiSS-1 protein.
The KiSS-1 investigated in our study is
the major cleavage product, consisting of
54 amino acids (kisspeptin-54). Kotani
et al. (2001) also discovered two other
naturally occurring smaller peptides in
the human placenta, termed kisspeptin-14
and kisspeptin-13. It was reported by Muir
et al. (2001) that KiSS-1 the main cleavage
product has greater stability than kisspep-
tin-14 and kisspeptin-13, which exhibit
stronger receptor binding activity.
In our study we also demonstrated by
multivariate analysis that positive immuno-
reactivity of KiSS-1 in HCC is associated
with poorer clinical outcome concerning
disease-free (i.e., early local recurrence
of HCC, lymph node, and distant metas-
tases) as well as overall survival, whereas
AXOR12 expression is not associated
with these parameters. Ikeguchi et al.
(2003) also reported a worse prognosis
for HCC patients with increased KiSS-1
and AXOR12 mRNA levels. Martin et al.
(2005) described higher mRNA transcripts
in advanced stages of breast carcinoma.
On the contrary, the remainders of clini-
cal data on KiSS-1 and/or AXOR12 gene
expression in human malignancies have
shown a strong correlation between KiSS-1
and/or AXOR12 gene overexpression and
improved clinical course (Shirasaki et al.,
2001; Dhar et al., 2004).
Reasons why the expression of KiSS-1
gene product in HCC and advanced stages
of breast cancer had a converse effect on
clinical course compared to other malig-
nant tumors remain to be clarified. Ikeguchi
et al. (2003) postulated that the estrogen
330
hormone might influence the expression
patterns of KiSS-1 gene products. Patients
with HCC generally suffer from liver cir-
rhosis, which leads to a hyperestrogenic
state. The fact that the KiSS-1/AXOR
system in HCC might be influenced by a
hormonal disorder is supported by the find-
ings of Muir et al. (2001) and Horikoshi
et al. (2003), who found significantly higher
plasma levels of KiSS-1 in pregnant women
than in nonpregnant women. Interestingly,
in breast carcinoma, an association between
the increase of KiSS-1 and tumor progres-
sion was found (Martin et al., 2005). It is
well known that one of the most important
systemic therapies in breast cancer is the
blockade of the estrogen receptor path-
way. Although Chow et al. (2002) found
that antiestrogenic therapy did not prolong
survival in patients with inoperable HCC,
further studies are warranted to investigate
whether patients with overexpression of
KiSS-1 in HCC improve their clinical out-
come with antiestrogenic therapy.
In conclusion, KiSS-1 and AXOR12
are expressed in a subset of human HCC.
High KiSS-1 expression is an independ-
ent prognostic factor for disease free and
overall survival in patients with HCC.
Thus, KiSS-1 could serve as an addi-
tional molecular marker for assessing
the malignant potential of HCC. Former
reports have shown that estrogen levels
might modulate the expression patterns
of KiSS-1 gene products; therefore, we
speculate that patients with overexpression
of KiSS-1 in HCC might benefit from a
antiestrogenic therapy. However, further
studies are needed to clarify a possible link
of estrogen signalling and KiSS-1 expres-
sion in HCC tumors.
K. Schmid et al.
REFERENCES
Chow, P.K., Tai, B.C., Tan, C.K., Machin, D., Win, K.M.,
Johnson, P.J., and Soo, K.C. 2002. High-dose
tamoxifen in the treatment of inoperable hepa-
tocellular carcinoma: a multicenter randomized
controlled trial. Hepatology 36:1221–1226.
Dhar, D.K., Naora, H., Kubota, H., Maruyama, R.,
Yoshimura, H., Tonomoto, Y., Tachibana, M.,
Ono, T., Otani, H., and Nagasue, N. 2004.
Downregulation of KiSS-1 expression is respon-
sible for tumor invasion and worse prognosis in
gastric carcinoma. Int. J. Cancer 111:868–872.
Horikoshi, Y., Matsumoto, H., Takatsu, Y., Ohtaki,
T., Kitada, C., Usuki, S., and Fujino, M. 2003.
Dramatic elevation of plasma metastin concen-
trations in human pregnancy: metastin as a novel
placentaderived hormone in humans. J. Clin.
Endocrinol. Metab. 88:914–919.
Ikeguchi, M., Hirooka, Y., and Kaibara, N. 2003.
Quantitative riverse transcriptase polymer-
ase chain reaction analysis for KiSS-1 and
orphan G-protein-coupled receptor (hOT7T175)
gene expression in hepatocellular carcinoma.
J. Cancer Res. Clin. Oncol. 129:531–535.
Ikeguchi, M., Yamaguchi, K., and Kaibara, N. 2004.
Clinical significance of the loss of KiSS-1 and
orphan G-protein-coupled receptor (hOT7T175)
gene expression in esophageal squamous cell
carcinoma. Clin. Cancer Res. 10:1379–1383.
Island, E.R., Pomposelli, J., Pomfret, E.A., Gordon,
F.D., Lewis, W.D., and Jenkins, R.L. 2005.
Twenty-year experience with liver transplanta-
tion for hepatocellular carcinoma. Arch. Surg.
140:353–358.
Kostadima, L., Pentheroudakis, G., and Pavlidis, N.
2007. The missing kiss of life: transcriptional
activity of the metastasis suppressor gene KiSS1 in
early breast cancer. Anticancer Res. 27:2499–2504.
Kotani, M., Detheux, M., Vandenbogaerde, A.,
Communi, D., Vanderwinden, J.M., Le Poul, E.,
Brézillon, S., Tyldesley, R., Suarez-Huerta, N.,
Vandeput, F., Blanpain, C., Schiffmann, S.N.,
Vassart, G., and Parmentier, M. 2001. The
metastasis suppressor gene KiSS-1 encodes
kisspeptins, the natural ligands of the orphan G
proteincoupled receptor GPR54. J. Biol. Chem.
276:34631–34616.
25. Hepatocellular Carcinoma: KiSS-1 Overexpression as a Prognostic Factor
Lee, J.H., Miele, M.E., Hicks, D.J., Phillips, K.K.,
Trent, J.M., Weissman, B.E., and Welch, D.R.
1996. KiSS-1, a novel human malignant
melanoma metastasis-suppressor gene. J. Natl.
Cancer Inst. 88:1731–1737.
Lee, J.H., and Welch, D.R. 1997. Suppression
of metastasis in human breast carcinoma
MDA-MB-435 cells after transfection with the
metastasis suppressor gene, KiSS-1. Cancer
Res. 57:2384–2387.
Martin, T.A., Watkins, G., and Jiang, W.G. 2005.
KiSS-1 expression in human breast cancer. Clin.
Exp. Metastasis 22:503–511.
Masui, T., Doi, R., Mori, T., Toyoda, E., Koizumi, M.,
Kami, K., Ito, D., Peiper, S.C., Broach, J.R.,
Oishi, S., Niida, A., Fujii, N., and Imamura, M.
2004. Metastin and its variant forms suppress
migration of pancreatic cancer cells. Biochem.
Biophys. Res. Commun. 315:85–92.
Muir, A.I., Chamberlain, L., Elshourbagy, N.A.,
Michalovich, D., Moore, D.J., Calamari, A.,
Szekeres, P.G., Sarau, H.M., Chambers, J.K.,
Murdock, P., Steplewski, K., Shabon, U., Miller, J.E.,
Middleton, S.E., Darker, J.G., Larminie, C.G.,
Wilson, S., Bergsma, D.J., Emson, P., Faull,
R., Philpott, K.L., and Harrison, D.C. 2001.
AXOR12, a novel human G protein-coupled
receptor, activated by the peptide KiSS-1.
J. Biol. Chem. 276:28969–28975.
Ohtaki, T., Shintani, Y., Honda, S., Matsumoto, H.,
Hori, A., Kanehashi, K., Terao, Y., Kumano, S.,
331
Takatsu, Y., Masuda, Y., Ishibashi, Y., Watanabe, T.,
Asada, M., Yamada, T., Suenaga, M., Kitada, C.,
Usuki, S., Kurokawa, T., Onda, H., Nishimura, O.,
and Fujino, M. 2001. Metastasis suppressor
gene KiSS-1 encodes peptide ligand of
a G-proteincoupled receptor. Nature 411:
613–617.
Pokorny, H., Gnant, M., Rasoul-Rockenschaub, S.,
Gollackner, B., Steiner, B., Steger, G., Steininger,
R., and Muehlbacher, F. 2005. Does additional
doxorubicin chemotherapy improve outcome in
patients with hepatocellular carcinoma treated
by liver transplantation? Am. J. Transplant
5:788–794.
Sanchez-Carbayo, M., Capodieci, P., and Cordon-
Cardo, C. 2003. Tumor suppressor role of KiSS-1
in bladder cancer. Loss of KiSS-1 expression is
associated with bladder cancer progression and
clinical outcome. Am. J. Pathol. 162:609–617.
Schmid, K., Wang, X., Sieghart, W., Peck-
Radosavljevic, M., Bodingbauer, M., Rasoul-
Rockenschaub, S., and Wrba, F. 2007. KiSS-1
overexpression as an independent prognostic
marker in hepatocellular carcinoma: an immu-
nohistochemical study. Virchows Arch. 450:
143–149.
Shirasaki, F., Takata, M., Hatta, N., and Takehara,
K. 2001. Loss of expression of the metastasis
suppressor gene KiSS-1 during melanoma pro-
gression and its association with LOH of chromo-
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
HEPATOMA-DERIVED
GROWTH FACTOR
Hepatoma-derived growth factor (HDGF)
is a heparin-binding protein purified from
the conditioned medium of the human
well-differentiated hepatocellular carci-
noma (HCC) cell line, HuH-7, which can
proliferate autonomously in a serum-free
chemically-defined medium; this factor
is highly expressed in several cancer cells
(Nakamura et al., 1989, 1994; Mori et al.,
2004). Hepatoma-derived growth factor is
an acidic 26 kDa protein consisting of 230
amino acids with no hydrophobic signal
sequence in its N-terminus, and it has a
high affinity to the glycosaminoglycans
heparin and heparan sulphate (Nakamura et
al., 1994; Dietz et al., 2002). Exogenously
supplied HDGF stimulates the prolifera-
tion of fibroblasts, endothelial cells, vas-
cular smooth muscle cells, pulmonary
epithelial cells and hepatocytes, as well
as HCC, lung cancer, and colon cancer
cells. A possible receptor-binding site is
estimated to be residing at amino acid
residues 81–100 within the hath region
(described below) (Abouzied et al., 2005).
Indeed, exogenous HDGF stimulates the
Erk phosphorylation in endothelial cells
(Everett et al., 2004).
Hepatoma-derived growth factor is a
major member of the HDGF family of
proteins which consists of HDGF and five
HDGF-related proteins (HRP) (Izumoto
et al., 1997; Dietz et al., 2002). The
N-terminal region of HDGF is highly con-
served among the other five HDGF-related
proteins (HRP), and this region is called
hath (homologous to the amino terminus of
HDGF) region, which contains the PWWP
domain (Izumoto et al., 1997). The HDGF
family members are characterized based on
whether they contain the hath region and
nuclear localization signals (NLS) in their
gene-specific regions and are targeting to
the nucleus (Nakamura and Hada, 2004).
Hepatoma-derived growth factor contains
two nuclear localization signals (NLS) in
the molecule of HDGF; the first func-
tional nuclear localization signal (NLS1)
is in the hath region and the second NLS
(NLS2) is in gene-specific regions of the
C-terminal region (Kishima et al., 2002a).
Hepatoma-derived growth factor can traf-
fic to the nucleus using these NLSs, espe-
cially the NLS2 in its gene-specific region,
and this factor is dominantly localized in
333
334
the nucleus rather than in the cytoplasm
(Everett et al., 2001). The gene-specific
region of HDGF, at least the bipartite NLS
sequence and both the N-and C-terminal
neighboring portions, is essential for the
mitogenic activity (Kishima et al., 2002a).
Thus, HDGF exerts its proliferating activity
via two different pathways: (1) via a putative
plasma membrane-located HDGF receptor
for which signaling depends on the hath
region, resulting in MAP kinase activation,
and (2) via targeting to the nucleus by NLS.
Developmentally Regulated Expression of
Hepatoma-Derived Growth Factor
Cancers develop by the accumulation of
dysregulated gene expression from multistep
genetic mutations of oncogenes and/or
suppressor genes. These oncogenic proteins
and tumor suppressor proteins generally
include growth factors, their receptors,
intracellular signal transduction molecules,
and transcriptional regulatory factors.
Genes that are reduced significantly during
organ development are frequently absent
or little expressed in the adult tissues,
and these developmentally regulated genes
may be reactivated in human cancers. In
the fetus, HDGF is abundantly expressed
in various organs including liver.
Hepatoma-derived growth factor is
highly expressed in fetal liver of the mid-
gestation stage, and is markedly decreased
near birth. The expression of this factor was
significantly decreased with differentiation
in fetal hepatocytes induced by oncostatin
M treatment in in vitro primary culture dif-
ferentiation system (Enomoto et al., 2002).
Adenoviral introduction of HDGF antisense
cDNA into the fetal hepatocytes suppresses
their proliferation, and the inhibitory effects
of the HDGF antisense virus were recov-
ered by exogenous HDGF. Furthermore,
H. Nakamura et al.
HDGF is highly expressed in oval cells,
that are progenitor cells with bipotential
activity for differentiating into hepatocytes
and bile duct cells. These findings sug-
gest that HDGF plays important roles in
the proliferation of immature hepatocytes
and hepatic progenitor cells including oval
cells, showing significant involvement of
HDGF in liver development.
In the fetus, HDGF was also abundantly
expressed in other organs. Hepatoma-
derived growth factor is highly expressed
in fetal rat aorta, conotruncus, and heart
(Everett et al., 2000, 2001), and is widely
distributed in the renal anlage at the early
stages of renal development and disap-
peared from the adult kidney (Oliver and
Al-Awqati, 1998). Hepatoma-derived
growth factor is highly expressed in the
endothelial cells of non-muscularized
forming blood vessels of the fetal lung
(Everett et al., 2004), and its expression is
enhanced dominantly in the bronchial and
alveolar epithelial cells including type II
pneumocytes by bleomycin-instillation in
mice (Mori et al., 2004). This growth fac-
tor is also highly expressed in hind gut epi-
thelia as well as atrial myocytes in the fetal
stage. Furthermore, HDGF is expressed in
the nucleus of the colonic epithelial cells,
dominantly in the bottom of the intestinal
crypts by immunohistochemical analysis
(Nakamura et al., 2002). Thus, HDGF
is one of the developmentally regulated
genes which are abundantly expressed in
cancer cells.
Role in Hepatocarcinogenesis
The Fatty Liver Shionogi (FLS) mouse is
an inbred strain that develops spontaneous
fatty liver without obesity, and results in
liver tumor development to ~ 45% in 52
weeks and 90% at 72 weeks after birth
26. Hepatocellular Carcinoma: Prognosis Using Hepatoma-Derived Growth Factor
in male mice; these tumors have been
histologically diagnosed as hepatocellular
adenoma and carcinoma. In the liver of
FLS mice, HDGF expression has already
increased at an early stage before the
tumors developed microscopically in the
liver (Yoshida et al., 2003). Hepatoma-
derived growth factor may be induced and
secreted or released from hepatocytes and
macrophages, enhancing the cell cycle
progression of hepatocytes, inducing their
transformation, and promoting the pro-
liferation of HCC cells in an intracrine,
autocrine and/or paracrine manner.
Furthermore, HDGF expression was
strongly detected in an oval cell line,
Oc 15–5, by immunohistochemistry, which
is established from the liver of Long-
Evans-Cinnamon rats. Hepatocellular car-
cinoma is considered to be developed
from oval cells induced in the regen-
eration process after liver injury. By dif-
ferential subtractive chain reaction from
strong anchorage-independent growth to
its negative HCC cells, HDGF is cloned
as one of the genes related to anchorage-
independency (Huang et al., 2004). These
findings strongly suggest that HDGF poten-
tially participates in hepatocarcinogenesis
and in the early stage of HCC.
Hepatoma-derived growth factor-over-
expressing NIH3T3 cells generate sarco-
matous tumors in nude mice. In our study,
HDGF-over-expressing NIH3T3 cells did
not show significant anchorage-independ-
ent growth in soft agar assay; however,
HDGF-over-expressing NIH3T3 cells
developed more small colonies in soft agar
than parent or neomycin-resistant cells
(Okuda et al., 2003). Hepatoma-derived
growth factor was more highly expressed
in colon cancer cells than non-transformed
intestinal epithelial cells, and its expression
335
was dramatically increased in human colorec-
tal cancers, especially in tumors proficient
in DNA mismatch repair (Lepourcelet et
al., 2005). Conversely, down regulation of
HDGF by use of HDGF-siRNA had mini-
mal effect on anchorage-dependent growth
but significantly reduced anchorage-
independent growth of non-small cell lung
cancer (NSCLC) cells in soft agar (Zhang
et al., 2006). Thus, these findings suggest
that HDGF is an oncogenic protein, and
plays a role in cancer development.
Role in Cancer Progression and
Angiogenesis
Hepatoma-derived growth factor protein
is abundantly expressed in various human
cancers. In liver, HDGF expression is
higher in HCC than in the adjacent tissues
(Yoshida et al., 2003, 2006). The HDGF-
over-expressing hepatoma cell line HepG2
proliferates more rapidly and produces
larger tumors, showing more rapid growth
in nude mice than neomycin-resistant cells
in vivo. Recombinant HDGF stimulates
the growth of cancer cells, while antisense
HDGF oligonucleotides or polyclonal anti-
HDGF antibody suppress their growth
(Kishima et al., 2002b; Nakamura et al.,
2002). Exogenously supplied HDGF pro-
motes the proliferation of bronchial squa-
mous cell carcinoma cell line, A549 cells
(Mori et al., 2004). Knock-down expres-
sion of HDGF by small interfering RNA
(siRNA) in NSCLC cells significantly
shows slower growth, less colony for-
mation in soft agar, and lesser in vitro
invasion activity (Zhang et al., 2006).
By proteomic differential display analysis
and mass spectrometry, HDGF is down-
regulated in regressive cancer cells as
compared with that in inflammatory cell-
promoting progressive cells of the murine
336
fibrosarcoma cell line (Hayashi et al.,
2005). The higher expression of HDGF
shows more malignant potentials, suggest-
ing that HDGF is a candidate factor for
cancer progression.
Hepatocellular carcinoma is a hyper-
vascular tumor. Hepatoma-derived growth
factor is intrinsically related to angiogen-
esis and vasculogenesis and its protein
expression is induced in the regenerating
process of vascular vessels in wound repair
(Everett et al., 2000). Additionally, HDGF
is a candidate endothelial growth factor
for renal glomerulus formation (Oliver and
Al-Awqati, 1998).
Tumors developed from HDGF-over-
expressing NIH3T3 cells inoculated
in nude mice are macroscopically red-
colored and histologically rich in vascu-
lature, and immunohistochemical analysis
by anti-CD31 antibody shows prominent
new vessel formation (Okuda et al., 2003).
Hepatoma-derived growth factor protein
stimulates the proliferation and tubule for-
mation of human umbilical vein endothe-
lial cells and the migration of human
pulmonary microvascular endothelial cells.
Using chick chorioallantoic membrane
(CAM) assay, recombinant HDGF stimu-
lates blood vessel formation, and stimu-
lates cellular reorganization within the
CAM from a loose network into a more
compact, linear alignment reminiscent
of tube formation (Everett et al., 2004).
The more potent growth stimulating activ-
ity of HDGF in vivo than in vitro must
be brought on by both the direct cell
growth activity and its angiogenic activity.
Furthermore, HDGF induces vascular
endothelial growth factor (VEGF) gene
expression, and is a potent angiogenic fac-
tor that functions by its own direct stimu-
lation of the proliferation of endothelial
cells and vascular smooth muscle cells
H. Nakamura et al.
and through the induction of VEGF in the
nucleus (Okuda et al., 2003).
IMMUNOHISTOCHEMICAL
AND ANALYTICAL
METHODS
Materials
1. Rabbit polyclonal antibody raised
against C-terminal amino acids (amino
acids 231–240) of the human HDGF.
2. Dulbecco’s phosphate-buffered saline
without magnesium and calcium – Tween
20 (PBS-T): 200 mg potassium chloride,
200 mg monobasic potassium phosphate,
8 g sodium chloride, and 2.9 g dibasic
sodium phosphate·12 H2O, bring volume
to 11 ml with deionized distilled water, pH
7.4. Add 0.1 ml Tween 20 and mix well.
3. Fixative: 10% formalin in distilled water.
4. 10 mM citric acid buffer (pH 6.0): 294.10 g
trisodium citrate·2 H2O, bring volume to
11 ml with deionized distilled water to
prepare 1 M trisodium citrate. 210.14 g
citric acid·H2O, bring volume to 11 ml
with deionized distilled water to prepare
1 M citric acid. Add 5 ml of 1 M trisodium
citrate and 0.8 ml of 1 M citric acid to
494.2 ml of deionized distilled water
to prepare 10 mM citric acid, pH 6.0.
5. Methanol containing 0.2% hydrogen
peroxide: 9 ml of 31% hydrogen perox-
ide in 150 ml methanol.
6. Blocking serum (goat serum) diluted at
1/50 in PBS-T.
7. Primary antibody (rabbit polyclonal
anti-HDGF antibody) diluted at 1/5,000
in PBS-T.
8. Secondary antibody (biotinylated goat
polyclonal anti-rabbit antibody) diluted
at 1/100 in PBS-T.
9. Avidin–biotin-complex solution: Add
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

Figure 26.1. The immunohistochemical staining of hepatoma-derived growth factor. Representative

pictures of HDGF immunostaining for HDGF L1 level 1 and HDGF L1 level 2 are shown. In HDGF L1
level 1, as a negative HDGF index, most HCC cells do not express HDGF. In HDGF L1 level 2, as a positive
HDGF index, more than 90% of HCC cells show positive staining both for nucleus and cytoplasma
26. Hepatocellular Carcinoma: Prognosis Using Hepatoma-Derived Growth Factor 339

Figure 26.2. Hepatoma-derived growth factor immunohistochemistry in hepatocellular carcinoma cells

and normal hepatocytes. HDGF is more strongly and frequently expressed in the nucleus and cytoplasma
of HCC cells than in the adjacent normal hepatocytes, by using specific anti-C terminus of HDGF
antibody at a dilution of 1:5,000

higher in well-differentiated carcinomas HDGF expression in HCC showed an

than in poorly-differentiated carcinomas
in our study (Yoshida et al., 2006). In
contrast, HDGF was reported to be more
highly expressed in poorly-differentiated
HCC than in well-differentiated HCC
(Hu et al., 2003). One possible explana-
tion for the discrepancy between the two
groups may be due to the specificity of
the anti-HDGF antibody used for immu-
nostaining. However, a more satisfactory
explanation will be shown by a larger
scale study. Conversely, in both our and
their studies, the patients with higher
earlier recurrence and a poorer over-
all survival rate than those with lower
expression after hepatectomy for HCC
(Hu et al., 2003; Yoshida et al., 2006).
In patients with differentiated HCC except
for Edmondson’s grade IV, the prognostic
significance of the HDGF index was
strong. By multivariate analysis, HDGF
expression was an independent prog-
nostic factor for disease-free and overall
survival in patients who underwent a
hepatectomy for HCC. These findings
suggest that HDGF is a candidate for use
340
as a prognostic factor for disease-free and
overall survival of patients with HCC.
Besides HCC, HDGF is a prognsostic
factor of pancreatic ductal cancer, lung
cancer, gastric cancer, and esophageal
cancer. In patients with HCC or early
stage NSCLC who underwent curative sur-
gery, the disease-free and disease-specific
survival and overall survival were lower
in those with higher HDGF expression
indexes than in those with lower HDGF
indexes (Ren et al., 2004; Iwasaki et al.,
2005). The high expression of HDGF
showed distant metastasis and shortened
survival time in patients with NSCLC. In
gastric cancer, the patients with abundant
expression of HDGF showed significantly
higher rates of infiltrative tumor growth
and vascular and lymphatic invasion, as
compared to those with lower expression
(Yamamoto et al., 2006). Pancreatic ductal
cancers with higher HDGF expression in
the nucleus showed a poorer 5-year sur-
vival rate than those with lower expression
(Uyama et al., 2006). These clinical find-
ings and in vitro data of pleiotropic HDGF
activities described above suggest that the
higher expression of HDGF shows more
malignant potential including invasive and
metastatic activity.
In conclusion, HDGF is a unique nuclear
targeting growth factor, which is expressed
abundantly in cancer cells, promotes their
malignant potential, and generates tumors
and promotes their growth in vivo via its
mitogenic activity and angiogenic activity
deriving from both its own direct ang-
iogenic activity and the induction of VEGF.
Multivariate analysis of the relationship
of HDGF expression and recurrence-free
and overall survival in patients with HCC
confirmed that HDGF has the potential to
become a significantly efficacious prog-
H. Nakamura et al.
nostic marker for cancer patients. It is
expected in the future that any tool regu-
lating HDGF expression or HDGF signal
pathways may be a useful candidate for
the suppression of carcinogenesis and can-
cer progression.
REFERENCES
Abouzied, M.M., El-Tahir, H.M., Prenner, L.,
Haberlein, H., Gieselmann, V., and Franken, S.
2005. Hepatoma-derived growth factor.
Significance of amino acid residues 81–100 in
cell surface interaction and proliferative activity.
J. Biol. Chem. 280: 10945–10954.
Dietz, F., Franken, S., Yoshida, K., Nakamura, H.,
Kappler, J., and Gieselmann, V. 2002. The family of
hepatoma-derived growth factor proteins:
characterization of a new member HRP-4 and
classification of its subfamilies. Biochem. J.
366: 491–500.
Enomoto, H., Yoshida, K., Kishima, Y., Kinoshita, T.,
Yamamoto, M., Everett, A.D., Miyajima, A., and
Nakamura, H. 2002. Hepatoma-derived growth
factor is highly expressed in developing liver
and promotes fetal hepatocyte proliferation.
Hepatology 36: 1519–1527.
Everett, A.D., Lobe, D.R., Matsumura, M.E.,
Nakamura, H., and McNamara, C.A. 2000.
Hepatoma-derived growth factor stimulates smooth
muscle cell growth and is expressed in vascular
development. J. Clin. Invest. 105: 567–575.
Everett, A.D., Stoops, T., and McNamara, C.A.
2001. Nuclear targeting is required for hepato-
ma-derived growth factor-stimulated mitogen-
esis in vascular smooth muscle cells. J. Biol.
Chem. 276: 37564–37568.
Everett, A.D., Narron, J.V., Stoops, T., Nakamura, H.,
and Tucker, A. 2004. Hepatoma-derived growth
factor is a pulmonary endothelial cell-expressed
angiogenic factor. Am. J. Physiol., Lung Cell.
Mol. Physiol. 286: L1194–1201.
Hayashi, E., Kuramatsu, Y., Okada, F., Fujimoto, M.,
Zhang, X., Kobayashi, M., Iizuka, N., Ueyama,
Y., and Nakamura, K. 2005. Proteomic profil-
ing for cancer progression: differential display
analysis for the expression of intracellular pro-
teins between regressive and progressive cancer
cell lines. Proteomics 5: 1024–1032.
26. Hepatocellular Carcinoma: Prognosis Using Hepatoma-Derived Growth Factor
Hu, T.H., Huang, C.C., Liu, L.F., Lin, S.Y.,
Chang, H.W., Changchien, C.S., Lee, C.M.,
Chuang, J.H., and Tai, M.H. 2003. Expression
of hepatoma-derived growth factor in hepatocel-
lular carcinoma. Cancer 98: 1444–1456.
Huang, J.S., Chao, C.C., Su, T.L., Yeh, S.H., Chen,
D.S., Chen, C.T., Chen, P.J., and Jou, Y.S. 2004.
Diverse cellular transformation capability of
overexpressed genes in human hepatocellular
carcinoma. Biochem. Biophys. Res. Commun.
315: 950–958.
Iwasaki, T., Nakagawa, K., Nakamura, H.,
Takada, T., Matsui, K., and Kawahara, K. 2005.
Hepatoma-derived growth factor as a prognostic
marker in completely resected non-small-cell
lung cancer. Oncol. Rep. 13: 1075–1080.
Izumoto, Y., Kuroda, T., Harada, H., Kishimoto, T., and
Nakamura, H. 1997. Hepatoma-derived growth
factor belongs to a gene family in mice showing
significant homology in the amino terminus.
Biochem. Biophys. Res. Commun. 238: 26–32.
Kishima, Y., Yamamoto, H., Izumoto, Y., Yoshida, K.,
Enomoto, H., Yamamoto, M., Kuroda, T., Ito, H.,
Yoshizaki, K., and Nakamura, H. 2002a.
Hepatoma-derived growth factor stimulates cell
growth after translocation to the nucleus by
nuclear localization signals. J. Biol. Chem. 277:
10315–10322.
Kishima, Y., Yoshida, K., Enomoto, H.,
Yamamoto, M., Kuroda, T., Okuda, Y., Uyama,
H., and Nakamura, H. 2002b. Antisense oli-
gonucleotides of hepatoma-derived growth
factor (HDGF) suppress the proliferation of
hepatoma cells. Hepatogastroenterology 49:
1639–1644.
Lepourcelet, M., Tou, L., Cai L., Sawada, J.,
Lazar, A.J.F., Glickman, J.N., Williamson,
J.A., Everett, A.D., Redston, M., Fox, E.A.,
Nakatani, Y., and Shivdasani, R.A. 2005.
Insights into development mechanisms and can-
cers in the mammalian intesitine derived from
serial analysis of gene expression and study of
the hepatoma-derived growth factor (HDGF).
Development 132: 415–427.
Mori, M., Morishita, H., Nakamura, H., Matsuoka, H.,
Yoshida, K., Kishima, Y., Zhou, Z., Kida, H.,
Funakoshi, T., Goya, S., Yoshida, M., Kumagai,
T., Tachibana, I., Yamamoto, Y., Kawase, I., and
Hayashi, S. 2004. Hepatoma-derived growth fac-
tor is involved in lung remodeling by stimulating
341
epithelial growth. Am. J. Respir. Cell Mol. Biol.
30: 459–469.
Nakamura, H., Kambe, H., Egawa, T., Kimura,
Y., Ito, H., Hayashi, E., Yamamoto, H., Sato, J.,
and Kishimoto, S. 1989. Partial purification and
characterization of hepatoma-derived growth
factor. Clin. Chim. Acta 183: 273–284.
Nakamura, H., Izumoto, Y., Kambe, H., Kuroda,
T., Mori, T., Kawamura, K., Yamamoto, H.,
and Kishimoto, T. 1994. Molecular cloning
of complementary DNA for a novel human
hepatoma-derived growth factor. J. Biol. Chem.
269: 25143–25149.
Nakamura, H., Yoshida, K., Ikegame, K., Kishima,
Y., Uyama, H., and Enomoto, H. 2002.
Antibodies against hepatoma-derived growth
factor and mucosal repair in ulcerative colitis.
J. Gastroenterol. 37 (Suppl. 14): 8–14.
Nakamura, H., and Hada, T. 2004. Hepatoma-
derived growth factor in ontogeny and tumori-
genesis. Recent Res. Devel. Biophys. Biochem.
4: 17–27.
Okuda, Y., Nakamura, H., Yoshida, K., Enomoto, H.,
Uyama, H., Hirotani, T., Funamoto, M., Ito,
H., Everett, A.D., Hada, T., and Kawase, I.
2003. Hepatoma-derived growth factor induces
tumorigenesis in vivo through both direct
angiogenic activity and induction of vascular
endothelial growth factor. Cancer Sci. 94:
1034–1041.
Oliver, J.A., and Al-Awqati, Q. 1998. An endothe-
lial growth factor involved in rat renal develop-
ment. J. Clin. Invest. 102: 1208–1219.
Ren, H., Tang, X., Lee, J.J., Feng, L., Everett, A.D.,
Hong, W.K., Khuri, F.R., and Mao, L. 2004.
Expression of hepatoma-derived growth factor
is a strong prognostic predictor for patients with
early-stage non-small cell lung cancer. J. Clin.
Oncol. 22: 3230–3237.
Uyama, H., Tomita, Y., Nakamura, H., Nakamori, S.,
Zhang, B., Hoshida, Y., Enomoto, H., Okuda, Y.,
Sakon, M., Aozasa, K., Kawase, I., Hayashi, N.,
and Monden, M. 2006. Hepatoma-derived
growth factor is a novel prognostic factor for
patients with pancreatic cancer. Clin. Cancer
Res. 12: 6043–6048.
Yamamoto, S., Yasuhiko, T., Hoshida, Y., Takiguchi,
S., Fujiwara, Y., Yasuda, T., Doki, Y., Yoshida, K.,
Aozasa, K., Nakamura, H., and Monden, M.
2006. Expression of hepatoma-derived growth
342
factor is correlated with lymph node metastasis and
prognosis of gastric carcinoma. Clin. Cancer
Res. 12: 117–122.
Yoshida, K., Nakamura, H., Okuda, Y., Enomoto,
H., Kishima, Y., Uyama, H., Ito, H., Hirasawa,
T., Inagaki, S., and Kawase, I. 2003. Expression
of hepatoma-derived growth factor in hepato-
carcinogenesis. J. Gastroenterol. Hepatol. 18:
1293–1301.
Yoshida, K., Tomita, T., Okuda, Y., Yamamoto, S.,
Enomoto, H., Uyama, H., Ito, H., Hoshida, Y.,
H. Nakamura et al.
Aozasa, S., Nagano, H., Sakon, M., Kawase, I.,
Monden, M., and Nakamura, H. 2006. Hepatoma-
derived growth factor is a novel prognostic fac-
tor for hepatocellular carcinoma. Ann. Surg.
Oncol. 13: 159–167.
Zhang, J., Ren, H., Yuan, P., Lang, W., Zhang, L.,
and Mao, L. 2006. Down-regulation of
hepatoma-derived growth factor inhibits
anchorage-independent growth and invasion
of non-small cell lung cancer cells. Cancer
Res. 66: 18–23.
27
Hepatitis C Virus-Related Human
Hepatocellular Carcinoma: Predictive
Markers Using Proteomic Analysis
(Methodology)
Yasuhiro Kuramitsu and Kazuyuki Nakamura
HEPATITIS C VIRUS-RELATED
HUMAN HEPATOCELLULAR
CARCINOMA (HCV–HCC)
It was reported by El-Serag and Mason
(1999) that hepatocellular carcinoma
(HCC) is one of the most common fatal
cancers world-wide. From the repots by
Okuda (1992, 2000) and Seow et al.
(2001), the most apparent risk factor for
HCC is thought to be the chronic infec-
tion with hepatitis B or C virus (HBV
or HCV). Although in the other Asian
countries HCCs developed from HBV-
infection, the characteristic of HCCs in
Japan is that they develop from the liver
tissues infected with HCV or of liver cir-
rhosis. From the reports by Yano et al.
(2003) and Yatsuhashi (2004), 80% of the
death cases of HCC patients have HCV-
infection, and the number of HCV-carriers
in Japan is 2 million. The number of
patients with HCC is still on the increase,
although diagnostic and therapeutic tech-
niques have become remarkably advanced.
It was reported by Yatsuhashi (2004) that
the incidence of HCC is increasing in
many developed countries in parallel to
an increase in chronic HCV infection.
Therefore, it is important to identify the
predictive marker proteins by which we
can predict the carcinogenesis of HCC in
HCV-infected patients.
In a recent study, we analyzed proteome
in HCV–HCC tissues and identified many
proteins whose expression in cancer tis-
sues were different from corresponding
noncancerous tissues by using two-dimen-
sional gel electrophoresis (2-DE) and
mass spectrometry (MS). Furthermore, we
identified the autoantibodies recognizing
HSP70.1, Mn-SOD or peroxiredoxin from
the sera of HCC-patients.
PROTEOMICS
The term of ‘proteome’ was named
by the group of Ian Humphery-Smith
(Wasinger et al., 1995). ‘Proteome’ was
defined as the total protein complement
of a genome, and ‘proteomics’ means the
analysis for the total protein complement
of a genome. The combination of two-
dimensional gel electrophoresis (2-DE)
and mass spectrometry (MS) is a popular
method of ‘proteomics’. By means of
2-DE we can separate many proteins
343
344
according to their charges and their
size in isoelectoric focusing (IEF) gels
and sodium dodecyl sulfate (SDS) gels,
respectively. With 2-DE we can exam-
ine the expressions of a huge number of
proteins simultaneously. There are many
reports of cancer proteomics obtained by
means of 2-DE. Simpson et al. (2000)
reported the proteomic analysis of the
human colon carcinoma cell line LIM
1215 by means of 2-DE. Ha et al. (2002)
reported the proteome analysis of human
stomach tissues. They separated soluble
proteins by 2-DE and identified by mass
spectrometry. Steel et al. (2001) reported
a proteomic approach for the discovery of
early detection markers of hepatocellular
carcinoma by using 2-DE. Presently, MS
has become the first choice for the deter-
mination of protein sequences. We can
determine the masses of peptides with
much accuracy by MS, and we can iden-
tify the protein from the huge data base
of genomics.
PROTEOMICS FOR
HEPATOCELLULAR
CARCINOMA TISSUES
Using the proteomic technique many pro-
teins have been reported as candidates for
new diagnostic biomarkers for HCCs. Li
et al. (2005) reported on proteomic analysis
of HBV–HCCs without HCV infection in
ten patients, and identified many upregu-
lated (enolase1, GRP 75, galactokinase1,
HSP27, HSP70.5, PCNA, stathmin, etc.)
and downregulated proteins (aldolase B,
annexin V, fructose-bisphosphatase, glutath-
ione peroxidase, nucleophosmin, SOD 1,
triosephosphate isomerase, vimentin, etc.)
Y. Kuramitsu and K. Nakamura
in cancerous tissues compared to non-
cancerous tissues. Kim et al. (2003)
performed proteomics for HCC tissues
including HBV–HCC patients, HCV–
HCC patients and NBNC-HCC patients,
and reported upregulated and downregu-
lated proteins in HCC cancerous tissues.
In HCV–HCC tissues, both 4-aminobu-
tyrate aminotransferase and aflatoxin B1
aldehyde reductase were upregulated,
and cytosol aminopeptidase or acyl-CoA
dehydrogenase, or glutamate dehydroge-
nase was downregulated.
PROTEOMICS FOR SERA
FROM HEPATOCELLULAR
CARCINOMA PATIENTS
Although we can understand the patho-
genesis of HCC from the proteomics
for HCC tissues, it is not expected to
discover biomarkers for HCC from the
results of proteomics for HCC tissues.
To discover predictive biomarkers, we
need to perform the proteomics for sera
from HCC patients. Some proteomic studies
for sera from HCC patients have been
reported. Paradis et al. (2005) reported
proteomics for sera from patients with
cirrhosis without HCC and patients with
cirrhosis without HCC, and identified a
catalytic fragment of vitronectin as a new
marker of HCC. Steel et al. (2003) ana-
lyzed sera from patients with HBV–HCC,
and identified down regulated carboxyl
terminal fragment of complement C3 and
an isoform of apolipoprotein A1 in sera
from HBV–HCC patients. It is neces-
sary to repeat these types of studies and
confirm the results for the clinical trial as
biomarkers.
27. Hepatitis C Virus-Related Human Hepatocellular Carcinoma: Predictive Markers 345

PROTEOMICS FOR HEPATITIS Two-Dimensional Gel Electrophoresis

C VIRUS-RELATED (2-DE)
HEPATOCELLULAR Isoelectric focusing (IEF)
CARCINOMA TISSUES The first-dimensional isoelectric focusing
(IEF) on 7-cm, immobilized, pH 3–10
Preparation of HCV–HCC Tissue linear gradient (IPG) strips (Bio-Rad) was
Samples performed with 300 μg of each tissue
Cancerous and corresponding noncan- sample at 20°C and 50 μA/strip. Isoelectric
cerous liver tissues were obtained from focusing was run in three steps: 500 V for
patients who were diagnosed with HCV– 1 h, 1,000 V for 1 h, and 8,000 V for 3 h.
HCC and underwent surgical liver resec-
tion. The liver tissues were suspended Sodium Dodecyl Sulfate-Polyacrilamide
and homogenized in lysis buffer (50 mM Gel Electrophoresis (SDS-PAGE)
Tris-HCl, pH 7.5, 165 mM NaCl, 10 mM Figure 27.2 shows the experimental
NaF, 1 mM sodium vanadate, 1 mM phe- design of 2-DE. The second-dimensional
nylmethylsulfonyl fluoride, 10 mM EDTA, run was performed on precast polyacry-
10 μg/ml aprotinin, 10 μg/ml leupeptin, lamide gels (2-D homogeneous 12.5;
and 1% NP40). Suspensions were soni- GE Healthcare) in two steps: 600 V,
cated and centrifuged with 15,000 g for 20 mA for 30 min and 600 V, 50 mA for
30 min at 4°C. The supernatants were 70 min. After electrophoresis the gels
taken as samples. Figure 27.1 shows a were stained with CBB R250 or SYPRO
workflow of our experiment. Ruby (Invitrogen) for 24 h. Gels stained

Figure 27.1. Workflow of proteomics for HCV–HCC

346 Y. Kuramitsu and K. Nakamura

Mass Spectrometry Analysis

In-gel Digestion
For staining gels, CBB dye was removed
by rinsing three times in 60% methanol,
50 mM ammonium bicarbonate, and 5 mM
dithiothreitol (DTT) for 15 min and twice
in 50% acetonitrile, 50 mM ammonium
bicarbonate, and 5 mM DTT for 10 min.
Gel pieces were dehydrated three times
in 100% acetonitrile for 30 min and then
Figure 27.2. Scheme of two-dimensional gel
rehydrated in an in-gel digestion reagent
electrophoresis. The technique of 2-DE is able to containing 10 μg/ml sequencing-grade
differentiate proteins according to both their charge modified trypsin in 30% acetonitrile,
in isoelectoric focusing (IEF) gels and their size in 50 mM ammonium bicarbonate, and 5 mM
sodium dodecyl sulfate (SDS) gels DTT. In-gel digestion was performed over-
night at 30°C. The samples were rinsed
in 30% acetonitrile, 50 mM ammonium
with CBB R250 were destained with bicarbonate, and 5 mM DTT for 2 h and
10% acetic acid in water containing lyophilized overnight at −30°C.
30% methanol for 30 min and then with
7% acetic acid for 3 h and used for in-
Amino Acid Sequencing by Liquid
gel digestion. Gels stained with SYPRO
Chromatography–Tandem Mass
Ruby were washed with 7% acetic acid
Spectrometry (LC-MS/MS)
in water containing 10% methanol for
30 min and washed twice with mili-Q Lyophilized samples were dissolved in
water for 5 min, and then they were gen- 20 μl 0.1% formic acid and centrifuged
erated by scanning of gels at suitable at 15,000 g for 5 min. Peptide sequenc-
excitation and emission wavelengths. ing of identified protein spots was per-
formed by LC-MS/MS with a Spectrum
Image Analysis of the Gels Mill MS Proteomics Workbench (Agilent
Technologies).
The positions of the protein spots The different expression levels of identi-
were recorded using an image scan- fied proteins were confirmed by Western
ner (ProXPRESS 2D Proteomic Imaging blotting using specific antibodies.
System, Perkin-Elmer), and the posi-
tions and intensities of the spots were
Proteins Whose Expression Was
analyzed individually with computer
Increased in HCV–HCC Tissues
software (Progenesis, Perkin-Elmer).
Spots stained at different intensities were By means of the analysis with 2-DE, 11
excised from the gels and identified by spots, whose expression intensity was
peptide sequence trap (PST) with liquid increased more than twofold in cancerous
chromatography-tandem mass spectro- tissues than in noncancerous tissues, were
metry (LC-MS/MS). detected, and identified as GRP75, GRP78,
27. Hepatitis C Virus-Related Human Hepatocellular Carcinoma: Predictive Markers
HSC71, HSP70.1, HSP60, glutamine syn-
thetase, triosephosphate isomerase, ATP
synthetase beta chain and alpha-enolase
(Takashima et al., 2003, 2005; Kuramitsu
et al., 2006). The functions of these protein
spots are described below.
(1) GRP75
Glucose-regulated 75 kDa protein (GRP75)
is located in the mitochondria, and inhibits
cell division and controls cell proliferation
and cellular aging.
(2) GRP78
Glucose-regulated 78 kDa protein (GRP78)
is a molecular chaperone. This molecule is
involved in the unfolded protein response,
and plays a role in facilitating the assembly
of multimeric protein complexes inside the
endoplasmic reticulum.
(3) HSC70
Heat shock cognate 71 kDa protein
(HSC70) is a constitutively expressed
chaperone that is also inducible by heat
shock stress. It binds to nascent polypep-
tides to facilitate correct folding and also
participates in the formation and repair
of higher order protein structures such as
that of ATPase.
(4) HSP70.1
Heat shock 70 kDa protein 1 (HSP70.1)
stabilizes preexistent proteins against
aggregation and mediates the folding of
newly translated polypeptides within the
cytosol as well as within in cooperation
with other chaperones.
(5) HSP60
Although heat shock 60 kDa protein
(HSP60) is one of the molecular chaper-
ones, this molecule does not belong to the
HSP70 family.
(6) Glutamine synthetase
Glutamine synthetase is an enzyme to
synthesize glutamine from glutamate and
NH3. This protein was identified from the
347
three protein spots of Mr 42,000 and pI
6.4–6.8 on the 2-DE gel.
(7) Triosephosphate isomerase
Triosephosphate isomerase is an enzyme of
glycolytic pathway, and catalyzes the reac-
tion between D-glyceraldehyde-3-phosphate
and dihydroxyacetone phosphate.
(8) Alpha-enolase
Alpha-enolase is an isoenzyme of eno-
lase, a key protein that catalyzes the
conversion of 2-phosphoglycerate to
phosphoenolpyruvate in the glycolytic
pathway. Three enolase isoenzymes
have been identified. The alpha form is
present in most embryonic tissues, the
beta form is expressed in muscle tissues,
and the gamma form is found only in
neuronal tissues. Alpha enolase has been
detected not only in the cytoplasm but also
at the membrane surface. Although the
mechanism of surface expression and the
orientation on the membrane are not well
understood, it is known that the C-terminal
amino acid, lysine, is exposed at the cell
surface and is involved in binding to
plasminogen, which is then activated and
converted to plasmin. Plasmin, stabilized
at the cell surface, induces fibrinolysis.
(9) Adenosine triphosphate (ATP) syn-
thase beta chain
Adenosine triphosphate synthase is an
enzyme that synthesizes ATP from adeno-
sine diphosphate (ADP) and phosphate.
Beta chain is a subunit of this enzyme.
Proteins With Decreased Expression
By means of the analysis with 2-DE, 11
spots with expression intensity decreased
less than half in cancerous tissues than
noncancerous tissues were detected, and
identified as aldolase, arginase 1, enoyl-CoA
hydratase, ketohexokinase, smoothelin,
348
tropomyosin beta chain, ferritin light
chain, and serum albumin (Yokoyama
et al., 2004). The functions of these protein
spots are described below.
(1) Aldolase B (liver type)
This protein was identified from the four
protein spots with different pI on the 2-DE
gel. Aldolase is an enzyme to perform
the sixth step of glycolysis, and has three
isozymes. Aldolase A exists in muscle,
aldolase B exists in liver, and aldolase C
exists in brain.
(2) Arginase 1
Arginase 1 is an enzyme that performs the
first step of urea cycle, and is rich in liver.
(3) Enoyl-CoA hydratase
Enoyl-CoA hydratase is an enzyme that
performs the second step of beta oxidation
pathway of fatty acid.
(4) Ketohexokinase
Ketohexokinase is an enzyme that performs
the first step of fructose metabolic pathway.
(5) Smoothelin
Smoothelin is a cytoskeletal protein exist-
ing in visceral and vascular smooth muscle
cells as 59 kDa smoothelin-A and 110 kDa
smoothelin-B.
(6) Tropomyosin beta chain
Tropomyosin is a heterodimer of alpha and
beta chains. Tropomyosin binds to actin
filaments and regulates muscle contraction
in conjunction with troponin. Tropomyosin
stabilizes of actin filaments.
(7) Ferritin light chain
Ferritin is composed of light chain and
heavy chain, and an iron-storage molecule,
which acts as an internal iron reserve and
protects against iron toxicity.
(8) Serum albumin
Serum albumin is the most abundant plasma
protein produced in hepatocytes, and main-
tains the osmotic pressure of body fluid.
Y. Kuramitsu and K. Nakamura
PROTEOMICS FOR AUTO-
ANTIBODIES IN THE SERA
OF HEPATOCELLULAR
CARCINOMA PATIENTS
Serum samples. The autologous serum
samples were obtained from each patient
and sera from patients with normal liver
tissues were also obtained as control sera.
Two-dimensional immunoblot analy-
sis. To detect autoantibodies expected as
cancer biomarkers for HCC, we tried to
identify serum autoantibodies against pro-
teins in HCC cancer tissue. Tissue proteins
were separated by 2-DE, transferred onto
PVDF membranes, and immunoblotted
with autologous sera. For the blotting-
gels, separated proteins were transferred
onto PVDF membranes (Immobilon-P;
Millipore) electrophoretically, and the
membranes were blocked for 1 h at 4°C
with TBS containing 1% skimmed milk.
They were subsequently incubated over-
night at 4°C with autologous serum (1:50
dilution), washed four times with TBS
containing 0.05% Tween 20, and incu-
bated for 1 h at 4°C with horseradish per-
oxidase-conjugated secondary antibody
(1:1,000). The reaction was visualized
with a chemiluminescence reagent (ECL;
GE Healthcare).
By comparing each immunoblot pattern,
we identified spots with stronger staining
intensity in cancer tissues than in non-
cancerous tissues. Matched proteins
on 2-DE gels were identified by MS.
Immunoreactivity of autologous serum
autoantibodies to tissue proteins was
assessed in samples of HCC cancer tissues
and noncancerous tissues. Four immuno-
reactive spots were detected that showed
increased intensity in cancer tissues
27. Hepatitis C Virus-Related Human Hepatocellular Carcinoma: Predictive Markers 349

Figure 27.3. Detection for autoantibodies from HCC patients.

Experimental procedure for the detection of autoantibodies against the proteins in HCC cancerous
tissues.

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
Paradis, V., Degos, F., Dargere, D., Pham, N.,
Belghiti, J., Degott, C., Janeau, J.L., Bezeaud,
A., Delforge, D., Cubizolles, M., Laurendeau,
I., and Bedossa, P. 2005. Identification of a new
marker of hepatocellular carcinoma by serum
profiling of patients with chronic liver diseases.
Hepatology 41: 40–47.
Seow, T. K., Liang, R. C. Y., Leow, C. K., and
Chung, M. C. M., 2001. Hepatocellular carci-
noma: from bedside to proteomics. Proteomics
1: 1249–1263.
Simpson, R.J., Connolly, L.M., Eddes, J.S.,
Pereira, J.J., Moritz, R.L., and Reid, G.E. 2000.
Proteomic analysis of the human colon carci-
noma cell line (LIM 1215): development of a
membrane protein database. Electrophoresis 21:
1707–1732.
Steel, L.F., Mattu, T.S., Mehta, A., Hebestreit,
H., Dwek, R., Evans, A. A., London, W.T., and
Block, T. 2001. A proteomic approach for the
discovery of early detection markers of hepato-
cellular carcinoma. Dis. Markers 17: 179–189.
Steel, L.F., Shumpert, D., Trotter, M., Seeholzer,
S.H., Evans, A.A., London, W.T., Dwek, R., and
Block, T.M. 2003. A strategy for the compara-
tive analysis of serum proteomes for the discov-
ery of biomarkers for hapatocellular carcinoma.
Proteomics 3: 601–609.
Takashima, T., Kuramitsu, Y., Yokoyama, Y., Iizuka,
N., Toda, T., Saiakida, I., Okita, K., Oka, M., and
Nakamura, K. 2003. Proteomic profiling of heat
Y. Kuramitsu and K. Nakamura
shock protein 70 family members as biomarkers
for hepatitis C virus-related hepatocellular carci-
noma. Proteomics 3: 2487–2493.
Takashima, T., Kuramitsu, Y., Yokoyama, Y.,
Iizuka, N., Fujimoto, M., Nishisaka, T., Sakaida,
I., Okita, K., Oka, M., and Nakamura, K. 2005.
Overexpression of alpha enolase in hepatitis C
virus-related hepatocellular carcinoma: asso-
ciation with tumor progression as determined by
proteomic analysis. Proteomics 5: 1686–1692.
Takashima, T., Kuramitsu, Y., Yokoyama, Y.,
Iizuka, N., Fujimoto, M., Sakaida, I., Okita, K.,
Oka, M., and Nakamura, K. 2006. Proteomic
analysis of autoantibodies in patients with hepa-
tocellular carcinoma. Proteomics 6: 3894–3900.
Wasinger, V.C., Cordwell, S.J., Cerpa-Poljak,
A., Yan, J.X., Gooley, A.A., Wilkins, M.R.,
Duncan, M.W., Harris, R., Williams, K.L., and
Humphery-Smith, I. 1995. Progress with gene-
product mapping of the Mollicutes: Mycoplasma
genitalium. Electrophoresis 16: 1090–1094.
Yano, K., Yatsuhashi, H., and Yano, M., 2003.
Hepatitis C. Nippon Rinsho 61: 241–244.
Yatsuhashi, H., 2004. General remarks on chronic
hepatitis C. Nippon Rinsho 62: 400–404.
Yokoyama, Y., Kuramitsu, Y., Takashima, T., Iizuka,
N., Toda, T., Terai, S., Sakaida, I., Oka, M.,
Nakamura, K., and Okita, K. 2004. Proteomic
profiling of proteins decreased in hepatocellular
carcinoma from patients infected with hepatitis
C virus. Proteomics 4: 2111–2116.
A. Diagnosis
28
Liver Metastases from Colorectal
Cancer: Ultrasound Imaging
Søren R. Rafaelsen
INTRODUCTION
Colorectal cancer is the second largest
cause of cancer-related deaths in the United
States.
Approximately 15–20% of the patients
with cancer of the colon and rectum have
liver metastases at the time of primary sur-
gery (Faivre et al., 2003). Assessment of
liver metastases is essential for the appro-
priate management of colorectal cancer.
Surgical resection is an effective treatment
for hepatic metastases of colorectal cancer.
However, this option is available only
to a small group of patients. Preoperative
portal vein embolization for hypertrophy
of future hepatic resection and other pio-
neering multidisciplinary modalities and
the aggressive surgical approach has been
adopted to extend the frontiers of surgi-
cal therapy (Khatri et al., 2005). For the
majo-rity of patients surgery is not a cura-
tive option, and only palliative therapy is
available resulting in no long-term survi-
vors. If left untreated the median survival
is 6 months, but with modern combina-
tion chemothe-rapy a median survival of
24 month is possible. In a French study
(Guyot et al., 2005) the proportion of
patients with liver metastases resected with
curative intent increased over time from
6% (1976–1984) to 30% (1994–2003).
No existing form of systemic therapy can
effectively eradicate diffuse metastatic
disease, but a partial response to chemothe-
rapy can be achieved to downstage the liver
disease, which would then be amenable to a
surgical resection. Neoadjuvant chemothe-
rapy is able to increase the tumor resectability
by 10%. This increased number of curative
resections was accompanied by a relatively
low complication rate and a 5-year survival
that is within the range of the survival rates
with initially resectable lesions (Adam et al.,
2001). Patients with advanced rectal cancer
as well as limited liver metastases may ben-
efit from a combination of radio frequency
ablation (RFA) and local radiotherapy prior
to rectal surgery. In patients with limited but
inoperable colorectal metastases RFA has
shown a 5-year survival up to 30%. Radio
frequency ablation provides the opportunity
for localized tissue destruction of limited
volumes of tumor; it can be offered to non-
surgical candidates and used in conjunc-
tion with systemic therapy and/or surgery
(Gillams, 2005).
Presently, multiple imaging modalities
are available to evaluate liver metastases.
Despite the powerful imaging modalities,
355
356
such as computed tomography (CT), mag-
netic resonance imaging (MRI) and positron
emission tomography (PET), ultrasound
has continued to play an important role.
Ultrasound has no ionizing radiation and
can be performed at low cost compared
to the former modalities. With the advent
of intraoperative- and larparoscopic ultra-
sound, tissue harmonic imaging, color-,
power Doppler, contrast enhancement, pulse
inversion imaging, microflow imaging, 3D,
and 4D ultrasound still have an important
role in diagnosing liver metastases as well
as response evaluation, ultrasound guided
biopsy and RFA.
ULTRASOUND SCANNING
TECHNIQUE
Hepatic ultrasound is usually performed
with the patient in the supine position.
Turning the patient with the right side
up may also be helpful in some cases.
Images are obtained in sagital, transverse,
and oblique planes. The transducer is
positioned in the right subcostal margin
and by angling or moving the transducer
it is possible to obtain images in various
planes. Deep inspiration is often helpful
by displacing the liver caudally inferior
to the curvature. Intercostal scanning is
also used as a visualization technique.
Harmonic waves are generated from non-
linear distortion of an acoustic signal as an
ultrasound wave insonates tissues in the
body. These harmonic ultrasound waves
are numeral multiples of a fundamental
transmitted frequency. Potential advan-
tages of harmonic imaging include better
axial resolution due to higher frequencies
and improved lateral resolution due to nar-
rower beams. Decreasing noise from side
S.R. Rafaelsen
lobes provides a better signal-to-noise ratio
and reduces artifacts. Destructive sound-
wave effects from the body wall are also
reduced. Choudhry et al. (2000) found that
in patients with a body mass index of 30 or
more tissue harmonic imaging was clearly
better for lesion visibility and confidence
of diagnosis. The authors recommend rou-
tine use of harmonic imaging for abdominal
ultrasound studies in all adult patients.
However, for patients with diffuse fatty
infiltration in the liver, penetration of the
low frequency ultrasound waves is often
better on conventional ultrasound images
than when using tissue harmonic imaging.
Ultrasound guided biopsy is often used
to obtain cytological and histological veri-
fication of liver metastases from colorectal
cancer prior to chemotherapy. Because of
the risk of needle-track tumor seeding,
liver surgeons advise against percutane-
ous biopsy. A recent study has described
a negative effect on patient long-term
survival after liver resection. The 4-year
survival was 32% in the group of patients
who had a preoperative liver biopsy versus
47% in those without (Jones et al., 2005).
In addition to biopsy assistance, ultrasound
is used as needle guidance of percutaneous,
intraoperative and larparoscopic RFA.
Realtime virtual sonography enables
the realtime demonstration of multiplanar
reconstructed CT imaging corresponding
to the realtime ultrasound image side-by-
side on the screen. The system uses an
electromagnetic position sensor mounted
on the transducer. This can be helpful
in ultrasound-guided RFA and biopsies.
The sonographical visualization of focal
lesions in a steatotic liver can be difficult.
The association between hepatic steatosis
and oral 5-fluorouracil (5-FU) agents
is well described (Miyake et al., 2005).
28. Liver Metastases from Colorectal Cancer 357

5-fluorouracil is frequently used in CRC GRAYSCALE ECHO PATTERN

patients and “realtime virtual sonography”
is helpful in ultrasound guided interven- Isoechoic and hyperechoic metastases are
tions within steatotic livers. often surrounded by a low echoic halo,
which is highly predictive of malignancy,
(Figure 28.1 [Bruneton et al., 1996]). There
ADVANTAGES OF DIAGNOSTIC is a considerable difference in the size of
HEPATIC ULTRASOUND the metastases. A mean size of 4.5 cm is
reported. The mean thickness of the halo
Although non-enhanced ultrasound has in colorectal liver metastases is publicized
lower sensitivity compared to CT espe- to 2.6 mm, range: 1–5 mm (Rafaelsen and
cially in obese patients, ultrasound can be Solvig, 2004).
performed bedside, is low cost and has no Calcification is described in 12–27% of
radiation. hepatic colorectal metastases. Calcification
Small liver cysts may be mistakenly occurred independent of the degree of tumor
diagnosed as metastatic disease on diag- differentiation, the presence of mucinous
nostic CT and screening CT colonogra- adenocarcinoma, or the hepatic tumor burden.
phy. Sonography may be useful in cancer The presence of calcification within a
patients with average body habitus to char- colorectal liver metastasis appears to imply
acterize small (0.6–1.5-cm) indeterminate a significantly better prognosis (Easson
liver lesions detected on CT (Eberhardt et al., 1996). Overall survival difference was
et al., 2003). The partial volume effect on compared between hyper- and hypoechoic
CT scan can cause small benign cysts to metastases on an intention-to-treat basis
mimic solid liver metastases. Since the dis-
tinction between cysts and metastases has
important clinical implications, sonography
should be used whenever the true nature of
these lesions cannot be determined by CT.
Hemangiomas can also be clarified by con-
trast enhanced sonography, showing centrip-
etal progression of the enhancement. There is
no contraindication using contrast-enhanced
ultrasound in patients with renal failure or
Metformin medication, as is the case in
contrast-enhanced CT scanning. Gallstones
can easily be detected using ultrasound and
evidence for an association between gall-
stones and colorectal cancer is confirmed.
This group of cancer patients had a signifi-
cantly higher prevalence of gallstone disease
than the control population (odds ratio = Figure 28.1. Ultrasound image showing a charac-
1.59; 95% confidence limits: 1.04–2.45) teristic hepatic metastasis. The thin halo is clearly
(Jorgensen and Rafaelsen, 1992). visible
358
in a study of 212 consecutive patients
with unresectable hepatic metastases from
colorectal carcinoma. A significant survival
benefit was observed in patients having
hyperechoic lesions compared to hypoe-
choic lesions (16 vs. 12 months median
survival) (Gruenberger et al., 2002).
DETECTION OF LIVER
METASTASES
In an early study, high-end ultrasound scan-
ning in experienced hands equalled the
sensitivity of CT in detecting small metas-
tases (Green et al., 1977). Ultrasound is
still more operator-dependent than CT and
yields results that are only as good as the
capabilities of the sonographer and has lim-
itations due to body habitus and overly-ing
gas. The diagnostic accuracy increased with
an increase in the number of metastases
and with an increase in size of the lesions.
A prospective comparative investigation of
71 colorectal patients demonstrated a sensiti-
vity of 71% using ultrasound and 82% using
CT scanning (Lundstedt et al., 1987).
The clinical value of preoperative ultra-
sonography in screening for synchronous
liver metastases was prospectively evalu-
ated in 338 patients with colorectal cancer.
Synchronous liver metastases were observed
at laparotomy in 11.5% (39/338) of the
patients. The sensitivity of preoperative
ultrasound was 76%. In detecting liver
metastases the results were superior to those
of biochemical blood tests and measure-
ments of carcinoembryonic antigen serum
levels. The accuracy of the ultrasonography
was also superior to that of these other tests
combined (Yoshida et al., 1989).
In a prospective blinded patient by
patient study of preoperative ultrasound,
S.R. Rafaelsen
295 consecutive patients with colorectal
cancer showed a sensitivity of 70% versus
97% with intraoperative ultrasound. Sub-
analysis on a segmental level showed a
high sensitivity of preoperative ultrasound
of metastases in segments III and VI, but
in the other liver segments intraoperative
ultrasound had statically higher sensitivity
(Rafaelsen et al., 1995). An earlier study
also found preoperative ultrasonography
less powerful for examination of the pos-
terior segment of the liver (Gunven et al.,
1985). A comparative study of preopera-
tive ultrasound, CT, and CT portography at
the time of primary resection of colorectal
cancer showed that preoperative percuta-
neous ultrasound was the best non-invasive
imaging method. The accuracy of CT porto-
graphy was worst because of many false
positive results (Nies et al., 1996).
Intraoperative ultrasonography (IOUS)
reduces the number of patients with liver
metastases from being subjected to super-
fluous or even harmful liver surgery and
it may increase the number in whom liver
surgery will prolong life. IOUS accu-
racy may be further improved by con-
trast enhanced -IOUS with an impact on
surgical strategy for metastatic tumors.
However, the new technique has a learn-
ing curve and requires additional costs
(Torzilli et al., 2005). Because conven-
tional palpation of the liver is not possible
during laparoscopic surgery, the routine
use of laparoscopic ultrasound should also
be considered for laparoscopic oncologic
colorectal surgery. This method enables
an earlier detection of liver metastases and
may have an impact on the survival rate
of a limited subgroup of colorectal cancer
patients (Milsom et al., 2000).
Using contrast enhanced sonography,
an increased sensitivity is reported. Pulse
28. Liver Metastases from Colorectal Cancer
inversion technique at low mechanical
index performed with microbubble con-
trast agents shows that noninvasive char-
acterization of liver masses is possible.
A hypoechoic mass seen during the portal
venous phase, regardless of its appearance
during the arterial phase, is suspicious for
malignancy. Metastases have consistently
shown less enhancement compared to the
liver during the portal venous phase of
contrast-enhanced ultrasound (Brannigan
et al., 2004). Several studies have shown
that the use of contrast agents increases
the sensitivity and specificity of liver
metastases. One study showed an increase
in sensitivity from 69% on non-enhanced
ultrasound to 90% using contrast enhance-
ment (Oldenburg et al., 2005). Missed
metastases using conventional ultrasound
is of small size or isoechoic.
In an animal study no significant differ-
ence was found between contrast-enhanced
ultrasound and MR imaging in the detec-
tion of hepatic tumors, whereas contrast-
enhanced ultrasound had the highest
accuracy (92%) of the modalities studied
(Forsberg et al., 2002a). However, these
results have to be proven in a clinical set-
ting, taking into consideration that the blind
angles and subcutaneous fat can hamper
diagnostic hepatic ultrasound scanning.
In another animal study the size of the
implanted tumor nodules measured by his-
topathology correlated closely (correlation
coefficient: 0.996) with those measured by
contrast enhanced sonography (Maruyama
et al., 2005).
A short hepatic contrast transit time
(< 27 s) appears to be found only in patients
with liver disease shown in a small series
by Bang et al. (2001). A sonographic
measurement of the transit time of an
echo enhancer from the hepatic artery
359
to the hepatic vein seems to be able to
discriminate between patients with and
without liver metastases. In a small series,
the transit times in 14 patients with liver
metastases were ≤ 10 s, while in 50 patients
without metastases the times were ≥ 12 s
(Bernatik et al., 2004).
DOPPLER FLOW PATTERN
The hepatic metastases from colorectal
cancer often have a characteristic sono-
graphic appearance, typically showing a
hypoechoic halo. The nature of this halo
is poorly understood and opinions differ.
Some authors have suggested that the halo
represents an oedema or compressed peri-
tumoral liver tissue (Marchal et al., 1985),
while others have shown active tumor tissue
in an in-vitro study (Wernecke et al., 1992).
Color Doppler can visualize blood flow
in tissue but the technique is not particu-
larly efficient for detection of slow flow
in small vessels. Power Doppler outper-
forms color Doppler for the detection of
blood flow. Although echo-enhancers have
been reported to improve the signal-to-
noise-ratio, these agents have only been
used for evaluation of liver metastases
from colorectal cancer in patients using
color Doppler. Power Doppler in com-
bination with ultrasound contrast agents
has been found most useful for detection
of vascular flow (Goldberg et al., 1996).
To our knowledge, a systematic, contrast-
enhanced power Doppler study of liver
metastases from colorectal cancer has not
previously been conducted. An under-
standing of the tumor margin is essential
in relation to the increasing use of thermal
ablation. The vascular nature of the metas-
tases also has definite clinical implications
360
because anti-angiogenic tumor therapy,
i.e., bevacizumab, represents a promising
new strategy for cancer treatment, and
the side-effect profile makes it a suitable
adjunct to standard chemotherapy in set-
tings where efficacy has been demonstrated.
Bevacizumab is now approved for use in
the USA and the European Union (Gordon
et al., 2005).
Rafaelsen and Solvig (2004) described
the ability of color Doppler, power Doppler
and contrast-enhanced flow imaging to
detect Doppler signals in liver metastases
from colorectal cancer and to evaluate a
possible connection between local recur-
rence and the vascular architecture in liver in focal lesions with power Doppler upon
metastases as shown by power Doppler.
The superiority of contrast-enhanced
power Doppler imaging over native power
Doppler examination of liver metastasis
halos from colorectal cancer (P < 0.005)
supports the assumption of peripheral vas-
cularity. However, it is emphasized that the which reported distended sinusoidal spaces
technique could not differentiate between
tumor and non-tumor vascularization,
Figure 28.2.
Figure 28.2. Power Doppler image showing flow 1994)
within the halo
S.R. Rafaelsen
An in vivo study of eight colon tumors
compared contrast-enhanced Doppler
ultrasonography with histological slides
for evaluation of angiogenesis and found
that the detection threshold after Levovist
allowed visualization of small vessels
(Lassau et al., 2001). Leen et al. (1994)
have previously described that Levovist
improved the sensitivity of color Doppler
flow imaging of the vascular appear-
ance of colorectal cancer metastases. A
study deploying contrast-enhanced power
Doppler in the examination of heterogene-
ous liver lesions in 32 patients detected an
increased number of intra-tumoral vessels
administration of contrast medium (Hosten
et al., 1999). This also applied to liver
metastases from colorectal cancer. The
vascularity of the halo in liver metastases
has also been demonstrated in an animal
study of hepatic colon cancer metastases,
that gave rise to new tumor-penetrating
vessels as well as tumor cells (Kruskal
et al., 2000). This supports the assumption
that the peritumoral vascularity represents
neoangiogenesis rather than simply dis-
placed liver vessels. A peripheral arterial
ring enhancement of liver metastases from
colorectal cancer has also been described
using contrast agent and pulse inversion
harmonic imaging (Tanaka et al., 2001).
Detection of flow signal in the center was
possible using contrast-enhanced power
Doppler in 21 of 43 of the metastases
(Rafaelsen and Solvig, 2004). The study
seems to be the first reporting this observa-
tion in liver metastases. None or few signals
were observed in the core of the lesions in
the above-mentioned study (Leen et al.,
using enhanced color Doppler. This
finding could be explained by the increased
28. Liver Metastases from Colorectal Cancer
ability of power Doppler to detect flow
in comparison with color Doppler. In our
flow study larger metastases were found
more frequently to have a central Doppler
signal than smaller lesions. This supports
the hypothesis that there is an associa-
tion between metastases showing a central
power Doppler flow and local tumor recur-
rence. However, the chance of detecting a
signal is greater in a large volume than in
a small volume. Moreover, the larger the
metastasis, the more advanced the disease,
and therefore the higher the likelihood of
disease with recurrence elsewhere. Thus,
it is implied that central signal detection is
merely a function of larger masses.
Among 12 patients with power Doppler,
flow signals within the center eight were
observed in association with a local recur-
rence, whereas none without central flow
had local recurrence, which indicates that
a central flow may be associated with
even more advanced disease (Figure 28.3).
Figure 28.3. Power Doppler image showing a central Doppler signal
361
Doppler signals tended to be detected
more frequently in patients with multiple
liver metastases than in patients with fewer
than five lesions (50% vs. 23%), which
may also signal a more aggressive beha-
vior. Synchronous liver disease and lymph
node metastases seemed to be associated
with a slight increase in Doppler signals
(Rafaelsen and Solvig, 2004). These find-
ings need to be further confirmed in future
large-scale studies. Patients with synchro-
nous disease have been shown to do signifi-
cantly worse than those with metachronous
hepatic resections (Jenkins et al., 1997).
Important new information may be derived
from studying the relationship between
flow pattern and tumor grade.
A recent study of 152 colorectal cancer
patients showed a significant correlation
between a poorer 5-year survival rate and
vascular endothelial growth factor C and
histological microvessel density (Furudoi
et al., 2002). Tumor necrosis factor has
362
exhibited specific tumoricidal activity and
may damage the blood vessels within the
tumor (Smyth et al., 2003). However,
it might be too early to conclude that a
natural tumor necrotic factor could be an
explanation for a correlation between the
registered few central Doppler signals and
a low incidence of local recurrence.
A large-scale follow-up study is required
to answer the important question of a pos-
sible relationship between flow pattern and
growth and response to chemotherapy.
In the future, power Doppler examina-
tion of neovascularity of hepatic metastases
could be a possible adjunct in the clinical
evaluation of antiangioneogenic therapy
(Du et al., 2005). Contrast-enhanced ultra-
sound has currently been described as a tool
for assessing differences in the microcircu-
lation of animal tumors in studies of antian-
giogenic agents (Pollard et al., 2002).
Bevacizumab (Avastin), the first approved
therapy designed to inhibit tumor angio-
genesis, has significant clinical benefits in
the management of colorectal cancer (Diaz-
Rubio and Schmoll, 2005).
With the advent of three-dimensional
power Doppler ultrasound it is now pos-
sible to visualize all detectable vessels of
the metastasis in one 3-D image and thus
reduce operator dependability at the same
time using contrast media (Forsberg et al.,
2002b). Furthermore, new contrast specific
methods and contrast agents are now avail-
able. Phase inversion imaging and flash-
echo imaging with intravenously infused
microbubble contrast have shown promis-
ing results in measuring vascular flow and
tissue perfusion (Lucidarme et al., 2004).
The new perfluro gasses replacing air and
multipulse scanning technique are very
sensitive in detecting the microbubble sig-
nals (Cosgrove and Blomley, 2004).
S.R. Rafaelsen
At present the measurements of the size
of the metastases should incorporate the
halo in the total diameter. It is essential
to include the halo in resection of liver
metastases. In addition, the halo must be
encompassed in the ablation field in inter-
stitial therapy studies. The neovascularity
in solid tumors can hinder thermal ablation
as well as large vessels surrounding the
lesion. The hypervascularity present in the
margin of the metastasis can act as a heat
sink that limits the heating of tissue used in
radiofrequency ablation. In the flow study
local recurrence was observed in 66% of
the patients with a power Doppler signal in
the center but in none of the patients who
had no such power Doppler signal. The
clinical conclusion of this could be that
attention should be paid to the possibility
of local recurrence where hepatic resection
or tissue ablation is considered.
It is concluded that the peripheral halo
in hepatic metastases indicates tissue with
increased vascularity. A central power
Doppler signal within the liver metas-
tases implies the co-existence of more
advanced disease. Further survival studies
are needed to assess the prognostic value
of this new finding and its usefulness as a
tool for monitoring the possible effect of
anti-angiogenic therapy.
An in vitro study showed that ultra-
sound contrast microbubbles targeted
via linkage with the tripeptide arginine-
arginine-leucine could specifically adhere
to tumor angiogenic endothelium (Weller
et al., 2005). Endothelial cells of angiogenic
tumor vasculature are characterized by
altered expression of molecular markers on
their surface. Targeted microbubbles may
thus offer a noninvasive contrast-enhanced
ultrasound imaging technique for the func-
tional imaging of tumor neovascularization
28. Liver Metastases from Colorectal Cancer
and may have further implications for thera-
peutic tumor targeting.
ULTRASOUND IN measuring the largest diameter, three-
POSTOPERATIVE FOLLOW-UP dimensional high-frequency ultrasound
The aim of follow-up programs in cancer
patients is to detect relapse or metastases
in an early asymptomatic and potential
curable stage. The 5-year cumulative risk
of metachronous liver metastases is 16%
(Faivre et al., 2003). Computed tomog-
raphy scanning is used in follow-up
programs. Ultrasound of the liver has also
been suggested to detect metacroneous
liver metastases in patients with high risk
of recurrence (de Goede et al., 1998). The
median time to development of liver metas-
tases detected by prospective serial ultra-
sound scanning of the liver at 3-monthly
intervals of ultrasound scanning was 59
weeks in the control group and 91 weeks
in the group treated with chemotherapy
(Taylor et al., 1984). Little is known of
the frequency of follow-up, and presently
there are no controlled trials randomising
patients to high- or low-volume control
programs. However, follow-up programs
for patients with curatively resected color-
ectal cancer do improve survival but it is
not clear which type of imaging or frequency
of visits is optimal (Figueredo et al., 2003).
Doppler perfusion index (DPI) measure-
ments for an early identification of patients
at high risk for recurrent disease have been
suggested by Leen et al. (1993). Patients
with normal DPI remained disease-free
and patients with an abnormally high DPI
at the time of primary resection later devel-
oped liver metastases. A recent study could
not confirm the usefulness of the method
(Roumen et al., 2005).
363
FUTURE POTENTIAL
ADVANCEMENTS
New results have shown that, rather than
imaging may be particularly well-suited for
the quantitative assessment of metastatic
progression and the evaluation of chemo-
therapeutics in preclinical liver metastasis
models. The traditional two-dimensional
method often gives large overestimation or
underestimation of the tumor volume com-
pared to the three-dimensional method
(Graham et al., 2005).
Conventional ultrasound systems acquire
ultrasound data sequentially one image
line at a time. The configuration of these
systems is, therefore, also sequential in
nature and processes most of the data
in a sequential pipeline. The possibility
of sending out arbitrary signals and the
storage of data from all transducer ele-
ments for 5–10 s allows clinical evaluation
of synthetic aperture and 3-D imaging.
Jensen et al. (2005) described a new high-
speed real-time ultrasound system with
16 Gbytes RAM especially designed for
research purposes.
Tele-ultrasound consultation systems
via high-speed commercial telecommuni-
cation lines are presently being tested, and
Yoo et al. (2004) suggested that a rate of
more than 0.6 Mbit/s at 30 frames/s is suffi-
cient for the maintenance of diagnostic
image quality. This will make it possible to
send “cine clips” via teleradiology and ask
an ultrasound expert for a second opinion.
With a new tele-operated robotic chain for
real-time ultrasound image, it might be
possible to hold and move a real probe on
a distant patient according to the expert ges-
ture and permit an image acquisition using
364
a standard ultrasound device. The remote
ultrasound scanning system agreed in at
least 80% of the 52 cases with the diagnosis
obtained from conventional scanning
(Courreges, 2005).
Computer-aided detection may have the
potential to increase the accuracy of diag-
nosis of focal liver lesions in ultrasound
images, although the literature is very
scarce (Yoshida, 2003).
The development of new contrast agents
and ultrasound software still continues,
i.e., micro flow imaging. In the future,
ultrasound might be adopted in various
drug delivery and other therapeutic appli-
cations. Microspheres and nanoparticles of
different sizes with or without drugs, stem
cells, and/or radioisotopes, can be used
for drug delivery or diagnostic purposes.
Ultrasound will then facilitate the deli-
very of chemotherapeutic drugs into liver
metastases and promote gene therapy to
targeted tissues under real-time imaging
(Mitragotri, 2005). Ultrasound will pro-
vide information that will benefit patient
management now and in the future.
REFERENCES
Adam, R., Avisar, E., Ariche, A., Giachetti, S.,
Azoulay, D., Castaing, D., Kunstlinger, F.,
Levi, F., and Bismuth, F. 2001. Five-year survival
following hepatic resection after neoadjuvant
therapy for nonresectable colorectal cancer. Ann.
Surg. Oncol. 8:347–353.
Bang, N., Nielsen, M.B., Rasmussen, A.N.,
Osterhammel, P.A., and Pedersen, J.F. 2001.
Hepatic vein transit time of an ultrasound con-
trast agent: simplified procedure using pulse
inversion imaging. Br. J. Radiol. 74:752–755.
Bernatik, T., Becker, D., Neureiter, D., Hansler,
J.H., Frieser, M., Schaber, S., Wein, A., Hahn,
E.G., and Strobel, D. 2004. Hepatic transit
time of an echo enhancer: an indicator of meta-
static spread to the liver. Eur. J. Gastroenterol.
Hepatol. 16:313–317.
S.R. Rafaelsen
Brannigan, M., Burns, P.N., and Wilson, S.R.
2004. Blood flow patterns in focal liver lesions
at microbubble-enhanced US. Radiographics
24:921–935.
Bruneton, J.N., Raffaelli, C., Balu-Maestro, C.,
Padovan I, B., Chevallier, P., and Mourou, M.Y.
1996. Sonographic diagnosis of solid liver nodules
in cancer patients. Eur. Radiol. 6:439–442.
Choudhry, S., Gorman, B., Charboneau, J.W.,
Tradup, D.J., Beck, R.J., Kofler, J.M., and Groth,
D.S. 2000. Comparison of tissue harmonic imag-
ing with conventional US in abdominal disease.
Radiographics 20:1127–1135.
Cosgrove, D., and Blomley, M. 2004. Liver tumors.
Abdom. Imaging 29:446–454
Courreges, F., Vieyres, P., Istepanian, R.S., Arbeille, P.,
and Bru C. 2005. Clinical trials and evaluation
of a mobile, robotic tele-ultrasound system. J.
Telemed. Telecare 11 Suppl 1:46–49.
Diaz-Rubio, E., and Schmoll, H.J. 2005. The future
development of bevacizumab in colorectal can-
cer. Oncology 69 Suppl 3:34–45.
Du, W.H., Yang, W.X., Xiong, X.Q., Wang, X.,
Zhou, Y., and Wang H. 2005. Contrast-enhanced
ultrasonographic imaging diagnosis on assess-
ment of vascularity in liver metastatic lesions.
World J. Gastroentol. 11:3610–3613.
Easson, A.M., Barron, P.T., Cripps, C., Hill, G.,
Guindi, M., and Michaud, C. 1996. Calcification
in colorectal hepatic metastases correlates with
longer survival. J. Surg. Oncol. 63:221–225.
Eberhardt, S.C., Choi, P.H., Bach, A.M., Funt,
S.A., Felderman, H.E., and Hann, L.E. 2003.
Utility of sonography for small hepatic lesions
found on computed tomography in patients with
cancer. J. Ultrasound Med. 22:335–343.
Faivre, J., Manfredi, S., and Bouvier, A.M. 2003.
Epidemiology of colorectal cancer liver metas-
tases. Bull. Acad. Nath. Med. 187:815–823.
Figueredo, A., Rumble, R.B., Maroun, J., Earle,
C.C., Cummings, B., McLeod, R., Zuraw, L.,
and Zwaal, C. Gastrointestinal Cancer Disease
Site Group of Cancer Care Ontario’s Program
in Evidence-based Care. 2003. Follow-up of
patients with curatively resected colorectal can-
cer: a practice guideline. Bio. Med. Central.
Cancer 6;3:26
Forsberg, F., Piccoli, C.W., Liu, J.B., Rawool, N.M.,
Merton, D.A., Mitchell, D.G., and Goldberg, B.B.
2002a. Hepatic tumor detection: MR imaging
28. Liver Metastases from Colorectal Cancer
and conventional US versus pulse-inversion
harmonic US of NC100100 during its reticu-
loendothelial system-specific phase. Radiology
222:824–829.
Forsberg, F., Rawool, N.M., Merton, D.A., Liu, J.B.,
and Goldberg, B.B. 2002b. Contrast enhanced
vascular three-dimensional ultrasound imaging.
Ultrasonics 40:117–122.
Furudoi, A., Tanaka, S., Haruma, K., Kitadai, Y.,
Yoshihara, M., Chayama, K., and Shimamoto, F.
2002. Clinical significance of vascular endo-
thelial growth factor C and angiogenesis at the
deepest invasive site of advanced colorectal
carcinoma. Oncology 62:157–166.
Gillams, A.R. 2005. The use of radiofrequency in
cancer. Br. J. Cancer 92:1825–1829.
de Goede, E., Filez, L., Janssens, J., and Van
Cutsem, E. 1998. Follow-up of colon cancer:
detection of liver metastases: benefit and perio-
dicity. Acta Gastroenterol. Belg. 61:8–10.
Goldberg, B.B., Merton, D.A., Forsberg, F., Liu,
J.B., and Rawool, N. 1996. Color Amplitude
Imaging: Preliminary results using vascular
sonographic contrast agents. J. Ultrasound Med.
15:127–134.
Gordon, M.S., and Cunningham, D. Managing
patients treated with bevacizumab combination
therapy. 2005. Oncology 69 Suppl 3:25–33.
Graham, K.C., Wirtzfeld, L.A., MacKenzie, L.T.,
Postenka, C.O., Groom, A.C., MacDonald, I.C.,
Fenster, A., Lacefield, J.C., and Chambers, A.F.
2005. Three-dimensional high-frequency ultra-
sound imaging for longitudinal evaluation of
liver metastases in preclinical models. Cancer
Res. 65:5231–5237.
Green, B., Bree, R.L., Goldstein, H.M., and Stanley,
C. 1977. Gray scale ultrasound evaluation of
hepatic neoplasms: patterns and correlations.
Radiology 124:203–208.
Gruenberger, T., Zhao, J., King, J., Chung, T, Clingan
P.R., and Morris, D.L. 2002. Echogenicity of
liver metastases from colorectal carcinoma is
an independent prognostic factor in patients
treated with regional chemotherapy. Cancer
94:1753–1759.
Gunven, P., Makuuchi, M., Takayasu, K.,
Moriyama, N., Yamasaki, S., and Hasegawa, H.
1985. Preoperative imaging of liver metastases.
Comparison of angiography, CT scan, and ultra-
sonography. Ann. Surg. 202:573–579.
365
Guyot, F., Faivre, J., Manfredi, S., Meny, B.,
Bonithon-Kopp, C., and Bouvier, A,M. 2005.
Time trends in the treatment and survival of
recurrences from colorectal cancer. Ann. Oncol.
16:756–761.
Hosten, N., Puls, R., Lemke, A.J., Steger, W.,
Zendel, W., Zwicker, C., and Felix, R. 1999.
Contrast-enhanced power Doppler sonography:
improved detection of characteristic flow pat-
terns in focal liver lesions. J. Clin. Ultrasound
27:107–115.
Jenkins, L.T., Millikan, K.W., Bines, S.D., Staren,
E.D., and Doolas, A. 1997. Hepatic resec-
tions for metastatic colorectal cancer. Am. Surg.
63:605–610.
Jensen, J.A., Holm, O., Jensen, L.J., Bendsen, H.,
Nikolov, S.I., Tomov, B.G., Munk, P., Hansen, M.,
Salomonsen, K., Hansen, J., Gormsen, K.,
Pedersen, H.M., and Gammelmark, K.L. 2005.
Ultrasound research scanner for real-time synthetic
aperture data acquisition. IEEE. Trans. Ultrason.
Ferroelectr. Freq. Control 52:881–891.
Jones, O.M., Rees, M., John, T.G., Bygrave, S., and
Plant, G. 2005. Biopsy of resectable colorectal
liver metastases causes tumour dissemination
and adversely affects survival after liver resec-
tion. Br. J. Surg. 92:1165–1168.
Jorgensen, T., and Rafaelsen, S. 1992. Gallstones
and colorectal cancer – there is a relationship,
but it is hardly due to cholecystectomy. Dis.
Colon Rectum 35:24–28.
Khatri, V.P., Petrelli, N.J., and Belghiti, J. 2005.
Extending the frontiers of surgical therapy for
hepatic colorectal metastases: is there a limit?
J. Clin. Oncol. 23:8490–8499.
Kruskal, J.B., Thomas, P., Nasser, I., Cay, O., and
Kane, RA. 2000. Hepatic colon cancer in mice:
Dynamic in vivo correlation with hypoechoic
rims visible at US. Radiology 215:852–857.
Lassau, N., Koscielny, S., Opolon, P., De Baere, T.,
Peronneau, P., Leclere, J., and Roche, A. 2001.
Evaluation of contrast-enhanced color Doppler
ultrasound for the quantification of angiogenesis
in vivo. Invest. Radiol. 36:50–55.
Leen, E., Angerson, W.J., Warren, H.W., O’Gorman, P.,
Moule, B., Carter, E.C., and Mc Ardel, C.S.
1994. Increased sensitivity of color Doppler
flow imaging of colorectal hepatic metastases
using galactose microparticles: a preliminary
report. Br. J. Surg. 8:252–254.
366
Leen, E., Goldberg, J.A., Robertson, J., Angerson,
W.J., Sutherland, G.R., Cooke, T.G., and
McArdle, C.S. 1993. Early detection of occult
colorectal hepatic metastases using duplex colour
Doppler sonography. Br. J. Surg. 80:1249–1251.
Lucidarme, O., Nguyen, T., Kono, Y., Corbeil, J.,
Choi, SH., Varner, J., and Mattrey, R.F. 2004.
Angiogenesis model for ultrasound contrast
research. Acad. Radiol. 11:4–12.
Lundstedt, C., Ekberg, H., Hederstrom, E.,
Stridbeck, H., Torfason, B., and Transberg, K.G.
1987. Radiologic diagnosis of liver metastases
in colo-rectal carcinoma. Prospective evaluation
of the accuracy of angiography, ultrasonography,
computed tomography and computed tomo-
graphic angiography. Acta Radiol. 28:431–438.
Marchal, G.J., Plyser, K., Tshibwabwa-Tumba,
E.A., Verbeken, E.K., Oyen, R.H., Baert, A.L.,
and Lauweryns, J.M. 1985. Anechoic halo in
solid liver tumors. Radiology 156:479–483.
Maruyama, H., Matsutani, S., Saisho, H., Mine, Y.,
Kamiyama, N., Hirata, T., and Sasamata, M.
2005. Real-time blood-pool images of contrast
enhanced ultrasound with Definity in the detec-
tion of tumour nodules in the liver. Br. J. Radiol.
78:512–518.
Milsom, J.W., Jerby, B.L., Kessler, H., Hale,
J.C., Herts, B.R., and O’Malley, C.M. 2000.
Prospective, blinded comparison of laparoscopic
ultrasonography vs. contrast-enhanced com-
puterized tomography for liver assessment in
patients undergoing colorectal carcinoma surgery.
Dis. Colon Rectum 43:44–49.
Mitragotri, S. 2005. Healing sound: the use of
ultrasound in drug delivery and other therapeutic
applications. Nat. Rev. Drug Discov. 4:255–260.
Miyake, K., Hayakawa, K., Nishino, M.,
Morimoto, T., and Mukaihara, S. 2005. Effects
of oral 5-fluorouracil drugs on hepatic fat con-
tent in patients with colon cancer. Acad. Radiol.
12:722–727.
Nies, C., Leppek, R., Sitter, H., Klotter, H.J., Riera, J.,
Klose, K.J., Schwerk, W.B., and Rothmund, M.
1996. Prospective evaluation of different diag-
nostic techniques for the detection of liver metas-
tases at the time of primary resection of colorectal
carcinoma. Eur. J. Surg. 162:811–816.
Oldenburg, A., Hohmann, J., Foert, E., Skrok, J.,
Hoffmann, C.W., Frericks, B., Wolf, K.J., and
Albrecht, T. 2005. Detection of hepatic metastases
S.R. Rafaelsen
with low MI real time contrast enhanced sonogra-
phy and SonoVue. Ultraschall Med. 26:277–284.
Pollard, R.E., Sadlowski, A.R., Bloch, S.H., Murray,
L., Wisner, E.R., Griffey, S., and Ferrara, K.W.
2002. Contrast-assisted destruction-replenishment
ultrasound for the assessment of tumor microvas-
culature in a rat model. Technol. Cancer Treat.
1:459–470.
Rafaelsen, S.R., Kronborg, O., Larsen, C., and
Fenger, C. 1995. Intraoperative ultrasonography
in detection of hepatic metastases from colorec-
tal cancer. Dis. Colon Rectum 38:355–360.
Rafaelsen, S.R. and Solvig, J. 2004. Ultrasound
imaging of flow patterns in liver metastases
from colorectal cancer. Scand. J. Gastroenterol.
39:761–765.
Roumen, R.M., Scheltinga, M.R., Slooter, G.D.,
and van der Linden, A.W. 2005. Doppler per-
fusion index fails to predict the presence of
occult hepatic colorectal metastases. Eur. J. Surg.
Oncol. 31:521–527.
Smyth, M.J., Takeda, K., Haykawa, Y., Peschon,
J.J., van den Brink, M.R., and Yagita, H. 2003.
Nature´s TRAIL – on a path to cancer immuno-
therapy. Immunity 18:1–6.
Tanaka, S., Ioka, T., Oshikawa, O., Hamada, Y.,
and Yoshioka, F. 2001. Dynamic sonography of
hepatic tumors. Am. J. Radiol. 177:799–805.
Taylor, R.H., Gilbert, J.M., Evans, M.G., Lane,
H.G., and Cassell, P.G. 1984. Prospective serial
liver ultrasound scanning in resectable colorec-
tal cancer treated with adjuvant razoxane. Clin.
Exp. Metastasis 2:321–331.
Torzilli, G., Del Fabbro, D., Palmisano, A., Donadon,
M., Bianchi, P., Roncalli, M., Balzarini, L.,
and Montorsi, M. 2005. Contrast-enhanced
intraoperative ultrasonography during hepatec-
tomies for colorectal cancer liver metastases.
J. Gastrointest. Surg. 9:1148–1154.
Weller, G.E., Wong, M.K., Modzelewski, R.A.,
Lu, E., Klibanov, A.L., Wagner, W.R., and
Villanueva, F.S. 2005. Ultrasonic imaging
of tumor angiogenesis using contrast micro-
bubbles targeted via the tumor-binding pep-
tide arginine-arginine-leucine. Cancer Res.
65:533–539.
Wernecke, K., Henke, L., Vassaloo, P., von
Bassewitz, D.B., Diederich, S., Peters, P.E., and
Edel, G. 1992. Pathologic explanation for hyp-
oechoic halo seen on sonograms of malignant
28. Liver Metastases from Colorectal Cancer
liver tumors: An in vitro correlative study. Am.
J. Radiol. 159:1011–1116.
Yoo, S.K., Kim, D.K., Jung, S.M., Kim, E.K.,
Lim, J.S., and Kim, J.H. 2004. Performance of a
web-based, realtime, tele-ultrasound consultation
system over high-speed commercial telecommu-
nication lines. J. Telemed. Telecare 10:175–179.
Yoshida, H., Casalino, D.D., Keserci, B., Coskun,
A., Ozturk, O., and Savranlar, A. 2003. Wavelet-
367
packet-based texture analysis for differentiation
between benign and malignant liver tumours
in ultrasound images. Phys. Med. Biol. 48:
3735–753.
Yoshida, T., Matsue, H., Suzuki, M., Okazaki, N.,
Yoshino, M., Moriya, Y., and Hojo, K. 1989.
Preoperative ultrasonography screening for liver
metastases of patients with colorectal cancer.
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
INTRODUCTION
Metastasis, the spread of a primary can-
cer to distant organs, continues to be the
most significant cause of cancer mortality.
Isolated primary tumors can often be treated
surgically with a relatively high success rate.
However, if the primary tumor has invaded
the surrounding tissue and metastasized
to secondary sites in the body, treatment
options are often limited to systemic chemo-
therapies with much lower success rates and
greater toxicity. Thus, it is imperative that a
greater understanding of the biology of the
metastatic process be acquired in order to
achieve a significant reduction in the mor-
bidity and mortality associated with cancer
diagnosis.
The process of metastasis consists of
multiple biological steps governed by
a wide range of molecular processes
(Chambers et al., 2002). The cells in a
developing primary tumor must invade
the surrounding tissue and gain access
to a blood or lymphatic vessel to facili-
tate dissemination. Once the metastatic
cell has arrived at a secondary site, the
cell must arrest in the vascular system,
survive, undergo cell division in the new
microenvironment, and eventually recruit
new blood vessels to allow for continued
development. Although many cells initiate
this sequence of events by gaining access
to the vascular system, < 1% of these cells
are able to complete all of the steps to
form overt metastases (Chambers et al.,
2002). The multi-step nature and bio-
logical and molecular complexity of the
metastatic process have necessitated that a
variety of research tools be used to effec-
tively model this process. In vitro models
have allowed for a greater understanding
of how tumor cells circumvent normal cell
growth and survival regulations. In vitro
models are essential to isolate the contri-
bution of specific molecular pathways to
the development of a metastatic cell, but
fail to capture the complexity of the entire
metastatic process that exists in the in vivo
situation. Thus, in order to study the com-
plete metastatic process it is necessary to
develop and utilize animal models. Animal
models are used to study the interactions
of a tumor cell with a changing micro-
environment as it progresses through the
metastatic process, and are often used to
evaluate novel therapeutics and treatment
strategies.
Historically, preclinical mouse mod-
els of cancer have been developed from
369
370
transplantable tumors of mouse origin in
immune competent (syngeneic) animals,
or of human origin (xenograft) in immune
deficient animals. In both xenograft and
syngeneic models the method of inoculat-
ing the cancer cells into the mice is cho-
sen based on the biological question that
is being asked. The tumor cells may be
introduced into the mice through surgical
implantation of cells or a tumor tissue frag-
ment into the orthotopic site. The orthotopic
site is the natural site of development for a
tumor type. For instance, the orthotopic
site for implanting colon tumor cells would
be into a mouse colon. This method reca-
pitulates many steps of metastatic develop-
ment including primary tumor growth and
invasion of the surrounding tissue and vas-
cular system. A drawback of this system
is that the primary tumors often develop
rapidly, inducing animal morbidity or mor-
tality that obscures the study of secondary
metastases or requires the surgical resec-
tion of the primary tumor. Furthermore,
there is a lack of control over the number
of metastases that form and the timing of
their development in orthotopic models.
For these reasons, experimental metastasis
models are often preferred (Welch, 1997).
In experimental metastasis models cancer
cells are grown in culture and then injected
intravenously to form metastases at specific
sites. In this procedure, the primary tumor
growth and invasion stages are bypassed,
but the researcher has greater control over
the number of potential metastatic clones
in an organ and the timing in which overt
metastases may form.
Recent advances in the development
of transgenic animals have created ani-
mal models that carry specific genetic
defects that increase their suscep-
tibility to developing particular cancers.
K.C. Graham et al.
Transgenic animals produce tumors that
develop over a time scale and in a pro-
gression series that more closely resem-
bles the clinical situation (Khanna and
Hunter, 2005). Transgenic models are
expensive to produce and maintain, and
because of variable tumor incidence
and infrequent development of metas-
tases these models are often limited to
studies with specific objectives that pre-
clude the use of transplantable or experi-
mental models.
One feature that all metastasis models
have in common is that the development
of the metastases is difficult to quantify
non-invasively. Traditionally, metastasis
models have been studied by using the
simple approach of analyzing groups of
experimental animals at a common end-
point. To study the process of metastasis a
large group of animals was used and then
sub-groups were sacrificed and analyzed at
specific timepoints throughout the experi-
ment. This method is prohibitively expen-
sive and relies on the assumption that the
variability in tumor development is small
enough that an accurate impression of
the kinetics of metastatic progression can
be gleaned by analyzing small groups of
mice. In many models, particularly geneti-
cally engineered models, the variability in
metastasis development is sufficiently high
to prevent statistically meaningful results
from the evaluation of small subgroups of
animals (Khanna and Hunter, 2005). To
address this problem, non-invasive imag-
ing technologies have been developed spe-
cifically for small animal research. The
ability to non-invasively image metasta-
sis models would decrease experimental
variability by allowing individual mice,
and the individual metastases developing
within them, to be studied longitudinally.
29. Preclinical Liver Metastases: Three-Dimensional High-Frequency Ultrasound Imaging
The variability in the timing of metastatic
development, while still present, is no
longer a factor in a longitudinal experiment
because the researcher can study metastatic
progression within each individual mouse
and gather data or initiate a treatment
based on the unique stage of progression
in each mouse. By decreasing the impact
of experimental variability, smaller groups
of animals are required to achieve a statis-
tically significant result, thereby reducing
both the cost and time required to perform
an experiment.
The importance of small animal imaging
in metastasis research is reflected by the
number of technologies dedicated to this task.
Many of the technologies commonly used
clinically, such as magnetic resonance
imaging (MRI), x-ray computed tomogra-
phy (CT), and positron emission tomogra-
phy (PET), have been redesigned to give
spatial resolutions that are adequate for
small animal imaging (Weissleder, 2002).
Other imaging technologies, because of
their inherent limitations in imaging depth
or their requirement for the genetic mani-
pulation of the cell line being studied, are
not widely used clinically, but have been
developed for and utilized by small animal
researchers. These technologies include
many of the optical techniques based on
fluorescent or bioluminescent proteins.
The general attributes that make an
imaging modality attractive for metastasis
research include high spatial resolution to
allow metastatic progression to be tracked
from early stages, fast image acquisition
to increase animal throughput, affordable
capital and operating costs, and inherent
contrast between the metastases and the
normal parenchyma. No single imaging
modality can fulfill all of these criteria, so
the researcher must evaluate the particular
371
needs of each research project when select-
ing an imaging modality. Criteria specific
to each research project include the type of
metastasis model being used, the anatomical
location at which metastases are expected,
the desired resolution, the availability of
equipment and funding, the necessity and
availability of contrast agents, and access
to personnel skilled in the acquisition and
interpretation of the images.
For researchers who have access to MRI
facilities, it offers high-resolution imag-
ing, inherent soft tissue contrast, and the
availability of a variety of contrast agents
for vascular imaging (Evelhoch et al.,
2000; Gillies et al., 2000). These features
have led to the selection of MRI for pre-
clinical studies involving the quantifica-
tion of tumor burden in liver metastasis
models (Cai et al., 2005; Wu et al., 2003)
and in the measurement of tumor blood
volume and vascular permeability through
dynamic contrast enhanced (DCE) MRI
(Leach, 2001). Current studies are extend-
ing the use of MRI to study metastasis to
the brain (Heyn et al., 2006). Widespread
use of MRI to noninvasive study of cancer
models is largely hampered by long image
acquisition and processing times, and the
high costs associated with equipment pur-
chase and operation (Weissleder, 2002).
X-ray CT offers high-resolution imag-
ing and inherent contrast between bone
and soft tissue, but poor contrast within
the various soft tissues themselves. Due
to these characteristics, CT has been pre-
dominately selected for studies of bone
metastasis (Paulus et al., 2000). The devel-
opment of novel iodinated contrast agents
which provide liver parenchyma – liver
metastases contrast has introduced the
possibility of using CT to study liver
metastasis models (Weber et al., 2004).
372
This development may make CT useful
for endpoint studies of tumor burden, but
longitudinal imaging may be limited due
to repeated ionizing radiation exposure
during imaging.
Positron emission tomography is
extremely sensitive, yet suffers from poor
resolution. The strength of PET imag-
ing lies in the use of the specific radi-
oisotope-labeled tracers that can serve
as reporters for specific molecular proc-
esses (Gambhir, 2002). As opposed to
the anatomical imaging of MRI and CT,
PET is able to image biological processes
such as glucose metabolism or cellular
proliferation by linking the radioisotope
to the appropriate tracer. By combining
PET and CT, imaging information can be
gleaned on both the anatomical location
and the biological activity of tumor cells
(Gambhir, 2002).
Similar to PET imaging, fluorescent
and bioluminescent imaging have limited
spatial resolution, but are extremely ver-
satile in that the reporter gene or protein
can be manipulated to provide a signal
in response to a variety of biological
processes. Fluorescent and bioluminescent
imaging technologies are widely available
and relatively inexpensive to purchase and
operate. For these reasons, optical imag-
ing has found a wide range of applications
in the study of cancer models from the
quantification of tumor burden to monitor-
ing the delivery of cancer gene therapies
(Choy et al., 2003; Lyons, 2005).
A large portion of the metastasis research
in our laboratory has focused on the biology
of the metastatic process with the liver as
our model organ (Chambers et al., 1995,
2002). Liver metastases are common from
primary tumors in the colon and breast,
and are a frequent contributor to both
K.C. Graham et al.
morbidity and mortality in cancer patients.
Preclinical models of liver metastasis are
invaluable for our understanding of the
metastatic process and in the development
of novel therapeutics to target this process.
We have developed an experimental
metastasis model, injecting cultured tumor
cells into the mesenteric vein, to produce
metastases in the liver of mice (Morris
et al., 1993). This method can be utilized
with a variety of cell lines, of both mouse
and human origin, from a variety of tumor
types. We opted to pursue an experimental
metastasis model to be able to control the
number and timing of cells delivered to
the liver. Since the timing of delivery was
known, we were able to study the earliest
steps in the metastatic process. This led us
to identify a number of key biological and
molecular processes that govern metastatic
efficiency in a secondary site (Chambers et
al., 1995, 2002). The resolution required
to study the early stages of metastasis was
on the single cell level, which necessitated
the use of high resolution in vivo videomi-
croscopy (IVVM). This is an inherently
invasive procedure that is terminal for the
animal being studied.
Expanding on these studies, we are work-
ing to identify key biological and molecu-
lar processes that govern tumor growth at
later stages of metastasis, with the ultimate
goal of using these discoveries to help
identify effective therapeutic strategies.
With the focus on following the develop-
ment of micrometastases into tumors of a
clinically relevant size, our studies require
an imaging modality capable of longitu-
dinal imaging at sub-millimeter resolu-
tions. The imaging modality that best fit
the biological requirements of these studies
was high-frequency ultrasound. High-
frequency ultrasound offers cost effective,
29. Preclinical Liver Metastases: Three-Dimensional High-Frequency Ultrasound Imaging
high-resolution imaging with rapid image
acquisition and processing, and allows for
the non-invasive longitudinal study of liver
metastasis models without the requirement
for contrast agents.
METHOD: HIGH-FREQUENCY
ULTRASOUND
Ultrasound images are produced by send-
ing short duration acoustic pulses from
the ultrasound transducer into the tissue of
interest. When the sound wave encounters
an interface between two tissues hav-
ing different acoustic properties, some of
the energy is reflected back towards the
transducer. The time it takes for the echo
to return to the transducer allows for the
calculation of the distance to this interface.
The resulting image displays the location
and magnitude of sound reflection at these
interfaces.
In ultrasound imaging there is an
inherent tradeoff between the resolution
and depth of penetration. As the ultra-
sound frequency increases, the resolution
increases, while the imaging penetration
decreases. For this reason, clinical ultra-
sound is often performed in the range of
3–12 MHz, allowing for a resolution of
0.5–2.0 mm and an imaging depth in the
range of 5–10 cm. For preclinical imag-
ing, mouse organs have linear dimen-
sions that are approximately tenfold
smaller than humans (Foster et al., 2002).
To resolve the same tissue characteristics
seen in the clinical setting, the preclinical
ultrasound system must achieve a resolu-
tion that is approximately tenfold greater
than the clinical systems. This increase
in resolution is achieved by increasing
the frequency of the sound waves into
373
the 20–60 MHz range. Increasing the fre-
quency restricts imaging penetration to
5–15 mm, which is sufficient for many
preclinical applications in mouse models.
Imaging penetration does limit clinical
use of high-frequency ultrasound to very
specialized applications, such as ocular
or intravascular imaging, where the ultra-
sound transducer can be placed in close
proximity to the tissue of interest (Foster
et al., 2000, 2002).
The exact specifications of a high-fre-
quency ultrasound system will depend on
the design and vary slightly between sys-
tems. The system that we have utilized for
preclinical cancer studies is the Vevo 660
produced by VisualSonics Inc. The design
and performance of this scanner has been
previously reviewed (Foster et al., 2002).
In brief, the Vevo 660 transducer operates
at a 40 MHz center frequency with a 6 mm
focal distance. The spatial resolution at the
focus is 40 × 80 × 80 μm3.
Similar to the clinical situation, the
acquisition of high quality preclinical
images with high-frequency ultrasound
depends on appropriate preparation of the
subject to be imaged and the technical
experience of the sonographer. In order
to image the mouse liver, the abdomen
must first be depilated with commercial
hair removal cream. The mouse must be
lightly anesthetized to keep the animal
immobile during imaging, a task most
easily accomplished with an inhaled anes-
thetic such as 1.5% isoflurane in oxygen.
Unlike clinical imaging, which uses a
very thin layer of ultrasound contact gel
to allow transmission of sound waves into
the body, when imaging a small animal a
0.5 cm thick layer of ultrasound contact
gel is used. This allows the transducer
to slide easily over the curvature of the
374
abdomen without losing contact with the
gel. Ultrasound is strongly reflected by
bone and air-filled structures, which hin-
ders imaging of any tissues located below
bony structures or the lungs. In the evalu-
ation of the mouse liver, we have found
that the liver is partially obscured by the
ribcage, but significant portions of the left
lateral, left medial, and right medial liver
lobes are accessible for imaging (Graham
et al., 2005). The liver lobes move in rela-
tion to the bony anatomy and in relation
to each other over time. This movement
necessitates landmarks internal to each
liver lobe to be noted during longitudinal
imaging. For liver metastases, landmarks
such as the specific liver lobe in which
the metastasis develops, tumor shape, and
proximity to readily identifiable structures
such as major blood vessels and liver lobe
edges may be used to reliably identify an
individual metastasis over multiple imag-
ing sessions.
THREE-DIMENSIONAL
IMAGING AND VOLUME
CALCULATION
A calculation of the volume of a region
of interest from two-dimensional ultra-
sound images involves an assumption of
a defined three-dimensional shape. For
instance, when calculating the volume of
a tumor in preclinical models, the tumor is
often assumed to have an ellipsoid volume
that can be calculated after measurement
of the length and diameter of the tumor.
The accuracy of this method is dependent
on how closely the tumor of interest
matches the assumed shape. The ability to
acquire three-dimensional images elimi-
nates the need to assume a predetermined
K.C. Graham et al.
shape and thus would be expected to give
a more accurate measurement of tumor
volume. Indeed, previous work with a
clinical ultrasound system has shown that
three-dimensional imaging is more accu-
rate than two-dimensional imaging for
phantoms of regular and irregular shapes
(Tong et al., 1998; Xu et al., 2003).
To acquire three-dimensional ultrasound
images with the Vevo 660, the transducer
is fixed to a stepper motor and acquires
two-dimensional images at 30 μm inter-
vals as it moves in the out-of-plane dimen-
sion. The two-dimensional images are
then interpolated and a three-dimensional
image is reconstructed on-line (Fenster
et al., 2001). A typical 3-D image of
the mouse liver containing a metastasis
of interest would take ∼20 s to acquire
and reconstruct. This rapid processing of
the imaging data is advantageous to the
researcher as it allows the acquired image
to be evaluated and approved before mov-
ing on to the next subject. Such rapid
analysis is not possible with many other
imaging modalities.
We have compared the calculated volume
of preclinical liver metastases after two-
dimensional measurement and assumption
of an ellipsoid volume with the volume
calculated after segmentation of three-
dimensional images. We have shown that,
for murine B16F1 melanoma liver metas-
tases, the two-dimensional method on
average yields a smaller value than the
three-dimensional method with a mean
percent difference of 8.8 ± 23.5% (Graham
et al., 2005). The large standard deviation
indicates that the two-dimensional method
can yield large overestimations or under-
estimations of tumor volume in compari-
son to the three-dimensional method. To
determine which method is more accurate,
29. Preclinical Liver Metastases: Three-Dimensional High-Frequency Ultrasound Imaging
the true tumor volume of each metastasis
would have to be known. There is currently
no method of measuring tumor volume that
is accepted as the “gold standard” to which
other methods can be compared. Tumor
dimensions seen through non-invasive
imaging may be compared to serial histo-
logical sections to gain a sense of whether
the two methods yield similar results, but
values obtained through histology cannot
be considered the true values because the
extent of tissue distortion during histologi-
cal processing is unknown. Without such
a “gold standard”, the accuracy of any
volume measurement cannot be known, but
it is reasonable to assume that the three-
dimensional method is more accurate in
measuring tumor volume. The improved
accuracy of the three-dimensional method
reflects the ability to measure complex
shapes, and the three-dimensional segmen-
tation is done offline, providing more time
for the operator to evaluate the images. In
the two-dimensional method, the operator
must quickly decide which imaging plane
contains the maximum tumor diameter, a
decision that may change between observers
or even for the same observer over multiple
imaging sessions.
For a three-dimensional imaging modal-
ity to be useful for longitudinal studies
of tumor progression, the tumor volume
measurement must be accurate and con-
sistent. The variability in repeated volume
measurements of the same tumor will
dictate the minimum change in volume
that can be measured. We have measured
the intraobserver variability in repeated
volume measurements of murine B16F1
melanoma and human HT-29 colon car-
cinoma cells growing as liver metastases.
Intraobserver measurement variability is
frequently reported by the coefficient of
375
variation (COV), the standard deviation
divided by the mean volume measure-
ment, for each set of volume measure-
ments. For B16F1 liver metastases with
volumes less than 2 mm3, the COV was
6.8 ± 2.6% (Hastie et al., 2004). For
HT-29 liver metastases, with volumes
less than 2 mm3, the COV was 13 ± 5.7%
(Hastie et al., 2004). The larger measure-
ment variability in the HT-29 model, in
comparison to the B16F1 model, may
reflect the more complex geometries of
the HT-29 metastases and the superior
delineation of tumor boundaries in the
B16F1 model (Hastie et al., 2004).
Since a measured change in tumor vol-
ume is only significant if the change is
greater than the measurement variation,
knowledge of measurement variability and
tumor growth rate can be used to ensure
that imaging is only performed when
significant changes in tumor volume are
expected. In the case of B16F1 metastases,
with an average doubling time of 1.2 days
(Graham et al., 2005) and a measure-
ment variability of 6.8 ± 2.6%, signifi-
cant changes in tumor volume could be
detected with daily imaging sessions. The
HT-29 metastases show variable growth
rates, with doubling times between 3.7 and
4.8 days (Graham et al., 2005) and a meas-
urement variation of 13 ± 5.7%. In this
case, daily imaging may be unnecessary,
with more time necessary to ensure sig-
nificant changes in tumor volume between
imaging sessions.
In studies of tumor growth or treatment
response, it is advantageous to use an
imaging modality and volume measure-
ment method with the smallest possible
measurement variability so that small
changes in volume can be distinguished.
The reported coefficients of variation for
376
tumor volume measurement with high-
frequency ultrasound, between 6.8% and
13% depending on the tumor model, are
similar to the clinical results obtained for
the measurement of liver tumor volumes
with MRI. It has been reported that quan-
tification of tumor volumes from MRI by
manual segmentation had an intraobserver
COV of 9.3%, while volume measurement
by point counting and stereological cor-
rection had a COV of 8.4% (Mazonakis
et al., 2004). Studies of measurement vari-
ation among preclinical imaging modali-
ties are notably absent from the literature,
which is an issue that needs to be addressed
to ensure the proper use of these emerging
technologies.
HIGH-FREQUENCY
ULTRASOUND IMAGING
OF PRECLINICAL LIVER
METASTASES
Using experimental metastasis models, we
have shown that high-frequency ultrasound
can reliably detect murine liver metastases
(Graham et al., 2005). We have shown that
ultrasound detection of liver metastases
shows excellent agreement with post-mor-
tem analysis, including gross pathology
and histological sections (Figure 29.1).
Ultrasound detection of liver metastases
was demonstrated with four cell lines of
different primary tumor origin, includ-
ing B16F1 (murine melanoma), HT-29
(human colon carcinoma), MDA-MB-435/
HAL (human breast carcinoma), and PAP2
(murine oncogene-transformed fibrob-
lasts). With all four models, liver metas-
tases were hypoechoic, appearing darker
than the surrounding liver parenchyma,
K.C. Graham et al.
on the ultrasound images. The minimum
metastasis size that has been detected
by ultrasound imaging, and confirmed
through histological methods is (maxi-
mum diameter → volume) 0.22 mm →
0.01 mm3, 0.47 mm → 0.03 mm3, 0.66 mm
→ 0.08 mm3 and 0.78 mm → 0.17 mm3
for B16F1, HT-29, MDA-MB-435/HAL
and PAP2 tumor, respectively (Graham
et al., 2005). For the B16F1 model a
large number of mice have been imaged
throughout the early stages of metastasis
development, thus the reported detection
threshold may be at or near the true detec-
tion limit. For the other three models the
absolute detection limit may be less than
the reported values, because only a limited
number of animals were imaged over vary-
ing time frames.
The experimental detection limits of
high-frequency ultrasound compare favo-
rably to many of the other non-invasive
imaging modalities. It has been reported
that contrast-enhanced micro-CT and
contrast-enhanced micro-MRI may detect
liver metastases as small as 0.3 mm in
diameter (Kobayashi et al., 2001; Weber
et al., 2004). Non-invasive optical imaging
with green fluorescent protein may detect
extremely small tumors, but imaging sen-
sitivity is dependent on tissue depth. For
liver metastasis models, we identified the
majority of the liver metastases at the
2–3 mm depth range. It has been reported
that for green fluorescent protein (GFP)-
labeled tumors at a depth of 2.2 mm, the
minimum tumor size detected is 1.86 mm
(Yang et al., 2000).
It is important to note that unlike micro-
CT, micro-MRI or optical methods, the
imaging protocol to identify liver tumors
by high-frequency ultrasound does not
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
inherent tumor to liver parenchyma con-
trast seen in ultrasound imaging offers
significant advantages in terms of animal
throughput and in the number of animal
models able to be used. For ultrasound
imaging, each animal can be anesthetized,
prepared, imaged, the images reconstructed
and the content reviewed, and the animal
recovered in a time window of 15 min
per animal. The use of a contrast agent,
usually delivered intravenously, may be
technically difficult and time consum-
ing, which decreases animal throughput.
The clearance kinetics and toxicity of a
contrast agent may also limit longitudi-
nal imaging. In the case of genetically
encoded marker proteins, such as GFP, the
necessity of introducing a foreign reporter
gene into each cell line to be studied may
limit the number of animal models that
can be utilized. Additionally, the expres-
sion of a foreign reporter protein can lead
to increased immunogenicity of the tumor
cell line being studied (Steinbauer et al.,
2003).
The high-throughput, detection sensitivity
and the inherent tumor to liver parenchyma
contrast makes high-frequency ultrasound
uniquely suited for studying the longitudinal
378
growth of preclinical liver metastases.
We have demonstrated the longitudinal
tracking of individual metastases in the
B16F1, HT-29 and MDA-MB-435/HAL
models (Graham et al., 2005). Individual
metastases were not monitored in the
PAP2 model since in this highly aggres-
sive model, individual metastases grow
and fuse very rapidly. In this case, tumor
growth may be evaluated by assessing
total tumor burden at various time points.
It is reasonable to speculate that the tumor
burden seen in the percentage of the liver
visible for ultrasound imaging reflects
the status of the entire liver, but this has
not yet been shown. For studies in which
analysis of individual metastases is dif-
ficult or not required, it may be more
appropriate to use an imaging modality
that gives a rapid quantification of total
tumor burden, such as bioluminescence or
one that gives three-dimensional images
unimpeded by the ribcage, such as con-
trast-enhanced CT.
Although imaging modalities that give a
rapid assessment of tumor burden, such as
bioluminescence, are invaluable for many
high throughput studies, they are unable
to resolve individual metastases in close
proximity. The ability to image and study
individual liver metastases through high-
frequency ultrasound imaging is a power-
ful new tool that can be used to address
many unanswered questions regarding the
metastatic process. For example, through
longitudinal study of liver metastases we
have shown that individual metastases
within a single animal may have widely
varied growth rates. The most striking
example of this is with the HT-29 (colon
carcinoma) liver metastasis model. In this
model a single animal developed two liver
metastases that were proliferating with a
K.C. Graham et al.
volume doubling time between 4 and 5
days (Graham et al., 2005). A third liver
metastasis, imaged during the same 10-day
time period, did not appreciably increase in
volume, remaining at a constant 0.03 mm3
(Figure 29.2). The longitudinal study of
individual metastases allows the identifi-
cation of metastases with variable growth
patterns for further investigation.
From a clinical perspective, the identi-
fication and study of dormant metastases,
as defined by their constant volume, would
be of great interest. The mechanisms by
which metastatic growth is controlled to
produce dormant metastases could poten-
tially be harnessed to produce a clinically
relevant treatment to slow or halt meta-
static progression. Alternatively, the trig-
gers and mechanisms which lead a once
dormant metastasis to reinitiate progres-
sive growth are also of great clinical interest.
In some cases metastases have been
known to occur many years after initial
diagnosis, with tumor dormancy being
a likely explanation for the large time
interval (Demicheli et al., 1998; Karrison
et al., 1999). It has been proposed that
Figure 29.2. Tracking the growth of individual
liver metastases by non-invasive ultrasound imag-
ing. Two HT-29 liver metastases are rapidly grow-
ing, as a third metastasis, in the same animal,
remains at a constant volume. (Modified with
permission from Graham et al., 2005.)
29. Preclinical Liver Metastases: Three-Dimensional High-Frequency Ultrasound Imaging
tumor dormancy may be the result of a
dynamic balance between apoptosis and
proliferation in preangiogenic microme-
tastases (Holmgren et al., 1995; Udagawa
et al., 2002). In this theory, a micrometas-
tasis is able to escape dormancy and switch
to progressive growth by recruiting a func-
tional blood supply. Longitudinal imaging is
imperative in order to evaluate this theory in
the context of metastatic models.
Apart from the issue of tumor dormancy,
the study of treatment response and treatment
resistance would also benefit from non-inva-
sive imaging. The ability to analyze growth
rates of individual metastases, with image-
guided isolation of the tumors of interest,
would enable many studies on the molecular
basis of tumor response and tumor resistance,
including an assessment of heterogeneity
in responsiveness of individual metastases.
Although the biological question may vary,
the unifying theme in all of these studies
is that non-invasive longitudinal imaging,
capable of resolving individual micrometas-
tases, is imperative for preclinical study of
many clinically relevant issues.
OTHER APPLICATIONS
OF HIGH-FREQUENCY
ULTRASOUND IMAGING
TO PRECLINICAL CANCER
RESEARCH
The characteristics of high-frequency
ultrasound imaging – high spatial resolu-
tion, rapid image acquisition, economical
capital and maintenance costs, and good
soft tissue contrast without the requirement
for contrast agents – make ultrasound an
appealing modality for the study of many
preclinical cancer models. Ultrasound may
379
be used to image tumor development in
any anatomical site that is not obscured by
air-filled structures, such as lungs, or by
bony structures, such as the rib cage.
High-frequency ultrasound has been
used to identify and track the longitudi-
nal growth of prostate tumors in a trans-
genic model of prostate adenocarcinoma
(Wirtzfeld et al., 2005). Non-invasive
imaging is particularly useful for trans-
genic prostate tumor models because the
study of tumor progression or therapeutic
response in these models is hindered by
the wide variability in the timing of tumor
development. This variability in tumor
development necessitates repeated and
long-term imaging of the experimental
animals, which in turn makes ultrasound
imaging, with rapid image acquisition and
non-ionizing radiation, particularly suited
for the evaluation of these models.
Non-invasive imaging is the only method
that enables longitudinal study of tumor
growth in internal organs. This is not the
case in preclinical models that utilize
tumors grown subcutaneously, intrader-
mally, or in the mammary fat pad, for
which calipers may be used to acquire
longitudinal measurements. Although con-
venient, the ease of caliper measurement
must be balanced against the volume over-
estimation that arises from including the
skin and other tumor surrounding tissue
layers in the caliper measurement and
against the fact that many tumors do not
conform to the assumed ellipsoid shape
for volume calculation. For these reasons,
high-frequency ultrasound has been pro-
posed as a time efficient imaging modality
to increase the accuracy of these preclini-
cal studies (Cheung et al., 2005; Turnbull
et al., 1996). With detection of tumors as
small as 0.25 mm in width, ultrasound can
380
also detect and track tumor development
from a much earlier stage than caliper
measurement (Cheung et al., 2005). The
continued development of tumors past
a size of 1 mm is thought to rely on the
formation of a neovasculature through the
process of angiogenesis. Thus, the ability
to detect and track small tumors may be
useful for studying biological processes,
such as angiogenesis, that occur early in
tumor development.
Beyond the study of tumor growth,
angiogenesis research would benefit from
non-invasive methods of quantifying vas-
cular parameters. For determination of
red blood cell velocity, Doppler ultra-
sound may be used. Doppler ultrasound is
based on the principle that when incident
ultrasound waves hit a moving object the
frequency of the return signal is shifted in
proportion to the velocity of the object. In
color Doppler ultrasound, the frequency
shift in the return signal from moving red
blood cells is displayed by mapping the
frequency shifts to a blue-red color map
with positive frequency shifts being red
and negative frequency shifts being blue
in the resulting image. In an alternative
method of processing Doppler frequency
shift data, termed Power Doppler, the
magnitude of frequency shift, and hence
red blood cell velocity, is ignored. In this
method, all frequency shifts contribute
equally to the display data, meaning the
displayed value is related to the number
of red blood cells moving through a given
area. The utility of high-frequency power
Doppler ultrasound in the non-invasive
evaluation of anti-vascular therapies has
been demonstrated in an intradermal
melanoma model (Goertz et al., 2002). It
is likely that Power Doppler will be fur-
ther developed for high-frequency systems
K.C. Graham et al.
(Goertz et al., 2003), as this technique is
less sensitive to image artifacts than color
Doppler (Lencioni et al., 1996).
Many cancer therapies induce apoptosis,
an active response to cellular injury that
results in characteristic changes to the cell
nucleus and membrane, eventually leading
to cell death. The ability to non-invasively
quantify apoptosis would be useful in pre-
clinical studies to assess the efficacy of
novel cancer therapeutics. To date, quan-
tification of apoptosis is largely limited to
biochemical analysis on harvested tissues,
although single photon emission computed
tomography with radiolabeled annexin V
has shown promise for non-invasive in vivo
analysis in the clinical setting (Blankenberg
et al., 1998; Narula et al., 2001). For preclin-
ical research, the low resolution of SPECT
imaging may be limiting, so alternative
methods of quantifying apoptosis may be
required. In a series of in vitro experiments,
utilizing cell pellets of human acute myeloid
leukemia cells, it has been demonstrated
that the structural changes associated with
apoptosis may be detectable by analyzing
the radiofrequency backscatter signal in
high-frequency ultrasound (Kolios et al.,
2002; Tunis et al., 2005). It is not yet known
if the in vitro results will be translatable to
the more complex in vivo model, where the
presence of tissue structures other than the
tumor cells may affect the method of analysis
(Kolios et al., 2002).
FUTURE DEVELOPMENTS –
ULTRASOUND CONTRAST
AGENTS
The use of high-frequency ultrasound in
the identification of primary and metastatic
tumors, and in the longitudinal evaluation
29. Preclinical Liver Metastases: Three-Dimensional High-Frequency Ultrasound Imaging
of tumor progression, is well documented
(Graham et al., 2005; Wirtzfeld et al.,
2005). These studies utilized the inherent
ultrasound contrast, present between tumor
and normal tissue, to delineate the tumors.
As outlined in this review, the longitudinal
evaluation of tumor growth allows for the
study of a variety of processes, including
tumor dormancy and treatment response,
and allows the researcher greater con-
trol when using models that exhibit large
variability in the timing of tumor develop-
ment. Although the ability to study tumor
growth based on inherent ultrasound con-
trast is clearly a significant development
in its own right, the current development
and application of novel contrast agents is
poised to expand the utility of preclinical
ultrasound imaging in the fields of vas-
cular imaging, molecular imaging, and
targeted drug delivery.
Inherent ultrasound contrast can be
enhanced by adding an agent to the imag-
ing field of view that has an acoustic
impedance that differs from the tissue
of interest. Intravenously injected micro-
bubbles are commonly used to achieve
this objective. Microbubbles are typically
composed of a gaseous perfluorocarbon
core surrounded by an outer phospholipid
or albumin shell. The large difference in
acoustic impedance between the perfluor-
ocarbon gas and the surrounding tissue
leads to greatly enhanced ultrasound signal
reflection and the ability to detect small
numbers of microbubbles in the ultrasound
image. Microbubbles are commonly pro-
duced with diameters between 1 and 5 μm
to allow them to pass freely through the
circulatory system without being arrested
by size restriction. Recently, the prop-
erties of much smaller contrast agents,
liquid filled nanoparticles with diameters
381
between 250 and 1,000 nm, have also been
investigated (Couture et al., 2006; Lanza
et al., 1997).
Sensitivity of microbubble detection
may be further enhanced based on the
non-linear reflection of ultrasound by
microbubbles. This characteristic allows
the echoes produced from microbubbles
to be differentiated from those of normal
tissue using appropriate signal processing
techniques. On clinical ultrasound systems
this is often accomplished by transmitting
at one frequency and receiving signals
at integer multiples of the transmit fre-
quency, referred to as harmonic imaging.
The strength of a nonlinear echo at a given
frequency depends on the transmitted fre-
quency and on the size of the microbubble.
For preclinical high-frequency ultrasound,
with a 40 MHz transmitting frequency, a
microbubble will produce a much stronger
nonlinear echo at one-half of the trans-
mit frequency, the subharmonic, than at
the higher harmonic frequencies. For this
reason, subharmonic imaging may be
more sensitive than harmonic imaging
for detecting microbubble contrast agents
with preclinical imaging systems (Goertz
et al., 2005).
While Doppler ultrasound can provide
valuable information about tumor blood
flow, the tissue motion artifacts created
by the movement of internal organs cur-
rently limit the utility of Doppler sys-
tems. This limitation may be overcome
by the development of array transducers
for preclinical systems or through the use
of intravenously injected microbubbles.
Microbubbles enhance the signal from
blood in B-mode images, and hence are
less susceptible to artifacts from tissue
motion. In addition, continuous imaging
after the contrast agent is administered
382
can provide information about perfusion
dynamics, including blood volume and
flow rates, that is not available from con-
ventional Doppler imaging. Perfusion
dynamics have been used clinically to
differentiate between different types of
liver lesions (Albrecht et al., 2004), while
preclinically, where the lesion is purposely
produced, perfusion dynamics hold prom-
ise as a method of evaluating the efficacy
of novel anti-angiogenic therapeutics (Krix
et al., 2003).
After intravenous injection, microbub-
bles are eventually sequestered in the liver
by functional hepatocytes. This results in
enhancement of the normal liver tissue,
but not primary liver tumors or metas-
tases, thereby providing enhanced liver
parenchyma to liver tumor contrast. This
contrast enhancement may be useful for
the delineation of small or poorly defined
tumors (Albrecht et al., 2004).
Ultrasound contrast agents may extend
the utility of ultrasound from anatomical
imaging to targeted or molecular imag-
ing (Dayton and Ferrara, 2002; Lindner,
2004). As with all targeted contrast agents,
the first step is the identification of cell
surface receptors that are specific to the
tissue of interest. Once a unique receptor
is identified, targeted microbubbles may
be created by conjugation of the receptor-
specific ligand to the outer surface of the
microbubble. After intravenous injection
the microbubbles will circulate the body
and adhere to their target receptors or be
cleared by the liver. Once the majority of
free microbubbles are cleared from the
circulation, the presence of the cell recep-
tor can be imaged indirectly by detect-
ing the presence of the microbubbles.
This approach has been used to image
angiogenesis in rat gliomas by using the
K.C. Graham et al.
ligand echistatin to target microbubbles
to the neovasculature specific integrin
receptor αvβ3 (Ellegala et al., 2003), and
to image the vasculature in subcutane-
ous tumors with a cyclic peptide ligand,
arginine–arginine–leucine, which binds to
a yet unidentified receptor on angiogenic
endothelium (Ellegala et al., 2003; Weller
et al., 2005).
Microbubbles have the potential to be
used for both therapeutic and diagnostic
applications. As toxicity is a major prob-
lem limiting the dose of chemotherapeutic
agents, the ability to specifically target
chemotherapy delivery to the tumor, thereby
limiting systemic exposure, is a promising
avenue for research. Chemotherapeutic
compounds can be attached to the outer
membrane of the microbubble, imbedded
within the membrane, or loaded into the
interior of the microbubble, preventing
the activity of the compound when the
microbubble is in the systemic circula-
tion (Unger et al., 2002). Focused high-
pressure ultrasound waves can be used
to rupture the microbubbles, a process
called cavitation, and thereby release the
chemotherapeutic. There is emerging evi-
dence that microbubble cavitation can lead
to local damage of cellular membranes
and disrupt capillary integrity (Miller and
Quddus, 2000; Price et al., 1998). The cel-
lular damage that results from cavitation
may be detrimental for some applications
of contrast-enhanced ultrasound, but for
targeted cancer therapy this effect may
increase the interstitial concentration of
the chemotherapeutic within the tumor,
while sparing the normal tissue that is not
exposed to the ultrasound pulse (Unger
et al., 2002).
The principles that are being developed
for targeted chemotherapeutic delivery
29. Preclinical Liver Metastases: Three-Dimensional High-Frequency Ultrasound Imaging
may also be applied to targeted gene Anwer, K., Kao, G., Proctor, B., Anscombe, I.,
delivery. For gene therapy, the applica-
tion of ultrasound is known to increase
the transfection efficiency of traditional
cationic lipid based transfection complexes
(Anwer et al., 2000), although performing
ultrasound in the presence of microbubbles,
causing cavitation and disruption of cellular
membranes may be more effective than
ultrasound alone (Lawrie et al., 2000). In
both of these procedures the application of
ultrasound can be localized to a target site
to enhance therapeutic effect, while mini-
mizing systemic toxicities.
CONCLUSIONS
High-frequency ultrasound offers cost
effective, non-invasive, high-resolution
imaging with rapid image acquisition and
processing. These attributes make high-
frequency ultrasound an attractive option
for the preclinical study of primary and
metastatic cancer models. The longitudi-
nal study of primary tumor development
and metastatic progression may provide
further insight into dynamic processes
such as tumor dormancy and angiogen-
esis. With the development of ultrasound
contrast agents for diagnostic and thera-
peutic purposes, high-frequency ultra-
sound will be an invaluable tool in the
preclinical characterization and evalua-
tion of these agents, eventually leading to
the clinical application of these emerging
technologies.
REFERENCES
Albrecht, T., Hohmann, J., Oldenburg, A., Skrok,
J., and Wolf, K.J. 2004. Detection and charac-
terisation of liver metastases. Eur. Radiol. 14
Suppl. 8:P25–P33.
383
Florack, V., Earls, R., Wilson, E., McCreery,
T., Unger, E., Rolland, A., and Sullivan, S.M.
2000. Ultrasound enhancement of cationic
lipid-mediated gene transfer to primary tumors
following systemic administration. Gene Ther.
7:1833–1839.
Blankenberg, F.G., Katsikis, P.D., Tait, J.F., Davis,
R.E., Naumovski, L., Ohtsuki, K., Kopiwoda,
S., Abrams, M.J., Darkes, M., Robbins, R.C.,
Maecker, H.T., and Strauss, H.W. 1998. In vivo
detection and imaging of phosphatidylserine
expression during programmed cell death. Proc.
Natl. Acad. Sci. U S A 95:6349–6354.
Cai, S.R., Garbow, J.R., Culverhouse, R., Church,
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.
Chambers, A.F., MacDonald, I.C., Schmidt, E.E.,
Koop, S., Morris, V.L., Khokha, R., and Groom,
A.C. 1995. Steps in tumor metastasis: new con-
cepts from intravital videomicroscopy. Cancer
Metastasis Rev. 14:279–301.
Chambers, A.F., Groom, A.C., and MacDonald,
I.C. 2002. Dissemination and growth of can-
cer cells in metastatic sites. Nat. Rev. Cancer
2:563–572.
Cheung, A.M., Brown, A.S., Hastie, L.A., Cucevic,
V., Roy, M., Lacefield, J.C., Fenster, A., and
Foster, F.S. 2005. Three-dimensional ultrasound
biomicroscopy for xenograft growth analysis.
Ultrasound Med. Biol. 31:865–870.
Choy, G., Choyke, P., and Libutti, S.K. 2003.
Current advances in molecular imaging: non-
invasive in vivo bioluminescent and fluores-
cent optical imaging in cancer research. Mol.
Imaging 2:303–312.
Couture, O., Bevan, P.D., Cherin, E., Cheung, K.,
Burns, P.N., and Foster, F.S. 2006. Investigating
perfluorohexane particles with high-frequency
ultrasound. Ultrasound Med. Biol. 32:73–82.
Dayton, P.A., and Ferrara, K.W. 2002. Targeted
imaging using ultrasound. J. Magn. Reson.
Imaging 16:362–377.
Demicheli, R., Terenziani, M., and Bonadonna,
G. 1998. Estimate of tumor growth time for
breast cancer local recurrences: rapid growth
after wake-up? Breast Cancer Res. Treat. 51:
133–137.
384
Ellegala, D.B., Leong-Poi, H., Carpenter, J.E.,
Klibanov, A.L., Kaul, S., Shaffrey, M.E., Sklenar,
J., and Lindner, J.R. 2003. Imaging tumor ang-
iogenesis with contrast ultrasound and micro-
bubbles targeted to alpha(v)beta3. Circulation
108:336–341.
Evelhoch, J.L., Gillies, R.J., Karczmar, G.S.,
Koutcher, J.A., Maxwell, R.J., Nalcioglu, O.,
Raghunand, N., Ronen, S.M., Ross, B.D., and
Swartz, H.M. 2000. Applications of magnetic
resonance in model systems: cancer therapeu-
tics. Neoplasia 2:152–165.
Fenster, A., Downey, D.B., and Cardinal, H.N.
2001. Three-dimensional ultrasound imaging.
Phys. Med. Biol. 46:R67–R99.
Foster, F.S., Pavlin, C.J., Harasiewicz, K.A.,
Christopher, D.A., and Turnbull, D.H. 2000.
Advances in ultrasound biomicroscopy.
Ultrasound Med. Biol. 26:1–27.
Foster, F.S., Zhang, M.Y., Zhou, Y.Q., Liu, G., Mehi,
J., Cherin, E., Harasiewicz, K.A., Starkoski,
B.G., Zan, L., Knapik, D.A., and Adamson, S.L.
2002. A new ultrasound instrument for in vivo
microimaging of mice. Ultrasound Med. Biol.
28:1165–1172.
Gambhir, S.S. 2002. Molecular imaging of cancer
with positron emission tomography. Nat. Rev.
Cancer 2: 683–693.
Gillies, R.J., Bhujwalla, Z.M., Evelhoch, J.,
Garwood, M., Neeman, M., Robinson, S.P.,
Sotak, C.H., and Van Der Sanden, B. 2000.
Applications of magnetic resonance in model sys-
tems: tumor biology and physiology. Neoplasia
2:139–151.
Goertz, D.E., Yu, J.L., Kerbel, R.S., Burns, P.N.,
and Foster, F.S. 2002. High-frequency Doppler
ultrasound monitors the effects of antivascu-
lar therapy on tumor blood flow. Cancer Res.
62:6371–6375.
Goertz, D.E., Yu, J.L., Kerbel, R.S., Burns, P.N.,
and Foster, F.S. 2003. High-frequency 3-D
color-flow imaging of the microcirculation.
Ultrasound Med. Biol. 29:39–51.
Goertz, D.E., Cherin, E., Needles, A., Karshafian,
R., Brown, A.S., Burns, P.N., and Foster, F.S.
2005. High frequency nonlinear B-scan imag-
ing of microbubble contrast agents. IEEE Trans.
Ultrason. Ferroelectr. Freq. Control 52:65–79.
Graham, K.C., Wirtzfeld, L.A., MacKenzie, L.T.,
Postenka, C.O., Groom, A.C., MacDonald, I.C.,
K.C. Graham et al.
Fenster, A., Lacefield, J.C., and Chambers, A.F.
2005. Three-dimensional high-frequency ultra-
sound imaging for longitudinal evaluation of
liver metastases in preclinical models. Cancer
Res. 65:5231–5237.
Hastie, L.A., Graham, K.C., Groom, A.C.,
MacDonald, I.C., Chambers, A.F., Fenster, A.,
and Lacefield, J.C. 2004. Variability of three-
dimensional high-frequency ultrasound measure-
ments of small tumor volumes. IEEE Ultrason.
Sympos. Proc. 3:2185–2188.
Heyn, C., Ronald, J.A., Mackenzie, L.T.,
MacDonald, I.C., Chambers, A.F., Rutt, B.K.,
and Foster, P.J. 2006. In vivo magnetic resonance
imaging of single cells in mouse brain with opti-
cal validation. Magn. Reson. Med. 55:23–29.
Holmgren, L., O’Reilly, M.S., and Folkman,
J. 1995. Dormancy of micrometastases: bal-
anced proliferation and apoptosis in the pres-
ence of angiogenesis suppression. Nat. Med. 1:
149–153.
Karrison, T.G., Ferguson, D.J., and Meier, P. 1999.
Dormancy of mammary carcinoma after mastec-
tomy. J. Natl. Cancer Inst. 91:80–85.
Khanna, C., and Hunter, K. 2005. Modeling metas-
tasis in vivo. Carcinogenesis 26:513–523.
Kobayashi, H., Saga, T., Kawamoto, S., Sato, N.,
Hiraga, A., Ishimori, T., Konishi, J., Togashi,
K., and Brechbiel, M.W. 2001. Dynamic micro-
magnetic resonance imaging of liver micrometas-
tasis in mice with a novel liver macromolecular
magnetic resonance contrast agent DAB-Am64-
(1B4M-Gd)(64). Cancer Res. 61:4966–4970.
Kolios, M.C., Czarnota, G.J., Lee, M., Hunt,
J.W., and Sherar, M.D. 2002. Ultrasonic spec-
tral parameter characterization of apoptosis.
Ultrasound Med. Biol. 28:589–597.
Krix, M., Kiessling, F., Vosseler, S., Farhan, N.,
Mueller, M.M., Bohlen, P., Fusenig, N.E.,
and Delorme, S. 2003. Sensitive noninvasive
monitoring of tumor perfusion during antiang-
iogenic therapy by intermittent bolus-contrast
power Doppler sonography. Cancer Res. 63:
8264–8270.
Lanza, G.M., Wallace, K.D., Fischer, S.E., Christy,
D.H., Scott, M.J., Trousil, R.L., Cacheris, W.P.,
Miller, J.G., Gaffney, P.J., and Wickline, S.A.
1997. High-frequency ultrasonic detection
of thrombi with a targeted contrast system.
Ultrasound Med. Biol. 23:863–870.
29. Preclinical Liver Metastases: Three-Dimensional High-Frequency Ultrasound Imaging
Lawrie, A., Brisken, A.F., Francis, S.E., Cumberland,
D.C., Crossman, D.C., and Newman, C.M. 2000.
Microbubble-enhanced ultrasound for vascular
gene delivery. Gene Ther. 7:2023–2027.
Leach, M.O. 2001. Application of magnetic reso-
nance imaging to angiogenesis in breast cancer.
Breast Cancer Res. 3:22–27.
Lencioni, R., Pinto, F., Armillotta, N., and
Bartolozzi, C. 1996. Assessment of tumor vascu-
larity in hepatocellular carcinoma: comparison
of power Doppler US and color Doppler US.
Radiology 201:353–358.
Lindner, J.R. 2004. Microbubbles in medical imag-
ing: current applications and future directions.
Nat. Rev. Drug Discov. 3:527–532.
Lyons, S.K. 2005. Advances in imaging mouse
tumour models in vivo. J. Pathol. 205:194–205.
Mazonakis, M., Damilakis, J., Mantatzis, M.,
Prassopoulos, P., Maris, T., Varveris, H., and
Gourtsoyiannis, N. 2004. Stereology versus
planimetry to estimate the volume of malignant
liver lesions on MR imaging. Magn. Reson.
Imaging 22:1011–1016.
Miller, D.L., and Quddus, J. 2000. Diagnostic
ultrasound activation of contrast agent gas bod-
ies induces capillary rupture in mice. Proc. Natl.
Acad. Sci. U S A 97:10179–10184.
Morris, V.L., MacDonald, I.C., Koop, S., Schmidt,
E.E., Chambers, A.F., and Groom, A.C. 1993.
Early interactions of cancer cells with the
microvasculature in mouse liver and mus-
cle during hematogenous metastasis: videomi-
croscopic analysis. Clin. Exp. Metastasis 11:
377–390.
Narula, J., Acio, E.R., Narula, N., Samuels,
L.E., Fyfe, B., Wood, D., Fitzpatrick, J.M.,
Raghunath, P.N., Tomaszewski, J.E., Kelly, C.,
Steinmetz, N., Green, A., Tait, J.F., Leppo,
J., Blankenberg, F.G., Jain, D., and Strauss,
H.W. 2001. Annexin-V imaging for noninvasive
detection of cardiac allograft rejection. Nat.
Med. 7:1347–1352.
Paulus, M.J., Gleason, S.S., Kennel, S.J., Hunsicker,
P.R., and Johnson, D.K. 2000. High resolution
X-ray computed tomography: an emerging tool
for small animal cancer research. Neoplasia
2:62–70.
Price, R.J., Skyba, D.M., Kaul, S., and Skalak, T.C.
1998. Delivery of colloidal particles and red
blood cells to tissue through microvessel ruptures
385
created by targeted microbubble destruction with
ultrasound. Circulation 98:1264–1267.
Steinbauer, M., Guba, M., Cernaianu, G., Kohl,
G., Cetto, M., Kunz-Schughart, L.A., Geissler,
E.K., Falk, W., and Jauch, K.W. 2003. GFP-
transfected tumor cells are useful in examining
early metastasis in vivo, but immune reaction
precludes long-term tumor development studies
in immunocompetent mice. Clin. Exp. Metastasis
20:135–141.
Tong, S., Cardinal, H.N., McLoughlin, R.F.,
Downey, D.B., and Fenster, A. 1998. Intra- and
inter-observer variability and reliability of pros-
tate volume measurement via two-dimensional
and three-dimensional ultrasound imaging.
Ultrasound Med. Biol. 24:673–681.
Tunis, A.S., Czarnota, G.J., Giles, A., Sherar, M.D.,
Hunt, J.W., and Kolios, M.C. 2005. Monitoring
structural changes in cells with high-frequency
ultrasound signal statistics. Ultrasound Med.
Biol. 31:1041–1049.
Turnbull, D.H., Ramsay, J.A., Shivji, G.S.,
Bloomfield, T.S., From, L., Sauder, D.N., and
Foster, F.S. 1996. Ultrasound backscatter micro-
scope analysis of mouse melanoma progression.
Ultrasound Med. Biol. 22:845–853.
Udagawa, T., Fernandez, A., Achilles, E.G.,
Folkman, J., and D’Amato, R.J. 2002. Persistence
of microscopic human cancers in mice: altera-
tions in the angiogenic balance accompanies loss
of tumor dormancy. FASEB J. 16:1361–1370.
Unger, E.C., Matsunaga, T.O., McCreery, T.,
Schumann, P., Sweitzer, R., and Quigley, R.
2002. Therapeutic applications of microbubbles.
Eur. J. Radiol. 42:160–168.
Weber, S.M., Peterson, K.A., Durkee, B., Qi,
C., Longino, M., Warner, T., Lee, F.T., Jr.,
and Weichert, J.P. 2004. Imaging of murine
liver tumor using microCT with a hepatocyte-
selective contrast agent: accuracy is dependent
on adequate contrast enhancement. J. Surg. Res.
119:41–45.
Weissleder, R. 2002. Scaling down imaging:
molecular mapping of cancer in mice. Nat. Rev.
Cancer 2:11–18.
Welch, D.R. 1997. Technical considerations for
studying cancer metastasis in vivo. Clin. Exp.
Metastasis 15:272–306.
Weller, G.E., Wong, M.K., Modzelewski, R.A.,
Lu, E., Klibanov, A.L., Wagner, W.R., and
386
Villanueva, F.S. 2005. Ultrasonic imaging of
tumor angiogenesis using contrast microbubbles
targeted via the tumor-binding peptide arginine-
arginine-leucine. Cancer Res. 65:533–539.
Wirtzfeld, L.A., Wu, G., Bygrave, M., Yamasaki,
Y., Sakai, H., Moussa, M., Izawa, J.I., Downey,
D.B., Greenberg, N.M., Fenster, A., Xuan,
J.W., and Lacefield, J.C. 2005. A new three-
dimensional ultrasound microimaging technol-
ogy for preclinical studies using a transgenic
prostate cancer mouse model. Cancer Res. 65:
6337–6345.
Wu, M., Mazurchuk, R., Chaudhary, N.D., Spernyak,
J., Veith, J., Pera, P., Greco, W., Hoffman, R.M.,
Kobayashi, T., and Bernacki, R.J. 2003. High-
resolution magnetic resonance imaging of the
K.C. Graham et al.
efficacy of the cytosine analogue 1-[2-C-cyano-
2-deoxy-beta-D-arabino-pentofuranosyl]-N(4)-
palmitoyl cytosine (CS-682) in a liver-metastasis
athymic nude mouse model. Cancer Res.
63:2477–2482.
Xu, H.X., Yin, X.Y., Lu, M.D., Liu, G.J., and Xu,
Z.F. 2003. Estimation of liver tumor volume
using a three-dimensional ultrasound volumetric
system. Ultrasound Med. Biol. 29:839–846.
Yang, M., Baranov, E., Jiang, P., Sun, F.X., Li, X.M.,
Li, L., Hasegawa, S., Bouvet, M., Al-Tuwaijri,
M., Chishima, T., Shimada, H., Moossa, A.R.,
Penman, S., and Hoffman, R.M. 2000. Whole-
body optical imaging of green fluorescent pro-
tein-expressing tumors and metastases. Proc.
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
INTRODUCTION
During the last years, 18F-fluorodeoxyglucose-
positron emission tomography (FDG-PET)
has been advocated as a useful tool in the
management of recurrent colorectal can-
cer, mainly in liver and pelvic recurrences,
whereas regarding the primary tumor,
FDG-PET has proven to show very poor
yield in comparison with conventional
diagnostic methods. In the first meta-
analysis of the literature for whole-body
FDG-PET detection of recurrent colorectal
cancer, FDG-PET had a positive impact
on surgical decision-making in 29% of
patients (Huebner et al., 2000). On the
other hand, in the beginning of FDG-PET,
most studies on colorectal liver metastases
have shown limitations. Survey of previ-
ous reports shows indeed heterogeneous
study designs that were mostly retrospec-
tive and simultaneously included various
subsets of circumstances, e.g., preopera-
tive staging of liver metastases together
with locoregional recurrence, unexplained
elevated serum carcinoembryonic antigen
(CEA) levels and/or symptoms with nega-
tive imaging. The methodology to define
the true lesion status of the FDG-PET
findings was also variable from one study
to another, including histopathologic
analy-sis (performed at surgery, biopsy,
and autopsy), intraoperative observation
(e.g., manual palpation or ultrasonogra-
phy) and/or 6-months or 1-year follow-
up. The rate of histological confirmation
that is the gold standard particularly in
the evaluation of false-positive and false-
negative readings varied from < 10% to
95% (Truant et al., 2005). A non-standard
method of image analysis, as well as
some degree of inequivalence in FDG-
PET and computed tomography (CT) per-
formance, may have led to a misjudgement
of their relative diagnostic performance,
as reflected by high rates of FDG-PET/
CT discordances (range, 10–31%). FDG-
PET was performed prospectively and the
interpretation of the FDG-PET images
was generally made with a priori knowl-
edge of results of conventional diagnostic
procedures. In contrast, CT was reviewed
retrospectively and performed at various
institutions, resulting in its sensitivity and
specificity rates being among the worst,
which were reported in general literature
regarding colorectal liver metastases. In
the two meta-analyses performed in the
387
388
beginning of the use of FDG-PET, only
few studies fulfilled the inclusion cri-
teria and the guidelines established by
the authors, reflecting the limitations of
published reports (Huebner et al., 2000;
Kinkel et al., 2002). Over time, however,
studies on FDG-PET have focused on
more definite settings and have included
more systematically histological verifica-
tions of FDG-PET findings, resulting in a
better assessment of FDG-PET accuracy
and impact. In this chapter, the yield of
FDG-PET in colorectal liver metastases
will be preferentially analyzed through
these latter studies.
PRINCIPLE OF FDG-PET
AND GENERAL PITFALLS
FDG-PET is based on the increased uptake
of glucose and, therefore, of FDG in
cancer cells. Many factors are involved
in FDG accumulation within a tumor or
metastases, such as the number of viable
tumor cells within a lesion, the amount of
necrotic tissue as well as regional blood
flow, and tumor hypoxia has proven to
increase FDG uptake. Besides, FDG is
not a specific tracer of malignant cells.
The uptake of this tracer may occur in
any condition associated with increased
tissue metabolism, such as inflammatory
cells – as those observed in nonneoplas-
tic stroma tissues within tumor tissue
– infection or healing. There also exist
many sources of physiologic increased
activity at FDG-PET, such as uptake in
brain, muscle, thyroid, digestive or urinary
tract. This lack of FDG specificity can
lead to its false-positive uptake in cancer
patients. Moreover, conditions that alter
S. Truant et al.
host cell and tumor glucose metabolism
may alter appearance of tumor on FDG-
PET. Patients with diabetes mellitus may
have an increased accumulation of FDG in
skeletal muscle related to insulin adminis-
tration and a subsequent decreased uptake
in malignant tissue, affecting both sensi-
tivity and specificity of FDG-PET imaging
in this patient population. In overweight
patients, the significant amount of soft tis-
sue can cause excessive scattering, and it
is recommended to increase the injected
activity and/or the acquisition time for
both emission and transmission scans
to improve the image quality. Systemic
therapy such as chemotherapy may also
deteriorate FDG-PET accuracy by signifi-
cantly decreasing the activity of the key
glycolytic enzyme hexokinase (Akhurst
et al., 2005). It is, therefore, recommended
to delay FDG-PET scanning for 2–3 weeks
after achieving chemotherapy. Mucinous
adenocarcinoma and thus mucinous-
differentiated metastases use 18F-FDG at
the same rate as normal surrounding tis-
sues owing to the relative abundance of
mucin and the subsequent relative hypocel-
lularity of such tumors. The sensitivity of
FDG-PET for detection of mucinous ade-
nocarcinoma is significantly lower than
the nonmucinous ones (Whiteford et al.,
2000). Finally, the risk of FDG-PET mis-
interpretation may be aggravated by the
lack of anatomic information. It has been
shown that the recently-developed hybrid
PET/CT machines, by better localizing
FDG uptakes, were able to reduce the rate
of inaccurate readings of FDG-PET alone.
The knowledge of these pitfalls empha-
sizes the importance of correlating FDG
uptake with clinical findings and results of
conventional imaging studies.
30. Colorectal Liver Metastases 389

COLORECTAL LIVER hepatic disease can be achieved. If isolated

METASTASES: YIELD hepatic metastases are suspected, patients
OF 18F-FLUORO- are screened with a CT of the chest and
abdomen to assess resectability and to iden-
DEOXYGLUCOSE-POSITRON tify any extrahepatic disease. Despite this
EMISSION TOMOGRAPHY extensive work-up, 20% patients prove to
have unresectable disease during laparot-
Approximately 50% of patients with color-
omy. Moreover, up to 60% of patients have
ectal cancer eventually develop liver metas-
cancer recurrence within 1 year of curative
tases. In 25% of cases liver metastases are
liver resection, with an equal distribution
synchronous and a further 25% will develop
between liver and extra-hepatic recurrences.
metachronously within 5 years. In 30–40%
This indicates the presence of hepatic and
of those patients, the metastases are still
extrahepatic tumor foci that were not identi-
confined to the liver at the time of detec-
fied by conventional diagnostic methods or
tion, but will be resectable in only 15–20%
at the time of liver surgery. In this setting, it
of patients, with partial hepatectomy being
was hoped that the application of FDG-PET
the only chance for cure. Postoperative
would lead to improved selection of patients
follow-up in association with CEA is man-
for partial hepatic resection by optimizing
datory afterwards to detect recurrences at
the detection of unsuspected metastases
an early and potentially curable stage. In
throughout the body. Contribution of FDG-
patients with nonresectable liver metas-
PET in the preoperative staging of colorectal
tases, systemic chemotherapy with irinote-
liver metastases is successively detailed for
can and/or oxaliplatin regimen, although
the liver, the abdominal cavity and the extra-
never curative, offers a median overall
abdominal organs, as well as specific pitfalls
survival of 2 years in responders or even
encountered in each site. Subsequent impact
more in secondarily-resectable patients. All
on patient’s management is then correlated
of these treatments can be associated with
with survival analysis.
a curative or palliative intent with local
ablative therapy such as radiofrequency
ablation and cryotherapy. The yield of Detection of Liver Metastases
FDG-PET in the management of colorectal
Interesting results have been obtained for
liver metastases has been assessed in these
the detection of metastases in the liver, with
different settings.
sensitivity and specificity of FDG-PET
ranging from 65% to 100% and from 57%
Preoperative Staging of Patients with to 100%, respectively. In the beginning of
Colorectal Liver Metastases FDG-PET use, the first meta-analysis of
A 35–40% 5-year survival rate can be 11 articles and 577 patients with recur-
anticipated after partial hepatectomy in rent colorectal cancer showed an average
patients with colorectal liver metastases, lesion-based sensitivity and specificity of
providing that a complete resection of FDG-PET for detection of liver metastases
liver disease with tumor-free margins and of 91% and 97%, respectively (Huebner
with absence of residual or metastatic extra- et al., 2000). Studies further reported a
390
higher accuracy for FDG-PET compared
with CT or other conventional diagnostic
methods in detecting liver metastases in
recurrent colorectal cancer (Vitola et al.,
1996a, b). In a second meta-analysis, the
mean sensitivity of FDG-PET (90%) was
significantly higher than that of ultrasound
(55%), CT (72%), and MRI (76%) in the
detection of liver metastases from color-
ectal, gastric, and esophageal cancer, but
any conclusion was difficult to draw due
to comparison of 15-year range of publi-
cation for CT with 3-year range for FDG-
PET (Kinkel et al., 2002). The sensitivity
of FDG-PET was comparable to that of the
most sensitive test for detection of hepatic
metastases, i.e., CT portography (90% vs.
97%), with a much higher specificity –
expected by the high rate of false-positive
findings of CT portography (100% vs. 9%)
(Vitola et al., 1996a, b). The adjunctive
of intraoperative sonography also limited
impact on treatment selection in patients
screened preoperatively by FDG-PET.
Over time however, with improve-
ment in study designs, the superiority
of FDG-PET over three phase CT or
contrast-enhanced MRI in the detection
of colorectal liver metastases was not
confirmed (Truant et al., 2005). This
was particularly true in studies using
a systematic histological verification of
PET/CT discordances and an analysis
by lesion that appeared to be more
accurate than an analysis by patient. In
lesion-by-lesion analyses of resected liver
specimens, FDG-PET demonstrated only
70–79% of liver metastases, namely not
better than conventional diagnostic meth-
ods (Truant et al., 2005). Rohren et al.
(2002), in comparing FDG-PET with sur-
gical inspection and intraoperative ultra-
sound, found that only 79% of hepatic
S. Truant et al.
metastatic lesions were identified by the
former. Accordingly, another meta-analy-
sis came to the conclusion that FDG-PET
had a significantly higher sensitivity for
detection of colorectal liver metastases on
a per-patient basis compared with that of
other modalities such as CT or MRI, but
showed comparable sensitivity on a per-
lesion basis (Bipat et al., 2005).
This poor sensitivity of the liver is partly
a result of the relatively high hexokinase
activity and thereby high FDG uptake
of normal hepatocytes. The liver usually
displays a faint and homogeneous FDG
uptake that is worsened by respiratory
motion leading to FDG smearing. Large
metastases are generally readily identified,
appearing as discrete foci of increased
activity in the liver or rings of increased
activity in the case of necrotic centers.
In contrast, small lesions < 10 mm can be
missed by FDG-PET due to its limited res-
olution. The sensitivity of FDG-PET in the
detection of liver metastases was confirmed
to be related to lesion size: only 14–25%
of hepatic lesions < 1 cm were detected
by FDG-PET in contrast with 80–100%
of lesions > 1 cm, resulting in an overall
rate of undetected lesions of 30% (Fong
et al., 1999). Preoperative chemotherapy
has proven to significantly alter tumor FDG
uptake, resulting in rates of undetected
lesions as high as 37% and possible missed
lesions > 3 cm (Akhurst et al., 2005).
FDG-PET is also unable to detect muci-
nous-differentiated metastases owing to
the relative hypocellularity of such tumors.
Another major drawback of FDG-PET
regarding liver metastases detection is the
lack of anatomic details, such as anatomi-
cal landmarks and relationship to hepatic
vasculature. This method is often unhelp-
ful in evaluating anatomical resectability of
30. Colorectal Liver Metastases
liver metastases. Inaccurate localization of
hepatic or extra-hepatic lesions positioned
in close proximity to the liver edge may
further lead to false-positive or false nega-
tive results, e.g., FDG-PET is unable to
discriminate between lesions close to each
other, an overlapping loop of bowel can
occasionally resemble a focal liver lesion.
On the other hand, false-positive findings
of FDG-PET in the liver may rarely occur,
for liver lesions that were found in the
published series to represent granuloma,
liver cyst, adenoma, fibronodular hyper-
plasia or even normal hepatic parenchyma.
For the latter circumstance, multiple fac-
tors have been involved, such as focal fat
infiltration, regionally altered blood flow,
intrinsic heterogeneity of hepatic paren-
chyma or image artifact. These benign
hepatic lesions have generally poor FDG
uptake, except for abscesses. The combi-
nation of the FDG-PET system with a CT
(integrated PET/CT system) was assumed
to avoid these limitations by better local-
izing FDG-PET abnormalities. Actually,
in 76 patients evaluated for resection of
liver metastases, PET/CT was equiva-
lent to contrast-enhanced CT for detect-
ing intrahepatic metastases, as expected,
as CT imaging without intravenous con-
trast does not allow visualization of many
hepatic metastases (Selzner et al., 2004).
However, PET/CT was superior in estab-
lishing the diagnosis of intrahepatic recur-
rences in patients with prior hepatectomy
(specificity 50% vs. 100%). In conclusion,
although the ability of FDG-PET to detect
hepatic lesions has been established, its
superiority to CT or MRI has not been
shown. Due to its limited anatomic reso-
lution, FDG-PET is unlikely to be a sub-
stitute for conventional imaging but may
rather complement it.
391
Detection of Extrahepatic Metastases
The preoperative work-up of presumed
resectable liver metastases generally
includes CT of the abdomen and the chest,
associated in case of symptoms with guided
investigations such as bone scintigraphy,
brain CT, transrectal endoscopic ultra-
sonography and/or pelvic MRI. Besides
a high accuracy in the detection of liver
and lung metastases, CT usually fails to
detect abdominal extrahepatic metastases
that may be a contraindication to hepate-
ctomy and are the main cause of unneces-
sary laparotomy in this patient population.
Another important issue is the diagnos-
tic of locoregional recurrence in patients
undergoing anterior or abdominoperineal
resection who develop a fibrotic mass in
the presacral operative bed aggravated
by preoperative radiation therapy. These
changes may appear on CT images for
many years and be indistinguishable from
tumor recurrence. With conventional imag-
ing, serial examinations are frequently
required, and the diagnosis of disease
recurrence may take 3–6 months or more
before slowly developing changes become
apparent.
Regarding these extrahepatic sites of
recurrence, FDG-PET has the advantage
over anatomic imaging by providing a
functional screening of the whole-body
and by detecting metabolic changes irre-
spective of the morphological changes.
In the preoperative staging of colorectal
liver metastases, FDG-PET accuracy for
the detection of extrahepatic recurrences
was difficult to assess. This was due to the
variability of lesion incidence among the
studied population, of lesion(s) size and
localization and of methodology to define
the true lesion status from one study to
another. Nevertheless, it is now established
392
that FDG-PET is a valuable tool for the
detection of extrahepatic metastases.
Abdominal Cavity
In overall abdomen, FDG-PET has shown
high rates of sensitivity (63–100%), specifi-
city (98–100%), and accuracy (93–100%)
(Flamen et al., 1999; Truant et al., 2005).
FDG-PET was more sensitive than CT
(sensitivity 25–57%), while specificity
rates were comparable (CT specificity of
91–99%). These results were conditioned
by the prior knowledge of the specific
pitfalls and artifacts encountered in the
abdominal cavity owing to the many sources
of physiologic and benign increased activ-
ity at FDG-PET, such as uptake in urinary
and gastrointestinal tract.
Physiologic and Benign Increased Activity of
FDG in the Abdominal Cavity
Increased activity can be seen in most
segments of the digestive tract. The
mechanism underlying this phenomenon
is unclear, partly due to smooth muscle
activity associated with peristalsis, meta-
bolically active mucosa, and bacterial
uptake. Gastric activity can be seen and
has a characteristic J-shape. Uptake along
the esophagus is also common, especially
in the distal portion and at the gastro-
esophageal junction; the esophagus is best
identified on frontal and sagittal views. A
physiologic uptake can also occur in the
small bowel and is usually identified by
its pattern that has often a linear “string of
beads” appearance. However, more limited
bowel uptake, which forms discrete foci
of increased activity, can be impossible
to distinguish from tumor deposits. The
colonic activity is also variable. Uptake
in the caecum and right colon is usually
higher than in the rest of the colon because
S. Truant et al.
of the abundance of the lymphocyte cells
in the region, which are very glucose avid.
Benign causes of FDG accumulation also
exist, such as fistulas, abscesses, diver-
ticulitis, inflammatory bowel disease, and
adenoma. As a rule, diffuse colonic FDG
uptake pattern, regardless of degree, is
predictive of normal findings at colonos-
copy, whereas a segmental high uptake
may imply an inflammatory condition.
In contrast, focal intense activity in the
colon is predictive of pathologic find-
ings and should be further evaluated with
colonoscopy to exclude a neoplastic proc-
ess. Moderate focal FDG uptake can also
be seen at stomas sites or in laparotomy
scars. Such increased FDG uptake associ-
ated with healing may persist for several
months. Because FDG is excreted in the
urine, intense FDG activity is encountered
in the intrarenal collecting systems, ure-
ters, and bladder. Other benign uptakes
are possible, such as increased activity of
FDG in the uterus during menstruation,
inflammatory changes within the ovary
during ovulation, hyperventilation-induced
uptake in the diaphragm, or aortic FDG
accumulation due to atherosclerosis. All of
these sites have potential for tumor recur-
rence, and in the case of abdominal 18F-
FDG uptake, correlation with clinical and
CT findings, as well as consideration for
biopsy, are mandatory. In case of inflam-
matory or infectious processes, repeated
FDG-PET, in two to four months or after
a course of appropriate therapy (e.g., anti-
biotics), will often show an absence of or
substantial decrease in uptake intensity at
the site. The use of laparoscopy to assess
intra-abdominal burden before laparotomy
should also be recommended but is not
always possible, given existing adhesions
and the extent of previous surgery.
30. Colorectal Liver Metastases
Contribution of FDG-PET
In the abdominal cavity, FDG-PET has
proven to be especially useful in areas in
which CT has failed to demonstrate high
accuracy due to limited anatomic resolu-
tions, such as locoregional recurrence of
rectal cancer, peritoneal metastases or dis-
tant lymph nodes metastases.
Pelvic Recurrence of Rectal Cancer
In spite of radiation therapy and total
mesorectal excision, locoregional recur-
rence occurs in 10–30% of patients and
is asymptomatic in up to 40% of patients.
Locoregional recurrences are a contraindi-
cation to surgical resection of liver metas-
tases except when resectable, although
attempted curative resection often causes
extreme morbidity with a low chance of
long-term survival. In contrast to CT that
has low specificity, FDG-PET has proven
to be particularly valuable in distinguish-
ing between local recurrence and scar tis-
sue after rectal excision (Figure 30.1).
Most series in recurrent cancer reported
high rates of sensitivity and specificity of
FDG-PET for the diagnosis of locoregional
recurrence, ranging from 84% to 100%
and from 80% to 100%, respectively
(Moore et al., 2003). The sensitivity of
FDG-PET in detecting locoregional recur-
rence compared favorably with that of CT
and was even higher than that for CT plus
colonoscopy (90% vs. 71%, respectively)
(Whiteford et al., 2000). Additional diag-
nostic value of FDG-PET in the detection
of pelvic recurrence was found in 56%
of patients if only CT findings were con-
sidered and in 20% of patients if both CT
and transrectal endoscopic ultrasonography
findings were considered with the subgroup
of patients with inconclusive conventional
393
imaging findings mainly profiting (Flamen
et al., 1999). Importantly especially for this
particular setting, FDG-PET has also shown
good positive and negative predictive val-
ues (76% and 92%, respectively) (Moore
et al., 2003).
One must nevertheless keep in mind the
specific pitfalls and limitations of FDG-
PET for diagnosis of pelvic recurrences.
FDG can not differentiate malignant tis-
sues from inflammatory process, such as
postoperative chronic fistula, suggesting
to delay data of acquisition to aid differ-
entiation of inflammatory from malignant
lesions. Increased FDG uptake following
radiation may also be due to inflamma-
tory changes and is not always associ-
ated with residual tumor. The time course
of postirradiation FDG activity has not
been studied systematically; however, it
is generally accepted that FDG activity
present six months after completion of
radiation therapy most likely represents
tumor recurrence, while positive predic-
tive value and accuracy have proven to
improve in FDG-PET imaging performed
more than 12 months after completion of
radiation (Moore et al., 2003). Similar to
anatomic imaging modalities, the altered
pelvic anatomy associated with surgical
removal of primary rectal tumors may
have an effect on the interpretation of
18
F-FDG-PET images, especially in the
differentiation between tumoral 18F-FDG
uptake and physiologic 18F-FDG uptake in
the gastrointestinal or genitourinary tract.
For optimal evaluation of the pelvis, the
bladder should be empty. The use of furo-
semide or even bladder catheterization can
occasionally be helpful in avoiding focal
ureteral activity. The integrated PET/CT
system may overcome the difficulties of
FDG-PET image interpretation associated
394 S. Truant et al.

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
patients, sensitivity, specificity, positive
predictive value, negative predictive value
and accuracy for differentiating malignant
from benign 18F-FDG uptake in the pelvis
were 98%, 96%, 90%, 97%, and 93% for
PET/CT and 82%, 65%, 73%, 75%, and
74% for PET, respectively (Even-Sapir
et al., 2004). In that study, the most common
cause for false-positive interpretation of
30. Colorectal Liver Metastases 395

PET findings was physiologic 18F-FDG value of FDG-PET in the preoperative

uptake in displaced pelvic organs. Thus, staging of liver metastases, avoiding futile
although FDG-PET is useful in distin- laparotomy, though there is no specific
guishing viable tumor from necrosis or study on FDG-PET accuracy in this set-
fibrosis in residual masses, the decision ting. As metastatic lymph nodes are fre-
to administer salvage therapy should be quently small and often contain small
made after the positive FDG-PET finding quantities of tumor cells, it is likely that
has been confirmed by biopsy. results may depend on their localization
and their size. False-negative results may
Peritoneal Metastases occur but are difficult to identify given the
Peritoneal dissemination is a major cause impossibility of a systematic histological
of unresectability of liver metastases, but verification. Nevertheless, studies on over-
is not easily detected at an early stage by all recurrent colorectal cancer have shown
anatomic imaging modalities. Few studies sensitivity rates of FDG-PET ranging from
have focused on FDG-PET contribution 65% to 100% for the detection of lymph
in the detection of peritoneal metastases. nodes metastatic to the groin, retroperi-
The largest one which included 23 patients toneum or portal triad (Lai et al., 1996).
with recurrent colorectal cancer was carried FDG-PET sensitivity was two to fourfold
out by Tanaka et al. (2002). In this study, that of CT in association with specificity
sensitivity and diagnostic accuracy of FDG- as high as 90–100%. The sensitivity of
PET (88% and 78%, respectively) were PET/CT (77%) was also superior to CT
about twofold those of CT (38% and 44%, alone (46%) in 76 patients with metastatic
respectively). However, FDG-PET failed colorectal cancer of the liver (Selzner
to show peritoneal lesions < 10–15 mm in et al., 2004). Importantly, as benign or
diameter and detected only some parts of post-therapy inflammatory processes may
the peritoneal deposits, resulting in unusu- occur in lymphoid tissue and cause false-
ally patchy FDG uptake that was sometimes positive results, histological confirmation
difficult to differentiate from benign uptake should be obtained before depriving a
in normal small bowel tissue (Tanaka et al., patient of potentially curative surgery.
2002). In 40 patients being considered Although it is likely that FDG-PET accu-
for hepatic resection but at high risk for racy regarding distant lymph node metas-
unresectable disease, 10% had peritoneal tases has been overestimated, its additional
disease that was not detected by FDG-PET diagnostic value has been established.
(Fong et al., 1999). Depending on the size
of the peritoneal deposits, sensitivity of
FDG-PET for detection of carcinosis varied Extra-Abdominal Organs
greatly from one study to another (range, Colorectal cancer may disseminate in lung
15–80%) but was higher than that obtained (25–30%), which is the second metastastic
by CT (Flamen et al., 1999) (Figure 30.2). site after the liver, and at later stages in
mediastinal lymph nodes, bone, and brain.
Distant Lymph Nodes Metastases
Except for brain, which shows physio-
Metastatic lymph nodes have contributed logic FDG uptake, whole-body FDG-PET
to some extent to the additional diagnostic has proven to be valuable in identifying
396 S. Truant et al.

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

extraabdominal metastatic spread. As for in the proliferative glandular breast tissue,

abdominal cavity, there are possibilities resulting in generally diffuse FDG uptake.
of normal FDG uptake in extra-abdominal
organs that must be considered in FDG- Contribution of FDG-PET
PET imaging interpretation. Some studies report a significant additional
diagnostic value of FDG-PET for the diag-
Physiologic and Benign Increased Activity of
nosis of unsuspected extra-abdominal sites,
FDG in the Extra-abdominal Organs
mainly in the lung. For this latter site
To avoid problems with image display, the of recurrence, sensitivity rates of FDG-
brain is excluded from the reconstructed PET (94–100%) were equivalent or slightly
whole-body images because of significant higher than those reported for CT (sensitiv-
accumulation of FDG and the fact that ity 50–100%). FDG-PET and CT showed
cerebral metastases may not be detected. similar specificity rates of 90–100%.
There can be an increased diffuse 18F-FDG Nevertheless, in a few of the prior studies
uptake in the bone marrow in patients of FDG-PET scanning, chest CT was rou-
undergoing hemopoietic stimulation ther- tinely used for preoperative staging. Some
apy or showing acute and severe bleeding. authors have further considered as a posi-
Conversely, focal activity within bone tive impact lung metastases detected by
marrow is always suspicious for an abnor- FDG-PET but not by chest X-rays a posi-
mality, although a healing bone, e.g., after tive impact, even though secondarily con-
rib fractures, or a degenerative or inflam- firmed by subsequent chest CT, resulting in
matory joint disease, can also give rise to additional disease in up to 31% of patients.
elevated FDG uptake. Vigorous exercise Other studies have shown that FDG-PET
in the days just before a scan should be was not superior to chest CT for detection
avoided as it can cause intense muscular of lung metastases (Lai et al., 1996; Truant
uptake in the muscle that usually has the et al., 2005). Detection of lung nodules is
distinctive pattern of a symmetric uptake. size dependent and small lung metastases
Stress-induced muscle tension often seen may be overlooked by FDG-PET that does
in the trapezius and paraspinal muscles can not match the spatial resolution of CT in the
be prevented by pretreatment with muscle chest. On the other hand, FDG-PET may
relaxants. Asymmetrical FDG uptake in help to differentiate benign from malignant
muscles and metabolically active fatty tis- lesions. Occasionally, other sites of extra-
sue (brown fat) is more difficult to interpret.abdominal recurrences have been reported
FDG uptake in residual thymic tissue has in patients being considered for hepatic
to be distinguished from mediastinal aden- resection (Figure 30.3), but additional diag-
opathy. Increased FDG uptake which most nostic value of FDG-PET may be difficult to
likely attributable to inflammation may assess due to the small number of patients.
also be seen in noninvolved lymph nodes
in patients with a medical history of silicosis,
bronchitis or granulomatous diseases such Impact of FDG-PET Findings on Patient’s
as tuberculosis or sarcoidosis. Benign Management
distribution of FDG occurs frequently in FDG-PET may have various types of
normal or goitrous thyroid, in heart and impact, mainly avoidance of laparotomy
398
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
unresectable, reducing futile morbidity or
seldom performance of more extensive sur-
gery by disproving a positive CT finding.
It may lead to commencement of palliative
chemotherapy, resulting in improved out-
come or much better palliation. Integration
of FDG-PET in the presurgical evaluation
of patients with liver metastases from
colorectal cancer has also been shown to
substantially reduce overall costs. Due to
the lack of randomized controlled clinical
trial to assess FDG-PET impact, only stud-
ies that focused on preoperative staging of
liver metastases have been considered for
this part of the analysis.
Additional Disease and Change of
Management
When FDG-PET was added to a complete
conventional work-up in preoperative
planning for patients with hepatic metas-
tases, additional sites of disease were
identified in 9.5–25% of patients (Truant
et al., 2005). FDG-PET contribution came
essentially from the discovery of extra-
hepatic sites, mainly in the abdomen,
S. Truant et al.
while FDG-PET did not confer any advan-
tage over conventional diagnostic methods
in the assessment of hepatic metastases.
Consequently, FDG-PET improved
resectability rates. In two series (Strasberg
et al., 2001; Truant et al., 2005), 95% of
patients who underwent laparotomy actu-
ally underwent hepatic resection, similar
to that achieved by the use of staging
laparoscopy to detect unresectable tumors
before laparotomy. In comparison, 20% of
patients who underwent exploration on the
basis of conventional imaging were found
to have unresectable disease at the time
of surgery. On the other hand, in these
series, FDG-PET also falsely upstaged and
understaged up to 7% and 8% of patients,
respectively (Topal et al., 2001; Truant
et al., 2005) and showed very poor yield
in patients who were on chemotherapy
at the time of FDG-PET (Akhurst et al.,
2005). In the preoperative staging of liver
metastases judged resectable on the basis
of conventional imaging findings, FDG-
PET resulted in a change of management
in 10–25% (Ruers et al., 2002). Selzner
et al. (2004) studied whether the novel
PET/CT imaging modality impacted the
30. Colorectal Liver Metastases
treatment of patients with metastatic color-
ectal cancer of the liver. As for FDG-PET
alone, the main impact of PET/CT came
from the discovery of extrahepatic disease.
New findings in the PET/CT resulted in a
change in the therapeutic strategy in 21%
of the patients. Therefore, the impact of
FDG-PET in the particular setting of pre-
sumed resectable liver metastases seems
to be less than the mean impact of 29%
reported by Huebner et al. (2000) in their
meta-analysis on FDG-PET contribution
in overall recurrent colorectal cancer.
The major part of the additional diag-
nostic value of FDG-PET resided in
the subgroup of patients at high risk for
unresectable disease, as predicted by the
Memorial Sloan Kettering Cancer Centre
clinical risk score based on five criteria:
(1) node positive primary, (2) disease-free
interval from primary to liver metastases
< 12 months, (3) largest hepatic tumour
> 5 cm, (4) number of hepatic tumours > 1
(5) CEA level >200 ng/ml (Fong et al.,
1999). Among patients with a low clinical
risk score, 0–3% of patients had additional
disease detected by FDG-PET, compared
with 14–17% for patients with a high clini-
cal risk score (Schussler-Fiorenza et al.,
2004; Truant et al., 2005). Thus, patients
with large (> 5 cm), multiple and synchro-
nous metastases were more likely to have
extrahepatic disease detected by FDG-PET.
In 40 patients being considered for hepatic
resection but with a high clinical risk score,
FDG-PET altered or influenced the clinical
management in 16 patients (40%) (Fong et
al., 1999). This FDG-PET impact is higher
than that reported in the aforementioned
series without consideration of the clini-
cal risk score. Some authors now consider
that patients at low risk of metastatic
disease should not undergo FDG-PET
399
before surgical exploration because these
patients have a low rate of detection of
extrahepatic disease and even a non-
negligible rate of false positive findings
(Schussler-Fiorenza et al., 2004; Truant
et al., 2005). As FDG-PET is costly and
not widely available, FDG-PET should be
restricted to high risk patients in centers
with low availability of FDG-PET in order
to delay a surgery.
Survival Impact
The addition of FDG-PET imaging to pre-
surgical evaluation may improve survival
by appropriately eliminating ineffective
surgery in patients with inoperable dis-
ease. Results are, however, highly contro-
versial regarding the survival benefit of
FDG-PET. It has been reported that the
use of preoperative FDG-PET lessened
the recurrence rate in patients undergo-
ing hepatic resection for colorectal liver
metastases by detection of disease not
found on conventional imaging. This had a
significant impact on the survival of these
patients. In 100 patients having hepatic
resection between March 1995 and June
2002, the 5-year disease-free and actu-
arial overall survival in patients screened
with FDG-PET was 34.8% and 58%,
respectively, which was higher than previ-
ously reported (Fernandez et al., 2004).
However, this series included 63% patients
with a single tumor while in other series,
resections of solitary liver metastases rep-
resented the minority of cases and the
resectioning of ten or more tumors was not
uncommon. Another study did not confirm
this survival benefit: 3-year overall and
disease-free survival rates after exclud-
ing the patients who did not undergo liver
resection were comparable with those for
400
all patients (51% vs. 47% and 27% vs.
24%, respectively) (Topal et al., 2001).
Most patients selected for surgery by
FDG-PET indeed still had recurrence of
disease because of the presence of dis-
ease that was not detected by FDG-PET.
In the study by Fong et al. (1999), after
a relatively short median follow-up (7
months), 40% patients had a recurrence
after liver resection, whereas other series
have reported recurrences as early as 3
months after resection (Strasberg et al.,
2001). Although a fraction of these find-
ings are likely due to disease that is below
the level of detection by FDG-PET, at least
some of these early recurrences are due to
true–false negative findings on FDG-PET.
On the other hand, patients in whom FDG-
PET provided additional information had a
dismal prognosis with a 3-year overall sur-
vival of 0% that could be explained by the
extent of their disease (Topal et al., 2001).
This may be the true survival impact of
FDG-PET: identifying the small number
of patients that have no chance for cure
despite reassuring conventional imaging.
Postoperative Follow-Up After Resection
of Colorectal Liver Metastases and
Rising Carcinoembryonic Antigen
(CEA)
After hepatic resection, recurrences occur
in up to 60% of patients, mostly within 2
years of surgery and with an equal distri-
bution between hepatic and extrahepatic
sites. It is assumed that recurrences diag-
nosed by systematic follow-up of asymp-
tomatic patients are more often resectable
with a curative intent. Serial determina-
tions of the plasma CEA concentration is
the most frequently used method for the
detection of such asymptomatic recurrences.
S. Truant et al.
However, rising CEA often occurs in asso-
ciation with negative or equivocal results
of conventional imaging studies, leading
to the diagnosis of tumor recurrences at
a non-resectable stage. The advantage
of FDG-PET over anatomic imaging to
screen the whole-body in one examination
session was assumed to be of particular
value in this setting. Due to the lack of
specific study, FDG-PET contribution in
the postoperative follow-up of resected
liver metastases can only be extrapolated
from studies on patients with suspected
recurrent colorectal cancer on the basis of
rising CEA.
In asymptomatic patients with a rising
CEA and nondiagnostic radiologic workup,
the sensitivity of FDG-PET for detection of
recurrence has been reported as 79–100%,
specificity as 50–83% and accuracy as
74–93% on a patient-by-patient analysis
(Flanagan et al., 1998). FDG-PET was
more sensitive than both CT and serum
CEA for detection of disease recurrence,
except for the liver, and the specificity of
FDG-PET was equivalent to that of CEA.
FDG-PET showed high positive predictive
value (89–97%) and negative predictive
value (71–100%) (Flanagan et al., 1998).
FDG-PET was also more accurate than
conventional diagnostic methods in pre-
dicting the resectability or non-resectabil-
ity of the recurrence, though it did show
a low sensitivity on a lesion-by-lesion
analysis (57%) in 28 patients who under-
went second-look laparotomy (Libutti
et al., 2001). By contrast, in patients with
normal CEA levels and symptoms, FDG-
PET showed a high rate of false-positive
results and was inaccurate in predicting
recurrence. In the postoperative follow-up
of liver metastases, specific pitfalls have
to be known such as FDG uptake in the
30. Colorectal Liver Metastases
scar of previous partial hepatectomy in
the case of important inflammatory reac-
tion, e.g., postoperative biloma. Studies
are now needed to address the impact of
early diagnosis by FDG-PET on survival
and its cost-effectiveness in the postopera-
tive follow-up of resected liver metastases
before large-scale introduction in clinical
practice can be advocated.
FDG-PET for Monitoring the Response
to Systemic or Regional Therapy of
Colorectal Liver Metastases
Hepatic metastases can be treated with
a curative or palliative intent with sys-
temic therapy (chemotherapy, new tar-
geted therapies) or regional therapy to
the liver, including local ablative therapy
(radiofrequency ablation, cryoablation),
chemotherapy administered through the
hepatic artery using infusion pomps, selec-
tive chemoembolization or radiolabeled
90
Y-microspheres.
Anticancer Systemic Therapy
Systemic Chemotherapy
Most patients with colorectal cancer meta-
static to the liver will receive chemotherapy,
using mostly 5-fluorouracil in combina-
tion chemotherapy regimen including
oxaliplatin (FOLFOX) and/or irinotecan
(FOLFIRI). When liver metastases are
non-resectable, chemotherapy alone may
supply a significant survival benefit pro-
viding that patient is a responder. Accurate
prediction of treatment failure may allow
the oncologist to apply an alternative treat-
ment regimen without subjecting the patient
to the toxicity of the full treatment scheme.
Monitoring with radiologic imaging meth-
ods like CT and MRI cannot assess ther-
401
apy outcome earlier than 2 months after
onset to therapy because resolution of
therapy-induced anatomic changes lags
behind tumor cell mortality. As metabolic
changes induced by chemotherapy precede
the morphological changes, 18F-FDG-PET
was assumed to predict response to therapy
at an earlier stage than conventional diag-
nostic methods. Evaluation of response
to therapy has involved careful compari-
son of pretreatment and post-treatment
FDG uptake, as assessed by qualitative
visual interpretation, standardized uptake
value (SUV) or extensive kinetic analysis.
Studies performed in patients with hepatic
metastases from colorectal cancer demon-
strated a relatively rapid decline (within
72 h) in the SUV in responding tumors after
a single infusion of 5-fluorouracil and foli-
nic acid, whereas nonresponding tumors
showed an increase, no change, or only a
small decline in 18F-FDG uptake. Using
SUV in tumors 4–5 weeks after the start
of treatment with 5-fluorouracil (Findlay
et al., 1996) or after FOLFOX therapy
(Dimitrakopoulou-Strauss et al., 2003) fur-
ther predicted the nonresponse with a high
accuracy. Quantitative, dynamic FDG-PET
also identified patients not responsive to
treatment with the FOLFOX regimen,
although the aggressiveness of the method
(multiple arterial samplings), its sensitiv-
ity to the noise and the long time of the
dynamic acquisition (1 h on a limited field)
are limiting for its use in carcinologic
domain (Dimitrakopoulou-Strauss et al.,
2003).
FDG-PET negativity following exten-
sive chemotherapy does not, however,
necessarily mean complete tumor destruc-
tion because injured but alive tumor cells
may not accumulate FDG to the same
extent as untreated cells. Conversely,
402
a marked increase in 18F-FDG metabolism
has been described in lesions responding
after initiation of chemotherapy, because
of major inflammatory response and tran-
sient increase in stromal reaction, and
may result in overestimation of the frac-
tion of viable cells. In a tumor mouse
model treated by chemotherapy, the FDG
uptake was stabilized by day 8–10 despite
a further reduction in viable tumor cell
fraction (Spaepen et al., 2003). In 18
patients with colorectal liver metastases,
the reduction in FDG uptake 1–2 weeks
after onset of chemotherapy was not sig-
nificantly different between the respond-
ing and non-responding groups and was
false-positive for response in 20% of the
patients (Findlay et al., 1996). Optimal
timing of FDG-PET is thus crucial and
imaging 1–2 weeks after completion of
chemotherapy is recommended to avoid
transient fluctuations in tumor FDG
uptake confounding response assessment.
These data suggest that a late 18F-FDG
examination, namely at 4–5 weeks after
the start of treatment, may be better for
therapy outcome. FDG-PET may further
be of prognostic value in patients under-
going chemotherapy. The first FDG-PET
study prior to onset of chemotherapy was
predictive for the therapy outcome, with
pretreated tumors displaying a low FDG
uptake more likely to have a worse prog-
nosis after chemotherapy than those with
a high SUV (Dimitrakopoulou-Strauss et
al., 2003). An early response to treatment
with 5-fluorouracil or FOLFOX regimen,
as shown on FDG-PET, was also well cor-
related with the ultimate response seen in
clinical, radiographic or pathologic find-
ings a few weeks or months later.
Despite interesting findings of sev-
eral studies investigating FDG-PET for
S. Truant et al.
monitoring the response during the course
of therapy, further clinical trials are needed
to assess the beneficial effect of early
FDG-PET scanning on the treatment of
patients and the ultimate outcome. The
method of quantifying tumor 18F-FDG
uptake, through visual or semiquantitative
methods such as SUV or kinetic analysis,
and remain challenging, as well as the
exact timing of FDG-PET.
New Targeted Therapies (Monoclonal
Antibodies)
Systemic chemotherapy is now used in
association with new targeted therapies,
such as the monoclonal antibodies bev-
acizumab and cetuximab that specifi-
cally block the receptors of the vascular
endothelial growth factor (VEGF) and
of the epidermal growth factor (EGF),
respectively. The VEGF stimulates the
proliferation and migration of vascularly
derived endothelial cells and is overex-
pressed in colon cancer. Several studies
have suggested that FDG-PET could allow
follow-up of patients treated by an angio-
genesis inhibitor. Endothelial cells have
indeed proven to contribute to 18F-FDG
uptake, with a positive correlation between
SUV and microvessel density. Recently,
FDG-PET showed an increased metabolic
activity in colorectal liver metastases fol-
lowing removal of the primary tumor, as
reflected by a rise in FDG uptake. As this
phenomenon coincided with a decrease in
serum anti-angiogenic factor levels, the
authors came to the conclusion that the
primary tumor suppressed angiogenesis in
its distant metastases and that its removal
resulted in an increase in metabolic activity
in liver metastases, emphasizing the inter-
est of angiogenesis inhibitors in colorectal
30. Colorectal Liver Metastases
cancer (Peeters et al., 2005). In the only
series of seven patients treated with beva-
cizumab and irinotecan for colorectal liver
metastases, FDG-PET findings correlated
better than those of CT with pathology
(Goshen et al., 2006). Further studies are
needed to confirm these results, as well
as studies on FDG-PET in the evaluation
of treatment of colorectal liver metastases
with cetuximab.
Regional Therapies
Local Ablative Therapy (Cryosurgery
Ablation and Radiofrequency Ablation)
Local ablative techniques are used as
adjuncts to chemotherapy or hepatic resec-
tion in a curative or palliative intent when
complete resection of tumor in the liver
cannot be achieved. These techniques can
be performed laparoscopically or during
open surgery, or for radiofrequency abla-
tion alone by a percutaneous procedure
guided by CT or sonography. In absence
of evolving extrahepatic lesions, as con-
firmed by pretherapy FDG-PET, these
local ablative therapies have been associ-
ated with improved median survival rates.
However, the frequency of incomplete
treatment, and thus local recurrence, is still
relatively high due to limitations in ade-
quately monitoring the destruction process
at the time of treatment. Moreover, after
local ablative therapy, local recurrences
in the liver are not easily identified with
CT, except at advanced stage of relapse,
because of difficulties in differentiating
viable tumor tissue from adjacent necrosis
or therapy-induced hyperperfusion.
The time of detection of recurrence by
FDG-PET has proven to be earlier than
the detection by CT alone. Comparative
studies reported a fully accurate prediction
403
of recurrence with follow-up FDG-PET,
whereas MRI and CT showed only 43%
and 50% of these recurrences, respectively
(Anderson et al., 2003). In lesions with
a positive FDG-PET after treatment, the
positive predictive value for the detection
of local recurrence was as high as 80%;
the negative predictive value was 100%
(Langenhoff et al., 2002). The authors
advised to administer a second ablative
treatment on any lesions that remain posi-
tive. The dual-modality PET/CT, that was
comparable to FDG-PET alone in terms of
sensitivity (65%) in a series of 11 patients
(Veit et al., 2006), might be of great value
for guidance of reinterventions. However,
detection of residual tumor remains chal-
lenging, as increased glucose metabolism
due to tissue regeneration surrounding
the ablative necrosis or abscess formation
in the treated area may superimpose on
small areas of residual tumor, resulting
in false-negative FDG-PET results. Some
authors recommended follow-up FDG-
PET directly after radiofrequency abla-
tion or at least within 24 h to avoid these
artefacts (Veit et al., 2006). Further studies
are needed to confirm these results and to
determine the best timing of FDG-PET.
Hepatic Arterial Chemotherapy
Patients with unresectable liver metastases
can be proposed hepatic arterial infusional
chemotherapy through an hepatic artery
pump placed during a short laparotomy.
This treatment provides at least equivalent
or even increased response rate for color-
ectal liver metastases when compared to
intravenous chemotherapy. It is, however
contraindicated in the presence of extrahe-
patic disease due to limited systemic effi-
ciency. Because patients with unresectable
404
liver metastases have a high risk of occult
extrahepatic disease, FDG-PET may spare
patients an unnecessary laparotomy in this
particular setting.
Other Regional Therapies
Some authors have studied the effects of
regional chemoembolization therapy using
FDG uptake as a criterion and found that
decreased FDG uptake correlated with
response, while presence of residual uptake
was used to guide further regional therapy
(Vitola et al., 1996a, b). A few studies have
also suggested the effectiveness of FDG-
PET in monitoring the therapeutic response
of hepatic metastases to 90Y-glass micro-
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,
CONCLUSION
In the preoperative staging of presumed
resectable colorectal liver metastases, the
best results of FDG-PET are obtained in the
detection of extra-hepatic disease, mainly
in the abdomen, while high-quality CT may
be as accurate as FDG-PET in depicting
liver or lung metastases. FDG-PET may be
especially valuable in patients with high
clinical risk scores or synchronous metas-
tases, which represent a greater proportion
of patients referred for surgery. Conversely,
in those instances in which FDG-PET
findings result in the patient being denied
potentially curative surgery, further inves-
tigations are needed, particularly in low
risk patients. The main impact involves
patients that are contraindicated before
undergoing futile surgery, whereas survival
S. Truant et al.
benefit of resected patients needs to be
confirmed. In the postoperative follow-up
of resected liver metastases, FDG-PET
may be useful in locating recurrences in
asymptomatic patients with a rising CEA
and nondiagnostic radiologic workup,
although a specific study is still needed in
this setting, as well as an analysis of the
impact of early diagnosis by FDG-PET
on survival. Results of preliminary studies
have also suggested the relevance of
FDG-PET in monitoring the response to
systemic or regional therapy of colorectal
liver metastases. The improvement of ther-
apeutic management and the cost savings
associated with the use of FDG-PET justify
its generalization in clinical practice.
H., Riedel, E. R., Burt, B. M., Blumgart, L.,
Jarnagin, W., Larson, S. M., and Fong, Y. 2005.
Recent chemotherapy reduces the sensitivity
of [18F]fluorodeoxyglucose positron emission
tomography in the detection of colorectal metas-
tases. J. Clin. Oncol. 23: 8713–8716.
Anderson, G. S., Brinkmann, F., Soulen, M. C.,
Alavi, A., and Zhuang, H. 2003. FDG positron
emission tomography in the surveillance of
hepatic tumors treated with radiofrequency abla-
tion. Clin. Nucl. Med. 28: 192–197.
Bipat, S., van Leeuwen, M. S., Comans, E. F.,
Pijl, M. E., Bossuyt, P. M., Zwinderman, A. H.,
and Stoker, J. 2005. Colorectal liver metastases:
CT, MR imaging, and PET for diagnosis meta-
analysis. Radiology 237: 123–131.
Dimitrakopoulou-Strauss, A., Strauss, L. G., and
Rudi, J. 2003. PET-FDG as predictor of therapy
response in patients with colorectal carcinoma.
Q. J. Nucl. Med. 47: 8–13.
Even-Sapir, E., Parag, Y., Lerman, H., Gutman, M.,
Levine, C., Rabau, M., Figer, A., and Metser, U.
2004. Detection of recurrence in patients with
rectal cancer: PET/CT after abdominoperineal or
anterior resection. Radiology 232: 815–822.
30. Colorectal Liver Metastases
Fernandez, F. G., Drebin, J. A., Linehan, D. C.,
Dehdashti, F., Siegel, B. A., and Strasberg, S.
M. 2004. Five-year survival after resection of
hepatic metastases from colorectal cancer in
patients screened by positron emission tomogra-
phy with F-18 fluorodeoxyglucose (FDG-PET).
Ann. Surg. 240: 438–447; discussion 447–450.
Findlay, M., Young, H., Cunningham, D., Iveson,
A., Cronin, B., Hickish, T., Pratt, B., Husband,
J., Flower, M., and Ott, R. 1996. Noninvasive
monitoring of tumor metabolism using fluorode-
oxyglucose and positron emission tomography
in colorectal cancer liver metastases: correlation
with tumor response to fluorouracil. J. Clin.
Oncol. 14: 700–708.
Flamen, P., Stroobants, S., Van Cutsem, E., Dupont,
P., Bormans, G., De Vadder, N., Penninckx,
F., Van Hoe, L., and Mortelmans, L. 1999.
Additional value of whole-body positron emis-
sion tomography with fluorine-18–2-fluoro-2-
deoxy-D-glucose in recurrent colorectal cancer.
J. Clin. Oncol. 17: 894–901.
Flanagan, F. L., Dehdashti, F., Ogunbiyi, O. A.,
Kodner, I. J., and Siegel, B. A. 1998. Utility of
FDG-PET for investigating unexplained plasma
CEA elevation in patients with colorectal cancer.
Ann. Surg. 227: 319–323.
Fong, Y., Saldinger, P. F., Akhurst, T., Macapinlac,
H., Yeung, H., Finn, R. D., Cohen, A., Kemeny,
N., Blumgart, L. H., and Larson, S. M. 1999.
Utility of 18F-FDG positron emission tom-
ography scanning on selection of patients for
resection of hepatic colorectal metastases. Am.
J. Surg. 178: 282–287.
Goshen, E., Davidson, T., Zwas, S. T., and Aderka,
D. 2006. PET/CT in the Evaluation of Response
to Treatment of Liver Metastases from Colorectal
Cancer with Bevacizumab and Irinotecan.
Technol. Cancer. Res. Treat. 5: 37–44.
Huebner, R. H., Park, K. C., Shepherd, J. E.,
Schwimmer, J., Czernin, J., Phelps, M. E., and
Gambhir, S. S. 2000. A meta-analysis of the
literature for whole-body FDG PET detection
of recurrent colorectal cancer. J. Nucl. Med. 41:
1177–1189.
Kinkel, K., Lu, Y., Both, M., Warren, R. S., and
Thoeni, R. F. 2002. Detection of hepatic metas-
tases from cancers of the gastrointestinal tract
by using noninvasive imaging methods (US, CT,
405
MR imaging, PET): a meta-analysis. Radiology
224: 748–756.
Lai, D. T., Fulham, M., Stephen, M. S., Chu, K. M.,
Solomon, M., Thompson, J. F., Sheldon, D. M.,
and Storey, D. W. 1996. The role of whole-body
positron emission tomography with [18F]fluoro-
deoxyglucose in identifying operable colorectal
cancer metastases to the liver. Arch. Surg. 131:
703–707.
Langenhoff, B. S., Oyen, W. J., Jager, G. J., Strijk,
S. P., Wobbes, T., Corstens, F. H., and Ruers,
T. J. 2002. Efficacy of fluorine-18-deoxyglu-
cose positron emission tomography in detecting
tumor recurrence after local ablative therapy for
liver metastases: a prospective study. J. Clin.
Oncol. 20: 4453–4458.
Libutti, S. K., Alexander, H. R., Jr., Choyke, P.,
Bartlett, D. L., Bacharach, S. L., Whatley,
M., Jousse, F., Eckelman, W. C., Kranda, K.,
Neumann, R. D., and Carrasquillo, J. A. 2001.
A prospective study of 2-[18F] fluoro-2-de-
oxy-D-glucose/positron emission tomography
scan, 99mTc-labeled arcitumomab (CEA-scan),
and blind second-look laparotomy for detecting
colon cancer recurrence in patients with increas-
ing carcinoembryonic antigen levels. Ann. Surg.
Oncol. 8: 779–786.
Moore, H. G., Akhurst, T., Larson, S. M., Minsky,
B. D., Mazumdar, M., and Guillem, J. G. 2003.
A case-controlled study of 18-fluorodeoxyglu-
cose positron emission tomography in the detec-
tion of pelvic recurrence in previously irradiated
rectal cancer patients. J. Am. Coll. Surg. 197:
22–28.
Peeters, C. F., de Geus, L. F., Westphal, J. R., de
Waal, R. M., Ruiter, D. J., Wobbes, T., Oyen,
W. J., and Ruers, T. J. 2005. Decrease in circu-
lating anti-angiogenic factors (angiostatin and
endostatin) after surgical removal of primary
colorectal carcinoma coincides with increased
metabolic activity of liver metastases. Surgery
137: 246–249.
Rohren, E. M., Paulson, E. K., Hagge, R., Wong, T.
Z., Killius, J., Clavien, P. A., and Nelson, R. C.
2002. The role of F-18 FDG positron emission
tomography in preoperative assessment of the
liver in patients being considered for curative
resection of hepatic metastases from colorectal
cancer. Clin. Nucl. Med. 27: 550–555.
406
Ruers, T. J., Langenhoff, B. S., Neeleman, N., Jager,
G. J., Strijk, S., Wobbes, T., Corstens, F. H., and
Oyen, W. J. 2002. Value of positron emission
tomography with [F-18]fluorodeoxyglucose in
patients with colorectal liver metastases: a pro-
spective study. J. Clin. Oncol. 20: 388–395.
Schussler-Fiorenza, C. M., Mahvi, D. M.,
Niederhuber, J., Rikkers, L. F., and Weber, S. M.
2004. Clinical risk score correlates with yield
of PET scan in patients with colorectal hepatic
metastases. J. Gastrointest. Surg. 8: 150–157;
discussion 157–158.
Selzner, M., Hany, T. F., Wildbrett, P., McCormack,
L., Kadry, Z., and Clavien, P. A. 2004. Does the
novel PET/CT imaging modality impact on the
treatment of patients with metastatic colorectal
cancer of the liver? Ann. Surg. 240: 1027–1034;
discussion 1035–1036.
Spaepen, K., Stroobants, S., Dupont, P., Bormans,
G., Balzarini, J., Verhoef, G., Mortelmans, L.,
Vandenberghe, P., and De Wolf-Peeters, C. 2003.
[(18)F]FDG PET monitoring of tumour response
to chemotherapy: does [(18)F]FDG uptake cor-
relate with the viable tumour cell fraction? Eur.
J. Nucl. Med. Mol. Imaging 30: 682–688.
Strasberg, S. M., Dehdashti, F., Siegel, B. A.,
Drebin, J. A., and Linehan, D. 2001. Survival of
patients evaluated by FDG-PET before hepatic
resection for metastatic colorectal carcinoma:
a prospective database study. Ann. Surg. 233:
293–299.
Tanaka, T., Kawai, Y., Kanai, M., Taki, Y.,
Nakamoto, Y., and Takabayashi, A. 2002.
Usefulness of FDG-positron emission tomogra-
phy in diagnosing peritoneal recurrence of color-
ectal cancer. Am. J. Surg. 184: 433–436.
Topal, B., Flamen, P., Aerts, R., D’Hoore, A.,
Filez, L., Van Cutsem, E., Mortelmans, L.,
and Penninckx, F. 2001. Clinical value of
whole-body emission tomography in potentially
S. Truant et al.
curable colorectal liver metastases. Eur. J. Surg.
Oncol. 27: 175–179.
Truant, S., Huglo, D., Hebbar, M., Ernst, O.,
Steinling, M., and Pruvot, F. R. 2005. Prospective
evaluation of the impact of [18F]fluoro-2-deoxy-
D-glucose positron emission tomography of
resectable colorectal liver metastases. Br. J.
Surg. 92: 362–369.
Veit, P., Antoch, G., Stergar, H., Bockisch, A.,
Forsting, M., and Kuehl, H. 2006. Detection of
residual tumor after radiofrequency ablation of
liver metastasis with dual-modality PET/CT:
initial results. Eur. Radiol. 16: 80–87.
Vitola, J. V., Delbeke, D., Meranze, S. G., Mazer,
M. J., and Pinson, C. W. 1996a. Positron emis-
sion tomography with F-18-fluorodeoxyglucose
to evaluate the results of hepatic chemoemboli-
zation. Cancer 78: 2216–2222.
Vitola, J. V., Delbeke, D., Sandler, M. P., Campbell,
M. G., Powers, T. A., Wright, J. K., Chapman,
W. C., and Pinson, C. W. 1996b. Positron emis-
sion tomography to stage suspected metastatic
colorectal carcinoma to the liver. Am. J. Surg.
171: 21–26.
Whiteford, M. H., Whiteford, H. M., Yee, L. F.,
Ogunbiyi, O. A., Dehdashti, F., Siegel, B. A.,
Birnbaum, E. H., Fleshman, J. W., Kodner, I. J.,
and Read, T. E. 2000. Usefulness of FDG-PET
scan in the assessment of suspected metastatic
or recurrent adenocarcinoma of the colon and
rectum. Dis. Colon Rectum 43: 759–767; discus-
sion 767–770.
Wong, C. Y., Salem, R., Qing, F., Wong, K. T.,
Barker, D., Gates, V., Lewandowski, R., Hill,
E. A., Dworkin, H. J., and Nagle, C. 2004.
Metabolic response after intraarterial 90Y-glass
microsphere treatment for colorectal liver metas-
tases: comparison of quantitative and visual
analyses by 18F-FDG PET. J. Nucl. Med. 45:
1892–1897.
A. Diagnosis
31
Biliary Cystic Tumors:
Clinicopathological Features
Yasuni Nakanuma, Hiroko Ikeda, Yasunori Sato, Kenichi Harada,
Koichi Nakamura, and Yoh Zen
INTRODUCTION
Several types of biliary cystic tumors
or tumorous lesions have been reported
(Nakanuma et al., 2000; Portman and
Nakanuma, 2006). Currently, with
advanced imaging techniques, increasing
numbers of cystic diseases including rare
types are being found, and their clinico-
pathological features are now being char-
acterized (Okuda, 2001). Hepatobiliary
cystadenoma and cystadenocarcinoma are
a prototype of biliary cystic tumors, and
intraductal papillary neoplasm of the bile
duct, particularly with cystic dilatation, is
a recently proposed disease entity (Chen
et al., 2001). However, the exact defini-
tion and clinicopathological features of
these biliary cystic tumors, particularly
for differential diagnosis, remain to be
established because the number of cases is
small. While the cystic lesions themselves
are rather easily detectable by imaging
modalities, an exact diagnosis based on
pathological characteristics is needed at
present.
In this chapter, the clinicopathologi-
cal features of biliary cystic tumors,
such as their walls being lined by biliary
or other epithelial cells, are reviewed.
The tumors can be divided into non-
neoplastic and neoplastic categories. The
former includes hamartomatous lesions
or developmental anomalies. Recently,
diagnostic criteria for intraductal papil-
lary mucinous neoplasm (IPMN) and
mucinous cystadenoma and cystadeno-
carcinoma, which are also called muci-
nous cystic neoplasm of the pancreas
(pancreatic MCN), have been established
with reference to the presence of com-
munication with the duct system and
also the presence of ovarian-like stroma
in the cystic walls (Longnecker et al.,
2000; Zamboni et al., 2000). Recent
studies showed that two similar types of
neoplastic diseases are also present in the
hepatobiliary system (Chen et al., 2001;
Zen et al., 2006b; Abraham et al., 2002,
2003). In this review, these biliary cystic
tumors are discussed and compared with
pancreatic cystic tumors from the stand-
point of their clinicopathological simi-
larities. Solitary hepatic cyst, polycystic
liver, and cystic dilatation of the bile duct
will not be discussed here.
411
412 Y. Nakanuma et al.

ANATOMICAL complex (VMC) and multicystic biliary

CLASSIFICATION hamartoma, are included in this category.
OF THE BILIARY TREE A majority of them may be ectopic lesions
or malformation of the biliary tree, and
The biliary tree is divided into the intrahe- some are acquired lesions.
patic, hilar (right and left hepatic duct) and
extrahepatic bile duct (EHBD) (Nakanuma Peribiliary Cysts
et al., 1997). The intrahepatic bile duct is Peribiliary Cysts in the Large
a branch of the right or left hepatic bile Portal Tracts and Hepatic Hilum
duct, and the first to third branches of both
hepatic ducts are known as the intrahepatic Peribiliary cysts (the latest discovery in
large bile duct. The extrahepatic bile duct, the field of hepatic cystic diseases or
hilar bile duct, and intrahepatic large bile lesions) are grossly visible and multi-
ducts are referred to collectively as the large ple serous cysts with thin walls are
bile duct in this chapter. Around the large found along the fourth-order bile duct
bile ducts, there are physiologically peri- branch up to the hepatic hilum (Figure
biliary glands with serous and mucinous 31.1B) (Nakanuma et al., 1984; Terada
acini in various combinations in the peri- and Nakanuma, 1990a). Peribiliary cysts
ductal connective tissue shown by Terada are usually tiny, < 10–15 mm in diameter,
et al. (1987), which are different from and do not communicate with the lumen
biliary hamarotoma (see below). These of the adjoining large bile ducts. When
glands are connected with bile duct lumen these cysts are clustered, they appear as a
via their own conduits. Occasionally, pan- cystic neoplasm, particularly resembling
creatic exocrine acini are admixed within a multilocular hepatobiliary cystadenoma.
these glandular lobules. These peribiliary Histologically, these cysts are located in
glands are regularly distributed along the the fibrous connective tissue of the portal
intrahepatic large bile ducts, while their tract and hepatic hilum. The lining of the
distribution is rather irregular or sparse at the cysts is a single layer of columnar or cuboi-
extrahepatic bile ducts. Intrahepatic small dal biliary epithelium, and the cysts do not
bile ducts are classified into septal and have their own fibrous wall (Figure 31.1C).
interlobular bile ducts, and bile ductules The latter is a differential histological point
are located at the periphery of the portal distinguishing from the adjoining, original
tracts. The biliary tree is covered by a layer bile ducts with its own fibrous wall. Some
of biliary epithelial cells or cholangiocytes cysts show a stratification of lining epithe-
and the large bile duct and septal bile ducts lial cells reflecting proliferation. The cysts
have their own dense fibrous wall. are admixed with lobules and conduits
of the peribiliary glands, some of which
show varying degrees of luminal dilatation
NON-NEOPLASTIC BILIARY (microcystic to macrocystic). These find-
CYSTIC LESIONS ings strongly suggest that peribiliary cysts
are derived from the cystic dilatation of
Peribiliary cysts, hepatic foregut cysts, bil- pre-existing peribiliary glands of the large
iary hamartoma such as von Meyenburg’s intrahepatic bile ducts.
31. Biliary Cystic Tumors: Clinicopathological Features 413

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-
biliary cysts were originally found in
cases of advanced liver disease such as
liver cirrhosis and other portal hyper-
tensive diseases, and in patients with
other hepatobiliary diseases (Nakanuma,
2004) (see below). Peribiliary cysts are
clinically suggested by computed tomog-
raphy (CT) in ~ 10% of cirrhotic livers
and more in autopsied cirrhotic livers
by histopathological studies. In these
diseases, peribiliary cysts may develop
secondarily, possibly related to disturbed
circulatory changes. Peribiliary cysts in
cases of these chronic liver diseases are
not infrequently associated with inflam-
matory and edematous changes (Terada
and Nakanuma, 1990a).
Peribiliary cysts are frequently found
in adult polycystic liver disease with
autosomal-dominant polycystic kidney
disease or alone, in addition to the cysts
in the parenchyma (parenchymal cysts)
which arise from the dilatation of VMC
or biliary microhamartoma (Kida et al.,
1992; Qian et al., 2003). Parenchymal
cysts are predominant in some cases,
while peribiliary cysts are predominant
in others. Peribiliary cysts are reportedly
identified by computed tomography (CT)
in 73% of adult polycystic liver disease.
That is, in adult polycystic liver disease,
intrahepatic peribiliary glands show fre-
quent and variable cystic dilatation. Such
peribiliary cysts may comprise a propor-
tion of numerous cysts of adult polycystic
liver. The genetic events responsible for the
development of liver cysts in the paren-
chyma may also be a factor in the develop-
ment of peribiliary cysts from peribiliary
glands. Peribiliary cysts also develop to a
lesser degree and frequency in livers with
solitary nonparasitic cysts. In this context,
peribiliary cysts could be a component of
the hepatobiliary fibropolycystic disease
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
2001). Therefore, unenhanced CT demon-
strates the lesion to be a high-density or
iso-density lesion relative to the surround-
ing liver parenchyma. Magnetic resonance
imaging (MRI) demonstrated the lesion to
be hyperintense on T1-weighted image.
These cystic lesions do not show a contrast
enhancement. When a subcapsular nodular
lesion shows the above-mentioned find-
ings, a diagnosis of CHFC should be con-
sidered (Kadoya et al., 1990). This type of
cyst may be mistaken for a hypovascular
solid tumor.
Variants: Recently, a histologic variant of
CHFC without a smooth muscle layer or
with inconspicuous muscle elements has
also been reported. It was < 2.0 cm in diam-
eter, and was located subcapusularly in the
medial segment (S4) of the liver. It had an
epithelial lining of ciliated pseudostratified
cells with occasional goblet cells, and lacked
a smooth muscle layer. The lining epithe-
lium contained cells positive for immunohis-
tochemical staining of surfactant apoprotein
A, suggesting an embryonic foregut origin
of the cysts and differentiation toward bron-
chiolar structures (Sato et al., 2006).
Biliary Hamartoma
The following two types are representative
of biliary hamartoma.
Von Meyenburg Complexes
Von Meyenburg complexes (VMCs), also
known as biliary microhamartoma or mul-
tiple bile duct hamartoma, are benign liver
malformations (Okuda, 2001). These com-
plexes are usually multiple, though their
number varies on the individual basis. They
are present in ~ 90% of livers with adult
polycystic liver disease, in ~ 20% of livers
with solitary nonparasitic cysts, and inciden-
415
tally in 1–6% of reported autopsy series. Von
Meyenburg complex (VMC) is a develop-
mental malformation that is thought to rep-
resent persistence of the embryonic ductal
plate. The histological appearance is quite
characteristic. Individual lesions are usually
< 5 mm in diameter, but sometimes up to
10 mm. Von Meyenburg complexes (VMC)
consist of a variable number of irregular bile
ductular structures embedded in a fibrocol-
lagenous stroma and are located along the
portal tracts. They are typically adjacent to
a portal area and consist of a fibrous stroma
that contains irregularly shaped duct-like
structures lined by a low cuboidal epithe-
lium. The ducts are usually somewhat dilated
and often U-shaped or branching. They may
contain proteinous or bile-stained secretions.
In some lesions, one of the ducts may dilate
sufficiently to form a macroscopic cyst.
Multiple VMCs are considered part of the
spectrum of hepatic fibropolycystic disease.
It is generally believed that both solitary bile
duct cyst and cysts of polycystic disease
arise from the dilatation of VMC.
Clinically, VMCs are symptomless
and occasionally misdiagnosed as malig-
nant liver disease including metastasis.
Imaging findings of VMC are nonspecific.
Ultrasonography (US), CT, and MRI show
multiple periportal small nodular or dot-like
lesions. Von Meyenburg complex (VMC)
shows multiple, or numerous tiny hypoat-
tenuating lesons, which are small hypoe-
choic structures occasionally associated
with comet-tail echoes on US, and hyperin-
tesne lesions on T2-weighted MRI (Okuda,
2001).
Multicystic Biliary Hamartoma
This lesion is a solitary and localized
cystic lesion ~ 2–5 cm in diameter, which
when located around the hepatic capsule
416 Y. Nakanuma et al.

of the hepatic lobe, characteristically pro-
trudes from the liver (Figure 31.2a). Some
cases not associated with inflammatory
changes have a honeycomb appearance.
Histologically, these nodular lesions consist
of ductal structures with ductal epithelium
resembling biliary epithelium and periduc-
tal glands resembling peribiliary glands,
and fibrous connective tissues containing
blood vessels (Zen et al., 2006c). Biliary
epithelial cells and glandular epithelial
cells express biliary-type cytokeratins such
as cytokeratins 7 and 19. Smooth muscle
bundles focally surround ductal structures
in some cases (Figure 31.2B). Bile-like
materials are observed within some ducts.
Two cases are associated with xanthogran-
ulomatous inflammation around bile-like
materials. This cystic lesion shares several
pathologic characteristics of CHFCs, such
as their location (around the falciform
ligament) and periductal smooth muscle
bundles, but does not fulfill the diagnostic
criteria (no ciliated cells and multilocular
lesions). These lesions might be related to
developmental abnormalities of the biliary
tract or embryonal foregut.
Ultrasonography (US) and CT reveal rel-
atively well-circumscribed nodules in the
medial segment of the liver, some of which
are protruding from the liver. This nodule
is enhanced on CT by contrast medium
and sustained until the delayed phase.
Computed tomography (CT) reveales a
relatively well-circumscribed nodule just
beneath the hepatic capsule. Angiographies
revealed that this lesion is supplied by the
hepatic artery. This nodule shows a multi-
locular cystic lesion containing many small
cystic spaces on radiologic images.
BILIARY CYSTIC NEOPLASM
Biliary cystic neoplasms are heterogene-
ous and can be divided into several cat-
egories. Hepatobiliary cystadenoma and
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

plastic biliary cystic disease, and intraductal

ring in the liver and along the extrahepatic
papillary neoplasm of the bile duct (IPNB),
bile ducts or gallbladder. They account
particularly with cystic dilatation of the
for < 5% of reported cystic lesions of the
affected bile ducts. While recent clinico-
liver (Devaney et al., 1994; Wheeler and
pathological studies suggested that these
Edmondson, 1985). They are solitary, usu-
two are different from each other, the simi-
ally multilocular and occasionally uniloc-
larities and differences between these two
ular, and the cyst wall is composed of
categories of biliary cystic neoplasms have
irregularly thickened fibrous tissue (Figure
not been well characterized or discussed
31.3B). In cystadenocarcinoma, tumor
in detail, and therefore, there may be some
infiltration into the adjacent parenchyma
confusion, especially among clinicians and
is found. Tumor sizes are variable and
pathologists who are not familiar with hepa-
may be as large as 28.0 cm in diameter.
tobiliary diseases. In addition, there could
Histologically, hepatobiliary cystadenoma/
be other uncategorized hepatobiliary cystic
cystadenocarcinoma are nearly always
neoplasms sharing some features of these
mucinous, and these tumors are composed
two categories of cystic neoplasms. of two components: an inner epithelial
layer and an outer fibrous layer. The
Hepatobiliary Cystadenoma most common type is mucinous cystad-
and Cystadenocarcinoma enoma with ovarian-like stroma occur-
ring in women (Devaney et al., 1994).
Clinicopathological Features
The cysts are lined by a mucin-secreting
Hepatobiliary cystadenoma and cystad- epithelium which may be flat or may
enocarcinoma are rare neoplasms occur- form papillary or polypoid projections,

a b c

Figure 31.3. Radiological and pathological features of hepatobiliary cystadenoma/cystadenocarcinoma.

(a) A multilocular cystic lesion with focal calcification is observed in the liver. (b) Macroscopically, the
inner surface of this lesion is rough, and associated with hemorrhage and fibrin exudates. (c) The epithe-
lium covering this lesion is made up of columnar cells containing intracytoplasmic mucin. A ovarian-like
stroma is observed beneath the epithelium (H&E staining, ×200)
418
pseudostratification and crypt-like invagi-
nation (Figure 31.3C). The epithelial cells
are usually columnar, resembling a biliary
or gastric foveolar epithelium with muci-
carmine and alcian blue-positive cyto-
plasmic mucin. Intestinal metaplasia with
goblet cells and neuroendocrine cells is
not infrequently present. According to the
grade of intraepithelial neoplasia, tumors
may be classified as adenoma, or border-
line (low-grade malignant), non-invasive,
or invasive adenocarcinoma, suggesting
a multi-step progression of hepatobiliary
cystadenoma. Distinguishing benign cys-
tadenoma from cystadenocarcinoma is
often difficult in tumors without distinct
invasion.
As for the outer layer, highly cellular
mesenchymal tissue resembling ovarian
stroma is characteristically seen in almost
all cases (Figure 31.3C). For this feature,
the name “cystadenoma with mesenchymal
stroma” has been used. The ovarian-like
stroma is found only in females, and the
stroma cells in the ovarian-like stroma are
positive for the estrogen receptor and pro-
gesterone receptor.
As secondary changes, the epithelium
may become ulcerated with extravasation
of cystic fluid into the stroma or wall, or
the lesions may be traumatized secondary
to their large size, so that inflammation,
xanthogranulomatous reaction, scarring,
and sometimes calcification are frequently
present in the tumors. While most reported
cases of hepatobiliary cystadenoma/ade-
nocarcinoma, particularly those with a
ovarian-like stroma, had no luminal com-
munication between the cystic tumor and
the adjacent bile duct, imaging modalities
revealed occasionally direct luminal com-
munication (Hamazaki et al., 1996; Sato
Y. Nakanuma et al.
et al., 2003). Such communication might
have developed secondarily after the inva-
sion of biliary cystadenocarcinoma.
Clinically, these tumors preferentially
affect women, especially middle-aged
women. The patients complain of an
enlarging abdominal mass. Occasionally,
a spontaneous rupture of cystic neoplasms
is reported. The background or preced-
ing conditions leading to hepatobiliary
MCNs are not known. Differentiating these
tumors from benign hepatic cysts may be
difficult. Because of their rarity, a diag-
nosis is often delayed and may result in
inaccurate treatment, resulting in unnec-
essary morbidity and mortality. Surgical
resectability is high, and the post-operative
progress is favorable, especially in MCNs
without invasion. Aspirated fluid of hepa-
tobiliary MCNs frequently shows high
levels of tumor markers such as CEA and
CA19-9, but it does not allow a diagnosis
of malignancy.
Radiologically, these tumors appear
as large multilocular or unilocular cysts
with internal septa and sometimes calci-
fication (Okuda, 2001) (Figure 31.3A).
Ultrasonography (US) shows a multilocu-
lar cyst with papillary projections that
are enhanced by dynamic CT. Hepatic
angiography demonstrates tumor stains
corresponding to the projected papillary
portions and displacement of vessels.
T1-weighted images show multilocules
with low signal intensity, unless hemor-
rhage or concentrated mucous fluid is
present. Most T2-weighted images dem-
onstrate hyperintensity within the cystic
areas, but the signal intensity sometimes
varies, depending on the fluid contents.
These above-mentioned findings are often
diagnostic for hepatobiliary MCNs.
31. Biliary Cystic Tumors: Clinicopathological Features
Comparison with Mucinous Cystic Neoplasm
of the Pancreas
Pancreatic MCN to which mucinous cys-
tadenoma and cystadenocarcinoma of the
pancreas belong in addition to MCN of
the retroperitoneum are cystic epithelial
neoplasms occurring almost exclusively in
women. They are composed of a colum-
nar, mucin-producing epithelium, sup-
ported by ovarian-type cellular stroma.
Tumors may be classified as adenoma,
borderline (low grade malignant), and
non-invasive or invasive adenocarcinoma.
Pancreatic MCN shows no communica-
tion with the pancreatic ductal system.
Interestingly, hepatobiliary cystadenoma/
cystadenocarcinoma also presents these
clinicopathological features of pancreatic
MCN. In this context, hepatobiliary cysta-
denoma/adenocarcinoma are a counterpart
of pancreatic MCN and could be called
hepatobiliary mucinous cystic neoplasm
(hepatobiliary MCN) (Zen et al., 2006a).
As for stromal components of these
MCNs, they might be derived from the
ovarian primodium. The ovarian-like
stroma in mucinous cystadenomas can
originate from gonadal epithelium during
early embryonic development. In human
embryos, preceding their ‘descent’, the
gonads are situated directly under the
diaphragm, dorsal to the liver, the tail of
the pancreas and the spleen, but separated
from these organs by the peritoneal cavity.
In contrast to the peritoneal epithelium
elsewhere, the cells covering the gonads
show an activated morphology. The ovar-
ian stroma in mucinous cystadenomas of
the liver and pancreas suggests a com-
mon origin in epithelial cells that cover
the embryonic gonads in early fetal life.
Cholangiocarcinoma arising in liver with
419
cystic disease is also often mistaken for
hepatobiliary cystadenocarcinoma.
Intraductal Papillary Neoplasm
of the Bile Duct (IPNB)
Clinicopathological Features
Biliary papillary tumors differing in malig-
nant potential and biological behavior from
ordinary tubular cholangiocarcinomas are
known to develop in the intrahepatic and
extrahepatic bile ducts (Adbores-Saavedra
et al., 2000), though their clinicopatho-
logic features remain to be clarified in
detail. As described below, these biliary
papillary tumors share clinicopathologi-
cal features of pancreatic IPMN. We pro-
posed referring to these papillary biliary
tumors collectively as intraductal papil-
lary neoplasm of the bile duct (IPNB) as a
biliary counterpart of pancreatic IPMN
(Chen et al., 2001; Zen et al., 2006b).
Both IPNBs and pancreatic IPMN
present with dilatation or multilocular
cystic changes of the affected ducts and
the pancreatic ducts, respectively.
(1) Classification into three subgroups
based on clinicopathological features
Our recent studies showed that IPNB is
characterized by three key pathologic
lesions in various components: (1) papillary
proliferation of neoplastic biliary epithelial
cells (dysplastic or adenocarcinomatous)
with delicate fibrovascular stalks within the
bile ducts, (2) mucin hypersecretion (mac-
roscopic mucin is recognizable in some
cases), and (3) variable dilatation or mul-
tilocular cystic changes of the affected bile
ducts (Figure 31.4B). Some IPNB cases
have been referred to as biliary papilloma
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)

intraductal growth-type of cholangiocarci- bile ducts. All cystic tumors contained

noma. Multiple and recurring biliary papil-
lomas are known as biliary papillomatosis
(BP). Biliary papilloma (papillomatosis) is
thought to be a pre-malignant or borderline
malignant lesion followed by progression
to papillary cholangiocarcinoma (Lee et al.,
2004). Other subgroups of IPNB are
characterized by tumors with clinically
recognizable mucin overproduction (IPNB
with mucin overproduction), and “mucin-
producing bile duct tumors” is an alterna-
tive term for this subgroup. The cystic,
aneurismal, or fusiform dilatation of the
affected bile ducts is prominent in another
subgroup of IPNB. These tumors appear
multilocular or unilocular. The average size
of such a dilated IPNB was 6.1 cm (range:
2.5–20 cm) in our series. Multilocular cystic
tumors might have been formed secondar-
ily by twisting or strangulation or dilatation
due to mucin overproduction, disturbed bile
flow, and biliary obstruction, in the affected
mucinous fluid. Except for mural nodules,
the internal surface of cystic tumors was
generally fine granular. Papillary mural
nodules, ranging from 0.5 to 3.2 cm, were
observed in all cases.
Clinically, IPNB affects both genders
and most cases arise in adults. IPNB
develops in an apparently normal liver
and biliary tree, and also occasionally in
cases of hepatolithiasis, primary scleros-
ing cholangitis, Caroli’s disease, choledo-
chal cysts, parasite infection, and liver
cirrhosis related to viral hepatitis. The left
hepatic lobe is more commonly affected.
All cystic tumors are located within the
liver, and rarely multilocular cystic tumors
protrude into hepatic hilum. Symptoms
result from biliary obstruction and its
complication. Mucinous discharge from
Vater’s ampulla, which is frequent in pan-
creatic IPMN, is occasionally observed in
IPNB on endoscopic examination during
31. Biliary Cystic Tumors: Clinicopathological Features 421

endoscopic retrograde cholangiography tumors or saccular or aneurysmal bile

(ERC). We have recently experienced one duct dilatation in the liver (Figure 31.4A).
autopsy case of IPNB in which bile duct In the cystic tumors, septation and lobula-
necrosis due to mucin overproduction by tion are identifiable. Intraductal papillary
IPNB showing biliary papillomatosis was neoplasm of the bile duct (IPNB) itself
the first clinical sign. In many cases, sur- may appear as a small and flat or polypoid
gical resections are performed based on a mass or nodule in the dilated bile ducts or
radiological diagnosis of biliary MCNs, cysts (Figure 31.4B). Large tumors may
and a pathological diagnosis of IPNB be depicted by US, CT, endoscopic ret-
is eventually made postoperatively. They rograde cholangiography (ERC) or MRI
show a rather favorable prognosis after (Okuda, 2001). However, tumors that are
total surgical excision, when compared relatively small or sessile or that spread
to ordinary tubular cholangiocarcinoma. along the mucosal surface, are difficult to
Curative resection in cases with multifo- visualize with imaging modalities. Mural
cality of the disease is difficult because nodules are also detectable on radiological
of frequent recurrence, and invasive carci- images. Endoscopic retrograde cholangi-
noma frequently develops in IPNB. ography (ERC) and intraoperative cholan-
giography from cystic tumors (cystgraphy)
(2) Communication to bile duct lumen
frequently reveal luminal communication
A majority of IPNBs show direct luminal between cystic tumors and the adjacent
communication between cystic spaces of the bile ducts. With US, mucin is usually ane-
tumors and adjoining bile ducts on preop- choic, like bile. In the future, advances in
erative or intraoperative cholangiographies, diagnostic techniques, particularly imag-
and also such communication is occasion- ing modalities and widespread recognition
ally demonstrable in surgically resected of this disease entity, may allow easy and
specimens, macroscopically. Intraepithelial correct recognition and diagnosis of IPNB,
neoplasm is also observed within adjacent and may lead to increased numbers of
non-dilated bile ducts, suggesting a direct IPNB patients.
in-situ like spread of neoplastic epithelial
cells from the affected bile ducts to the (4) Histologic grade of the tumor and
adjacent bile ducts or multiple occurrences progression to invasive carcinoma
of such neoplastic processes along the bil-
Commonly in these subgroups of IPNB, the
iary tree. More importantly, the walls of
affected bile ducts are lined by a papillary
these dilated lesions are microscopically
epithelium (micropapillary to grossly vis-
bile duct fibrous walls themselves, andible papillary) with dysplastic epithelium,
ovarian-like stroma is not observed, and
borderline malignancy to well-differenti-
immunoreactivity for estrogen receptor or
ated papillary adenocarcinoma (Zen et al.,
progesterone receptor are not detected in
2006b) (Figure 31.4C). Interestingly, many
the walls at all. of these cases are of borderline malignancy,
or invasive or non-invasive well-differenti-
(3) Radiological features
ated papillary adenocarcinoma, and only a
Computed tomography (CT) and MRI few are categorized as dysplasia. Intraductal
show multilocular or unilocular cystic papillary neoplasm of the bile duct (IPNB)
422
of either of these subgroups could progress
to two types of invasive carcinoma; tubular
adenocarcinoma or mucinous carcinoma.
These two histological types of invasive
lesions are closely related to the post-
operative survival status of patients. That
is, the latter shows a favorable prognosis in
comparison with the former. Interestingly,
phenotypic changes of MUC occurring
during tumor progression also relate to
the histological types of invasive lesions.
Mucus core protein 2 (MUC2) expression
is frequently observed in IPNB, including
non-invasive tumors and invasive lesions,
whereas MUC1 is expressed in almost
all cases of tubular adenocarcinoma, and
rarely in non-invasive IPNB and mucinous
carcinoma.
(5) Phenotypes
The neoplastic cells of IPNB are categorized
into three phenotypes (Zen et al., 2006b):
(1) the pancreaticobiliary type, composed
of columnar cells with an eosinophilic
cytoplasm and round nucleus; (2) the intes-
tinal type, characterized by stratified tall
columnar cells with some goblet cells,
resembling intestinal adenoma or adenocar-
cinoma; and (3) the gastric type, composed
of columnar cells with an abundant intracy-
toplasmic mucin. In addition, the oncocytic
type, which is characterized by abundant
eosinophilic cytoplasm and round nucleus,
is also reported as a variant of the pancreati- of the pancreatobiliary type and the other
cobiliary type (Sudo et al., 2001; Zen et al., half is of the intestinal type. The gastric
2006b). These three phenotypes including
the oncocytic type are also well known in
pancreatic IPMN.
(6) Mucin phenotypes and mucin
hypersecretion
Intraductal papillary neoplasm of the rise to invasive adenocarcinomas. Unlike
bile duct (IPNB) frequently presents as pancreatic IPMNs, little is known about
Y. Nakanuma et al.
gastrointestinal metaplasia with aberrant
expression of MUC2 and MUC5AC, while
the expression of MUC2 is not seen in con-
ventional ICC (Ishikawa et al., 2004a). The
nuclear expression of CDX2, a homeobox
gene involved in the regulation of intestinal
development and differentiation, is closely
associated with the expression of MUC2
in IPNB, suggesting its role in intestinal
metaplasia and its possible association
with carcinogenesis in IPNBs. Such fea-
tures are also reported in pancreatic IPMN.
As mentioned above, mucus hypersecre-
tion is a key feature of IPNBs. Mucin is
macroscopically identifiable (macroscopic
mucin) on the tumor surface and mac-
roscopic oversecretion of mucin into the
ductal lumen occurred in one fourth of
IPNB cases examined in our study, whereas
abundant mucin-production is observed in
almost all cases of pancreatic IPMN (Zen
et al., 2006d). Intraductal papillary neo-
plasm of the bile duct (IPNB) of the intes-
tinal phenotype has many goblet cells (one
of the mucin producing cells) and IPNB
commonly expresses the secretory type of
mucin core proteins, such as MUC2 and
MUC5AC, as seen in pancreatic IPMN.
The difference in the frequency of mac-
roscopic mucin hypersecretion between
IPNB and pancreatic IPMN might be
related to the extent of mucin-production.
Approximately, one half of IPNB cases is
type is rare in IPNB.
Molecular and Genetic Aspects of IPNB
Like pancreatic IPMN, IPNB can con-
sists of purely intraductal lesions with
step-wise histologic grades or can give
31. Biliary Cystic Tumors: Clinicopathological Features
the molecular and genetic pathogenesis
and progression processes of IPNB. As
for microsatellite instability, high-level
instability was present in 12 % of IPNB
cases, and low-level instability was present
in 35 % of IPNB cases (Abraham et al.,
2002). No IPNB cases showed methyla-
tion of hMLH1. These results indicate
that microsatellite instability is a relatively
frequent event in IPNB but is not associated
with hMLH1 promoter hypermethylation.
Interestingly, the pattern of allelic shifts
was frequently not identical between the
intraductal and invasive cholangiocarci-
noma components of IPNB, suggesting that
there is significant genetic heterogeneity
within these neoplasms.
E-cadherin is one subtype of trans-
membrane glycoprotein, whose cytoplas-
mic domain is bound to α- and β-catenin
(Ashida et al., 1998), and β-catenin acts
as a key structural molecule in E-cadherin-
mediated cell-to-cell adhesion. Membranous
β-catenin and E-cadherin expression is well
preserved in non-neoplastic epithelium.
Expression levels of both molecules are
significantly decreased in parallel in inva-
sive IPNB as compared with non-invasive
counterparts, respectively, suggesting that
a loss or reduction of the membranous
expression of β-catenin and E-cadherin is
involved in the progression of IPNB. A
reduction of β-catenin in the cell membrane
is presumed to disrupt cell to cell adhesive
stability, and therefore, such a reduction
may be followed by invasion and then
metastasis of carcinoma cells. Matrix met-
alloproteinase (MMP) and tissue inhibitor
of metalloproteinases (TIMP)-2 also play
important roles in the degradation of the
basement membrane during tumor inva-
sion. It is well known that the expression or
activities of matrix metalloproteinases such
423
as MMP-2, -7 and MT1-MMP in IPNB are
low or weakly positive in comparison with
conventional cholangiocarcinoma (CC) and
a balance of expression of MMP-2/tissue
inhibitor of matrix metalloproteinase-2
(TIMP-2) is also significantly decreased in
conventional CC compared to IPNB. The
participation of MMP/TIMP in the devel-
opment and progression of IPNB remains
to be clarified.
As for K-ras gene, activating muta-
tions in codon 12 of the K-ras oncogene
are present in 30 % of IPNB cases and
identical K-ras mutations are present in
both the intraductal and invasive com-
ponents, suggesting that these mutations
arise early in the tumorigenesis of IPNB
(Abraham et al., 2003). This frequency
of K-ras gene mutations is lower than
that previously reported for pancreatic
IPMNs. The expression of cyclin D1,
p21WAF1/CIP, pRb, p53, and MDM2 is
gradually increased with the progression
of IPNB. p53 expression is found in 40 %
of dysplastic IPNB and the expression
rate reaches a plateau in borderline malig-
nant or non-invasive IPNB and invasive
IPNB. Cyclin D1 and c-myc, a target of
Wnt signaling, are frequently present in
IPNB, and interestingly, nuclear β-catenin
staining, reflecting activation of Wnt sign-
aling, is observed in 25 % of non-invasive
IPNB and also in invasive IPNB without
a significant difference. This suggests
that nuclear staining of β-catenin, prob-
ably reflecting stimulation of the Wnt
signaling pathway, is probably involved
at an early stage in the progression of
IPNB. However, no β-catenin gene muta-
tions have been demonstrated in IPNB.
Interestingly, p53 expression, K-ras muta-
tion, and the nuclear accumulation of
β-catenin are greater in conventional CC
424
than IPNB, suggesting that the carcino-
genesis is different in these two types of
biliary neoplasm.
Inactivation of p16INK4A, a cancer
suppressor gene, is reportedly involved
in the carcinogenesis in several organs.
Our recent studies showed that while
p16INK4A is frequently expressed in
non-neoplastic, stone-containing bile
ducts, its expression is markedly reduced
in low grade IPNB to invasive IPNB in
hepatolithiasis (Ishikawa et al., 2004b).
Methylation-specific PCR revealed that
54.6 % of 44 IPNB foci harbored p16INK4a
promoter hypermethylation, and such foci
are significantly correlated with decreased
expression of p16INK4, suggesting that
p16INK4a inactivation mainly due to its
promoter hypermethylation, is a frequent
and early event of IPNB. DeHaan et al.
(2007) recently reported that a homozygous
9p21 deletion is frequently detected in
cholangiocarcinoma with complete loss of
p16INK4a. And this may be the case in
IPNB. This inactivation may be partially
responsible for the step-wise increase in the
cell proliferative activities of biliary epithe-
lial cells during the progression of INPB,
followed by further genetic and epigenetic
alterations of other cell cycle regulators
shown in IPNB. DPC4, the gene encod-
ing Smad4/Dpc4, is a tumor-suppressor
gene located on chromosome 18q. Dpc4
protein is almost universally expressed in
pancreatic IPMN and pancreatic cancer
derived from pancreatic IPMN. Abraham
et al. (2003) revealed that Dpc4 expression
was intact in all cases of IPNB examined
by immunohistochemistry. However, in our
studies, Dpc4 expression was lost in 21.4 %
of invasive IPNB cases, while Dpc4 expres-
sion was diffusely observed in all cases of
non-neoplastic biliary epithelium in their
Y. Nakanuma et al.
cytoplasm. Our results of Dpc4 expression
are not in line with the results of previous
studies. This discrepancy might be depend-
ent on the cases examined. Our study con-
sisted of Asian cases, some of which were
associated with hepatolithiasis.
IPNB Corresponds to IPMN of Main
Pancreatic Duct Type
The disease entity of pancreatic IPMN is
well established, and it arises in the main
pancreatic duct (main-duct type) and also
in its major branches (branch-duct type).
The histologic grade and invasiveness, and
phenotypes of papillary epithelium compo-
nent, and the degree of mucin secretion and
mucin phenotypes are different depending
on these two types of pancreatic IPMN. For
example, while pancreatic IPMN is divided
into benign, borderline, and malignant non-
invasive or invasive lesions. Benign lesions
are common in the branch-duct IPMN and
borderline and malignant lesions are com-
mon in the main-duct type.
Interestingly, the gross and microscopic
characteristics and corresponding radio-
logic features of IPNB are similar to those
of pancreatic IPMN, particularly those
arising in the main pancreatic duct. The
affected ducts of IPNB and pancreatic
IPMN are cystically or cylindrically dilated
and some appear to be multilocular. They
are characterized by the prominent intra-
ductal papillary proliferation of biliary
lining cells. Phenotypically, IPNB and pan-
creatic IPMN show one of three epithelial
phenotypes; pancreaticobiliary, intestinal
and gastric types. In addition, an oncocytic
variant belonging to the pancreatobiliary
phenotype is also reported in IPNB and
pancreatic IPMN. Like pancreatic IPMN,
IPNB could progress to tubular adenocar-
cinoma or mucinous carcinoma. These
31. Biliary Cystic Tumors: Clinicopathological Features 425

clinicopathological characteristics and the were biliary cystadenoma, and 18 (26 %)

survival status of IPNB closely resemble were cystadenocarcinoma. Ovarian-like
those of pancreatic IPMN, suggesting that stroma was observed in 85 % (44 of 52
IPNB may be the biliary counterpart of cases) of hepatobiliary cystadenoma and
pancreatic IPMN. The simultaneous devel- 28 % (5 of 18 cases) of cystadenocarci-
opment of IPNB and IPMN is occasionally noma. So far, IPNBs and hepatobiliary
reported, suggesting a common pathogene- MCNs have been confused because they
sis of these neoplasms and field canceriza- are both cystic and have a similar epithelial
tion of the bile ducts and pancreatic ducts. component. As for the lining epithelial cells
There are, however, several pathological and their histologic grade, IPNB and hepa-
differences between IPNB and pancreatic tobiliary MCN range pathobiologically
IPMN. As for the histological grade of from adenoma or dysplasia, borderline
IPNB and pancreatic IPMN, almost all malignancy and non-invasive malignancy
cases of IPNB belong to non-invasive to invasive carcinoma. Intestinal or gas-
or invasive well-differentiated papillary tric metaplasias are found in both cystic
adenocarcinoma or borderline malignancy, lesions, while oncocytic changes are found
while dysplasia or adenoma is infrequent. in IPNB.
Pancreatic IPMN of the main-duct type IPNB, particularly with a multilocu-
also shows such features, while pancreatic lar appearance or cystic dilatation of the
IPMN of the branch-duct type is usu- affected bile ducts (a cystic variant of
ally an adenoma. Phenotypically, a great IPNB), may resemble hepatobiliary MCNs.
majority of IPNBs are of pancreatobiliary Prevalence of tumor locations (intrahepatic)
or intestinal type, and the gastric type is and the type of cyst (multilocular cyst) are
rare. This proportion is similar to the main similar in both tumor groups. It is hard
duct type of IPMN. In contrast, the major- to differentiate IPNB from MCN radio-
ity of the branch-duct type is gastric type. logically. As described above, it is becom-
While IPNB and pancreatic IPMN of the ing evident that IPNBs and hepatobiliary
main duct type, particularly those of intes- MCNs are distinct entities, and they can
tinal phenotype, are characterized by the be separated based on clinicopathological
common expression of MUC2, CDX2 and features. Most biliary cystic tumors with an
cytokeratin 20 (CK20), pancreatic IPMN ovarian-like stroma are hepatobiliary cysta-
of the branch type shows lower frequencies denoma and some are cystadenocarcinoma.
of CK20 expression and a higher percent- It is likely that many of the hepatobiliary
age of gastric-type phenotype and mucin, MCNs reported in men or those without
thus differing from IPNB. Taken together, mesenchymal ovarian like stroma in the
it is likely that IPNB is a counterpart of early literature could actually belong to
pancreatic IPMN of the main pancreatic IPNB, as demonstrated in pancreatic MCNs
duct-type (Zen et al., 2006b). and IPMN. In addition, the presence of
luminal communication with the bile duct
Differentiation from Hepatobiliary MCNs
favors the diagnosis of IPNB.
According to Devaney’s report (1994), However, luminal communication
which is the largest study of “hepatobiliary is occasionally reported in hepatobil-
MCN”, 52 out of 70 tumors (74 %) iary MCNs with ovarian-like stroma. It
426
remains to be clarified whether all hepato-
biliary “cystadenoma and cystadenocarci-
noma” with luminal communication to the
bile ducts and those without ovarian-like
stroma are IPNB, and also whether uncate-
gorized biliary neoplasm(s) may constitute
these unusual cystic tumors.
Other Types of Neoplastic Cystic
Diseases
There may be other types or categories of
biliary cystic tumors which are different
from hepatobiliary MCN or IPNB with
cystic dilatation or multilocular changes
of the affected bile ducts. For example,
cholangiocarcinoma is known to develop
in biliary diseases showing cystic dilata-
tion such as congenital choledochal cyst
and Caroli’s disease, and also in hepatic
cystic diseases such as adult polycystic
liver disease or solitary liver cysts, and
rarely in the peribiliary glands or cysts
(Nakanuma et al., 2000; Terada and
Nakanuma, 1990b). Cholangiocarcinoma
arising in these diseases may present cystic
changes within or adjacent to the tumor
tissue as suggested by Fujii et al. (2005).
In conclusion, biliary cystic tumors are
largely divided into non-neoplastic, largely
hamartomatous, and neoplastic etiologies.
The former includes peribiliary cysts and
multicystic biliary hamartoma. The lat-
ter includes hepatobiliary cystadenoma/
adenocarcinoma and IPNB, particularly
those with cystic dilatation or multilocular
cystic changes of the affected bile ducts
(cystic variant of IPNB). Intraductal pap-
illary neoplasm of the bile duct (IPNB)
may be a biliary equivalent of pancreatic
IPMN, and hepatobiliary cystadenoma/cys-
tadenocarcinoma share clinicopathological
features of pancreatic MCT, and therefore
can be called biliary MCT. An exact diag-
Y. Nakanuma et al.
nosis of and distinguishing between these
cystic lesions are clinically important in
the choice of treatment, and genetic and
molecular approaches are needed to clarify
the etiology and pathogenesis of these
cystic diseases.
REFERENCES
Abraham, S.C., Lee, J.H., Hruban, R.H., Argani, P.,
Furth, E.E., and Wu, T.T. 2003. Molecular and
immunohistochemical analysis of intraductal
papillary neoplasms of the biliary tract. Hum.
Pathol. 34: 902–910.
Abraham, S.C., Lee, J.H., Boitnott, J.K., Argani,
P., Furth, E.E., and Wu, T.T. 2002. Microsatellite
instability in intraductal papillary neoplasms of
the biliary tract. Mod. Pathol. 15: 1309–1317.
Adbores-Saavedra, J., Scoazec, J.C., Wittekind,
C., Sripa, B., Menck, H.R., Soehendra, N., and
Sriram, P.V.J. 2000. Carcinoma of the gallbladder
and extrahepatic bile ducts. In: Hamilton, S.R.,
Aaltonen, L.A., eds. World Health Organization
Classification of Tumours. Pathology and
Genetics of Tumours of the Digestive System.
Lyon: IARC Press, 207–214.
Ashida, K., Terada, T., Kitamura,Y., and Kaibara, N.
1998. Expression of E-cadherin, alpha-catenin,
beta-catenin, and CD44 (standard and variant iso-
forms) in human cholangiocarcinoma: an immu-
nohistochemical study. Hepatology 27:974–82.
Chen, T.C., Nakanuma, Y., Zen, Y., Chen, M.F.,
Jan, Y.Y., Yeh, T.S., Chiu, C.T., Kuo, T.T.,
Kamiya, J., Oda, K., Hamaguchi, M., Ohno, Y.,
Hsieh, L.L., and Nimura, Y. 2001. Intraductal
papillary neoplasia of the liver associated with
hepatolithiasis. Hepatology 34: 651–658.
DeHaan, R.D., Kipp, B.R., Smyrk, T.C., Abraham,
S.C., Roberts, L.R., and Halling, K.C. 2007. An
assessment of chromosomal alterations detected
by fluorescence in situ hybridization and p16
expression in sporadic and primary scleros-
ing cholangitis-associated cholangiocarcinomas.
Hum. Pathol. 38: 491–499.
Devaney, K., Goodman, Z.D., and Ishak, K.G.
1994. Hepatobiliary cystadenoma and cystade-
nocarcinoma. A light microscopic and immuno-
histochemical study of 70 patients. Am. J. Surg.
Pathol. 18: 1078–1091.
31. Biliary Cystic Tumors: Clinicopathological Features
Fujii, T., Harada, K., Katayanagi, K., Kurumaya, H.,
and Nakanuma, Y. 2005. Intrahepatic cholangi-
ocarcinoma with multicystic, mucinous appearance
and oncocytic change. Pathol. Int. 55: 206–209.
Hamazaki, K., Gochi, A., Shimamura, H., Kaihara,
A., Maruo, Y., Doi, Y., Orita, K., Lygidakis, N.J.
1996. Serum levels of circulating intercellular
adhesion molecule 1 in hepatocellular carci-
noma. Hepatogastroenterology 43: 229–234.
Ishikawa, A., Sasaki, M., Ohira, S., Ohta, T., Oda,
K., Nimura, Y., Chen, M.F., Jan, Y.Y., Yeh, T.S.,
and Nakanuma, Y. 2004a. Aberrant expression
of CDX2 is closely related to the intestinal
metaplasia and MUC2 expression in intraductal
papillary neoplasm of the liver in hepatolithiasis.
Lab. Invest. 84: 629–638.
Ishikawa, A., Sasaki, M., Sato, Y., Ohira, S., Chen,
M.F., Huang, S.F., Oda, K., Nimura, Y., and
Nakanuma, Y. 2004b. Frequent p16ink4a inac-
tivation is an early and frequent event of intra-
ductal papillary neoplasm of the liver arising in
hepatolithiasis. Hum. Pathol. 35: 1505–1514.
Kadoya, M., Matsui, O., Nakanuma, Y., Yoshikawa,
J., Arai, K., Takashima, T., Amano, M., and
Kimura, M. 1990. Ciliated hepatic foregut cyst:
radiologic features. Radiology 175: 475–477.
Kida, T., Nakanuma, Y., and Terada, T. 1992.
Cystic dilatation of peribiliary glands in livers
with adult polycystic disease and livers with
solitary nonparasitic cysts: an autopsy study.
Hepatology 16: 334–340.
Lee, S.S., Kim, M.H., Lee, S.K., Jang, S.J., Song,
M.H., Kim, K.P., Kim, H.J., Seo, D.W., Song,
D.E., Yu, E., Lee, S.G., and Miin, Y.I. 2004.
Clinicopathologic review of 58 patients with bil-
iary papillomatosis. Cancer 100: 783–793.
Longnecker, D.S., Hruban, R.H., Adler, G., and
Kloppel, G. 2000. Intraductal papillary-mucinous
neoplasms of the pancreas. In: Hamilton, S.R.,
Aaltonen, L.A., eds. World Health Organization
Classification of Tumours. Pathology and
Genetics of Tumours of the Digestive System.
Lyon: IARC Press, 237–240.
Nakanuma, Y., Kurumaya, H., and Ohta, G.
1984. Multiple cysts in the hepatic hilum and
their pathogenesis. A suggestion of periductal
gland origin. Virchows. Arch. A. Pathol. Anat.
Histopathol. 404: 341–350.
Nakanuma, Y., Hoso, M., Sanzen, T., and Sasaki,
M. 1997. Microstructure and development of the
427
normal and pathologic biliary tract in humans,
including blood supply. Microsc. Res. Tech. 38:
552–570.
Nakanuma, Y., Leong, A.S.Y., Sripa, B.,
Ponchon, T., Vatanasapt, V., and Ishak, K.G.
2000. Intrahepatic cholangiocarcinoma. In:
Hamilton, S.R., Aaltonen, L.A., eds. World
Health Organization Classification of Tumours.
Pathology and Genetics of Tumours of the
Digestive System. Lyon: IARC Press, 173–180.
Nakanuma, Y. 2004. Peribiliary cysts have at least
two different pathogeneses. J. Gastroenterol. 39:
407–408.
Okuda, K. 2001. Okuda, K., Mitchell, D.G., Itai,
Y., Ariyama, J., eds. Hepatobiliary Diseases.
Oxford: Blackwell Science.
Portman, B.C., and Nakanuma, Y. 2006. Diseases
of bile ducts. Burt, A.D., Portman, B.C. Ferrell,
L.D., eds. MacSween’s Pathology of the Liver.
Churchill Livingstone 5th ed., pp. 517–581.
Qian, Q., Li, A., King, B.F., Kamath, P.S., Lager,
D.J., Huston, J., 3rd, Shub, C., Davila, S.,
Somlo, S., and Torres, V.E. 2003. Clinical profile
of autosomal dominant polycystic liver disease.
Hepatology 37: 164–71.
Sato, Y., Kitagawa, S., Zen, Y., Minato, H., and
Nakanuma, Y. 2006. Ciliated hepatic cyst with-
out smooth muscle layer: a variant of ciliated
hepatic foregut cyst? Pathol. Int. 56: 340–344.
Sato, M., Watanabe, Y., Tokui, K., Kohtani, T.,
Nakata, Y., Chen, Y., and Kawachi, K. 2003.
Hepatobiliary cystadenocarcinoma connected to
the hepatic duct: a case report and review of the lit-
erature. Hepatogastroenterology 50: 1621–1624.
Sudo, Y., Harada, K., Tsuneyama, K., Katayanagi,
K., Zen, Y., and Nakanuma, Y. 2001. Oncocytic
biliary cystadenocarcinoma is a form of intra-
ductal oncocytic papillary neoplasm of the liver.
Mod. Pathol. 14: 1304–1309.
Terada, T., and Nakanuma, Y. 1990a. Pathological
observations of intrahepatic peribiliary glands
in 1,000 consecutive autopsy livers. III. Survey
of necroinflammation and cystic dilatation.
Hepatology 12: 1229–1233.
Terada, T., Nakanuma, Y., and Ohta, G. 1987.
Glandular elements around the intrahepatic bile
ducts in man; their morphology and distribution
in normal livers. Liver 7: 1–8.
Terada, T., Nakanuma, Y., Kono, N., Ueda, K.,
Kadoya, M., and Matsui, O. 1990. Ciliated
428
hepatic foregut cyst. A mucus histochemical,
immunohistochemical, and ultrastructural study
in three cases in comparison with normal bronchi
and intrahepatic bile ducts. Am. J. Surg. Pathol.
14: 356–363.
Terada, T., and Nakanuma, Y. 1990b. Pathological
observations of intrahepatic peribiliary glands in
1,000 consecutive autopsy livers. II. A possible
source of cholangiocarcinoma. Hepatology 12:
92–97.
Wheeler, D.A., and Edmondson, H.A. 1985.
Cystadenoma with mesenchymal stroma (CMS)
in the liver and bile ducts. A clinicopathologic
study of 17 cases, 4 with malignant change.
Cancer 56: 1434–1445.
Zamboni, G., Longnecker, D.S., Kloppel, G., Adler,
G., and Hruban, R.H. 2000. Mucinous cystic neo-
plasm of the pancreas. In: Hamilton SR, Aaltonen
LA, eds. World Health Organization Classification
of Tumours. Pathology and Genetics of Tumours
of the Digestive System. Lyon: IARC Press, pp.
234–236.
Zen, Y., Fujii, T., Itatsu, K., Nakamura, K., Konishi,
F., Masuda, S., Mitsui, T., Asada, Y., Miura, S.,
Miyayama, S., Uehara, T., Katsuyama, T., Ohta,
T., Minato, H., and Nakanuma, Y. 2006a. Biliary
Y. Nakanuma et al.
cystic tumors with bile duct communication: a
cystic variant of intraductal papillary neoplasm
of the bile duct. Mod. Pathol. 19: 1243–1254.
Zen, Y., Fujii, T., Itatsu, K., Nakamura, K., Minato,
H., Kasashima, S., Kurumaya, H., Katayanagi,
K., Kawashima, A., Masuda, S., Niwa, H.,
Mitsui, T., Asada, Y., Miura, S., Ohta, T., and
Nakanuma, Y. 2006b. Biliary papillary tumors
share pathological features with intraductal
papillary mucinous neoplasm of the pancreas.
Hepatology 44: 1333–1343.
Zen, Y., Terahata, S., Miyayama, S., Mitsui, T.,
Takehara, A., Miura, S., Nobata, K., Kitao, A.,
Kakuda, K., Kiyohara, K., and Nakanuma, Y.
2006c. Multicystic biliary hamartoma: a hitherto
undescribed lesion. Hum. Pathol. 37: 339–344.
Zen, Y., Sasaki, M., Fujii, T., Chen, T.C., Chen,
M.F., Yeh, T.S., Jan, Y.Y., Huang, S.F., Nimura,
Y., and Nakanuma, Y. 2006d. Different expres-
sion patterns of mucin core proteins and
cytokeratins during intrahepatic cholangiocar-
cinogenesis from biliary intraepithelial neo-
plasia and intraductal papillary neoplasm of
the bile duct – an immunohistochemical study
of 110 cases of hepatolithiasis. J. Hepatol. 44:
350–358.
32
Cholangiocarcinoma: Intraductal
Sonography
Kazuo Inui, Hironao Miyoshi, and Junji Yoshino
INTRODUCTION
The recent development of magnetic reso-
nance cholangiopancreatography (MRCP)
and multiple-detector-row computed tom-
ography (MD-CT) facilitate diagnosis of
the obstruction site in patients with cholan-
giocarcinoma. However, the extension of
cholangiocarcinoma is still difficult to diag-
nose precisely. The correct diagnosis of
cancer extension, especially invasion to
the portal vein and pancreas, is necessary
to decide the operative procedures for a
complete resection. Intraductal ultrasonog-
raphy (IDUS) has been reported by several
authors, (Kuroiwa et al., 1998; Fujita et al.,
1998; and Tamada et al., 2001a), as a
reliable method for the evaluation of extra-
hepatic cholangiocarcinoma and aiding in
its accurate staging. We, Kanemaki et al.
(1997) and Inui et al. (1998), reported on the
usefulness of three-dimensional intraductal
ultrasonography (3D-IDUS) as a compre-
hensive image display, and the instruments
and processor systems of 3D-IDUS have
been developed. In this chapter, we introduce
the methodology of IDUS and 3D-IDUS for
the diagnosis of cholangiocarcinoma.
METHODS
Instruments
Several types of ultrasonic probes, a 2.5 mm
diameter miniature ultrasonic probe (UM-
2R or UM-3R, Olympus Medical Systems,
Tokyo, Japan) and a 2.1 mm diameter ultra-
sonic probe (UM-4R, Olympus Medical
Systems) have been developed. The UM-2R
ultrasonic probe incorporates a radial scan-
ning system with a frequency of 12 MHz, and
the UM-3R and UM-4R with 20 MHz. These
are connected to an endoscopic ultrasonic
observation unit (EU-M30 or EU-M2000,
Olympus Medical Systems) and incorpo-
rated by a probe-driving unit (MAJ 335 or
MAJ 2000, Olympus Medical Systems). The
probes are easily passed through the 2.8 mm
diameter biopsy channel of a duodenoscope
(JF-200, Olympus Medical Systems) or the
3.6 mm diameter biopsy channel of a new
electroduodenoscope (JF-V260, Olympus
Medical Systems).
We use the 3D-IDUS system developed
by Olympus Medical Systems (Figure
32.1); a 3D ultrasonic probe, UM-3D-2R
(12 MHz) or UM-3D-3R (20 MHz), with
429
430
Figure 32.1. Photograph of a 3D-IDUS system,
which includes a 3D ultrasonic probe and a probe
driving unit, the MAJ 2000. The ultrasonic probe
(arrow), a UM-DG20–35R, is a new type of ultra-
sonic probe with a ropeway system
a diameter of 3.4 mm; and a probe driving
unit, the MAJ 355 beginning in 1995 and
the MAJ 2000 since 2002. Recently, we
usually use a new type of ultrasonic probe
with a ropeway system, UM-DG20-35R
(Figure 32.1). These probes are connected
to an endoscopic ultrasonic observation
unit (EU-M30 or EU-M2000). The 3D
ultrasonic probe consists of an external
tube as an outer sheath and the probe itself
with a diameter of 2.4 mm and a 12 or
20 MHz radial scan transducer at its tip.
K. Inui et al.
Ultrasonography is performed while the
ultrasonic probe in the outer sheath is with-
drawn automatically. During scanning, the
outer sheath does not move. Linear recon-
struction (longitudinal) images are pro-
duced from integrating 40 serial radial
images. We can obtain from 40 to 118
serial radial images. The length of the
longitudinal images can be set at 10, 20,
30, or 40 mm, and in pitches of 0.25, 0.5,
0.75, or 1.0 mm. We routinely obtain 118
serial images at the length of 40 mm and
in pitches of 0.5 mm. For 3D-IDUS, we
use an ultrasound image-processing unit
(EU-IP2, Olympus Medical Systems) to
produce the reconstruction images such as
the dual plane reconstruction images.
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-
mide intramuscularly, and 10 mg diazepam
intravenously. After insertion of the elec-
troduodenoscope into the second portion
of the duodenum, endoscopic retrograde
cholangiopancreatography (ERCP) is
performed. After ERCP, we insert the
ultrasonic probe into the stricture of the
bile duct through the biopsy channel of
the electroduodenoscope (Figure 32.2).
Ultrasonography is performed, and the
area observed by the probe is confirmed
by fluoroscopy.
Percutaneous Transhepatic Approach
In patients with obstructive jaundice, per-
cutaneous transhepatic biliary drainage
(PTBD) is performed immediately after
admission to our hospital. We dilate the
32. Cholangiocarcinoma: Intraductal Sonography
Figure 32.2. An ultrasonic probe is inserted into
the bile duct through the biopsy channel of a duo-
denoscope along with a guide-wire
sinus tract of the PTBD to 16 F and per-
form percutaneous transhepatic cholangi-
oscopy (PTCS), IDUS, and/or 3D-IDUS 1
week after PTBD.
The patient is placed on the fluoroscopy
table in the supine position. Premedication
consists of 0.5 mg atropine sulfate, 15 mg
pentazocine intramuscularly, and 10 mg
diazepam intravenously. We observe the
biliary stricture with PTCS and take biop-
sies from the stricture. After PTCS, a
431
16 F-PTBD drainage catheter is inserted
into the sinus tract along with a guide-
wire. The ultrasonic probe is inserted into
the stricture of the bile duct through the
PTBD drainage catheter (Figure 32.2).
Intraductal ultrasonography or 3D-IDUS
is performed, and the area observed by the
probe is confirmed by fluoroscopy. For
better images, 2 mg scopolamine butyl-
bromide is given intramuscularly before
scanning to stop the peristalsis of the
duodenum.
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
with cholangiocarcinoma had 3D-IDUS
after treatment by Nd-YAG laser irradia-
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
approach.
Intraductal ultrasonography and
3D-IDUS may be indicated for the diag-
nosis of remnant stones after endoscopic
lithotripsy of common bile duct stones.
They also may be indicated for the
diagnosis of biliary strictures and the
precise extension of cancer in patients
with cholangiocarcinoma and intraduc-
tal papillary mucinous neoplasia of the
pancreas.
432
IMAGES AND CLINICAL
UTILITY
Intraductal Sonography
Findings of Cholangiocarcinoma
The normal bile duct wall is observed as
two to three layers. On the other hand,
IDUS reveals a carcinoma lesion as an
image of low echoic, irregular thickness.
When the tumor echogram extends into
the high echoic layer of the bile duct, we
diagnose the tumor as invading the subse-
rosa. When the tumor echogram extends to
the parenchyma of the pancreas, we diag-
nose the tumor as invading the pancreas.
The important demarcation of tumor inva-
sion into the pancreas is the disappearance
of the high echoic layer between the bile
duct and the pancreas. When the tumor
echogram extends to the high echoic wall
of the portal vein, we diagnose the tumor
as invading the portal vein. Longitudinal
tumor spread is known to be one of
the characteristics of tumor extension of
cholangiocarcinoma. Its diagnosis is an
important issue for a curable resection.
If an irregular thickness of the bile duct
is observed consecutively and away from
the main lesion, we diagnose longitudinal
spread of the tumor.
Precise Diagnosis of Cancer Extension
The results of IDUS and pathological stud-
ies for tumor extension in 13 patients with
bile duct carcinoma were investigated (Inui
et al., 2002). The overall accuracy for the
depth of tumor invasion was 84.6 %, for
tumor invasion of the pancreas was 88.9 %,
and invasion of the portal vein was 92.3 %.
Oblique Reconstruction Images
The results of IDUS and the pathological
diagnosis for longitudinal tumor spread The oblique reconstruction images (Figure
were studied in 13 patients. Eight patients 32.3b) and surface rendering reconstruction
K. Inui et al.
without intraductal spread were diagnosed
correctly and three of five patients with intra-
ductal spreading were diagnosed correctly
with IDUS. The overall accuracy of IDUS
for intraductal tumor spreading was 84.6 %.
Three-Dimensional Intraductal
Sonography
One of the advantages of 3D-IDUS is that
the time required for the examinations is
reduced compared with that required for
conventional IDUS, because conventional
IDUS needs to clarify the relationship
between the lesions and the surrounding
organs and vessels. Three types of images
can be produced using the functions of the
3D-IDUS system: dual plane reconstruc-
tion (DPR) images; oblique reconstruction
images; and surface rendering reconstruc-
tion images.
Dual Plane Reconstruction Images
Dual plane reconstruction images, includ-
ing radial and longitudinal images, are
useful in assessing the tumor extension
and the tumor’s relationship with the sur-
rounding organs (Figure 32.3). The lon-
gitudinal images can be rotated without
changing the radial images, and the radial
images can also be changed without alter-
ing the linear images. Because the longi-
tudinal reconstruction images are similar
to an X-ray cholangiogram, it is easy to
recognize the relationship between the
lesions and the surrounding organs. When
rotated in an anti-clockwise direction,
tumor invasion into the pancreas can be
clearly observed.
32. Cholangiocarcinoma: Intraductal Sonography
Figure 32.3. An ultrasonic probe is inserted into
the stenosis of the bile duct through the PTBD
drainage catheter
images can be produced automatically as ste-
reographic views in real time. It is useful to
understand the relationship between tumors
and the surrounding organs or vessels as
a comprehensive image display. However,
the 3D-images are not yet satisfactory when
compared with those of a CT scan or MRI.
The technology will likely be improved for
clinical use in the near future.
Diagnositc Accuracy for Cancer Extension
To determine the diagnostic ability of
3D-IDUS with regard to tumor infiltration
into the pancreas, the 3D-IDUS findings
in ten surgical cases were compared with
the findings of other imaging modalities
as well as the histological findings. The
433
accuracy rate of 3D-IDUS was 90 %, and
the accuracy rate of conventional IDUS was
80 %. On the other hand, the accuracy rates
for CT scans and angiography were 62.5 %
and 50%, respectively. Three-dimensional
IDUS is useful for the precise diagnosis of
cancer extension in cholangiocarcinoma,
especially for assessing invasion into the
portal vein and pancreas (Figure 32.4).
Volume Measurement of Tumors
Tumor volume could be calculated within
a few minutes using the trace areas of the
serial radial images. With EU-IP2, a new
type of image-processing, it is easier to
calculate tumor volume than with EU-IP.
In 11 of the 15 patients (73.3 %), the whole
outline of the tumor could be visualized,
and we could measure the tumor volumes.
We treated high-risk patients with bile
duct carcinoma using Nd-YAG laser irra-
diation under PTCS. Cholangiography
revealed complete recanalization of the
biliary stenosis. We performed 3D-IDUS
before and after laser therapy. The tumor
volume ranged from 913 to 16,960 mm3
(mean ± SD, 5,922 ± 7,593) before treat-
ment, while it had decreased to 484–
10,060 mm3 (mean ± SD, 3,148 ± 4,621)
after treatment. The tumor echos were
decreased significantly after laser therapy
compared with those before therapy. The
results of cholangiography also differed
after laser therapy. The tumor reduction
rates were 20–73 % (mean ± SD, 45.3
± 21.7). The survival time of the four
patients was from 105 to 920 days (mean
± SD, 463 ± 352) after laser therapy. The
calculation of tumor volume may help
facilitate the evaluation of the efficacy
of laser therapy for patients with bile
duct cancer (Inui and Miyoshi, 2004)
(Figure 32.5).
434 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)

including us reported that it is a reliable

method for the evaluation of extrahepatic
cholangiocarcinoma. On the other hand,
the first report of 3D-IDUS was published
by Kallimanis et al. (1995). We also
reported the usefulness of 3D-IDUS as
a comprehensive image display in 1997
(Kanemaki et al., 1997).
The instruments for IDUS, including a
miniature ultrasonic probe, an observation
unit, the processing system, etc., have been
under development. At first, the minimum
diameter of a miniature probe was 3.6 mm,
while an ultrasonic probe 2.1 mm in diam-
eter is now available. Stavropoulos et al.
Figure 32.5. Oblique reconstruction images reveal
the tumor echo (arrow) which has not invaded the (2005) reported that the applicability of
portal vein (PV) IDUS is limited in tight stenosis compared
with endoscopic ultrasonography, although
IDUS using a high-frequency miniature
DISCUSSION ultrasound probe is accurate for the stag-
ing of bile duct cancers. For improved
Intraductal ultrasonography was reported resolution in cases where there is a limita-
by Silverstein et al. (1989) in an experi- tion in insertion, an ultrasonic probe with
mental study in canines. Many authors a ropeway system has been developed, and
32. Cholangiocarcinoma: Intraductal Sonography
an increased success rate using a trans-
papillary approach was reported by Fujita
et al. (2000). Using the ropeway system
with a guide-wire, scanning of the bile
duct was achieved in 98.2–99.2 % of the
patients, regardless of the site and length
of the biliary strictures. Once the guide-
wire is passed through the biliary stenosis,
no further technique is required to intro-
duce the ultrasonic probe.
The clinical indications for IDUS of
the biliary tract include the diagnosis of
choledocholithiasis (Palazzo, 1997), the
differential diagnosis of biliary strictures
(Menzel et al., 2000) and the local stag-
ing of cholangiocarcinoma. Wada et al.
(2000) reported that the diagnosis of bil-
iary metallic stent obstruction and evalu-
ation for microwave coagulation therapy
were new clinical applications of IDUS.
Menzel et al. (2000) proposed that
malignancy should be suspected for hyp-
oechoic masses with irregular margins and
inhomogeneous echo-poor areas invading
the surrounding tissue. Penetration was
defined as a continuation of the main
hypo-echoic tumor mass into adjacent
structures. On the other hand, homogene-
ous echo patterns and smoothly defined
margins were most likely not malignant.
They reported that the sensitivity, specifi-
city, and accuracy of IDUS were 91.1 %,
80 %, and 89.1 %, respectively. Tamada
et al. (1998) reported that if the inter-
ruption of the bile duct wall structure or
a lesion with a diameter of greater than
8.0 mm were diagnosed as malignant, the
accuracy of IDUS was 76%.
When a tumor protrudes into the pan-
creatic parenchyma, or if the outer echoic
layer of the intra-pancreatic bile duct is
interrupted by a tumor, invasion to the
pancreas can be diagnosed. The accuracy
435
of IDUS in assessing tumor invasion to the
pancreas was reported to be 88.9–100 %
(Inui et al., 2002). When the high-echoic
layer between the tumor and a vessel dis-
appears, it is diagnosed as positive vessel
invasion. The accuracy of IDUS in assess-
ing tumor invasion to the right hepatic
artery and portal vein was reported to be
92.3–100 % (Inui et al., 2002). IDUS is
useful for the staging of cholangiocarci-
noma, especially invasion to the portal
vein and pancreas. Menzel and Dosmschke
(1999) reported on the usefulness of IDUS
for diagnosis of cholangiocarcinoma, and
showed that tumor extent was diagnosed
correctly in 76.8 % of cases when using
IDUS and in 53.6 % cases in EUS.
Tamada et al. (2001b) reported that an
accurate distinction of T1 and T2 tumor
was difficult. In these cases, even if the
tumor was limited to the inside low-echoic
layer, it could be a T2 tumor (the tumor
invaded the peri-muscular connective tis-
sue) or a T1 tumor (the tumor was con-
fined to the fibromuscular layer).
The diagnosis of the longitudinal spread
of a tumor is the most important issue
for a curable resection, but this is diffi-
cult for IDUS. When the irregular thick-
ness of the bile duct can be observed
consecutively and away from the main
lesion, the longitudinal spread of tumor
can be diagnosed. The overall accuracy
of IDUS for intraductal tumor spread was
84.6%, and the result was not satisfactory.
However, Tamada et al. (2001b) reported
that transpapillary IDUS prior to biliary
drainage is a tactic that can minimize the
influence of the biliary drainage catheter.
Nevertheless, their results were the same
as our prior study (Inui et al., 2002) that
used a percutaneous approach, with an
accuracy of 85 %.
436
Three dimensional intraductal
Ultrasonography is more useful for the
precise diagnosis of cancer extension in
cholangiocarcinoma than conventional
IDUS. Visualization of the left hepatic
artery and proper hepatic artery was poor
(14 %/18 %) due to echo attenuation in
conventional IDUS as reported by Tamada
et al. (2001a). On the other hand, Kanemaki
et al. (1997) reported that 3D-IDUS was
useful to estimate tumor invasion to the
portal vein with an accuracy of 100 %.
However, Tamada et al. (2001b) described
same limitations of 3D-IDUS in the assess-
ment of the longitudinal spread of cancer.
Another limitation of both IDUS and
3D-IDUS in evaluating cholangiocarci-
noma is its difficulty in diagnosing lymph
node metastasis.
CONCLUSIONS
IDUS is a reliable method for the evalu-
ation and staging of cholangiocarcinoma.
DPR images produced by 3D-IDUS are
useful for assessing the tumor extension
and its relationship with the surrounding
organs. 3D-IDUS is useful for the precise
diagnosis of cancer extension in cholan-
giocarcinoma, especially for invasion into
the portal vein and pancreas. Volume cal-
culation should facilitate the evaluation of
the efficacy of non-surgical treatments.
REFERENCES
Fujita, N., Noda, Y., Kobayashi, G., Kimura,
K., Yago, A. 1998. Staging of bile duct carci-
noma by EUS and IDUS. Endoscopy. 30(S1):
A132–134.
Fujita, N., Noda, Y., Yokohata, K., Tanaka, M.,
Maguchi, H., Komatsu, Y., Omata, M., Inui, K.,
Nakazawa, S., Mukai, H., Yasuda, K. 2000.
K. Inui et al.
Newly developed ultrasonic probe with ropeway
system for transpapillary Intraductal ultrasonog-
raphy of the bilio-pancreatic ductal system. Dig.
Endosc. 12:250–254.
Inui, K., Nakazawa, S., Yoshino, J., Wakabayashi,
T., Okushima, K., Nakamura, Y., Hattori, T.,
Miyoshi, H. 1998. Ultrasound probes for biliary
lesions. Endoscopy 30(S1):A 120–123.
Inui, K., Yoshino, J., Okushima, K., Miyoshi,
H., Nakamura, Y. 2002. Intraductal EUS.
Gastrointest. Endosc. 4:S58–62.
Inui, K. Miyoshi, H. 2004. Cholangiocarcinoma
and intraductgal sonography. Gastrointest.
Endosc. N. Am. 15:143–155.
Kallimanis, G., Garra, B.S., Tio, T.L., Krasner, B.,
al-Kawas, F.H., Fleischer, D.E., Zeman, R.K.,
Nguyen, C.C., Benjamin, S.B. 1995. The fea-
sibility of three-dimensional endoscopic ultra-
sonography: a preliminary report. Gastrointest.
Endosc. 41:235–239.
Kanemaki, N., Nakazawa, S., Inui, K., Yoshino,
J., Yamao, K., Okushima, K. 1997. Three-
dimensional intraductal ultrasonography: pre-
liminary results of a new technique for the
diagnosis of diseases of the pancreatobiliary
system. Endoscopy 29:726–731.
Kuroiwa, M., Goto, H., Hirooka, Y., Furukawa,
T., Hayakawa, T., Naitoh, Y. 1998. Intraductal
ultrasonography for the diagnosis of proximal
invasion in extrahepatic bile duct cancer. J.
Gastroenterol. Hematol. 13:715–719.
Menzel, J. and Dosmschke, W. 1999. Intraductal
ultrasonography (IDUS) of the pancreato-biliary
duct system. Personal experience and review of
literature. Euro. J. Ultrasound 10:105–109.
Menzel, J., Poremba, C., Dietl, K.H., Domschke,
W. 2000. Preoperative diagnosis of bile duct
strictures- Comparison of intraductal ultra-
sonography with conventional endosonography.
Scand. J. Gastroenterol. 35:77–82.
Palazzo, L. 1997. Which test for common bile duct
stones? Endoscopic and intraductal ultrasonog-
raphy. Endoscopy 29:655–665.
Silverstein, F.E., Martin, R.W., Kimmey, M.B.,
Jiranek, G.C., Franklin, D.W., Proctor, A. 1989.
Experimental evaluation of an endoscopic ultra-
sound probe: in vitro and in vivo canine studies.
Gastroenterology 96:1058–1062.
Stavropoulos, S., Larghi, A., Verna, E., Battezzati,
P., Stevens, P. 2005. Intraductal ultrasound for
32. Cholangiocarcinoma: Intraductal Sonography
the evaluation of patients with biliary strictures
and no abdominal mass on computed tomogra-
phy. Endoscopy 37:715–721.
Tamada, K., Ueno, N., Tomiyama, T., Oohashi,
A., Wada, S., Nishizono, T., Tano, S.,
Aizawa, T., Ido, K., and Kimura, K. 1998.
Characterization of biliary structures using
intraductal Ultrasonography: comparison
with percutaneous cholangioscopic biopsy.
Gastrointest. Endosc. 47:341–349.
Tamada, K., Inui, K., Menzel, J. 2001a. Intraductal
ultrasonography of the bile duct system.
Endoscopy 33:878–885.
437
Tamada, K., Nagai, H., Yasuda, Y., Tomiyama, T.,
Ohashi. A., Wada, S., Kanai, N., Satoh, Y., Ido,
K., Sugano, K. 2001b. Transpapillary intraductal
ultrasonography prior to biliary drainage in the
assessment of longitudinal spread of extrahe-
patic bile duct carcinoma. Gastrointest. Endosc.
53:300–307.
Wada, S., Tamada, K., Tomiyama, T., Ohashi, A.,
Utsunomiya, K., Higashisawa, T., Satoh, Y.,
Miyata, T., Sugano, K. 2000. Endoscopic micro-
wave coagulation therapy for bile duct cancer with
intraductal ultrasonographic monitoring: brief case
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
INTRODUCTION
Bile duct carcinomas (cholangiocarci-
noma) are devastating tumors. Surgical
resection is the mainstay of current treat-
ment; however, it is only effective for a
limited number of patients diagnosed at
early stages of the disease. Unfortunately,
the vast majority of patients seek treat-
ment of the advanced disease making
them poor candidates for curative surgery.
Conventional chemotherapy and radiation
therapy have not been shown to be effec-
tive in prolonging long-term survival of
patients with cholangiocarcinoma. In addi-
tion, there are no sensitive and specific
tests to diagnose and exclude this disease
in high-risk populations. Thus, there is
an urgent need for novel biomarkers and
therapies based on a better understanding
of biliary tumorigenesis that would signifi-
cantly impact the disease outcome. This
chapter reviews current clinical data on the
p53 protein family and its role in biliary
tumorigenesis.
Epidemiology
Cholangiocarcinoma is a malignant neo-
plasm arising from the epithelium lining
the bile ducts. Based on the location of
the primary tumor, this disease is classi-
fied into intrahepatic bile duct carcinoma
(peripheral cholangiocarcinoma), perihilar
cholangiocarcinoma (Klatskin tumor), and
distal bile duct carcinoma. Extrahepatic
bile duct (EBD) carcinoma (including
perihilar and distal cholangiocarcinomas)
arises from the hepatic hilum to distal
bile duct, and constitutes approximately
80–90 % of all cholangiocarcinomas
(Malhi and Gores, 2006). It usually occurs
in patients over 60 years of age and is
more frequently observed in males than
females (Hong et al., 2005). The EBD
carcinoma is a relatively uncommon neo-
plasm in the United States and European
countries, while the prevalence in Asia,
including Korea, Japan, and Vietnam, is
higher (Hong et al., 2005; Malhi and
Gores, 2006; Nagakawa et al., 2002). In
general, the worldwide incidence of EBD
carcinoma is slightly decreasing, while
that of intrahepatic cholangio-carcinoma
is increasing.
Etiology
Several different risk factors may give rise
to the injury and proliferation of the bile
441
442
duct epithelial cells. These factors are asso-
ciated with the development of EBD carci-
noma, which included primary sclerosing
cholangitis associated with inflammatory
bowel disease, hepatolithiasis, infection
with liver flukes (Clonorchis sinensis and
Opisthorchis viverrini), familial adeno-
matous polyposis coli, von Meyenberg
complexes, anomalous pancreaticobiliary
junction, and choledochal cysts (Choi
et al., 2004).
Diagnosis
The patients’ symptoms, which include
abdominal pain, jaundice, and weight
loss, are non specific. Therefore, the
evaluation of patients in regard to the
extent of EBD carcinoma is determined
by several radiologic examinations, such
as ultrasonography, helical-CT, direct
cholangiography, and MRCP (magnetic
resonance cholangiopancreatography)
(Seyama and Makuuchi, 2007). In some
situations, it is difficult to make a definite
diagnosis with these imaging tools. Bile
cytology or brushing biopsies are the
tools of choice in this case (Seyama and
Makuuchi, 2007).
Pathology
EBD carcinoma usually appears as an ill-
demarcated, firm, whitish tan mass with
induration or ulceration of the bile duct
wall. Adenocarcinoma is the most com-
mon histologic type, and consists of about
80 % of EBD carcinomas. Other histo-
logic variants that compose the remain-
ing 20 % include: papillary carcinomas,
intestinal-type adenocarcinomas, adenos-
quamous carcinomas, mucinous carcino-
mas, clear cell carcinomas, signet ring
cell carcinomas, small cell carcinomas,
A.I. Zaika and S.-M. Hong
and undifferentiated carcinomas (Hong
et al., 2005). By tumor growth feature,
EBD carcinoma can be classified as poly-
poid, nodular, and infiltrating types. One
of the often seen histologic characteristics
of EBD is marked desmoplastic stro-
mal reaction, which is extensive stromal
fibrosis surrounding tumor cells caused
by cancer cell infiltration. A few malig-
nant tumors from other organs can show
desmoplastic stromal reaction. Other than
desmoplastic stromal reaction, micro-
scopic findings of EBD carcinoma are
similar to adenocarcinomas from other
organs. Usually tumor cells form tubular
glands. Based on the degree of glandular
formation (histologic grade), EBD carci-
noma can be divided into well, moderate,
or poorly differentiated adenocarcinomas.
For example, when more than 75 % of all
tumor cells make glandular structures, it
can be classified as a well-differentiated
carcinoma. On the other hand, when less
than 25 % of cancer cells form tubular
structures, EBD carcinoma can be called
poorly differentiated carcinoma. Most
EBD carcinomas are classified as well to
moderately differentiated adenocarcinomas
(Albores-Saavedra et al., 2000). Nuclei of
tumor cells show significant difference
in size (pleomorphism), are stained more
darkly by hematoxylin and eosin staining
(hyperchromasia), and demonstrate atypi-
cal, asymmetric mitotic features. In about
10 % of EBD carcinoma cases, dysplasia is
observed in the peritumoral noncancerous
epithelial cells (Albores-Saavedra et al.,
2000). At the time of diagnosis, many EBD
carcinoma cases show tumor cell invasion
into lymphatic system, blood vessels, nerve
fibers, lymph nodes, and other adjacent
organs such as pancreas, duodenum, and
liver.
33. Extrahepatic Bile Duct Carcinoma: Role of the p53 Protein Family
Prognosis
The prognosis of EBD carcinoma is very
poor. For patients who undergo surgi-
cal resection, the 5-year survival rate is
approximately 20 % (Nathan et al., 2007).
Several factors are related to the prognosis
of patients with EBD carcinomas. The
most important factors are the extent of
tumor cell invasion (T classification of
TNM staging system) and metastasis to
lymph nodes (N classification). In addi-
tion, several other factors, including his-
tologic type, histologic grade, presence of
tumor cells at resection margin, vascular,
lymphatic, and perineural invasion, are
also reported to show significant survival
difference in patients with EBD carcino-
mas (Green et al., 2002).
ROLE OF P53 IN MALIGNANT
TUMORS OF THE BILE DUCT
Extensive molecular studies of cholan-
giocarcinoma significantly improved our
understanding of the tumorigenic proc-
esses in the bile duct. Aberrant activation
of several key oncogenes, such as recep-
tor tyrosine kinases Erb-b2, IL6R/gp130,
c-Met, EGFR, signaling molecules Ki-ras,
BRAF; anti-apoptotic proteins Mcl-1, Bcl-
2, Bcl-xL, and potent inducers of cellular
proliferation, COX-2 and β-catenin, have
been found in these tumors. Inactivation of
tumor suppressor genes TP53, APC, p16/
INK4a, DPC4 by genetic and epigenetic
mechanisms have also been reported. p53
deserves special attention, as this protein
plays a pivotal role in tumor suppression.
p53 is a transcription factor that at the
cellular level, is involved in the regula-
tion of multiple intracellular processes
including apoptosis, cell cycle progression,
443
cellular senescence, autophagy, differen-
tiation, and repair from genotoxic damage.
Inactivation of p53 is strongly associated
with an increased susceptibility to tumors
in laboratory animals and humans. In
fact, chronic bile duct injury provokes
cholangiocarcinoma in p53-deficient mice
(Farazi et al., 2006).
On average, one-half of cholangiocar-
cinoma patients carry mutations in the
p53 gene. Depending on the study, its
range is from 19–86 % which may reflect
differences in etiology between studied
populations, site of the disease along the
biliary tree and detection techniques. A
comprehensive analysis of p53 by Khan
et al. (2005) concluded that the presence
and type of p53 alterations are dependent
on environmental factors and might be
caused by changes in the composition and
metabolism of cytotoxic biliary constitu-
ents. Tumor location may also play a role
as the rate of p53 mutations is higher in
EBD carcinomas compared with proximal
tumors (Argani et al., 2001; Diamantis
et al., 1995). p53 inactivation occurs rela-
tively late in neoplastic progression in the
biliary tract. Murata et al. (2002) and Tan
et al. (2004) have suggested that immu-
nohistochemical staining for p53 is use-
ful for distinguishing cholangiocarcinoma
from benign lesions and determination of
tumor spread using preoperative biopsies.
The role of p53 mutations in predicting
patients’ survival has been examined in
several studies of cholangiocarcinoma.
Some studies have reported favorable
prognosis for patients lacking p53 muta-
tions (Ahrendt et al., 2000; Diamantis
et al., 1995). However, other investigators
have not found such association (Argani
et al., 2001; Lee et al., 1996; Suto et al.,
1997).
444
Analysis of p53 Mutations
For the study of p53 in tumors, the scien-
tific community employs two broadly used
techniques: immunohistochemical staining
(IHC) and various methods of mutational
analysis (sequencing, single strand confor-
mation polymorphism (SSCP) and others).
Immunochistochemical detection of p53
is based on the assumption that mutant
p53 is stabilized in cancer cells and can be
detected by staining with a p53-specific
antibody, typically DO-7 or CM1. IHC is
a useful and cost-effective approach for
detection of p53 mutations. However, the
IHC results should be interpreted with
caution as positive IHC does not neces-
sarily imply that p53 is fully inactive, nor
does the absence of staining mean that
p53 is functionally active (Munro et al.,
2005). More accurate information about
p53 mutations can be achieved by direct
sequencing alone or combined with SSCP.
Typically, mutations are determined in
exons 5–8 of the p53 gene, since the
majority of cancer-related mutations are
located in this region. Exons 5–8 encode
a highly conserved DNA binding domain
that is critical for the biological functions
of p53. However, 10–25 % of all muta-
tions occur outside exons 5–8 and include
a prevalence of frameshift, nonsense and
splice site mutations. Therefore, when it is
feasible, complete sequencing of the p53
gene is the preferable strategy.
ROLES P73 AND P63
IN MALIGNANT TUMORS
OF THE BILE DUCT
p53 is the founding member of the protein
family, which also includes the recently
A.I. Zaika and S.-M. Hong
discovered p63 and p73 proteins. These
proteins share significant structural and
functional similarities with p53. The high-
est similarity is found in the DNA-binding
domain, in which p53, p63 and p73 share
approximately 60 % of amino acid identity.
p73 and p63 proteins are expressed as
a complex variety of protein isoforms
that originate from two p73 and p63 gene
promoters (P1 and P2) and extensive gene
splicing at the NH2- and COOH-termini.
These mechanisms lead to the generation
of multiple isoforms with and without
transactivation domain. The latter isoforms
are termed ΔN (or ΔTA) isoforms (transac-
tivation-deficient) while the former are TA
isoforms (Figure 33.1). Recently, it has
been confirmed that the principal organi-
zation of the p53 gene is similar to that of
the TP73 and TP63 genes, and multiple
p53 isoforms are expressed in human
tumors.
A number of activities attributed to p53
are shared by TAp73 or TAp63 proteins.
These include induction of apoptosis, cell
cycle arrest and cellular senescence. p73
and p63 can bind to the p53 consensus
binding site and activate transcription of
most of the p53-target genes that are
involved in cell cycle regulation and apop-
tosis. It has been demonstrated that the
blockade of TAp73 or TAp63 with a domi-
nant-negative mutant or a small-inhibitory
RNA (siRNA) suppresses apoptosis in
cancer cells, irrespective of p53’s status
(Vilgelm et al., 2008). The combined loss
of p63 and p73 results in the failure of
mouse embryonic fibroblasts containing
p53 to undergo apoptosis in response to
DNA damage.
A recent study of p73- and p63-deficient
mice found that p73+/− (and p63+/−) hetero-
zygous animals developed both malignant
33. Extrahepatic Bile Duct Carcinoma: Role of the p53 Protein Family 445

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

and benign lesions. Similar to the p53+/− However, analyses of ΔN isoforms

mice, tumors from these animals undergo
a loss of the remaining wild-type allele.
Moreover, loss of p73 and p63 can coop-
erate with the loss of p53 function. These
data are consistent with the idea that p73
and p63 function as tumor suppressors
(Flores et al., 2005).
show that these proteins can function as
transcriptional inhibitors of p53, TAp63,
and TAp73 and bona fide oncoproteins.
Overexpression of ΔNp63 leads to a sig-
nificant increase in colony growth in soft
agar and xenograft tumor formation in
nude mice. ΔNp63 has been identified as a
446 A.I. Zaika and S.-M. Hong

downstream target of the phosphoinositide levels in the normal biliary epithelium. In

3-kinase (PI3K) pathway, a cell survival
and proliferation pathway in cancer.
Similarly, ΔNp73 facilitates cell immor-
talization and cooperates with oncogenic
Ras in transformation of mouse embryonic
fibroblasts. Implantation of these cells
into immunocompromised mice results in
tumor development.
Data from multiple tumors show that
loss-of-function mutations in p73 and p63
open reading frames are rare. Moreover,
p73 and p63 are frequently overexpressed
in human tumors. These data are consistent
with an increased titer of p73 antibodies in
patients with various types of cancers. In
a number of human tumors high-global
p73 and p63 protein expression correlates
with reduced patient survival suggesting
that these proteins play an important role
in tumorigenesis and may serve as useful
tumor markers.
Currently, a limited number of studies
have investigated p63 and p73 expression
in cholangiocarcinoma. Immunohistoche-
mical analyses revealed that both p73 and
p63 are present at low or undetectable
contrast, approximately 20 % of biliary EBD
displasias are positive for p63 and p73. The
immunoreactivity for these proteins is pri-
marily confined to the cell nuclei (Figure
33.2). The staining is further increased
in adenocarcinoma with 26 % and 41 %
p63- and p73-positive tumors, respectively.
Surrounding normal tissue and fibrovascular
stroma within the tumor are uniformly nega-
tive (Hong et al., 2007).
p63 is a good epithelial basal cell marker
that is frequently expressed in squamous cell
carcinomas. In rare cases of adenosquamous
carcinoma of the bile duct, p63 staining
was only noticed in the areas with squa-
mous carcinoma and not in adenocarcinoma.
Expression of p63 was also found in the areas
of squamous metaplasia. p63 staining was
significantly higher in EBD tumors located
in the distal bile duct and cases with vascular
invasion (Hong et al., 2007). Increased p63
expression was also found in intrahepatic
cholangiocarcinoma accompanied by liver
cirrhosis (Nomoto et al., 2006).
A study conducted by Tannapfel et al.
(1999) found that p73 expression associated

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

with patients’ survival in intrahepaticp63 transcripts. However, two important

cholangiocarcinoma. The median survivalaspects need to be taken into considera-
time for patients with p73-positive tumors
tion before designing the experiment. p73
was 132 days, whereas patients without and p63 are expressed as a complex set of
detectable p73 protein had a survival time
isoforms. Past studies that have searched
of 320 days. Because the antibody used in
for transcripts have thus far identified
this study can potentially recognize multi-
nine TAp73, three TAp63 transcripts, at
ple p73 isoforms, it remains unclear which
least six p73 mRNAs, and three p63 iso-
p73 isoform, TA or ΔN, is associated with
forms that lack the entire N-terminus (ΔN
the patients’ prognosis. isoforms) or its part (Figure 33.1). As a
In a large study that encompassed 175
result, RT-PCR primers should be carefully
cases of EBD cholangiocarcinoma, neither
designed to avoid detection of unrelated
p73 nor p63 was correlated with 1-, 3- and
isoforms. Detailed information about p73
5-year survival rates (Hong et al., 2007).
and p63 isoforms are provided in Vilgelm
However, analysis of tumors co-express-et al. (2008), primer sequences and PCR
ing p63 and p73 has found that combinedconditions in Concin et al. (2004) and
p63 and p73 expression is a significantHong et al. (2007). It is also important to
prognostic factor (Hong et al., 2007). note that p73 and p63 proteins are control-
This interesting phenomenon may find a led at post-translation levels and for that
mechanistic explanation in studies of non-
reason mRNA might not directly reflect
biliary tumors. Ellisen’s group at Harvard
protein expression (Vilgelm et al., 2008).
Medical School found that co-expression Proteins: Routine immunohistochemical
of p63 isoform, ΔNp63, with TAp73 leds techniques are applicable to the detection
to the inhibition of anti-proliferative and
of p73 and p63 proteins in formalin-fixed,
pro-apoptotic potential of TAp73 (Leongparaffin-embedded tumor tissues. Detailed
et al., 2007). It is plausible to suggest that
immunostaining protocol for p73 and p63
TAp73 is inhibited in cholangiocarcinoma
is described by Hong et al. (2007). Of note,
and ΔNp63 expression is one of the mul-current immunohistochemical applications
tiple mechanisms of p73 inactivation. This
are limited by availability of isoform-
is consistent with data showing that the
specific antibodies.
promoter responsible for expression of Western blotting has also been shown to
TAp73 (P1) is hypermethylated in a subset
be useful for the detection of p73 and p63
of extrahepatic and intrahepatic cholangi-
proteins in tumor tissues. Although this
ocarcinomas (Yang et al., 2005). Loss of
approach permits evaluation of all known
heterozygosity on the chromosome 1p36 isoforms, it requires high quality clinical
region that includes the TP73 gene was materials.
also noted in some intrahepatic cholangi- In summary, current analysis suggests
ocarcinoma (Momoi et al., 2001). that expression of p63 and p73 proteins
may serve as useful prognostic markers in
cholangiocarcinoma. These proteins may
Analyses of p63 and p73 in Tumors
also play an important role in tumorigenic
mRNA: RT-PCR is a robust and effi- processes in the bile duct. However, cur-
cient approach for analyses of p73 and rent studies need to be expanded to more
448
patients and interpreted based on isoform-
specific analysis.
REFERENCES
Ahrendt, S.A., Rashid, A., Chow, J.T., Eisenberger,
C.F., Pitt, H.A., and Sidransky, D. 2000. p53
overexpression and K-ras gene mutations in
primary sclerosing cholangitis-associated biliary
tract cancer. J. HepatoBiliary-Pancreat. Surg. 7:
426–431.
Albores-Saavedra, J., Henson, D.E., and
Klimstra, D.S. 2000. Tumors of the Gall blad-
der, Extrahepatic Bile Ducts and Ampulla of
Vater. Atlas of Tumor Pathology, Armed Forces
Institute of Pathology, Washington, DC.
Argani, P., Shaukat, A., Kaushal, M., Wilentz,
R.E., Su, G.H., Sohn, T.A., Yeo, C.J., Cameron,
J.L., Kern, S.E., and Hruban, R.H. 2001.
Differing rates of loss of DPC4 expression
and of p53 overexpression among carcinomas
of the proximal and distal bile ducts. Cancer
91: 1332–1341.
Choi, B.I., Han, J.K., Hong, S.T., and Lee, K.H.
2004. Clonorchiasis and cholangiocarcinoma:
etiologic relationship and imaging diagnosis.
Clin. Microbiol. Rev. 17: 540–552.
Concin, N., Becker, K., Slade, N., Erster, S.,
Mueller-Holzner, E., Ulmer, H., Daxenbichler,
G., Zeimet, A., Zeillinger, R., Marth, C., and
Moll, U.M. 2004. Transdominant DeltaTAp73
isoforms are frequently up-regulated in ovarian
cancer. Evidence for their role as epigenetic p53
inhibitors in vivo. Cancer Res. 64: 2449–2460.
Diamantis, I., Karamitopoulou, E., Perentes,
E., and Zimmermann, A. 1995. p53 protein
immunoreactivity in extrahepatic bile duct
and gallbladder cancer: correlation with tumor
grade and survival, Hepatology 22: 774–779,
Baltimore, MD.
Farazi, P.A., Zeisberg, M., Glickman, J., Zhang, Y.,
Kalluri, R., and DePinho, R.A. 2006. Chronic
bile duct injury associated with fibrotic matrix
microenvironment provokes cholangiocarci-
noma in p53-deficient mice. Cancer Res. 66:
6622–6627.
Flores, E.R., Sengupta, S., Miller, J.B., Newman,
J.J., Bronson, R., Crowley, D., Yang, A., McKeon,
F., and Jacks, T. 2005. Tumor predisposition in
A.I. Zaika and S.-M. Hong
mice mutant for p63 and p73: evidence for
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.,
Balch, C.M., Haller, D.G., and Morrow, M.
2002. AJCC Cancer Staging Manual, Springer,
New York.
Hong, S.M., Cho, H., Moskaluk, C.A., Yu, E., and
Zaika, A.I. 2007. p63 and p73 expression in ext-
rahepatic bile duct carcinoma and their clinical
significance. J. Mol. Histol. 38: 167–175.
Hong, S.M., Kim, M.J., Pi, D.Y., Jo, D., Cho, H.J.,
Yu, E., and Ro, J.Y. 2005. Analysis of extrahe-
patic bile duct carcinomas according to the New
American Joint Committee on Cancer staging
system focused on tumor classification problems
in 222 patients. Cancer 104: 802–810.
Khan, S.A., Thomas, H.C., Toledano, M.B., Cox,
I.J., and Taylor-Robinson, S.D. 2005. p53
Mutations in human cholangiocarcinoma: a
review. Liver Int. 25: 704–716.
Lee, Y.C., Song, S.Y., Chung, J.B., Kang, J.K.,
and Park, I.S. 1996. p53 protein expression in
extrahepatic bile duct cancer. Yonsei Med. J. 37:
112–117.
Leong, C.O., Vidnovic, N., DeYoung, M.P., Sgroi,
D., and Ellisen, L.W. 2007. The p63/p73 net-
work mediates chemosensitivity to cisplatin in
a biologically defined subset of primary breast
cancers. J. Clin. Invest. 117: 1370–1380.
Malhi, H., and Gores, G.J. 2006.
Cholangiocarcinoma: modern advances in
understanding a deadly old disease. J. Hepatol.
45: 856–867.
Momoi, H., Okabe, H., Kamikawa, T., Satoh,
S., Ikai, I., Yamamoto, M., Nakagawara, A.,
Shimahara, Y., Yamaoka, Y., and Fukumoto,
M. 2001. Comprehensive allelotyping of human
intrahepatic cholangiocarcinoma. Clin. Cancer
Res. 7: 2648–2655.
Munro, A.J., Lain, S., and Lane, D.P. 2005.
P53 abnormalities and outcomes in colorectal
cancer: a systematic review. Br. J. Cancer 92:
434–444.
Murata, T., Nagasaka, T., Kamiya, J., Nimura,
Y., Wakai, K., Yoshida, K., and Nakashima, N.
2002. P53 labeling index in cholangioscopic
biopsies is useful for determining spread of
bile duct carcinomas. Gastrointest. Endos. 56:
688–695.
33. Extrahepatic Bile Duct Carcinoma: Role of the p53 Protein Family
Nagakawa, T., Kayahara, M., Ikeda, S., Futakawa,
S., Kakita, A., Kawarada, H., Matsuno, M.,
Takada, T., Takasaki, K., Tanimura, H., Tashiro,
S., and Yamaoka, Y. 2002. Biliary tract cancer
treatment: results from the Biliary Tract Cancer
Statistics Registry in Japan. J. HepatoBiliary-
Pancreat. Surg. 9: 569–575.
Nathan, H., Pawlik, T.M., Wolfgang, C.L., Choti,
M.A., Cameron, J.L., and Schulick, R.D. 2007.
Trends in survival after surgery for cholangi-
ocarcinoma: a 30-year population-based SEER
database analysis. J. Gastrointest. Surg. 11:
1488–1496; discussion: 1496–1487.
Nomoto, K., Tsuneyama, K., Cheng, C., Takahashi,
H., Hori, R., Murai, Y., and Takano, Y. 2006.
Intrahepatic cholangiocarcinoma arising in cir-
rhotic liver frequently expressed p63-positive
basal/stem-cell phenotype. Pathol. Res. Pract.
202: 71–76.
Seyama, Y., and Makuuchi, M. 2007. Current
surgical treatment for bile duct cancer. World J.
Gastroenterol. 13: 1505–1515.
Suto, T., Sugai, T., Nakamura, S., Uesugi, N.,
Sasaki, R., Kanno, S., and Saito, K. 1997.
449
Immunohistochemical detection of proliferating
cell nuclear antigen and p53 expression in car-
cinoma of the extrahepatic bile duct. Oncology
54: 407–413.
Tan, G., Yilmaz, A., De Young, B.R., Behling,
C., Lehman, A., and Frankel, W.L. 2004.
Immunohistochemical analysis of biliary tract
lesions. Appl. Immunohistochem. Mol. Morphol.
12: 193–197.
Tannapfel, A., Engeland, K., Weinans, L., Katalinic,
A., Hauss, J., Mossner, J., and Wittekind, C.
1999. Expression of p73, a novel protein related
to the p53 tumour suppressor p53, and apoptosis
in cholangiocellular carcinoma of the liver. Br. J.
Cancer 80: 1069–1074.
Vilgelm, A., El-Rifai, W., and Zaika, A. 2008.
Therapeutic prospects for p73 and p63: Rising
from the shadow of p53. Drug Resist. Updat.
11: 152–163.
Yang, B., House, M.G., Guo, M., Herman, J.G., and
Clark, D.P. 2005. Promoter methylation profiles
of tumor suppressor genes in intrahepatic and
extrahepatic cholangiocarcinoma. Mod. Pathol.
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
INTRODUCTION
Malignant tumors of the extrahepatic bile
ducts are relatively rare, although they
are one of the most common causes of
extrahepatic bile duct obstruction. Most
tumors of extrahepatic bile ducts are car-
cinomas, and they can develop at any
level of bile duct tree. Extrahepatic bile
duct carcinoma (EHBDC) is difficult to
treat because most patients with this car-
cinoma are not surgically resectable at the
time of diagnosis despite its small size.
A gradual improvement with better prog-
nosis has been obtained with a combined
modality approach consisting of surgery,
radiation therapy, and chemotherapy, but
survival rate is still low. A complete surgi-
cal resection offers the only possibility of
cure; however, many patients treated with
curative resection have a poor outcome.
Our previous immunohistochemical stud-
ies of EHBDC showed that MUC4 mucin
expression is a new independent factor for
poor prognosis and predicts the outcome
of patients with EHBDC (Tamada et al.,
2006), in addition to MUC1 mucin expres-
sion (Tamada et al., 2002).
MUCIN CHARACTERISTICS
In general, mucins remain at the api-
cal surface of epithelial cells and are
capable of providing a lubricant func-
tion by their abilities to retain water ions
and their entanglement through intra- or
intermolecular disulfide linkages. Their
functions are not only protection and
lubrication of epithelial surfaces, but
also signaling modulators and affect-
ing tumor cell phenotype. Mucins are
high molecular weight glycoproteins with
oligosaccharides attached to serine or
threonine residues of the mucin core
protein backbone by O-glycosidic link-
ages, which are produced by various
epithelial cells. In the past few years,
20 core proteins for human mucin genes
(MUC) have been identified and des-
ignated as MUC1, MUC2, MUC3A,
MUC3B, MUC4, MUC5AC, MUC5B,
MUC6-13, MUC15-17, MUC19, and
MUC20 (Hollingsworth and Swanson,
2004). Mucins are categorized into mem-
brane-associated mucins (MUC1, MUC3
MUC4, MUC12, MUC16 and MUC17),
gel forming mucins (MUC2, MUC5AC,
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
to extracellular matrix proteins. In addi-
tion, overexpression of SMC masks the
surface antigens on target tumor cells
and effectively suppresses tumor cell kill-
ing by cytotoxic lymphocytes (Komatsu
et al., 1999). These phenomena may be
related with the poor outcome of patients
with MUC4 overexpression, although
MUC4 expression is not related to the
morphological invasive parameters, such
as lymphatic invasion, venous invasion,
perineural invasion, and distant lymph
node metastasis.
However, in EHBDCs examined in
our study, ErbB2 expression was not a
significant prognostic factor (Tamada
et al., 2006). The combined evaluation
of MUC4 and ErbB2 expression did not
show significant results. Low expression
of p27 is reported to be a poor prognos-
tic factor in gastric cancer (Sgambato
et al., 2000), colorectal carcinoma (Chen
et al., 2002), and intrahepatic cholan-
giocarcinoma (Taguchi et al., 2001). In
EHBDCs examined in our study, how-
ever, the p27 expression itself was not an
independent prognostic factor (Tamada
et al., 2006). In the combined evaluation,
the patients with MUC4 overexpression
and p27 overexpression did not show
significantly worse outcome than those
with MUC4 low expression and p27 low
expression. There may be a difference
between the cascade reaction in vitro and
the immunohistochemical localization in
vivo tumor tissues. In conclusion, impor-
tant information on the degree of MUC4
expression is a useful indicator to predict
the outcome of patients with EHBDCs,
who had surgical resection of the tumor.
MUC4 could potentially be a useful
marker for the clinical management of
patients with EHBDCs.
453
ANTIBODIES AGAINST MUC4
We use two monoclonal antibodies against
MUC4. One of the antibodies, clone 8G7,
was generated and provided by Surinder
K. Batra, University of Nebraska Medical
Center, Omaha, NE. The 8G7 shows high
specificity and avidity in ELISA, Western
blotting, immunoprecipitation, and in
immunohistochemical examination. This
antibody recognizes an epitope on the
16-amino-acid tandem repeats, which are
flanked by unique sequences (Moniaux
et al., 2004). The other antibody is
purchased from Invitrogen, South San
Francisco, U.S. (previously distributed by
Zymed). This antibody, clone 1G8, recog-
nizes an epitope on the ASGP-2 subunit,
which is one of the main subunits that
comprise the MUC4 glycoprotein complex
(data sheet). We have shown the represent-
ative positive immunohistochemical stain-
ing of MUC4 (Figures 34.1 and 34.2). The
difference between these two antibodies
consists of antibodies’ reactivity with the
endothelial cells of small blood vessels and
duodenal mucosal epithelia. The stained
by the clone 1G8 is seen in the blood ves-
sels and duodenal mucosal epithelia, in
addition to the cancer cells, whereas it by
the clone 8G7 is not observed in them.
MUC1 MUCIN
MUC1 is a polymorphic, highly glycosy-
lated, type I transmembrane protein
expressed by ductal epithelial cells of many
organs including gastrointestinal tract, pan-
creas, breast, and airway. MUC1 and MUC4
present significantly different structures
although they both belong to the same mucin
subfamily (membrane-bound mucins).
454
a a
b b
Figure 34.1. Expression of MUC4 in EHBDC.
(a). MUC4, clone 8G7 (b) MUC4, clone 1G8. The
clone 8G7 antibody is expressed in the cytoplasm
and/or membrane of the carcinoma cells, and is
negative for normal peribilliary glands (arrow-
heads). The clone 1G8 is weakly expressed in the
membrane of the carcinoma cells, and also stains
the endothelial cells of the small blood vessels. This is
negative for normal peribilliary glands (arrowheads)
MUC1 has a relatively simple architec-
ture. The MUC1 gene codes a membrane-
associated protein with distinct domains;
an amino-terminal domain consisting of a
20-amino acid signal peptide and degener-
ate tandem repeats; a large central domain
(two-thirds of the core protein sequence)
made up of variable numbers of a 20-amino
acid repeat; a carboxyl terminus made of
degenerate tandem repeats; a hydrophobic
membrane-spanning domain of 31 amino
acids, and a cytoplasmic domain of 69
M. Higashi et al.
Figure 34.2. Expression of MUC4 in duodenal
mucosa. a. MUC4, clone 8G7 The clone 8G7 is
not expressed in the duodenal mucosa. b. MUC4,
clone 1G8 The clone 1G8 stains the luminal sur-
face of the duodenal epithelia
amino acids (Hollingsworth and Swanson,
2004). In normal individuals, there is a
variation in the number of repeats per allele
(from 25 to > 125 repeats), implying that
the length of the repetitive array may be
critical for the normal function of MUC1.
The large extracellular central domain is
heavily O-glycosylated. Glycosylation
varies among different tissues, and tumor-
associated MUC1 has been reported to be
either less glycosylated or more glyco-
sylated than the forms expressed by normal
tissue (Higashi et al., 1999; Horinouchi
et al., 2003). Another important domain of
MUC1 is the cytoplasmic domain, which is
highly conserved across mammalian species
34. Extrahepatic Bile Duct Carcinoma: Mucin 4, a Poor Prognostic Factor
and hypothesized to play a role in its post-
translational processing, subcellular locali-
zation, signal transduction, and intracellular
localization (Hollingsworth and Swanson,
2004).
ANTIBODIES AGAINST MUC1
The most extensively studied mucin is
MUC1. As a potential tool in the immuno-
detection and immunotherapy of cancer,
a large number of monoclonal antibodies
have been raised against the protein core
of MUC1. Fifty six antibodies were clas-
sified at a workshop of the International
Society for Oncodevelopmental Biology
and Medicine in San Diego in 1996, and
the results have been summarized else-
where (1998). The majority of antibod-
ies against MUC1 defined epitopes are
located within the 20-amino acid tandem
repeat sequence of the MUC1 protein
core. Of the remaining antibodies, there
was evidence for the involvement of car-
bohydrate residues in the epitopes.
Several monoclonal antibodies against
MUC1 are available, which are used for
detecting different glycoforms of MUC1.
In the following section, we focus on four
monoclonal antibodies, NCL-MUC-1-
CORE (MUC1/CORE), DF3(MUC1/DF3),
MY. 1E12(MUC1/MY.1E12), and HMFG-
1(MUC1/HMFG-1). MUC1/CORE recog-
nizes MUC1 mucin core peptide (TRPAPG)
(data sheet from Novocastra Laboratories
Ltd., New Castle, U.K.). The MUC1/ DF3
(Toray-Fuji Bionics, Tokyo, Japan) identi-
fies MUC1 core peptide (TRPAPGS), but
MUC1/DF3 binding to protein is thought
to be enhanced by the presence of carbo-
hydrates. MUC1/ MY. 1E12 (developed in
the laboratory of Dr. T. Irimura, University
of Tokyo, Tokyo, Japan) is specific for
455
sialylated-MUC1 mucin (Yamamoto
et al., 1996). The binding of MUC1/
HMFG-1 (provided by Dr. Joyce Taylor-
Papadimitriou, Imperial Cancer Research
Fund laboratories, London, U.K.) to core
protein epitopes is influenced by the car-
bohydrate chains, and MUC1/HMFG-1
detects fully glycosylated MUC1 mucin,
and its binding is particularly affected by
sialic acid (Burchell et al., 1983).
The expression of various glycoforms
of MUC1 mucins was also recognized in
other human carcinomas of the stomach
(Utsunomiya et al., 1998), intrahepatic bile
duct (Higashi et al., 1999), and pancreas
(Horinouchi et al., 2003). Our study of
EHBDC demonstrated that MUC1 expres-
sion of poorly glycosylated MUC1 (MUC1/
CORE and MUC1/DF3) and glycosylated
MUC1 (MUC1/MY.1E12) showed a sig-
nificant relationship with tumor progre-
sion factors such as poor differentiation,
deep invasion, lymph node metastasis,
lymphatic invasion, or perineural inva-
sion. In contrast, the expression of MUC1/
HMFG-1 (fully glycosylated MUC1) did
not show a significant relationship with
tumor progression factors (Tamada et al.,
2002). Figure 34.3 shows representative
positive immunohistochemical stainings
of MUC1 in EHBDC tissues.
METHODS
Tissue Sample Management
Appropriate fixation is very important for
handling tissue samples. Less than opti-
mal fixation may result in worse results
not only in the histological examina-
tion, but also in the immunohistochemi-
cal examination. The resected specimens
should be fixed as soon as possible. Both
456 M. Higashi et al.

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
only for histological diagnosis, but also
for immunohistochemical and genomic
examination.
Here, we describe the proper handling
procedure of pancreaticoduodenectomy
specimens. First of all, the orientation of
duodenum is recognized. There are two
points to distinguish the proximal and dis-
tal sides of the duodenum. One of the dis-
tinguishing features is at the proximal free
end, which is shorter than that at the distal
free end. The other is that the stomach
(pyloric region), is sometimes attached at
the proximal margin. Subsequently, bile
duct should be identified. When the gall-
bladder is attached to the resected speci-
mens, it is relatively easy to identify the
bile duct. We follow the gallbladder, and
the cystic duct to common bile duct ante-
gradely. When the gallbladder has been
removed, we have to find the suture thread
of bile duct ligation. Usually, the distal
side of the bile duct is dilated; and these
procedures are easy. But in some cases it is
difficult to find the bile duct. In such cases,
we follow the bile duct from the papilla of
Vater retrogradely after opening the duo-
denum. We have to identify the anterior
and posterior sides before opening the
duodenum. The superior mesenteric vessel
sometimes indents the uncinate process to
form a groove called the pancreatic notch.
The opening of the duodenum should be
performed after the correct orientation of
the above sites has been noted. It is better
to open the duodenal wall near the pancreas
at the posterior side than opening it at the
opposite side (lateral side), because with
the former method it is easy to observe the
papilla of Vater. The common bile duct is
also opened at the posterior side after open-
ing the duodenum. A long metal probe of
small diameter can be used as a guide wire.
457
We should avoid opening both common
bile duct and main pancreatic duct, because
the shape of the resected specimens may
be distorted and the orientation will be
lost. Only common bile duct thus should
be opened. It is better to stop opening the
bile duct before the papilla of Vater for the
preservation of its structure of the papilla
of Vater. Photographs with a scale bar are
taken, and fixation is completed.
The orientation and the shape should be
adequately kept when chemical fixation is
performed. A corkboard, and pins or nee-
dles are useful tools, because the mucosa
of the duodenum and bile duct is kept in
the correct position when the mucosa is
immobilized with these tools on the cork-
board. The corkboard may be replaced
by a board made of Styrofoam or rubber.
Injection of fixative solution into the pan-
creas parenchyma before attaching it on
the board is also recommended in order
to avoid autolysis. An 18 gauge is suitable
for this purpose. Injection of too much
fixative should be avoided especially for
adipose tissue because the adipose tissue
shows edematous change.
The cases of simple extrahepatic bile
duct excision are relatively simply handled
when compared to those of pancreati-
coduodenectomy. After the identification
of the anatomical orientation, the bile
duct is opened in a longitudinal direction.
Injection fixation is useful if the dissection is
not done. Pinning is necessary in all cases.
All specimens are fixed in 10 % buffered
formalin or 10 % formalin. Sampling for
histological examination should be done
within 48 h after fixation. Prolonged fixta-
tion may produce formalin pigments and
loss of the antigenecity. Further fixation is
expected by cutting out thin blocks from
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
K. L. 2002. Muc4/sialomucin complex, the
intra-membrane ErbB2 ligand, induces spe-
cific phosphorylation of ErbB2 and enhances
expression of p27(kip), but does not activate
mitogen-activated kinase or protein kinaseB/Akt
pathways. Oncogene 21: 7524–7532.
Komatsu, M., Yee, L., and Carraway, K. L. 1999.
Overexpression of sialomucin complex, a rat
homologue of MUC4, inhibits tumor killing by
lymphokine-activated killer cells. Cancer Res.
59: 2229–2236.
Moniaux, N., Escande, F., Batra, S. K., Porchet,
N., Laine, A., and Aubert, J. P. 2000. Alternative
splicing generates a family of putative secreted
and membrane-associated MUC4 mucins. Eur. J.
Biochem. 267: 4536–4544.
Moniaux, N., Nollet, S., Porchet, N., Degand, P.,
Laine, A., and Aubert, J. P. 1999. Complete
sequence of the human mucin MUC4: a putative
cell membrane-associated mucin. Biochem. J.
338: 325–333.
Moniaux, N., Varshney, G. C., Chauhan, S.
C., Copin, M. C., Jain, M., Wittel, U. A.,
Andrianifahanana, M., Aubert, J. P., and Batra,
S. K. 2004. Generation and characterization
of anti-MUC4 monoclonal antibodies reactive
with normal and cancer cells in humans. J.
Histochem. Cytochem. 52: 253–261.
Porchet, N., Nguyen, V. C., Dufosse, J., Audie, J.
P., Guyonnet-Duperat, V., Gross, M. S., Denis,
C., Degand, P., Bernheim, A., and Aubert, J.
P. 1991. Molecular cloning and chromosomal
localization of a novel human tracheo-bronchial
mucin cDNA containing tandemly repeated
sequences of 48 base pairs. Biochem. Biophys.
Res. Commun. 175: 414–422.
Price, M. R., Rye, P. D., Petrakou, E., Murray,
A., Brady, K., Imai, S., Haga, S., Kiyozuka, Y.,
Schol, D., Meulenbroek, M. F., Snijdewint, F.
G., von Mensdorff-Pouilly, S., Verstraeten, R. A.,
Kenemans, P., Brockzjil, A., Nilsson, K., Nilsson,
O., Reddish, M., Suresh, M. R., Koganty, R. R.,
Fortier, S., Baronic, L., Berg, A., Longenecker,
M. B., Hilgers, J. et al. 1998. Summary report
on the ISOBM TD-4 Workshop: Analysis of 56
monoclonal antibodies against the MUC1 mucin.
Tumour Biol. 19(Suppl 1): 1–20.
459
Ramsauer, V. P., Carraway, C. A., Salas, P. J.,
and Carraway, K. L. 2003. Muc4/sialomucin
complex, the intramembrane ErbB2 ligand,
translocates ErbB2 to the apical surface in
polarized epithelial cells. J. Biol. Chem. 278:
30142–30147.
Sgambato, A., Migaldi, M., Leocata, P., Ventura,
L., Criscuolo, M., Di Giacomo, C., Capelli, G.,
Cittadini, A., and De Gaetani, C. 2000. Loss
of p27Kip1 expression is a strong independent
prognostic factor of reduced survival in N0 gas-
tric carcinomas. Cancer 89: 2247–2257.
Singh, A. P., Chaturvedi, P., and Batra, S. K. 2007.
Emerging roles of MUC4 in cancer: a novel
target for diagnosis and therapy. Cancer Res. 67:
433–436.
Singh, A. P., Moniaux, N., Chauhan, S. C., Meza,
J. L., and Batra, S. K. 2004. Inhibition of MUC4
expression suppresses pancreatic tumor cell growth
and metastasis. Cancer Res. 64: 622–630.
Taguchi, K., Aishima, S., Asayama, Y., Kajiyama,
K., Kinukawa, N., Shimada, M., Sugimachi, K.,
and Tsuneyoshi, M. 2001. The role of p27kip1
protein expression on the biological behavior of
intrahepatic cholangiocarcinoma. Hepatology
33: 1118–1123.
Tamada, S., Goto, M., Nomoto, M., Nagata, K.,
Shimizu, T., Tanaka, S., Sakoda, K., Imai, K.,
and Yonezawa, S. 2002. Expression of MUC1
and MUC2 mucins in extrahepatic bile duct car-
cinomas: its relationship with tumor progression
and prognosis. Pathol. Int. 52: 713–723.
Tamada, S., Shibahara, H., Higashi, M., Goto, M.,
Batra, S. K., Imai, K., and Yonezawa, S. 2006.
MUC4 is a novel prognostic factor of extrahe-
patic bile duct carcinoma. Clin. Cancer Res. 12:
4257–4264.
Utsunomiya, T., Yonezawa, S., Sakamoto, H.,
Kitamura, H., Hokita, S., Aiko, T., Tanaka, S.,
Irimura, T., Kim, Y. S., and Sato, E. 1998.
Expression of MUC1 and MUC2 mucins in gas-
tric carcinomas: its relationship with the prognosis
of the patients. Clin. Cancer Res. 4: 2605–2614.
Yamamoto, M., Bhavanandan, V. P., Nakamori, S.,
and Irimura, T. 1996. A novel monoclonal anti-
body specific for sialylated MUC1 mucin. Jpn.
J. Cancer Res. 87: 488–496.
C. Treatment
35
Hilar Cholangiocarcinoma: Photodynamic
Therapy and Stenting
Marcus Wiedmann, Joachim Mössner, and Helmut Witzigmann
SUMMARY
Carcinomas of the biliary tree are rare tumors
of the gastrointestinal tract with rising inci-
dence during the last years. Endoscopic
therapy plays a central role in the preopera-
tive and palliative treatment of extrahepatic
cholangiocarcinoma (CC). In obstructive
jaundice, biliary drainage may cause only
few complications and relieves symptoms
reliably. It can prevent further complica-
tions and is indispensable in the treatment
of cholangitis. The principal drawbacks
of biliary stents are stent occlusion and
cholangitis. Prophylactic application of
antibiotics and/or ursodeoxycholic acid has
not been shown to be effective. Technical
details concerning material, number and
location of stents await further clarification.
There are no prospective randomized tri-
als investigating the effect of preoperative
biliary drainage on morbidity and mortality
in jaundized patients before liver resection.
Palliative surgical cholangio-jejunostomy
with the bile duct of liver segment III for
Klatskin-tumors is rarely performed due
to the high success rate of stent therapy.
Self-expandable metal stents (SEMS) are a
safe, efficacious and less invasive treatment
option for palliation of duodenal obstruc-
tion in biliary tract cancer. Technical com-
plications can be managed endoscopically
and the bile duct remains accessible for
endoluminal treatment. Palliative photo-
dynamic therapy (PDT) for non-resectable
hilar CC is effective in prolonging survival,
improving biliary drainage and quality of
life, and represents a promising adjunct to
palliative therapy. Early referral for PDT
may provide maximum survival benefit.
In contrast, neoadjuvant and adjuvant PDT
for hilar CC is a new approach that needs
further evaluation.
INTRODUCTION
According to de Groen et al. (1999),
67 % of biliary tract tumors are gallbladder
carcinomas, 22 % are hilar carcinomas
(Klatskin tumors), 8 % occur in the dis-
tal bile duct, and 3 % are intrahepatic
cholangiocarcinomas (ICC). Since 1970
the incidence of ICC is increasing in North
America and Western Europe (Welzel
et al., 2006; West et al., 2006). The cause
of this rise is unknown and does not appear
to be explained simply by improvements
in diagnosis or changes in coding practice.
Besides regional etiological factors such as
liver fluke infection and hepatolithiasis, it
was hypothesized by Khan et al. (2002) that
463
464
this increase may be the result of a rise in one
or several genotoxic environmental agents,
causing cholangiocyte DNA damage.
At present, complete resection is the
only potentially curative therapy, but most
of the patients with intra- and extrahepatic
CC present with already advanced disease.
Given the small size of available studies
and lack of randomized trials, there is no
established role for neoadjuvant and adju-
vant therapy. In the palliative setting, sys-
temic chemotherapy (chemoradiation) has
not been clearly proven to prolong survival
significantly as discussed by Lazaridis and
Gores (2005) in a review article.
PREOPERATIVE BILIARY
DRAINAGE
Liver dysfunction caused by obstructive
jaundice is a relevant risk factor in major
hepatic resections. There are no prospec-
tive randomized trials investigating the
effect of preoperative biliary drainage
on morbidity and mortality in jaundized
patients before liver resection. In non-
randomized studies such as by Cherqui
et al. (2000) no difference in mortality
was observed between patients with or
without biliary drainage. No difference
with respect to morbidity was seen in three
studies and in our own series of 60 patients
with hilar CC who underwent resection.
On the other hand, higher complication
rates after biliary stenting were reported
by other authors. Despite these inconclu-
sive data, the authors of the majority of
recent studies dealing with hilar CC recom-
mend biliary drainage in order to decrease
the serum bilirubin level below 2–5 mg/
dl before resection (Ebata et al., 2003;
Hemming et al., 2005; Kawasaki et al.,
2003; Kondo et al., 2004; Kosuge et al.,
M. Wiedmann et al.
1999; Miyazaki et al., 1998; Seyama et al.,
2003; Todoroki et al., 2000). We performed
preoperative biliary drainage according to
these criteria in our patients and obtained
a median preoperative bilirubin level of
2.1 mg/dl. Recovery of hepatic function
after biliary decompression takes 4 weeks,
so surgery should not be done too early.
PALLIATIVE BILE DUCT
STENTING
The goal of palliative treatment is to achieve
stable long-term biliary drainage in order to
improve quality of life and prolong survival.
Refractory obstructive cholestasis leads to
death owing to severe bacterial cholangitis
or liver failure within 3–6 months. Because
surgical drainage procedures do not reveal
any advantage to nonsurgical palliation
according to studies such as by Kosuge
et al. (1999), biliary stenting is regarded as
the palliative method of choice. Endoscopic
approach for biliary stent insertion is pre-
ferred over an ultrasound-guided percuta-
neous approach due to a lower complication
rate. Furthermore, cholestasis was more
efficiently reduced and 30-day and overall
mortality rates were lower in a randomized
study (Speer et al., 1987). Percutaneous
transhepatic cholangiodrainage (PTCD)
(Figure 35.1) should be used after failure
of endoscopy and for anatomically dif-
ficult situations, i.e., after partial stomach
resection or gastrectomies, especially with
Y-Roux-anastomosis (Doctor et al., 1999).
Often in these cases the papilla Vateri can-
not be reached endoscopically. In selected
cases both methods can be combined as a
“rendezvous-approach”. A new technique
of endoscopic biliary-enteric bypass was
introduced by Burmester et al. (2003),
which comprises endosonographically
35. Hilar Cholangiocarcinoma: Photodynamic Therapy and Stenting
Figure 35.1. Klatskin-tumor Bismuth-Corlette
stage IV. Right-sided external–internal percuta-
neous transhepatic cholangiodrainage (PTCD;
Yamakawa-endoprosthesis; thick white arrows) and
left-sided pigtail endoprosthesis (thin white arrows)
which was inserted endoscopically (* gallbladder)
guided bile duct puncture followed by bil-
iary drainage into the duodenum.
The major complication of endoprosthesis
implantation is stent occlusion with consec-
utive cholangitis. Light and electron micro-
scopy studies of blocked stents revealed that
the material blocking the lumina was com-
posed of a matrix of bacteria and their fibril-
lar anionic extracellular products. Crystals
of calcium bilirubinate, calcium palmitate,
and cholesterol were embedded within this
matrix. Bacteria were attached to the stent
surface by a fibrillar matrix, suggesting
that the initial event in stent clogging is
the development of an adherent bacterial
biofilm. Bacterial enzyme activity (beta-
glucuronidase and phospholipase) leads to
the deposition of crystals. Therefore, it was
assumed that the use of antibiotics and/or
ursodeoxycholic acid may reduce bacte-
rial adhesion and stent clogging. However,
prophylactic long-term antibiosis and/or
465
ursodeoxycholic acid were useful in several
studies.
Technical questions regarding stent mate-
rial, number, and location of stents have
not been answered definitely. Polyurethane
and polyethylene seemed to be slightly
better in comparison to Teflon because of
a four-times lower occlusion rate accord-
ing to preclinical studies.
In a prospective randomized trial of Teflon
versus polyethylene stents for distal malig-
nant biliary obstruction by van Berkel et
al. (1998) Teflon material did not improve
patency in biliary stents with an Amsterdam-
type design. In this study, 42 patients received
a Teflon stent and 42 patients a polyeth-
ylene stent between September 1995 and
November 1996. All patients had a distal
malignant biliary stricture without a previ-
ous drainage procedure. Diagnoses included
carcinoma of the pancreas (n = 76), papilla
(n = 1), bile duct (n = 5) and metastases
(n = 2). The internal and external diam-
eter (10 French (F) ), length (9 cm) and stent
design (a straight stent with two side flaps
and one side hole at each end) were similar
for both stents. As a result, a reduction in
bilirubin of more than 20 % within one week
was seen in 91 % of the patients. Early com-
plication rates were similar in both groups
(10 %). The median follow-up was 142 days.
Stent dysfunction occurred in 28 Teflon and
29 polyethylene stents. The 30-day mortality
was 14 % in both groups. Patient survival did
not differ significantly between the groups
(median survival: Teflon 165 days, polyeth-
ylene 140 days). The median stent patency
was 83 days for Teflon and 80 days for
polyethylene stents, and was not significantly
different.
To assess whether the use of large stents
is justified, Speer et al. (1988) retrospectively
reviewed the results of 8 F stents with pig-
tails and 10 F straight stents in the palliation
466
of biliary obstruction due to malignancy.
The incidence of cholangitis following
stent insertion was significantly lower with
10 F stents compared with 8 F stents (5 %
vs. 34 %, p < 0.001). Stent survival until
blockage was significantly longer for 10 F
compared with 8 F (median 32 weeks vs.
median 12 weeks, p < 0.001).
In addition, Kadakia and Starnes (1992)
retrospectively compared the efficacy and
complications of 10 F biliary stents with
11.5 F stents in the management of malig-
nant and benign biliary tract diseases.
They concluded that 10 F stents have the
same success rate and complication rate as
11.5 F stents in the management of biliary
tract diseases; thus, a stent size larger than
10 F offers no significant advantage.
A new approach may be the use of
an Olympus DoubleLayer™ stent (DLS;
perfluoro alkoxy, without sideholes). In a
prospective randomized comparison of this
stent with the standard polyethylene stent
(with sideholes), including a total of 120
patients with jaundice due to malignant
strictures of the middle to distal third of
the common bile duct, 28 DLS patients
(47 %) and 17 polyethylene stent patients
(29 %) died without clinical evidence of
stent occlusion after a mean of 114 and
105 days, respectively (p < 0.05). Twenty-
six DLS patients (43 %) and 38 polyeth-
ylene stent patients (63 %) had symptoms
of stent clogging after a mean of 144
and 99 days, respectively (p < 0.05). Stent
dysfunction (stent orifice impacted in the
bile duct or duodenal wall, stent migra-
tion) was recorded in 6 DLS patients
(10 %) and 5 polyethylene patients (8 %)
(n.s.). Kaplan-Meier analysis of DLS and
PE stent clogging-free survival showed a
significantly longer patency period with
the DLS stents (p = 0.0005). Tringali et al.
M. Wiedmann et al.
(2003) concluded that these results showed
that DoubleLayer™ stents have a longer
patency period than polyethylene stents.
Another new development, as described
by Raju et al. (2006), is a 10 F “winged”
stent without a lumen to prevent stent occlu-
sion. Finally, our own group, Wiedmann
et al. (2004b), studied in a small case
series, whether bacterial colonization and
incrustration of stents can be avoided
by using an external-internal Yamakawa-
endoprosthesis in combination with a sub-
cutaneously implanted port system. A total
of three patients with bile duct obstruc-
tion (1 × recurrence of gastric cancer,
1 × invasive gallbladder cancer, 1 × recur-
rence of bile duct cancer) were included.
In all patients open biliary drainage and
prevention of cholangitis were achieved.
Therefore, this method is currently being
evaluated in a larger number of patients.
Cost analyses according to studies by Prat
et al. (1998) and Wagner et al. (1993) in
patients with middle and distal bile duct or
ampullary cancer showed that metallic stents
were advantageous in patients surviving
more than 6 months, whereas a plastic stent
was advantageous in patients surviving 6
months or less (Figure 35.2). Self-expanding
metal stents (SEMS) which can be inserted
endoscopically and percutaneously, have
the advantage of a higher patency rate with
consecutive lower rates of reinterventions
and rehospitalizations in comparison to less
costly plastic stents (Deviere et al., 1988;
Wagner et al., 1993). The higher patency
rate of metallic stents is related to the wider
lumen and better drainage of segmental bile
duct branches (Figure 35.3). These data were
confirmed by a single-center prospective
randomized controlled trial by Soderlund
and Linder (2006) comparing the patency of
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
steel Wallstents™ in patients with nonresect-
able malignant common bile duct strictures.
As a result, median patency times were 1.8
and 3.6 months in the polyethylene stents
and Wallstents™ groups, respectively (p =
0.002). Interestingly, when overall costs of
therapy, determined by adding the cost of
hospitalization to the costs of the stent, were
compared, there was no difference between
both groups. The authors recommended
the more-effective SEMS in unresectable
patients with malignant common bile duct
strictures, who survive a median of 4.5
months and less costly plastic stents for
patients who have distant metastases.
Despite the higher patency rate of SEMS,
stent dysfuncion can be induced by tumor
ingrowth, overgrowth or stent disloca-
tion. Stent dysfunction can be solved by
implantation of additional plastic or metal
stents, laser treatment or balloon dilatation.
Whether stent occlusion may be prevented
by covered SEMS or brachytherapy, is
currently unknown. A prospective ran-
domized study by Isayama et al. (2004)
evaluating patients with distal malignant
bilary duct obstruction showed that a
polyurethane covered Diamond™-stent is
superior to a non-covered metallic stent in
terms of patency rate. However, cholecys-
titis and pancreatitis rates were higher in
the covered metallic stent group. Overall
survival was equivalent in both groups.
In contrast, a recent study by Yoon
et al. (2006) comparing polyurethane-
covered Wallstents™ (that were used in
36 patients) with uncovered Wallstents™
(that were used in 41 patients) found no
significant difference between stent paten-
cies. Cholecystitis occurred in 1 patient of
the covered Wallstent™ group but in none
of the uncovered Wallstent™ group. Stent
occlusion occurred after a mean of 398
M. Wiedmann et al.
days in the covered group and after 319
days in the uncovered Wallstent group (p >
0.05). Stent patency rates were 83, 78, 67,
and 54 % at 100, 200, 300, and 400 days,
respectively, in the covered group and 83,
66, 54, and 36 % in the uncovered group,
which was not significantly different.
However, it must be mentioned that this
study was not a prospective randomized
study and the sample size was not large.
A prospective randomized study by
De Palma et al. (2001) revealed a lower
cholangitis rate after unilateral stent
insertion in comparison to bilateral stent
insertion. Howewer, thus far general con-
sensus regarding unilateral versus bilateral
bile duct drainage has not been obtained
(Table 35.1). In general, at least 25 % of
liver tissue should be drained and con-
trast medium should be injected only
in those bile duct segments that will be
finally drained. There is no need to drain
a certain liver lobe. It is better to drain the
easiest accessable one. In case of concur-
rent single-sided liver atrophy as a result
of tumor induced blood vessel occlusion,
the primarily intact liver lobe has to be
drained. A recent study by Freeman and
Overby (2003) showed that MRCP-guided
drainage of the widest dilated bile ducts
can improve results of stent therapy.
Despite controversial data with respect
to the benefit of unilateral or bilateral
biliary drainage, in studies by Liu et al.
(1998), Ducreux et al. (1992), Polydorou
et al. (1989), and Polydorou et al. (1991)
successful endoscopic drainage rate with
plastic stents was 41 % (20 % reduc-
tion of bilirubin within the first week),
53 % (30 % reduction of bilirubin within
the first week), 74 % (30 % reduction of
bilirubin within the first month), and 80 %
(30 % reduction of bilirubin within the
35. Hilar Cholangiocarcinoma: Photodynamic Therapy and Stenting 469

Table 35.1. Uni-versus bilateral biliary drainage.

Success Cholangitis 30 day Median
Number of patients rate rate mortality survival Reference
48 (retrospective: unilateral 96% 38% vs. 17% 29% vs. 8% 119 vs. 176 days Deviere et al.
vs. bilateral; attn: bilateral (p < 0.01) (p < 0.01) (p < 0.01) (1988)
contrast injection)
141 (retrospective: group A ? 6% vs. 0% ? Bismuth types II and Chang et al.
(one lobe opacified with vs. 32% III: 145 vs. 225 vs. (1998)
same lobe drained), 46 days (p < 0.001
B (both lobes opacified for A vs. C and
with both lobes drained), B vs. C; p < 0.001
and C (both lobes for C vs. A plus
opacified with one lobe B and for B vs.
drained)) A plus C)
157 (prospective: unilateral 89% vs. 77% 9% vs. 17% 11% vs. 14% 140 vs. 142 days De Palma
vs. bilateral) (p = 0.041) (p = 0.013) (p = 0.638) et al. (2001)

first month), respectively. In contrast, the

ally represents a terminal event, but also
randomized study by Ortner et al. (2003), grossly impairs patients’ quality of life.
which contained 79 % Bismuth type IV We found an incidence of 25 % in our
tumors, showed a decrease of cholestasis own series of 80 biliary tumors. Relief
by at least 50 % within the first week in of obstruction to allow enteral feeding,
only 21 % with stenting alone. In our own thereby improving weight and possibly
study, Witzigmann et al. (2006), hyper- increasing life expectancy, is the primary
bilirubinemia was decreased by at least aim in these patients. Surgical bypass is the
50 % in 39 % of patients three months standard treatment with a reported success
after stenting despite a high percentage rate up to 90 %. The procedure carries a
of Bismuth-Corlette type IV tumors. Very high complication rate of 25–35 % and a
important for long-term success of stent perioperative mortality of 2 % determined
therapy is a close monitoring of patients in studies such as that by Lillemoe et al.
for early detection of latent biliary sepsis.
(1999). A procedure that prolongs hospital
As part of this monitoring plastic endo- stay and increases costs dramatically may
prostheses should be routinely exchanged not always be appropriate, especially in
every 3 months (Prat et al., 1998). terminally ill patients with a poor clinical
condition. Previous endoscopic interven-
tions like balloon dilatation or laser reca-
PALLIATIVE STENTING OF nalization provide only transient success,
MALIGNANT DUODENAL resulting in high reocclusion rates.
OBSTRUCTION Based on the experience in palliation of
obstructive esophageal and biliary tract can-
Malignant gastric outlet obstruction or cer, SEMS also have been used for malig-
duodenal obstruction is a complication of nant gastric outlet and duodenal obstructions
advanced hepatobiliary carcinomas. It usu- since the early 1990s. In contrast to surgical
470 M. Wiedmann et al.

bypass surgery, endoscopic placement of mostly tumor tissue. The mechanisms by

duodenal SEMS is a less invasive procedurewhich this localization occurs are complex
possibly resulting in shorter hospitalization
and not fully understood. High vascular
time and lower costs. However, some stud- permeability of the agents, their affin-
ies reported complications that may increase
ity for proliferating endothelium, and the
the length of hospital stay, especially tumor
lack of lymphatic drainage in tumors may
ingrowth, tumor overgrowth, and stent dis-contribute to an accumulation in tumors.
location. In our own prospective study, com-
Moreover, tumors might have increased
prising 20 patients with biliary-tract-cancer
lipid content, elevated numbers of low-
(7 × gallbladder, 13 × Klatskin tumors) density lipoprotein receptors, abnormal
successful stent placement was achieved invasculature, and decreased pH. In a second
all patients (Schiefke et al., 2003). An addi-
step, non-thermal laser light of a specific
tional stent was required in six cases (four
wavelength is applied, and adapted to
occlusions, two dislocations). Median sur-the absorption spectrum of the sensitizer
vival was 20.5 weeks, no treatment relatedto be excited. Following the absorption
deaths occurred. Twenty-eight biliary stent
of light, the compound is transformed
exchanges were performed in 13/20 (65 %) into a relatively long-lived electronically
patients. Erosive reflux esophagitis improved
excited state via a short-lived excited sin-
in 11/12 (92 %) cases. After 4 weeks, all 17
glet state.
surviving patients tolerated soft-solid or This so-called triplet state can undergo
solid food, while 13/17 (77 %) tolerated atwo kinds of reaction. In the type I reac-
more solid diet (p < 0.001, gastric-outlet-
tion, it can react directly with a substrate,
obstruction-scoring-system). 12/17 (71 %) such as the cell membrane or a molecule,
patients gained a median of 1.5 kg weight and transfer a hydrogen atom or electron
(p = 0.001). The median Karnofsky-Index to form radicals. The radicals interact with
increased from 50 % to 60 % in 13/17 (77 %)
oxygen to produce oxygenated products.
patients. In conclusion, SEMS is a safe, effi-
Alternatively, in the type II reaction (which
cacious and low invasive treatment option is more common), the triplet can transfer its
for palliation of duodenal obstruction in energy directly to oxygen to generate singlet
biliary tract cancer. Technical complications
oxygen (Figure 35.4). Therefore, the effects
can be managed endoscopically and the bileof both PDT reaction types are oxygen-
duct remains accessible for endoluminal dependent. Tumor destruction, mediated
treatment. by oxygenated products and singlet oxy-
gen, can be explained by several mecha-
nisms causing necrosis and apoptosis. First,
PDT kills tumor cells directly; secondly, it
PHOTODYNAMIC THERAPY damages the tumor-associated vasculature,
(PDT) leading to tumor infarction, and finally can
activate an immune response against tumor
Mechanism of Action
cells. In detail, PDT has been demon-
Photodynamic therapy is based on the strated to induce (1) direct cytotoxicity by
administration of a photosensitizer which degeneration of lipids in cell membranes,
localizes selectively within the target, (2) microvascular damage from platelet
35. Hilar Cholangiocarcinoma: Photodynamic Therapy and Stenting 471

Figure 35.4. Mechanism of action of photodynamic therapy (PDT)

activation and thromboxane-induced vessel
constriction and thrombus formation, prob-
ably mediated by nitric oxide, (3) cytosolic
and mitochondrial membrane damage with
cytochrome C release, bcl-2 depletion and
induction of apoptosis, and (4) release
of inflammatory mediators that induce
T-lymphocyte mediated cellular immune
response. The extent of photodamage and
cytotoxicity is multifactorial and depends
on the type of photosensitizer used, its
localization and administered dose, the
light source used, generating the light expo-
sure dose and light fluence rate, the oxygen
availability in the target, and finally the
time interval between the administration of
the photosensitizer and the light treatment.
Because PDT is a cold photochemical proc-
ess, there is no tissue heating, and connec-
tive tissues such as collagen and elastin are
largely unaffected.
Contraindications for Photodynamic
Therapy
Photodynamic therapy should not be per-
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-
menting the success of PDT performed via
percutaneous cholangioscopy in a single
patient with incompletely resected bile
duct carcinoma (McCaughan et al., 1991).
This type of treatment caused a prolonged
survival time of > 4 years in a tumor
entity with a median survival of only 4–6
months. Ortner et al. (1998) performed the
first prospective non-randomized single-
arm study including nine patients with
advanced Bismuth type III and IV hilar
CC, who showed no sufficient drainage
after endoscopic stent insertion. Two days
after intravenous application of porfimer
sodium (Photofrin™) at 2 mg/kg body
weight, intraluminal photoactivation at
630 nm was performed cholangioscopi-
cally with laser diffusing fibers. After PDT,
bilirubin serum levels declined signifi-
cantly (p = 0.004) with no increase during
the two monthly follow-ups. Quality of
life indices improved dramatically and
remained stable. Thirty-day mortality was
0 %, and median survival 439 days. Side
effects were limited to skin hyperpigmen-
tation, but no acute phototoxicity in all
patients. One patient also reported fever
and pain, self-limited, for 2 weeks after
PDT (Table 35.1). Ortner’s work (Ortner,
2000) was subsequently expanded to a
larger study of 21 patients (Table 35.3).
We performed our own study (Berr
et al., 2000b), including 23 patients with
hilar CC (Bismuth type III, n = 2; type IV;
n = 21) who were treated with a combina-
tion of bile duct stenting and endocopi-
cally performed Photofrin™-PDT (Figure
35.5). In detail, we intubated the tumor
stenosis with a translucent ERC-cannula
and inserted a 400 μm thin quartz fiber
with either a 2, 3, or 4 cm long cylindrical

Table 35.3. Palliative Photodynamic Therapy (PDT) in Hilar Cholangiocarcinoma.

Protocol n Photosensitizer Drainage Survival# Reference
Phase I 1 Dihemato- T-Drain 48 McCaughan et al. (1991)
porphyrin 2
Phase II 9 Photofrin II plastic stent 14.6 (3–19.2) Ortner et al. (1998)
Phase II 21 Photofrin II plastic stent 15 Ortner (2000)
Phase II 23 Photofrin II plastic stent 12.0 (10.4–13.6) Berr et al. (2000);
Wiedmann et al. (2004)
Phase II 8 Photosan-3 plastic stent 4.0 (1.7–14.8) Zoepf et al. (2001a)
Phase II 4 5-ALA and PTCD NA Zoepf et al. (2001b)
Phase II 6 Photofrin II plastic stent NA Rumalla et al. (2001)
Phase II 24 Photofrin II metallic stent 9.9 (6.4–13.4)+ Dumoulin et al. (2003)
Phase III 20 Photofrin II plastic stent 16.4** Ortner et al. (2003)
19 plastic stent 3.3
Phase II 8 Photofrin II plastic stent 9.2+ Harewood et al. (2005)
Phase II 36 Photofrin II plastic stent 11.2 (2.5–17.3) Pereira et al. (2005b)
Phase II 24 Hematoporphy- PTCD 18.6 (2.1–27) Shim et al. (2005)
rine derivative
Phase II 25 Photofrin II plastic stent 8.1 Prasad et al. (2005)
Phase III 16 Photosan-3 plastic stent 21* (3–30) Zoepf et al. (2005)
16 plastic stent 7 (1–24)
Phase II 68 Photofrin II plastic stent 12** (8.8–15.3) Witzigmann et al. (2006)
56 plastic stent 6.4 (4.2–8.5)
NA = non-applicable; #median (range) in months from diagnosis; *p < 0.05; **p < 0.01; +median (range) in months from therapy
35. Hilar Cholangiocarcinoma: Photodynamic Therapy and Stenting
Figure 35.5. Hilar cholangiocarcinoma Bismuth
type IV, (a) tumor imaging with endoscopic retro-
grade cholangiography (ERC; white arrows), (b)
reopened bile ducts 3 months after photodynamic
therapy (PDT) and stent insertion (white arrows)
light diffuser and radioopaque markers
(Lambda Plaus PDL-2, Coherent, Dieburg,
Germany). The power emmited by the
diffuser tip through the ERC cannula
had been calibrated to 0.400 W/cm before
PDT and checked thereafter using an inte-
475
grating sphere power meter. The tumor
stenoses were exposed to laser light for
10 min with a light dose of 242 J/cm2.
The patients showed a 6 month survival
rate of 91 % after diagnosis and 74 %
after start of PDT with a median local
tumor response of 74 %, 54 %, 29 %, and
67 % after the first, second, third, and
fourth PDT session, respectively (Table
35.3). Cholestasis, performance, and qual-
ity of life of the patients improved clearly.
Cholangitis rate with PDT was not higher
than in historical control patients with
Bismuth type III tumors and bile duct
drainage, only. After a 5 year follow-up,
median survival time after diagnosis was
18 months for patients without perito-
neal carcinosis (n = 19) and 12 months
for all patients (n = 23) (Wiedmann et al.,
2004a). Survival was 63 %, 26 %, 16 %,
and 5 % after one, two, three, and four
years, respectively. Of the patients who
died, 73.9 % (n = 17) died because of
tumor progression; 26.1 % died as a result
of cholangitis (n = 4), septic shock (n = 1),
or appendicitis/peritonitis (n = 1). For all
patients, except one with diffuse liver
metastases, there was improvement in
cholestasis, performance, and quality of
life, which was maintained for an extended
period of time. These data were confirmed
by a similar recent phase II study by Pereira
et al. (2005b) including 36 patients.
Zoepf et al. (2001a) described the use
of a novel laser diode PDT system in
the treatment of 8 patients with Bismuth
II-IV non-resectable CC using Photosan-3
(2 mg/kg; dihematoporphin ether) and a
light dose of 200 J/cm2 at a wavelength of
633 nm (Table 35.3). Four weeks after ini-
tial PDT, which was performed transpapil-
lary (n = 4) or percutaneously (n = 4), all
patients showed a marked reduction of bile
476
duct stenosis. The median serum bilirubin
value declined from 5.8 to 1.0 mg/dl. The
median survival time was 119 days (52–
443 days).
The therapeutic value of percutane-
ous transhepatic PDT in patients with
advanced bile duct cancer was the subject
of a prospective trial by Shim et al. (2005).
The utility of intraductal ultrasonography
(IDUS) for the assessment of responses
and for regular follow-up after PDT was
also examined. PTCD was initiated before
PDT. Following dilation and maturation
of the PTCD tract, percutaneous PDT was
performed. Intraluminal photoactivation
was carried out using percutaneous cholan-
gioscopy 2 days after intravenous applica-
tion of a hematoporphyrin derivative. All
patients were provided additionally with
percutaneous bile duct drainage catheters
after PDT. IDUS was conducted monthly
to measure the thickness of the tumor
mass before and after PDT. Twenty-four
patients with advanced cholangiocarcino-
mas (Bismuth IIIa, n = 4; IIIb, n = 10; IV,
n = 10) were treated with PDT. At 3 months
after PDT, the mean thickness of the tumor
mass had decreased from 8.7 ± 3.7 to 5.8 ±
2.0 mm (p < 0.01). At 4 months after PDT,
the thickness of the mass had increased
to 7.0 ± 3.7 mm. Quality of life indices
improved dramatically and remained sta-
ble 1 month after PDT; the Karnofsky
index increased from 39 to 58 points (p =
0.003). The 30-day mortality rate was 0 %,
and the median survival time was 558 days
(range 62–810 days).
Another single-arm study by Rumalla
et al. (2001) described a new method of
applying photodynamic therapy in the
biliary tract by using accessories avail-
able in the United States (Table 35.3).
Patients were injected with 2 mg/kg of
M. Wiedmann et al.
sodium porfimer, and 48 h later a com-
mercially available cylindrical diffusing
laser fiber was inserted into an 8 F biliary
catheter equipped with a 0.038 in. side-
hole at its distal tip. After positioning
the 0.035 inch guidewire proximal to
the biliary stricture, the preloaded cath-
eter was advanced over the guidewire
by using the monorail technique. Laser
light was applied at a power of 400 mW/
cm fiber for a total energy of 180 J/cm2.
By using the preloaded biliary catheter,
adequate positioning of the laser fiber
was achieved in all patients. A fracture of
the diffuser tip occurred during one of the
treatments. Two patients developed acute
cholangitis and two patients experienced
skin phototoxicity. The applicability of
this technique was confirmed in recent
studies by Harewood et al. (2005) and
Prasad et al. (2005) (Table 35.3). It was
concluded from the first study that (1)
PDT treatment every 3 months is applica-
ble, (2) patient selection for PDT need not
to be based on response to primary biliary
stenting, and (3) the average cost of treat-
ment per patient would be US$10,337,
which is based on direct medical costs
estimated from USA Medicare ambula-
tory patient classification (ABC, 2003)
plus professional fees for hospital-based
outpatient procedures, which can be com-
pared with the costs of management with
biliary stenting alone. The conclusion
from the second study was that early
treatment with PDT may lead to greater
preservation of liver function based on
a univariate and multivariate analysis,
showing a longer interval from diagnosis
to PDT treatment, which was significantly
associated with shorter survival following
PDT (HR 1.23, 1.02–1.47; p = 0.029 and
HR 2.38, 1.38–4.85; p = 0.0001).
35. Hilar Cholangiocarcinoma: Photodynamic Therapy and Stenting
Another prospective phase II study by
Dumoulin et al. (2003) investigated PDT
and consecutive metal stent insertion for
palliation of hilar CC. It was feasible but
there was only a modest benefit in overall
survival in comparison to a historical con-
trol group (Table 35.3). In addition, a small
study by Zoepf et al. (2001b) investigated
the use of 5-aminolevulinic acid (5-ALA)
for the palliative treatment of hilar CC.
Light activation was performed 5 to 7
hours after oral administration of 5-ALA.
All patients had an endoprosthesis placed
in the bile duct after PDT. However, 4
weeks after PDT, 5-ALA had failed to
significantly reduce malignant bile duct
obstruction, and thus cannot be recom-
mended for the palliative treatment of bile
duct cancer (Table 35.3). Finally, a pro-
spective analysis of resectional and pallia-
tive treatment of 184 patients with hilar CC
was reported from our hospital (60 patients
underwent resection, 8 after neoadjuvant
PDT, 68 had PDT in addition to stenting
and 56 had stenting alone) (Witzigmann
et al., 2006). During a 10-year period it was
revealed that PDT and stenting resulted in
longer median survival (12 vs., 6.4 months,
p < 0.01), lower serum bilirubin levels
(p < 0.05), and higher Karnofsky perform-
ance status (p < 0.01) as compared to
stenting alone. Median survival after PDT
and stenting, but not after stenting alone,
did not differ from that after both R1 and
R2 resection. However, these improve-
ments in palliative treatment by PDT will
not change the concept of an aggressive
resectional approach.
The first prospective, randomized multi-
center trial by Ortner et al. (2003), includ-
ing our hospital, confirmed preliminary
data from the above mentioned non-ran-
domized studies. Patients with hilar CC
477
fulfilling inclusion criteria (non-resectable
tumor, tumor diameter > 3 cm, tumor
clearly visible on CT and ERCP, unequiv-
ocal positive histology, no evidence of
cancer of another organ) were randomized
to group A (stenting and subsequent PDT)
and group B (stenting alone). For PDT,
Photofrin™ at 2 mg/kg body weight was
injected intravenously 2 days before intra-
luminal photoactivation (wavelength,
630 nm; light dose, 180 J/cm2). Further
treatments were performed in cases of
residual tumor in the bile duct. Oral cipro-
floxacin therapy with 250 mg twice daily
was started before ERCP and continued
for 14 days. The primary outcome param-
eter was survival time. Secondary outcome
parameters were cholestasis and quality
of life. PDT resulted in prolongation of
survival (group A: n = 20, median 493
days; group B: n = 19, median 98 days;
p < 0.0001). It also improved biliary
drainage and quality of life. The study
was terminated prematurely because PDT
proved to be so superior to simple stent-
ing treatment that further randomization
was deemed unnecessary. However, the
study was criticized for the short duration
of survival in the control group, mainly
caused by insufficient biliary drainage
by stenting alone. Nevertheless, it was
remarkable that an additional group of 31
patients who were excluded from rand-
omization because of a statistically signifi-
cant lower Karnovsky performance status,
but received PDT treatment voluntarily,
performed as well as the randomized PDT
group (Table 35.3). In summary, pallia-
tive PDT for non-resectable hilar CC was
effective in restoring biliary drainage and
improving quality of life, survival was
prolonged; thus, PDT represents a promis-
ing adjunct to palliative therapy.
478 M. Wiedmann et al.

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
penetration depth of 2 mm of 5-aminole-
vulinic acid (5-ALA, Levulan™), a prodrug
of heme biosynthesis and a photosensitizer
of the second generation. In the future,
further progress may be made by using
photosensitizers with a higher penetration
depth, such as Temoporfin (Foscan™) with
a strong absorption at 652 nm or tetrahy-
droporphyrin tetratosylat (THPTS) with a
strong absorption at 760 nm, in combina-
tion with cheaper laser systems and better
light applicators.
A first preliminary study by Pereira et
al. (2005a) tried to determine the safety
and efficacy of Foscan™-PDT in the treat-
ment of patients with irresectable malignant
biliary strictures and recurrent stent occlu-
sion. Thirteen patients with malignant
biliary obstruction due to carcinoma of
the biliary tract (n = 5), ampulla (n = 5),
pancreas (n = 2) or stomach (n = 1), were
studied. All had been initially palliated with
metal (n = 10) or polyethylene (n = 3)
biliary stents, and presented with recurrent
obstructive jaundice. Patients received
0.15 mg/kg Foscan™ intravenously 72 h
before endoluminal light activation with
an endoscopically-placed optical fiber
(4–5 cm diffusing tip, 652 nm light at
10–50 J, 100 mW/cm diffuser fiber), fol-
lowed by polyethylene stent insertion in
all but one patient. Before PDT, patients
had a median of 3 (0–5) stent occlusions
over 13 (2–27) months, with a median
patency of plastic stents placed inside
metal stents for recurrent stent occlusion
of 3.5 (0.5–13) months. After PDT, tumor
necrosis and/or metal stent recanalization
was seen in all patients, with a median of
0 (0–3) stent occlusions during 17 (1–78)
months follow-up. The median patency
of plastic stents inside metal bile duct
stents after PDT was 5 (1–52) months.
479
The median survival after diagnosis and
PDT administration was 21 (10–90) and
9 (1–27) months, respectively. PDT was
generally well-tolerated but two patients
developed significant hemobilia, one of
whom died with a gallbladder empyema 2
weeks after treatment.
Further modifications of Foscan™, like
a liposomal formulation (Foslip™) or
pegylation (Fospeg™), aim to reduce pho-
tosensitizer dose, side effects, and costs. In
addition to the development of new pho-
tosensitizers, other innovative ideas are
currently under investigation. For instance,
fractionated drug-dose PDT regimens were
reported to result in a superior therapeutic
effect, compared to single-dose regimens,
and were able to induce long-term tumor
growth control. Moreover, the use of spe-
cific targeting carriers, such as conjugated
antibodies directed to tumor-associated
antigens or vascular antigens are supposed
to direct the photosensitizer to a certain cell
type or compartment. Others suggest the
use of different advanced delivery systems,
such as ligand-based targeting with insu-
lin, epidermal growth factor or adenoviral
proteins, protease-mediated drug delivery,
photosensitizing adenoviruses, water-sol-
uble polymer carriers, and pH-responsive
polymeric micelles. A completely different
approach is the protection of normal tissue
with drugs like WR-2721 and WR-77913
from possible PDT side effects if the pho-
tosensitizer is used in a high dose.
REFERENCES
Berr, F., Tannapfel, A., Lamesch, P., Pahernik,
S., Wiedmann, M., Halm, U., Goetz, A. E.,
Mossner, J., and Hauss, J., 2000a. Neoadjuvant
photodynamic therapy before curative resection
of proximal bile duct carcinoma. J. Hepatol. 32:
352–357.
480
Berr, F., Wiedmann, M., Tannapfel, A., Halm, U.,
Kohlhaw, K. R., Schmidt, F., Wittekind, C.,
Hauss, J., and Mossner, J., 2000b. Photodynamic
therapy for advanced bile duct cancer: evidence
for improved palliation and extended survival.
Hepatology 31: 291–2988.
Burmester, E., Niehaus, J., Leineweber, T., and
Huetteroth, T., 2003. EUS-cholangio-drainage
of the bile duct: report of 4 cases. Gastrointest.
Endosc. 57: 246–251.
Chang, W. H., Kortan, P., and Haber, G. B.,
1998. Outcome in patients with bifurcation
tumors who undergo unilateral versus bilateral
hepatic duct drainage. Gastrointest. Endosc. 47:
354–362.
Cherqui, D., Benoist, S., Malassagne, B., Humeres,
R., Rodriguez, V., and Fagniez, P. L., 2000.
Major liver resection for carcinoma in jaundiced
patients without preoperative biliary drainage.
Arch. Surg. 135: 302–308.
de Groen, P. C., Gores, G. J., LaRusso, N. F.,
Gunderson, L. L., and Nagorney, D. M., 1999.
Biliary tract cancers. N. Engl. J. Med. 341:
1368–1378.
De Palma, G. D., Galloro, G., Siciliano, S., Iovino,
P., and Catanzano, C., 2001. Unilateral versus
bilateral endoscopic hepatic duct drainage in
patients with malignant hilar biliary obstruction:
results of a prospective, randomized, and control-
led study. Gastrointest. Endosc. 53: 547–553.
Deviere, J., Baize, M., de Toeuf, J., and Cremer,
M., 1988. Long-term follow-up of patients with
hilar malignant stricture treated by endoscopic
internal biliary drainage. Gastrointest. Endosc.
34: 95–101.
Doctor, N., Dick, R., Rai, R., Dafnios, N., Salamat,
A., Whiteway, H., Dooley, J., and Davidson, B.
R., 1999. Results of percutaneous plastic stents
for malignant distal biliary obstruction following
failed endoscopic stent insertion and comparison
with current literature on expandable metal-
lic stents. Eur. J. Gastroenterol. Hepatol. 11:
775–780.
Ducreux, M., Liguory, C., Lefebvre, J. F., Ink, O.,
Choury, A., Fritsch, J., Bonnel, D., Derhy, S.,
and Etienne, J. P., 1992. Management of malig-
nant hilar biliary obstruction by endoscopy.
Results and prognostic factors. Dig. Dis. Sci. 37:
778–783.
Dumoulin, F. L., Gerhardt, T., Fuchs, S., Scheurlen,
C., Neubrand, M., Layer, G., and Sauerbruch, T.,
M. Wiedmann et al.
2003. Phase II study of photodynamic therapy
and metal stent as palliative treatment for nonre-
sectable hilar cholangiocarcinoma. Gastrointest.
Endosc. 57: 860–867.
Ebata, T., Nagino, M., Kamiya, J., Uesaka, K.,
Nagasaka, T., and Nimura, Y., 2003. Hepatectomy
with portal vein resection for hilar cholangiocar-
cinoma: audit of 52 consecutive cases. Ann.
Surg. 238: 720–727.
Freeman, M. L. and Overby, C., 2003. Selective
MRCP and CT-targeted drainage of malignant
hilar biliary obstruction with self-expanding
metallic stents. Gastrointest. Endosc. 58:
41–49.
Harewood, G. C., Baron, T. H., Rumalla, A., Wang,
K. K., Gores, G. J., Stadheim, L. M., and de
Groen, P. C., 2005. Pilot study to assess patient
outcomes following endoscopic application of
photodynamic therapy for advanced cholan-
giocarcinoma. J. Gastroenterol. Hepatol. 20:
415–420.
Hemming, A. W., Reed, A. I., Fujita, S., Foley, D.
P., and Howard, R. J., 2005. Surgical manage-
ment of hilar cholangiocarcinoma. Ann. Surg.
241: 693–699; discussion 699–702.
Isayama, H., Komatsu, Y., Tsujino, T., Sasahira, N.,
Hirano, K., Toda, N., Nakai, Y., Yamamoto, N.,
Tada, M., Yoshida, H., Shiratori, Y., Kawabe, T.,
and Omata, M., 2004. A prospective randomised
study of “covered” versus “uncovered” diamond
stents for the management of distal malignant
biliary obstruction. Gut 53: 729–734.
Kadakia, S. C. and Starnes, E., 1992. Comparison
of 10 French gauge stent with 11.5 French gauge
stent in patients with biliary tract diseases.
Gastrointest. Endosc. 38: 454–459.
Kawasaki, S., Imamura, H., Kobayashi, A., Noike,
T., Miwa, S., and Miyagawa, S., 2003. Results of
surgical resection for patients with hilar bile duct
cancer: application of extended hepatectomy
after biliary drainage and hemihepatic portal
vein embolization. Ann. Surg. 238: 84–92.
Khan, S. A., Taylor-Robinson, S. D., Toledano,
M. B., Beck, A., Elliott, P., and Thomas, H. C.,
2002. Changing international trends in mortality
rates for liver, biliary and pancreatic tumours. J.
Hepatol. 37: 806–183.
Kondo, S., Hirano, S., Ambo, Y., Tanaka, E.,
Okushiba, S., Morikawa, T., and Katoh, H., 2004.
Forty consecutive resections of hilar cholangio-
carcinoma with no postoperative mortality and no
35. Hilar Cholangiocarcinoma: Photodynamic Therapy and Stenting
positive ductal margins: results of a prospective
study. Ann. Surg. 240: 95–101.
Kosuge, T., Yamamoto, J., Shimada, K., Yamasaki,
S., and Makuuchi, M., 1999. Improved surgical
results for hilar cholangiocarcinoma with pro-
cedures including major hepatic resection. Ann.
Surg. 230: 663–671.
Lazaridis, K. N. and Gores, G. J., 2005.
Cholangiocarcinoma. Gastroenterology 128:
1655–1667.
Lillemoe, K. D., Cameron, J. L., Hardacre, J. M.,
Sohn, T. A., Sauter, P. K., Coleman, J., Pitt, H.
A., and Yeo, C. J., 1999. Is prophylactic gastro-
jejunostomy indicated for unresectable periam-
pullary cancer? A prospective randomized trial.
Ann. Surg. 230: 322–328; discussion 328–330.
Liu, C. L., Lo, C. M., Lai, E. C., and Fan, S. T.,
1998. Endoscopic retrograde cholangiopancrea-
tography and endoscopic endoprosthesis inser-
tion in patients with Klatskin tumors. Arch. Surg.
133: 293–296.
McCaughan, J. S., Jr., Mertens, B. F., Cho, C.,
Barabash, R. D., and Payton, H. W., 1991.
Photodynamic therapy to treat tumors of the
extrahepatic biliary ducts. A case report. Arch.
Surg. 126: 111–113.
Miyazaki, M., Ito, H., Nakagawa, K., Ambiru, S.,
Shimizu, H., Shimizu, Y., Kato, A., Nakamura, S.,
Omoto, H., Nakajima, N., Kimura, F., and Suwa,
T., 1998. Aggressive surgical approaches to hilar
cholangiocarcinoma: hepatic or local resection?
Surgery 123: 131–136.
Nanashima, A., Yamaguchi, H., Shibasaki, S., Ide,
N., Sawai, T., Tsuji, T., Hidaka, S., Sumida,
Y., Nakagoe, T., and Nagayasu, T., 2004.
Adjuvant photodynamic therapy for bile duct
carcinoma after surgery: a preliminary study. J.
Gastroenterol. 39: 1095–1101.
Ortner, M. A., 2000. Photodynamic therapy of
cholangiocarcinoma cancer. Gastrointest.
Endosc. Clin. N. Am. 10: 481–486.
Ortner, M. A., Liebetruth, J., Schreiber, S., Hanft,
M., Wruck, U., Fusco, V., Muller, J. M., Hortnagl,
H., and Lochs, H., 1998. Photodynamic ther-
apy of nonresectable cholangiocarcinoma.
Gastroenterology 114: 536–542.
Ortner, M. E. J., Caca, K., Berr, F., Liebetruth,
J., Mansmann, U., Huster, D., Voderholzer,
W., Schachschal, G., Mössner, J., and Lochs,
H., 2003. Successful photodynamic therapy
for nonresectable cholangiocarcinoma: a ran-
481
domized prospective study. Gastroenterology
125: 1355–1363.
Pereira, S. P., Ayaru, L., Hatfield, A. R., Rogowska,
A., and Bown, S., 2005a. Photodynamic ther-
apy (PDT) of malignant biliary strictures using
meso-tetrahydroxyphenylchlorin (mTHPC). J.
Clin. Oncol. 23: abstract 4176 (ASCO).
Pereira, S. P., Ragunath, K., Devlin, J., and Meadows,
H. M., 2005b. Preliminary results of a phase II
trial to examine the safety and efficacy of por-
fimer sodium photodynamic therapy (PDT) in
locally advanced biliary tract carcinoma (BTC). J.
Clin. Oncol. 23: abstract 4180 (ASCO).
Polydorou, A. A., Cairns, S. R., Dowsett, J. F.,
Hatfield, A. R., Salmon, P. R., Cotton, P. B.,
and Russell, R. C., 1991. Palliation of proximal
malignant biliary obstruction by endoscopic
endoprosthesis insertion. Gut 32: 685–689.
Polydorou, A. A., Chisholm, E. M., Romanos, A.
A., Dowsett, J. F., Cotton, P. B., Hatfield, A. R.,
and Russell, R. C., 1989. A comparison of right
versus left hepatic duct endoprosthesis insertion
in malignant hilar biliary obstruction. Endoscopy
21: 266–271.
Prasad, G., Wang, K. K., Baron, T. H., Buttar, N. S.,
WongKeeSong, M., Borkenhagen, L., Lutzke, L.,
Papanfuss, S., Gores, G., and Roberts, L., 2005.
Factors predicting Survival in patients with cholan-
giocarcinoma treated with photodynamic therapy.
Gastroenterology: abstract T1090 (DDW).
Prat, F., Chapat, O., Ducot, B., Ponchon, T., Pelletier,
G., Fritsch, J., Choury, A. D., and Buffet, C.,
1998. A randomized trial of endoscopic drain-
age methods for inoperable malignant strictures
of the common bile duct. Gastrointest. Endosc.
47: 1–7.
Raju, G. S., Sud, R., Elfert, A. A., Enaba, M.,
Kalloo, A., and Pasricha, P. J., 2006. Biliary
drainage by using stents without a central lumen:
a pilot study. Gastrointest. Endosc. 63: 317–320.
Rumalla, A., Baron, T. H., Wang, K. K., Gores,
G. J., Stadheim, L. M., and de Groen, P. C.,
2001. Endoscopic application of photodynamic
therapy for cholangiocarcinoma. Gastrointest.
Endosc. 53: 500–504.
Schiefke, I., Zabel-Langhennig, A., Wiedmann,
M., Huster, D., Witzigmann, H., Mossner, J.,
Berr, F., and Caca, K., 2003. Self-expandable
metallic stents for malignant duodenal obstruc-
tion caused by biliary tract cancer. Gastrointest.
Endosc. 58: 213–219.
482
Seyama, Y., Kubota, K., Sano, K., Noie, T.,
Takayama, T., Kosuge, T., and Makuuchi, M.,
2003. Long-term outcome of extended hemi-
hepatectomy for hilar bile duct cancer with no
mortality and high survival rate. Ann. Surg. 238:
73–83.
Shim, C. S., Cheon, Y. K., Cha, S. W., Bhandari,
S., Moon, J. H., Cho, Y. D., Kim, Y. S., Lee, L.
S., Lee, M. S., and Kim, B. S., 2005. Prospective
study of the effectiveness of percutaneous tran-
shepatic photodynamic therapy for advanced
bile duct cancer and the role of intraductal ultra-
sonography in response assessment. Endoscopy
37: 425–433.
Soderlund, C. and Linder, S., 2006. Covered
metal versus plastic stents for malignant com-
mon bile duct stenosis: a prospective, rand-
omized, controlled trial. Gastrointest. Endosc.
63: 986–995.
Speer, A. G., Cotton, P. B. and MacRae, K. D.,
1988. Endoscopic management of malignant
biliary obstruction: stents of 10 French gauge
are preferable to stents of 8 French gauge.
Gastrointest. Endosc. 34: 412–417.
Speer, A. G., Cotton, P. B., Russell, R. C., Mason,
R. R., Hatfield, A. R., Leung, J. W., MacRae,
K. D., Houghton, J., and Lennon, C. A., 1987.
Randomised trial of endoscopic versus percu-
taneous stent insertion in malignant obstructive
jaundice. Lancet 2: 57–62.
Todoroki, T., Kawamoto, T., Koike, N., Takahashi,
H., Yoshida, S., Kashiwagi, H., Takada, Y.,
Otsuka, M., and Fukao, K., 2000. Radical resec-
tion of hilar bile duct carcinoma and predictors
of survival. Br. J. Surg. 87: 306–313.
Tringali, A., Mutignani, M., Perri, V., Zuccala, G.,
Cipolletta, L., Bianco, M. A., Rotondano, G.,
Philipper, M., Schumacher, B., Neuhaus, H.,
Schmit, A., Deviere, J., and Costamagna, G.,
2003. A prospective, randomized multicenter
trial comparing DoubleLayer and polyethylene
stents for malignant distal common bile duct
strictures. Endoscopy 35: 992–997.
van Berkel, A. M., Boland, C., Redekop, W. K.,
Bergman, J. J., Groen, A. K., Tytgat, G. N., and
Huibregtse, K., 1998. A prospective randomized
trial of Teflon versus polyethylene stents for
distal malignant biliary obstruction. Endoscopy
30: 681–686.
M. Wiedmann et al.
Wagner, H. J., Knyrim, K., Vakil, N., and Klose,
K. J., 1993. Plastic endoprostheses versus metal
stents in the palliative treatment of malignant
hilar biliary obstruction. A prospective and ran-
domized trial. Endoscopy 25: 213–218.
Welzel, T. M., McGlynn, K. A., Hsing, A. W.,
O’Brien, T. R., and Pfeiffer, R. M., 2006. Impact
of classification of hilar cholangiocarcinomas
(Klatskin tumors) on the incidence of intra- and
extrahepatic cholangiocarcinoma in the United
States. J. Natl. Cancer Inst. 98: 873–875.
West, J., Wood, H., Logan, R. F., Quinn, M., and
Aithal, G. P., 2006. Trends in the incidence
of primary liver and biliary tract cancers in
England and Wales 1971–2001. Br. J. Cancer.
94: 1751–1758.
Wiedmann, M., Berr, F., Schiefke, I., Witzigmann,
H., Kohlhaw, K., Mossner, J., and Caca, K.,
2004a. Photodynamic therapy in patients with
non-resectable hilar cholangiocarcinoma: 5-year
follow-up of a prospective phase II study.
Gastrointest. Endosc. 60: 68–75.
Wiedmann, M., Caca, K., Berr, F., Schiefke, I.,
Tannapfel, A., Wittekind, C., Mossner, J., Hauss,
J., and Witzigmann, H., 2003. Neoadjuvant pho-
todynamic therapy as a new approach to treating
hilar cholangiocarcinoma: a phase II pilot study.
Cancer 97: 2783–2790.
Wiedmann, M., Dietrich, A., Mossner, J.,
Witzigmann, H., and Caca, K., 2004b. Combined
percutaneous transhepatic biliary drainage with
port implantation for management of patients
with malignant biliary obstruction. Gastrointest.
Endosc. 60: 117–120.
Witzigmann, H., Berr, F., Ringel, U., Caca, K.,
Uhlmann, D., Schoppmeyer, K., Tannapfel, A.,
Wittekind, C., Mossner, J., Hauss, J., and Wiedmann,
M., 2006. Surgical and Palliative Management
and Outcome in 184 Patients With Hilar
Cholangiocarcinoma: Palliative Photodynamic
Therapy Plus Stenting is Comparable to R1/R2
Resection. Ann. Surg. 244: 230–239.
Yoon, W. J., Lee, J. K., Lee, K. H., Lee, W. J.,
Ryu, J. K., Kim, Y. T., and Yoon, Y. B., 2006. A
comparison of covered and uncovered Wallstents
for the management of distal malignant biliary
obstruction. Gastrointest. Endosc. 63: 996–1000.
Zoepf, T., Jakobs, R., Arnold, J. C., Apel, D., and
Riemann, J. F., 2005. Palliation of nonresectable
35. Hilar Cholangiocarcinoma: Photodynamic Therapy and Stenting 483

bile duct cancer: improved survival after pho-
todynamic therapy. Am. J. Gastroenterol. 100:
2426–2430.
Zoepf, T., Jakobs, R., Arnold, J. C., Apel, D.,
Rosenbaum, A., and Riemann, J. F., 2001a.
Photodynamic therapy for palliation of nonresect-
able bile duct cancer – preliminary results with
a new diode laser system. Am. J. Gastroenterol.
96: 2093–2097.
Zoepf, T., Jakobs, R., Rosenbaum, A., Apel,
D., Arnold, J. C., and Riemann, J. F., 2001b.
Photodynamic therapy with 5-aminolevulinic
acid is not effective in bile duct cancer.
Gastrointest. Endosc. 54: 763–766.
A. Diagnosis
36
Splenic Metastases: Diagnostic Methods
Eva Compérat and Frédéric Charlotte
INTRODUCTION
Metastases from solid tumors to the spleen due
to hematogenous dissemination are confined
to the splenic parenchyma, and should not be
confused with small superficial subcapsular
foci associated with peritoneal dissemination
observed in ovarian cancers (Tserkezoglou
et al., 2005). According to this definition,
splenic metastases were thought to be excep-
tional but the incidence of reported cases has
been increasing due to the improvement of
medical imaging and the long-term follow-
up of patients with cancer. Metastases to the
spleen generally occur in the context of mul-
tivisceral metastatic cancer at terminal stage.
Solitary metastases have also been reported
(Compérat et al., 2007), and are becoming a
diagnostic dilemma with primary tumors of
the spleen. After a brief description of epi-
demiology, pathologic features, pathogenesis
and clinical aspects of splenic metastases,
this review will outline diagnostic procedures
and the differential diagnoses.
EPIDEMIOLOGY
The prevalence of splenic metastases Splenic metastases can present as three
in large populations with cancer was main macroscopic patterns: macronodular,
obtained mainly from autopsy series pub-
lished before 1990 and ranged between
2.3 % and 7.1 % (Berge, 1974). The larger
autopsy series was reported by Berge
(1974) who found 312 splenic metastases
for 7,165 autopsies performed for cancer.
In this study, microscopic splenic metas-
tases were identified in 50 % of subjects
who had metastases in at least five organs.
After 1990, studies from different parts
of the world brought indications regard-
ing the prevalence of splenic metastases
in living patients. In a study from the
U.S.A., (Kraus et al., 2001), 1.3 % of
1,280 sequential tumors in splenectomy
specimens were metastatic. In the same
study, 9.8 % of 122 splenectomies per-
formed for diagnosis contained metastasis.
In a Japanese study (Ishida et al., 1997),
0.15 % of 24,761 patients examined by
ultrasonography had splenic metastasis.
A Chinese report (Lam and Tang, 2000)
found metastasis in 1.1 % of 1,743 sequen-
tial splenectomy specimens.
PATHOLOGIC FEATURES
489
490 E. Compérat and F. Charlotte

micronodular, and diffuse (Berge, 1974). In

the macronodular pattern, the splenic paren-
chyma is partially replaced by macroscopi-
cally discernible, solitary or multiple large
nodules of varying sizes (Figure 36.1). The
micronodular pattern is characterized by a
relatively uniform miliary aspect (Figure
36.2). The tumor nodules can be located in
the white pulp or the red pulp. In the diffuse
pattern, the splenic parenchyma is com-
pletely replaced by the tumor cells. In the
microscopic pattern, no macroscopic lesion
can be seen. In this pattern, the tumor cells
can be confined to the venous sinuses, the
red pulp, the white pulp, or the trabecular
vessels, or can be found in several of these
compartments (Berge, 1974). Whatever the

Figure 36.2. Macroscopic findings of micronodu-

lar aspects with uniform miliary pattern

pattern of splenic infiltration, the splenic

metastasis and primary tumor are generally
similar in terms of cytological and architec-
tural aspect.

Figure 36.1. Macroscopic findings: multiple large
nodules of varying sizes in spleen parenchyma
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
the splenic implantation of blood-borne
cancer cells, that is, the constant flow of
36. Splenic Metastases: Diagnostic Methods
blood through the spleen and the rhythmic
contraction of splenic capsule; the sharp
angle of splenic artery branching from the
celiac artery preventing large clumps of
tumor cells from passing through; the lack
of afferent lymphatic vessels limiting lym-
phogenic metastases, and (2) inhibitory
effect of splenic microenvironment on the
growth of metastatic cells (Berge, 1974).
The respective contribution of these factors
to the resistance of the spleen to metastatic
involvement is still debated. The mechani-
cal theory was proposed at the time the
dissemination from the primary tumor and
the implantation of tumor cells in distant
organs were regarded as a late event in
the multistep process of cancer. However,
the recent development of sensitive immu-
nologic and molecular methods has per-
mitted the detection of individual tumor
cells in regional lymph nodes, blood, or
distant organs at the time of primary tumor
diagnosis for various histological types
of cancer (Pantel and Brakenhoff, 2004).
This supports the concept that micrometa-
static dissemination occurs, in fact, early
in the course of malignant disease and is
not affected by mechanical factors. This
is also in accordance with the results of
experimental metastasis models using neo-
plastic cells labeled with green fluorescent
protein (Goodison et al., 2003).
From this recent advance in the knowl-
edge of metastatic process, it may be
assumed that the presence of disseminated
cancer cells in the spleen may not be
detected at the time of primary diagnosis
by conventional methods. It is noteworthy
that the expression of chemokines in the
spleen does not constitute an obstacle to
cancer cell homing because stromal-cell-
derived factor 1 (SDF-1) and the ligand
for CXC chemoreceptor-4 expressed by
491
various types of metastatic cells, are found
not only in the preferential sites of metas-
tases such as the bone marrow, lung and
lymph nodes, but also in the spleen where
it plays a crucial role in lymphocyte hom-
ing (Chambers et al., 2002). Therefore,
implantation of cancer cells in the splenic
parenchyma may occur, but clinically
detectable metastases are rare because
the splenic microenvironment may not
facilitate the growth of micrometastatic
foci. This probably explains the contrast
between the high prevalence of splenic
micrometastases at autopsy in patients
with multivisceral cancer and the rarity
of clinically detectable metastases. This
suggests that mechanical factor seems to
play a minor role in explaining the rarity
of clinical splenic metastases.
CLINICAL FEATURES
Most splenic metastases are a part of
multivisceral metastatic disease. In this
context, breast, lung, ovarian, colorectal, and
gastric carcinomas, and skin melanoma are
the most common primary sources (Berge,
1974; Lam and Tang, 2000). Skin melanoma
has the highest rate of splenic metastases
per primary tumor since more than 30 % of
patients with skin melanoma have splenic
metastasis at autopsy (Berge, 1974; Lam and
Tang, 2000). Splenic metastases also occur
as a solitary splenic mass, synchronous or
metachronous to the primary tumor. To our
knowledge, 93 well-documented cases of
solitary splenic metastases are known which
are listed in Table 36.1. From diagnosis of
primary tumor to the discovery of solitary
splenic metastasis it might take between
0 to 264 months with a median of 28
months. Colorectal and ovarian carcinomas
492 E. Compérat and F. Charlotte

Table 36.1. Summary of 93 case reports of solitary splenic metastases.

No. of
asymp- No. of cases
Interval (mo) tomatic with positive
Primary site No. Histological type (No.) range (median) cases serum marker References
Colon/rectum 20 Adenocarcinoma 0–132 (27) 18/20 13/20 Ishida et al., 1997
Kim et al., 2000
Agha-Mohammadi and Calne,
2001
Avesani et al., 2001
Okuyama T et al., 2001
Genna et al., 2003
Cavallaro et al., 2004
Ovary 18 Serous cystadeno- 18–183 (51) 15/17* 10/17* Tserkezoglou et al., 2005
carcinoma (17)
Mucinous cystaden- 1 1 1 Koh et al., 2004
ocarcinoma (1)
Lung 10 Adenocarcinoma (4) 0–25(12) 2/4 nd Agha-Mohammadi and Calne,
2001
Large cell carcinoma 0 1/3 nd Schmidt and Smith, 2004
(3)
Bronchioalveolar 20 0 nd
carcinoma (1)
Carcinoid (1) 96 0 nd
Squamous cell 14 1 nd
carcinoma (1)
Endometrium 9 Adenocarcinoma 11–120 (31) 3/9 2/9 Agha-Mohammadi and Calne,
2001
Giuliani et al., 1999
Hadjileontis et al., 2004
Kidney 9 Clear cell carcinoma 0–264 (48) 3/6* nd Ishida et al., 1997
Kugel et al., 2003
McGregor et al., 2003
Stomach 7 Adenocarcinoma 16–102 (48) 3/3* 3/3* Opocher et al., 2000;
Yamanouchi et al., 2002
Cervix 6 Squamous cell 49–69 (56) 1/4* 1/1* Agha-Mohammadi and Calne,
carcinoma (4) 2001
Adenocarcinoma (2) 14–60 (37) 1/2 nd Carvalho et al., 1997
Pang, 2004
Breast 3 Adenocarcinoma 0–108 (24) 0 nd Barreca et al., 2001
Iype S et al., 2002
Hoar et al., 2003
Prostate 3 Adenocarcinoma 0–60 (18) 2/3 2/3 Sharpe et al., 1993
Naseem et al., 1998
Compérat et al., 2007
Esophagus 3 Squamous cell 0–96 (6) 1/3 nd Agha-Mohammadi and Calne,
carcinoma 2001
Vyas et al., 2002;
Kimura et al., 2003
Fallopian tube 1 Adenocarcinoma 48 1 1 Dubecq-Princeteau et al., 1997
Skin 1 Melanoma 30 0 nd Agha-Mohammadi and Calne,
2001
Uveal tract 1 Melanoma 10 0 nd Tas et al., 2004
Thyroid 1 Undifferentiated 6 0 nd Mayayo et al., 2003
carcinoma
Unknown 1 Adenocarcinoma 0 0 1 Alici et al., 2003
Total 93 0–264 (28) 55/86 34/56 (61%)
(64%)
*
data not available for all cases; mo: months; nd: not described; ref: references.
36. Splenic Metastases: Diagnostic Methods 493

are the most common sources. On the
other hand, solitary splenic metastases
from breast carcinoma (Barreca et al., 2001;
Iype et al., 2002; Hoar et al., 2003) or skin
melanoma are rare. Exceptionally, a splenic
metastasis reveals a clinically occult primary
tumor (McGregor et al., 2003). Only one
report described a solitary splenic metas-
tasis from unknown primary cancer (Alici
et al., 2003).
Splenic metastases are most often
incidentally detected by medical imag-
ing techniques. When isolated, more than
60 % of splenic metastases are asympto-
matic (Table 36.1). Splenic metastasis can
also be revealed by fatigue (Ishida et al.,
1997), weight loss (Agha-Mohammadi and
Calne, 2001), fever (Kugel et al., 2003),
abdominal pain (Carvalho et al., 1997;
Agha-Mohammadi and Calne, 2001; Iype
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).
FDG-PET is to discriminate benign from
malignant splenic masses, the latter is char-
acterized by an increase in FDG uptake.
Moreover, FDG-PET coupled with CT can
also identify other tumor sites expressing
as hypermetabolic foci which are unsus-
pected on conventional imaging but may
be more accessible for biopsy.
Serum level of tumor markers utilized
in the follow-up of patient with cancer can
predict the appearance of solitary splenic
metastases before their clinical expression
or radiological detection and decreases
within normal limits after splenectomy
(Dubecq-Princeteau et al., 1997). For
example, Compérat et al. (2007) have
recently seen a solitary splenic metastasis
from prostate carcinoma occurring 5 years
after a radical prostatectomy (Figure 36.3).
A progressive increase in the serum PSA

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
level was observed before the tumor could
be detected by imaging technique. Serum
PSA level decreased within normal limits
after splenectomy and the patient remained
asymptomatic 13 months thereafter. This
case history along with other published
similar cases suggests that splenic metas-
tases might result from the growth of early
blood-borne disseminated cancer cells
within the spleen after a period of clinical
latency. This could explain why isolated
splenic metastases are not necessarily the
precursor sign of active metastatic can-
cer at terminal stage because long term
remission has been achieved in patients
treated with splenectomy alone (Agha-
Mohammadi and Calne, 2001). Further
follow-up studies of patients with solitary
splenic metastasis detected by imaging
techniques such as PET, are needed to
better understand the significance of this
clinical manifestation.
The finding of a splenic nodule is
regarded as metastatic in case of dissemi-
nated cancerous disease, and splenectomy
is generally not performed, as the diagno-
sis relies on biopsy of a more accessible
tumor site. Isolated splenic metastasis can
be diagnosed by fine-needle aspiration or
percutaneous biopsy in place of splenec-
tomy. The use of these imaging-guided
techniques to diagnose splenic metastases
has rarely been reported in the medical lit-
erature. This may be attributed to the fear
of hemorrhagic complication because the
spleen is highly vascular. However, these
imaging-guided techniques are less aggres-
sive than the surgical approach and permit
the avoidance of unnecessary splenectomy
for the diagnosis of benign splenic lesions
such as granulomatous disease. A success-
ful diagnosis can be achieved in ∼ 90 % of
the cases. Complications range between
E. Compérat and F. Charlotte
0 % and 2 % and mainly affect patients
with thrombocytopenia or vascular splenic
neoplasms.
DIFFERENTIAL DIAGNOSIS
Despite improvement of medical imag-
ing techniques such as PET, the clini-
cal diagnosis of splenic metastases is
largely dependent on patient history.
Radiologically, metastatic disease fre-
quently affects the spleen in a manner
that is indistinguishable from primary
splenic lesions. A splenic mass without
context of active cancerous disease is
more suggestive of a primary splenic
lesion such as lymphomas, vascular
tumors, or infectious lesions. Conversely,
a previous history of malignant disease
constitutes an indication for the diagno-
sis of splenic metastasis. The search for
additional clinical findings such as lym-
phadenopathy or disease in other organs
may aid in the diagnosis of lymphoma or
metastatic disease. Rarer splenic lesions
such as hamartoma and inflammatory
pseudotumor can clinically and radiologi-
cally mimic a metastatic tumor but there
is generally no confusion on histological
grounds.
Primary malignant lymphoma of the
spleen is defined as a disease limited to
the spleen and splenic hilum. The main
histological types include large B-cell
lymphoma, marginal zone B-cell lymphoma,
mantle cell lymphoma, hairy-cell leuke-
mia, and gamma-delta T-cell lymphoma
(Mollejo et al., 2003). Large B-cell lym-
phoma can be mistaken for metastasis from
undifferentiated carcinoma or melanoma.
Immunostaining with cytokeratins for car-
cinoma, common leukocyte antigen for
36. Splenic Metastases: Diagnostic Methods 495

lymphoma, and S-100 protein, HMB45, or

melan A for melanoma is useful in estab-
lishing a correct diagnosis (Figure 36.4).
Interestingly, two case reports described
the association between carcinoma and
lymphoma in the same spleen (Sharpe
et al., 1993; Hoar et al., 2003).
Vascular tumors of the spleen can pose
a diagnostic problem radiologically, but
they are easily distinguished from met-
astatic tumors on histological grounds.
Hemangioma is the most common primary
tumor of the spleen. Littoral cell angioma
is composed of vascular channels lined
by tall endothelial cells expressing both
endothelial (factor VIIIr) and histiocytic
markers (CD68). It is noteworthy that this
vascular tumor is frequently associated
with visceral malignancy such as colorec-
tal, lung, renal, and pancreatic carcinoma.
Primary angiosarcoma of the spleen, par- Figure 36.4. Metastasis of melanoma in spleen
ticularly in its pleomorphic or epithelioid parenchyma, H&E × 20
variant, can be a problem of differential
diagnosis with a metastasis from undif-
ferentiated carcinoma or melanoma. The
right diagnosis is based upon the expres- disorder. Fine needle aspiration is very
sion of vascular makers CD31, factor VIII, useful in the differential diagnosis between
or CD34 by the tumor cells. these infectious causes and neoplasm.
Nonneoplastic disorders of the spleen In conclusion, although classically
include infections such as disseminated regarded as very low, the prevalence of
fungal disease, or septic emboli, and gran- splenic metastases seems to increase with
ulomatous diseases such as sarcoidosis the improvement of medical imaging tech-
or tuberculosis. Findings of associated niques. Therefore, splenic metastases are
lymphadenopathies, pulmonary, and medi- becoming a diagnostic dilemma with pri-
astinal diseases should aid in the diagnosis mary lesions of the spleen, particularly for
of sarcoidosis. In immunocompromised solitary lesions in patients with a history
individuals, tuberculosis, Mycobacterium of malignant disease. The most common
avium intracellulare, Pneumocystis cari- sources of splenic metastases are breast,
nii, and disseminated Kaposi sarcoma may lung, colorectal and ovarian carcinoma,
cause multiple splenic masses. In these and melanoma. Fine-needle aspiration and
cases, the patient’s clinical condition and percutaneous biopsy of splenic lesions are
immune status will help in distinguishing useful in establishing the right diagnosis in
an infectious entity from a neoplastic most cases.
496
REFERENCES
Agha-Mohammadi, S., and Calne, R.Y. 2001.
Solitary splenic metastasis. Case report and
review of the literature. Am. J. Clin. Oncol.
24:306–310.
Alici, S., Kosem, M., and Kotan, C. 2003. Isolated
splenic metastases occurring as an unknown pri-
mary lesion. J. Postgrad. Med. 49:83–84.
Avesani, E.C., Cioffi, U., de Simone, M., Botti, F.,
Carrara, A., and Ferrero, S. 2001. Synchronous
isolated splenic metastasis from colon carci-
noma. Am. J. Clin. Oncol. 24:311–312.
Barreca, M., Angelini, D., Gallo, A., Puntillo, F.,
Amodio, P.M., and Fernandes, E. 2001. Single
asymptomatic splenic metastasis of breast car-
cinoma: report of a clinical case. G. Chir.
22:227–228.
Berge T. 1974. Splenic metastases. Frequencies
and patterns. Acta Pathol. Microbiol. Scand.
82:499–506.
Carvalho, L., Azevedo, I., Salgado, L., Ferreira,
E.S., Henrique, R., de Carvalho, R.G., and
Vieira, E. 1997. Squamous cell carcinoma of
the cervix metastatic to the spleen – case report.
Gynecol. Oncol. 67:107–110.
Cavallaro, A., Modugno, P., Specchia, M., Pontenza,
A.E., Loschiavo, V., Colli, R., Lauriola, L., and
Barone, C. 2004. Isolated splenic metastasis
from colon cancer. J. Exp. Clin. Cancer Res.
23:143–146.
Chambers, A.F., Groom, A.C., and MacDonald, I.C.
2002. Dissemination and growth of cancer cells
in metastatic sites. Nat. Rev. Cancer 2:563–572.
Compérat, E., Azzouzi, A.R., Chartier-Kastler
E., Ménégaux F., Capron, F., Richard, F., and
Charlotte, F. 2007. Late recurrence of a prostatic
adenocarcinoma as a solitary splenic metastasis.
Urol. Int. 78:86–88.
Dubecq-Princeteau, F., Lefranc, J.P., Touboul, E.,
Chauvel, A., and Le Charpentier, Y. 1997.
Adenocarcinoma of the fallopian tube: natural
history of an isolated splenic metastasis. Acta
Oncol. 36:221–223.
Genna, M., Leopardi, F., Valloncini, E., Molfetta,
M., De Manzoni, G., and Castelli, A. 2003.
Metachronous splenic metastasis of colon can-
cer. A case report. Minerva. Chir. 58:811–814.
Giuliani, A., Caporale, A., Di Bari, M., Demoro,
M., and Mingazzini, P. 1999. Isolated splenic
E. Compérat and F. Charlotte
metastasis from endometrial carcinoma. J. Exp.
Clin. Cancer Res. 18:93–96.
Goodison, S., Kawai, K., Hihara, J., Jiang, P.,
Yang, M., Urquidi, V., Hoffman, R.M., and
Tarin, D. 2003. Prolonged dormancy and site-
specific growth potential of cancer cells spon-
taneously disseminated from non metastatic
breast tumors as revealed by labeling with
green Fluorescent Protein. Clin. Cancer Res.
9:3808–3814.
Hadjileontis, C., Amplianitis, I., Valsamides, C.,
Harisis, G., Nepka, H., and Kafanas, A. 2004.
Solitary splenic metastasis of endometrial carci-
noma ten years after hysterectomy. Case report
and review of the literature. Eur. J. Gynaecol.
Oncol. 25:233–235.
Hoar, F.J., Chan, S.Y., Stonelake, P.S., Wolverson,
R.W., and Bareford, D. 2003. Splenic rupture as
a consequence of dual malignant pathology: a
case report. J. Clin. Pathol. 56:709–710.
Ishida, H., Konno, K., Ishida, J., Shirayama, K.,
Naganuma, H., Komatsuda, T., Hamashima, Y.,
and Masamune, O. 1997. Isolated splenic metas-
tases. J. Ultrasound Med. 16:743–749.
Iype, S., Akbar, M.A., and Krishna, G. 2002.
Isolated splenic metastasis from carcinoma of
the breast. Postgrad. Med. J. 78:173–174.
Kim, J.C., Jeong, C.S., Kim, H.C., Yu, C.S., Kang,
G.H., and Lee, M.G. 2000. Isolated splenic
metastasis from colorectal carcinoma: a case
report. J. Korean Med. Sci. 15:355–358.
Kimura, Y., Miyazaki, M., Saeki, H., Ohga, T.,
Nozoe, T., and Sugimachi, K. 2003. Solitary
splenic metastasis derived from esophageal can-
cer. Hepatogastroenterology 50:1336–1337.
Koh, Y.S., Kim, J.C., and Cho, C.K. 2004.
Splenectomy for solitary splenic metastasis of
ovarian cancer. BMC. Cancer 4:96
Kraus, M.D., Fleming, M.D., and Vonderheide,
R.H. 2001. The spleen as a diagnostic specimen:
a review of 10 years’ experience at two tertiary
care institutions. Cancer 91:2001–2009.
Kugel, V., Dekel, Y., Konichezky, M., Baniel,
J., Livne, P.M., and Koren, R. 2003. Unusual
splenic metastasis from renal cell carcinoma. A
case report and review of the literature. Pathol.
Res. Pract. 199:739–743.
Lam, K.Y., and Tang, V. 2000.Metastatic tumors
to the spleen. A 25-year clinicopathologic study.
Arch. Pathol. Lab. Med. 124:526–530.
36. Splenic Metastases: Diagnostic Methods
Mayayo, E., Blazquez, S., Gomez-Aracil, V., Sauri,
A., and Martinez, S. 2003. Spleen metastasis
from thyroid carcinoma. Report of a case with
diagnosis by fine needle aspiration cytology.
Acta Cytol. 47:1116–1118.
McGregor, D.H., Wu, Y., Weston, A.P., Mcanaw,
M.P., Bromfield, C., and Bhattatiry, M.M. 2003.
Metastatic renal cell carcinoma of spleen diag-
nosed by fine-needle aspiration Spleen. Am. J.
Med. Sci. 326:51–54.
Mollejo, M., Algara, P., Mateo, M., Menarguez,
J., Pascual, E., Fresno, M.F., Camacho, F., and
Piris, A. 2003. Large B-cell lymphoma present-
ing in the spleen: Identification of different clin-
icopathologic conditions. Am. J. Surg. Pathol.
27: 895–902.
Naseem, M.S., Jan, H.A., Britton, K.E., and
Nargund, V.H. 1998. Splenic metastasis from
adenocarcinoma of the prostate. Br. J. Urol.
82:597–598.
Okuyama, T., Oya, M., and Ishikawa, H. 2001.
Isolated splenic metastasis of sigmoid colon
cancer: a case report. Jpn. J. Clin. Oncol.
31:341–345.
Opocher, E., Santambrogio, R., Bianchi, P., Cioffi,
U., De Simone, M., Vellini, S., and Montorsi,
M. 2000. Isolated splenic metastasis from gastric
carcinoma. value of CEA and CA 19–9 in early
diagnosis: report of two case. Am. J. Clin. Oncol.
23:579–580.
Pang, L.C. 2004. Solitary recurrent metastasis of
squamous cell carcinoma of the uterine cer-
497
vix in the spleen: case report. South Med. J.
97:301–304.
Pantel, K., and Brakenhoff, R.H. 2004. Dissecting
the metastatic cascade. Nat. Rev. Cancer 4:
448–456.
Schmidt, B.J., and Smith, S.L. 2004. Isolated splenic
metastasis from primary lung Adenocarcinoma.
South. Med. J. 97:298–300.
Sharpe, R.W., Rector, J.T., Rushin, J.M., Garvin,
D.F., and Cotelingam, J.D. 1993. Splenic metas-
tasis in hairy cell leukemia. Cancer 71:2222–
2226.
Tas, F., Ustuner, Z., Buyukbabani, N., Tenekeci, N.,
and Topuz, E. 2004. Massive and isolated metas-
tases to spleen of uveal malignant melanoma.
Retina 24:170–172.
Tserkezoglou, A., Kontou, S., Hatjieleftheriou,
G., Nikolaidou, M.E., Plataniotis, G.,
Apostolikas, N., and Magiakos, G. 2005.
Solitary parenchymal splenic recurrence
of ovarian adenocarcinoma: a case report
and review of the literature. Anticancer Res.
25:1471–1476.
Vyas, S.J., Chitaleb, A.R., and Deshpande, R.K.
2002. Late splenic metastasis after curative
resection for oesophageal carcinoma. Eur. J.
Cardiothorac. Surg. 22:1011–1013.
Yamanouchi, K., Ikematsu, Y., Waki, S., Kida, H.,
Nishiwaki, Y., Gotoh, K., Ozawa, T., and
Uchimura, M. 2002. Solitary splenic metastasis
from gastric cancer: report of a case. Surg. Today
32:1081–1084.
Index
AASLD. See American Association for the Study
of Liver Diseases guidelines
Abdomen
magnetic resonance imaging of, 21
T2-weighted images of, 22
2DT-FSE, 22
lesion characterization with, 22–23
single-shot echo planar, 22, 23
single-shot fast spin-echo, 22, 23
steady-state free precession, 22
Abscess, characterization of, 167
arterial phase, 167
parenchymal phase, 167
portal vein phase, 167
typical features, 167
ActiTest, 139, 140
Adenoma, characterization of, 166–167
arterial phase, 167
atypical features, 167
parenchymal phase, 167
portal vein phase, 167
typical features, 167
Adenomatous hyperplasia, computed
tomography of, 255
Adenosine triphosphate (ATP) synthase
beta chain, hepatitis C virus-related
human hepatocellular carcinoma
and, 347
ADI. See Agent detection imaging
AFROC. See Alternative-free response receiver
operating characteristic
Agent detection imaging (ADI), 163
contrast-enhanced, 151–153
AI-700, 161
Albunex, 161
Alcohol. See Fatty liver
Aldolase B (liver type), hepatitis C virus-
related human hepatocellular
carcinoma and, 348
Alpha-enolase, hepatitis C virus-related
human hepatocellular carcinoma
and, 347
Alpha-fetoprotein (AFP), for hepatocellular
carcinoma screening, 138–139, 174
Alternative-free response receiver operating
characteristic (AFROC), 231
American Association for the Study of Liver
Diseases (AASLD) guidelines, 177,
262–263
Angiography-assisted computed tomography,
36–37, 253
Angiosarcoma, characterization of, 171
Antiangiogenesis therapy, evaluation of, future
perspectives of, 282–284
APRI. See AST to platelet ratio index
Arginase 1, hepatitis C virus-related
human hepatocellular carcinoma
and, 348
Arterial portal microfistulas, contrast-enhanced
ultrasound and, 164–165
Aspartate transaminase (AST)/alanine
transaminase (ALT) ratio, 139, 140
AST/ALT. See Aspartate transaminase/alanine
transaminase ratio
AST to platelet ratio index (APRI), 139, 140
ATP. See Adenosine triphosphate synthase beta
chain
Atypical adenomatous hyperplasia, computed
tomography of, 255
AXOR12, 325
499
500
Bile cyst, characterization of, 166
Bile duct, intraductal papillary neoplasm of,
biliary cystic neoplasm, 419–426
bile duct lumen communication, 421
classification, 419–421
genetic aspects of, 422–424
histologic grade of tumor, 421–422
invasive carcinoma progression, 421–422
molecular aspects of, 422–424
mucin hypersecretion, 422
mucin phenotypes, 422
phenotypes, 422
radiological features, 421
Biliary cystic neoplasm, 416–426
cystadenocarcinoma, 416–419
clinicopathologic features of, 416–419
pancreas mucinous cystic neoplasm
comparison, 419
hepatobiliary cystadenoma
clinicopathologic features of, 416–419
pancreas mucinous cystic neoplasm
comparison, 419
intraductal papillary neoplasm of bile duct,
419–426
bile duct lumen communication, 421
classification, 419–421
genetic aspects of, 422–424
histologic grade of tumor, 421–422
invasive carcinoma progression, 421–422
molecular aspects of, 422–424
mucin hypersecretion, 422
mucin phenotypes, 422
phenotypes, 422
radiological features, 421
Biliary hamartoma, non-neoplastic biliary cystic
lesions, 415–416
multicystic biliary hamartoma, 415–416
Von Meyenburg complexes, 415
Biliary tree, anatomical classification of, 412
Biopsy
fine needle aspiration, of combined hepatocellular
and cholangiocarcinoma, 242
of liver, 139, 174
needle, 252–253
Biosphere, 161
Breast, splenic metastases to, 492
Carcinoembryonic antigen (CEA), positron
emission tomography and, 57–58
CDS. See Cirrhosis discriminant score
Index
CE-IOUS. See Contrast-enhanced intraoperative
ultrasonography
Cervix, splenic metastases to, 492
CEUS. See Contrast-enhanced ultrasound
CHA. See Coded harmonic angio
cHCC-CC. See Combined hepatocellular and
cholangiocarcinoma
Chemotherapy
intraarterial, 287–289
monitoring of, 2-fluoro-2-deoxy-D-glucose-
positron emission tomography in, 14–15
CHFCs. See Ciliated hepatic foregut cysts
Cholangiocarcinoma
combined hepatocellular and
differential diagnosis of, 246–247
enhanced computed tomographic findings,
243–247
pathogenesis of, 242–243
endoscopic retrograde
cholangiopancreatography for, 10
2-fluoro-2-deoxy-D-glucose-positron emission
tomography for, 10–11
intraductal ultrasonography for
cancer extension precise diagnosis, 432
findings of, 432
indications, 431
insertion methods, 430–431
instruments, 429–430
methods, 429–431
percutaneous transhepatic approach, 430–431
three-dimensional, 432–433
transpapillary approach, 430
magnetic resonance cholangiopancreatography
for, 10
magnetic resonance imaging for, 10
percutaneous transhepatic cholangiography
for, 10
three-dimensional intraductal ultrasonography
diagnostic accuracy for extension, 433
dual plane reconstruction images, 432
oblique reconstruction images, 432–433
tumor volume measurement, 433
Ciliated hepatic foregut cysts (CHFCs), 414–415
Cirrhosis discriminant score (CDS), 139, 140
Cirrhosis, of liver
epidemiology of, 137–138
surrogate tests for, 139–141
tumor arising in, 35–36
Coagulative necrosis, 87
Coded harmonic angio (CHA), 150–151
Index
Colon/rectum, splenic metastases to, 492
Colorectal cancer
liver metastases and, 66–67
detection of, 358–359
diagnostic hepatic ultrasound advantages, 357
Doppler flow pattern, 359–363
future advancements in, 363–364
grayscale echo pattern, 357–358
postoperative follow-up ultrasound, 363
ultrasound scanning technique for,
356–357
liver metastases and, 18F-fluorodeoxyglucose
positron emission tomography, 389–404
abdominal cavity, 392–395
additional disease and, 398–399
anticancer systemic therapy, 401–402
benign increased activity of, 392, 397
change of management, 398–399
contribution of, 393, 397
cryosurgery ablation, 403
detection of, 389–391
distant lymph nodes metastases, 395
extra-abdominal organs, 395–397
extrahepatic metastases detection, 391–392
hepatic arterial chemotherapy, 403–404
on patient’s management, 397–398
pelvic recurrence of rectal cancer, 393–395
peritoneal metastases, 395
physiologic activity of, 392, 397
preoperative staging, 389
radiofrequency ablation, 403
regional therapy for, 401, 403
rising carcinoembryonic antigen, 400–401
survival impact, 399–400
systemic therapy for, 401
Combined hepatocellular and cholangiocarcinoma
(cHCC-CC)
differential diagnosis of, 246–247
enhanced computed tomographic findings,
243–247
pathogenesis of, 242–243
Computed tomography (CT)
of adenomatous hyperplasia, 255
of advanced hepatocellular carcinoma, 253–254
angiography, 36–37, 228–229
angiography combined, 253
of atypical adenomatous hyperplasia, 255
contrast material fixed injection duration, 239
contrast material fixed injection rate, 237–239
of early hepatocellular carcinoma, 254–255
501
enhanced findings, combined hepatocellular and
cholangiocarcinoma, 243–247
of localized fibrous tumors of liver, 17–19
multidetector, methods of, 253
of nodule-in-nodule hepatocellular
carcinoma, 255
spiral, for hepatocellular carcinoma, 265–266
vascular endothelial growth factor, Western
blotting measurement of, 202–209
Computed tomography during arterial portography
(CTAP), 37–39, 214, 253
Computed tomography during hepatic
arteriography (CTHA), 37, 39, 253
Contrast-enhanced agent detection imaging,
151–153
Contrast-enhanced coded harmonic sonography,
150–151
Contrast-enhanced color Doppler sonography,
146–147
Contrast-enhanced harmonic power Doppler
sonography, 147–148
Contrast-enhanced intraoperative ultrasonography
(CE-IOUS), 63–64, 65–66, 176
classification by, 66
colorectal cancer, liver metastases and, 66–67
liver metastases, colorectal cancer and, 66–67
tumor location by, 67–69
Contrast-enhanced power Doppler
sonography, 147
Contrast-enhanced pulse-inversion harmonic
sonography, 148–150
Contrast-enhanced sonography, hepatocellular
carcinoma and, 153–156
Contrast-enhanced ultrasound (CEUS). See also
Harmonic imaging; Nonlinear contrast
enhancement
arterial portal microfistulas and, 164–165
for hepatocellular carcinoma, 264–265
physical background of, 162–163
technological background of, 162–163
Contrast Pulse Sequencing, 163
Cryoablation, 85
Cryosurgery ablation, colorectal cancer, liver
metastases and, 18F-fluorodeoxyglucose
positron emission tomography, 403
CT. See Computed tomography
CTAP. See Computed tomography during arterial
portography
CTHA. See Computed tomography during hepatic
arteriography
502 Index

Cystadenocarcinoma, biliary cystic neoplasm, Endoscopic retrograde cholangiopancreatography

416–419 (ERCP), for cholangiocarcinoma, 10
clinicopathologic features of, 416–419 Enoyl-CoA hydratase, hepatitis C virus-related
pancreas mucinous cystic neoplasm human hepatocellular carcinoma and, 348
comparison, 419 ERCP. See Endoscopic retrograde
Cyst, characterization of, 166 cholangiopancreatography
arterial phase, 166 Esophagus, splenic metastases to, 492
bile, 166 Ethanol injection, 85
hydatid, 166 percutaneous, 153
nonvital parasitic, 166 European Association for the Study of Liver
parenchymal phase, 166 Disease (EASLD), 177
portal vein phase, 166 European Association for the Study of the Liver
typical features, 166 (EASL) consensus, 262–263
DCE-MRI. See Dynamic contrast-enhanced Extrahepatic bile duct carcinoma (EHBDC)
magnetic resonance imaging diagnosis of, 442
epidemiology of, 441
Definity, 161 etiology of, 441–442
Diabetes. See Fatty liver mucin characteristics
Diffuse-weighted magnetic resonance imaging, immunohistochemistry, 458
189–190 methods, 455–458
Diffusion-weighted imaging (DWI), hepatic vessel MUC1, 453–455
signal suppression, 214–215 MUC4, 452–453
Doppler flow pattern, liver metastases, colorectal pathology of, 442
cancer and, 359–363 prognosis of, 443
Doppler sonography role of p53 in malignant tumors, 443–444
contrast-enhanced color, 146–147 mutation analysis, 444
contrast-enhanced harmonic power, 147–148 role of p63 in malignant tumors, 444–448
contrast-enhanced power, 147 analysis of, 447–448
Doppler-ultrasonography, for hepatocellular role of p73 in malignant tumors, 444–448
carcinoma, 264–265 analysis of, 447–448
Drugs. See Fatty liver Fallopian tube, splenic metastases to, 492
Dual plane reconstruction images,
cholangiocarcinoma, three-dimensional Fast spin echo (FSE), 215–216
intraductal ultrasonography, 432 Fatty liver, characterization of, 165–166
Dynamic contrast-enhanced magnetic resonance arterial phase, 166
imaging (DCE-MRI), 283 focal fatty change, 166
Dynamic magnetic resonance imaging, 39–40 focal fatty sparing, 165–166
Dysplastic nodule, characterization of parenchymal phase, 166
arterial phase, 168 portal vein phase, 166
parenchymal phase, 168 typical features, 166
portal vein phase, 168 Fatty Liver Shionogi (FLS) mouse, 334
typical features, 168 FDG-PET. See 18F-fluorodeoxyglucose positron
EASL. See European Association for the Study emission tomography; 2-fluoro-2-
of the Liver consensus deoxy-D-glucose-positron emission
tomography
EASLD. See European Association for the Study Ferritin light chain, hepatitis C virus-related human
of Liver Disease hepatocellular carcinoma and, 348
18
Echovist, 161 F-fluorodeoxyglucose positron emission
EHBDC. See Extrahepatic bile duct carcinoma tomography (FDG-PET)
Endometrium, splenic metastases to, 492 colorectal liver metastases and, 389–404
Index 503

abdominal cavity, 392–395 portal vein phase, 167

additional disease and, 398–399 typical features, 167
anticancer systemic therapy, 401–402 adenoma, characterization of, 166–167
benign increased activity of, 392, 397 arterial phase, 167
change of management, 398–399 atypical features, 167
contribution of, 393, 397 parenchymal phase, 167
cryosurgery ablation, 403 portal vein phase, 167
detection of, 389–391 typical features, 167
distant lymph nodes metastases, 395 angiosarcoma, characterization of, 171
extra-abdominal organs, 395–397 cyst, characterization of, 166
extrahepatic metastases detection, 391–392 arterial phase, 166
hepatic arterial chemotherapy, 403–404 bile, 166
on patient’s management, 397–398 hydatid, 166
pelvic recurrence of rectal cancer, 393–395 nonvital parasitic, 166
peritoneal metastases, 395 parenchymal phase, 166
physiologic activity of, 392, 397 portal vein phase, 166
preoperative staging, 389 typical features, 166
radiofrequency ablation, 403 detection of, 171–174
regional therapy for, 401, 403 hepatic, 172
rising carcinoembryonic antigen, 400–401 hepatocellular carcinoma surveillance,
survival impact, 399–400 173–174
systemic therapy for, 401 oncologic patient follow-up, 172–173
patient outcome of, 54–55 staging, 172–173
pitfalls of, 388 dysplastic nodule, characterization of
principle of, 388 arterial phase, 168
Fibroscan, 139 parenchymal phase, 168
Fibrosis, of liver, surrogate tests for, 139–141 portal vein phase, 168
Fibrotest, 139, 140 typical features, 168
Fine needle aspiration biopsy (FNAB), fatty liver, characterization of, 165–166
of combined hepatocellular and arterial phase, 166
cholangiocarcinoma, 242 focal fatty change, 166
FLLs. See Focal liver lesions focal fatty sparing, 165–166
FLS. See Fatty Liver Shionogi mouse parenchymal phase, 166
2-fluoro-2-deoxy-D-glucose-positron emission portal vein phase, 166
tomography (FDG-PET) typical features, 166
for chemotherapy monitoring, 14–15 hemangioma, characterization of, 164–165
for cholangiocarcinoma, 10–11 arterial phase, 165
for gallbladder carcinoma, 11–12 atypical features, 165
for hepatocellular carcinoma, 12–14 capillary, 164
for metastatic liver disease, 7–10 cavernous, 164
false-negative findings, 8 focal liver lesions, nonlinear contrast
false-positive findings, 8 enhancement of, 164–165
FNAB. See Fine needle aspiration biopsy parenchymal phase, 165
(FNAB) portal vein phase, 165
Focal liver lesions (FLLs), 66 typical features, 165
Focal liver lesions (FLLs), nonlinear contrast hepatocellular carcinoma, characterization of
enhancement of arterial phase, 168
abscess, characterization of, 167 parenchymal phase, 168
arterial phase, 167 portal vein phase, 168
parenchymal phase, 167 typical features, 168
504
Focal liver lesions (FLLs), nonlinear contrast
enhancement of (Continued)
histiocytoma, characterization of, 171
inflammatory pseudo tumors, characterization
of, 167–168
leiomyoma, characterization of, 171
liver metastases, characterization of, 168–171
arterial phase, 170, 171
hyper-vascular, 171
hypo-vascular, 170
parenchymal phase, 170, 171
portal vein phase, 170, 171
typical features, 170, 171
lymphoma, characterization of, 171
percutaneous ablative treatment monitoring,
174–176
regenerative nodule, characterization of, 166
arterial phase, 166
parenchymal phase, 166
portal vein phase, 166
typical features, 166
sarcomas, characterization of, 171
Focal nodular hyperplasia, characterization of, 165
arterial phase, 165
atypical features, 165
parenchymal phase, 165
portal vein phase, 165
typical features, 165
Forns index, 139, 140
FSE. See Fast spin echo
Gadobenate dimeglumine (Gd-BOPTA), 89, 227
Gadolinium
hepatocellular carcinoma, enhancement of,
185–187
superparamagnetic iron oxide and, 187
Gadolinium chelates, 39–40, 88
Gadoxetic acid (Gd-ethoxybenzyldiethylenetriami
nepentaacetic acid; Gd-EOB-DTPA), 89,
94, 227
Gallbladder carcinoma, 2-fluoro-2-deoxy-D-
glucose-positron emission tomography
for, 11–12
Gd-BOPTA. See Gadobenate dimeglumine
Gd-EOB-DTPA. See Gadoxetic acid
Gd-ethoxybenzyldiethylenetriaminepentaacetic
acid. See Gadoxetic acid
Glucose-regulated 75 kDa protein (GRP75),
hepatitis C virus-related human
hepatocellular carcinoma and, 347
Index
Glucose-regulated 78 kDa protein (GRP78),
hepatitis C virus-related human
hepatocellular carcinoma and, 347
Glutamine synthetase, hepatitis C virus-related
human hepatocellular carcinoma and, 347
Gradient-recalled echo (GRE), 184
Grayscale echo pattern, liver metastases,
colorectal cancer and, 357–358
GRE. See Gradient-recalled echo
GRP75. See Glucose-regulated 75 kDa protein
GRP78. See Glucose-regulated 78 kDa protein
HAIC. See Hepatic arterial infusion
chemotherapy
HAP. See Hepatic arterial-dominant phase
scanning
Harmonic imaging, 160–161
advantage of, 160
disadvantage of, 160
HBV. See Hepatitis B virus
HCC. See Hepatocellular carcinoma
HCV. See Hepatitis C virus
HCV-HCC. See Hepatitis C virus-related human
hepatocellular carcinoma
HDGF. See Hepatoma-derived growth factor
Heat shock 60 kDa protein (HSP60), hepatitis
C virus-related human hepatocellular
carcinoma and, 347
Heat shock 70 kDa protein 1 (HSP70.1), hepatitis
C virus-related human hepatocellular
carcinoma and, 347
Heat shock cognate 71 kDa protein (HSC70),
hepatitis C virus-related human
hepatocellular carcinoma and, 347
Hemangioma, characterization of
arterial phase, 165
atypical features, 165
capillary, 164
cavernous, 164
focal liver lesions, nonlinear contrast
enhancement of, 164–165
parenchymal phase, 165
portal vein phase, 165
typical features, 165
Hepatic arterial-dominant phase (HAP)
scanning, 237
Hepatic arterial infusion chemotherapy (HAIC),
287, 289–290
Hepatic foregut cyst, non-neoplastic biliary cystic
lesions, 414–415
Index 505

Hepatic intraarterial catheters, perfusion territory characterization of

visualization, 291–298 arterial phase, 168
image analysis, 293–297 parenchymal phase, 168
imaging techniques, 291 portal vein phase, 168
single-photon emission computed tomography typical features, 168
with integrated computed tomography contrast-enhanced sonography and, 153–156
for port perfusion scintigraphy, 291–293 development of, 196–199
therapeutic consequences, 297–298 diagnosis accuracy of, 231–232
Hepatic steatosis. See Fatty liver diagnosis of, magnetic resonance imaging for,
Hepatitis B virus (HBV), 137–138, 183 267–268
Hepatitis C virus (HCV), 137–138, 183 early, computed tomography images of, 254–255
Hepatitis C virus-related human hepatocellular epidemiology of, 137–138
carcinoma (HCV-HCC), 343 2-fluoro-2-deoxy-D-glucose-positron emission
glucose-regulated 75 kDa protein, 347 tomography for, 12–14
glucose-regulated 78 kDa protein, 347 gadolinium enhancement of, 185–187
glutamine synthetase, 347 hepatitis C virus-related human, 343
heat shock 60 kDa protein, 347 hepatoma-derived growth factor
heat shock 70 kDa protein 1, 347 evaluation of, 337–338
heat shock cognate 71 kDa protein, 347 prognostic significance, 338–340
proteomics for, 345–348 homeoprotein six1 overexpression
amino acid sequencing by liquid cell lines in, 314, 315–317
chromatography-tandem mass clinical samples of, 314
spectrometry, 346 materials in, 314–315
gel image analysis, 346 methods in, 314–315
in-gel digestion, 346 results of, 315–319
isoelectric focusing, 345 reverse transcription-polymerase chain
mass spectrometry analysis, 346 reaction, 314–315
proteins expression decreased in, 347–348 statistical analysis, 315
proteins expression increased in, 346–347 Western blot, 315
sodium dodecyl sulfate-polyacrylamide gel immunohistochemistry, research on, 200–202
electrophoresis, 345–346 KiSS-1 overexpression
tissue sample preparation, 345 immunohistochemistry, 328
two-dimensional gel electrophoresis, 345 materials, 327
Hepatobiliary cystadenoma, biliary cystic methods, 327–328
neoplasm results of, 328–330
clinicopathologic features of, 416–419 tissue preparation, 327
pancreas mucinous cystic neoplasm magnetic resonance imaging
comparison, 419 appearance of, 184–185
Hepatocarcinogenesis, multistep progression features of, 229–230
of, 250–251 nodule-in-nodule, computed tomography of, 255
Hepatocellular carcinoma (HCC) postoperative interferon alpha treatment of
after percutaneous radiofrequency ablation clinicopathological data in, 305–306
therapy, 232–233 disease-free survival of, 306–307
after transcatheter arterial immunohistochemistry, 304
chemoembolization, 232 materials in, 304–305
alpha-fetoprotein and, 138–139, 174 methods in, 304–305
atypical appearances of, 188–189 overall survival of, 306–307
diffuse, 189 results of, 305–307
large, 188–189 scoring of p48 immunohistochemistry, 304–305
small, 188 statistical analysis, 305
506
Hepatocellular carcinoma (HCC) (Continued)
survival of, 306
tissue microarray, 304
proteomicss for, 344
antibodies and, 348–349
radiologic diagnosis history of, 194–196
resection of, 69–73
limited, 70–71
parenchymal dissection, 71–72
portal branch, compression of, 70
portal branch, hooking of, 70
portal branch, puncture of, 69–70
post-resectional control, 72–73
systematic segmentectomy, 69
screening benefit of, 138–139
cost-effectiveness analyses of, 139
ultrasonographic, 141–142
small nodules and, 268–270
staging of, 231
surveillance for
contrast-enhanced ultrasonography, 264–265
Doppler-ultrasonography, 264–265
power Doppler-ultrasonography, 264–265
radiologic techniques in, 263–264
spiral computed tomography, 265–266
ultrasonography, 264–265
thalidomide and, vascular response evaluation,
279–282
tumor vascularity evaluation, with power
Doppler ultrasound, 279
ultrasonography and, 138–139
vascular endothelial growth factor expression,
196–199
vasculature of, 196–199
Hepatoma-derived growth factor (HDGF),
333–336
analytical methods, 336–337
angiogenesis, role in, 335–336
cancer progression, role in, 335–336
developmentally regulated expression of, 334
hepatocarcinogenesis, role in, 334–335
in hepatocellular carcinoma, evaluation of,
337–338
in hepatocellular carcinoma, prognostic
significance, 338–340
immunohistochemical methods, 336–337
materials in, 336–337
High-frequency ultrasound
imaging of, preclinical liver metastases, 376–379
preclinical liver metastases, 373–374
Index
Hilar cholangiocarcinoma, palliation of,
photodynamic therapy for, 472–478
Histiocytoma, characterization of, 171
Hodgkin’s lymphoma, 171
Homeoprotein six1 overexpression, in
hepatocellular carcinoma, 314, 315–317
cell lines in, 314, 315–317
clinical samples of, 314
materials in, 314–315
methods in, 314–315
results of, 315–319
reverse transcription-polymerase chain reaction,
314–315
statistical analysis, 315
Western blot, 315
HSC70. See Heat shock cognate 71 kDa protein
HSP60. See Heat shock 60 kDa protein
HSP70.1. See Heat shock 70 kDa protein 1
Hydatid cyst, characterization of, 166
Hyper-vascular metastases, characterization of, 171
Hypo-attenuating nodular lesions
natural outcome of, 256
treatment of, 257
Hypo-vascular metastases, characterization of, 171
IEF. See Isoelectric focusing
Imagent, 161
iMRI. See Intraoperative magnetic resonance
imaging
Inflammatory pseudo tumors, characterization
of, 167–168
Interferon alpha treatment, postoperative, of
hepatocellular carcinoma
clinicopathological data in, 305–306
disease-free survival of, 306–307
immunohistochemistry, 304
materials in, 304–305
methods in, 304–305
overall survival of, 306–307
results of, 305–307
scoring of p48 immunohistochemistry, 304–305
statistical analysis, 305
survival of, 306
tissue microarray, 304
Intraarterial chemotherapy, 287–289
Intraductal papillary neoplasm of bile duct (IPNB),
biliary cystic neoplasm, 419–426
bile duct lumen communication, 421
classification, 419–421
genetic aspects of, 422–424
Index 507

histologic grade of tumor, 421–422 Laser-induced thermotherapy (LITT), 85

invasive carcinoma progression, 421–422 LC-MS/MS. See Liquid chromatography-tandem
molecular aspects of, 422–424 mass spectrometry
mucin hypersecretion, 422 Leiomyoma, characterization of, 171
mucin phenotypes, 422 LFTs. See Localized fibrous tumors
phenotypes, 422 Liquid chromatography-tandem mass
radiological features, 421 spectrometry (LC-MS/MS), amino acid
Intraductal ultrasonography (IDUS), sequencing by, for hepatitis C
cholangiocarcinoma virus-related human hepatocellular
cancer extension precise diagnosis, 432 carcinoma, 346
findings of, 432 LITT. See Laser-induced thermotherapy
indications, 431 Liver. See also Hepatocellular carcinoma
insertion methods, 430–431 biopsy of, 139, 174
instruments, 429–430 cirrhosis of
methods, 429–431 epidemiology of, 137–138
percutaneous transhepatic approach, 430–431 surrogate tests for, 139–141
three-dimensional, 432–433 tumor arising in, 35–36
transpapillary approach, 430 exploration of, intraoperative ultrasonography
Intraoperative magnetic resonance imaging for, 64–69
(iMRI), 81, 88 fibrosis of, surrogate tests for, 139–141
developments in, 96–97 localized fibrous tumors of, 17–19
radiofrequency ablation and, 90–98 noncirrhotic, tumor arising in, 36
conduct of, 92–96 radiofrequency ablation of, 14
future developments of, 96–98 transplant of, 83
potential indications for, 92 Liver metastases
rationale of, 90 characterization of, 168–171
technical developments in, 90–92 arterial phase, 170, 171
Intraoperative ultrasonography (IOUS) hyper-vascular, 171
contrast-enhanced, 63–64 hypo-vascular, 170
in hepatic surgery parenchymal phase, 170, 171
liver exploration in, 64–69 portal vein phase, 170, 171
technical aspects of, 63–64 typical features, 170, 171
tumor vessel classification, 68–69 colorectal cancer and, 66–67
IOUS. See Intraoperative ultrasonography detection of, 358–359
IPNB. See Intraductal papillary neoplasm diagnostic hepatic ultrasound advantages, 357
of bile duct Doppler flow pattern, 359–363
Isoelectric focusing (IEF), for hepatitis C virus-related future advancements in, 363–364
human hepatocellular carcinoma, 345 grayscale echo pattern, 357–358
postoperative follow-up ultrasound, 363
Ketohexokinase, hepatitis C virus-related human ultrasound scanning technique for,
hepatocellular carcinoma and, 348 356–357
Kidney, splenic metastases to, 492 colorectal cancer and, 18F-fluorodeoxyglucose
KiSS-1 overexpression, in hepatocellular positron emission tomography, 389–404
carcinoma abdominal cavity, 392–395
immunohistochemistry, 328 additional disease and, 398–399
materials, 327 anticancer systemic therapy, 401–402
methods, 327–328 benign increased activity of, 392, 397
results of, 328–330 change of management, 398–399
tissue preparation, 327 contribution of, 393, 397
Kupffer cells, of liver, 40–41 cryosurgery ablation, 403
508
Liver metastases (Continued)
detection of, 389–391
distant lymph nodes metastases, 395
extra-abdominal organs, 395–397
extrahepatic metastases detection, 391–392
hepatic arterial chemotherapy, 403–404
on patient’s management, 397–398
pelvic recurrence of rectal cancer, 393–395
peritoneal metastases, 395
physiologic activity of, 392, 397
preoperative staging, 389
radiofrequency ablation, 403
regional therapy for, 401, 403
rising carcinoembryonic antigen, 400–401
survival impact, 399–400
systemic therapy for, 401
2-fluoro-2-deoxy-D-glucose-positron emission
tomography for, 7–10
false-negative findings, 9
false-positive findings, 8
liver dissection and, 73–74
magnetic resonance imaging for, 7–8
positron emission tomography assessment
of, 104–105
post-resectional control of, 73–74
preclinical
bioluminescent imaging, 372
fluorescent imaging, 372
high-frequency ultrasound, 373–374
high-frequency ultrasound imaging of,
376–379
magnetic resonance imaging and, 371
positron emission tomography, 372
three-dimensional imaging, 374–376
volume calculation, 374–376
X-ray computed tomography, 371–372
resection of, 73–74
area definition in, 73
Liver tumors
cryoablation for, 85
ethanol injection for, 85
radiofrequency ablation for, 81–85
alternatives of, 85
indications for, 83–84
outcomes of, 84–85
principles of, 81–83
Localized fibrous mesothelioma. See Localized
fibrous tumors
Localized fibrous tumors (LFTs), 17–19
Low mechanical index imaging, 153
Index
Lung, splenic metastases to, 492
Lymphoma, characterization of, 171
MAA. See Macroaggregated albumin
Macroaggregated albumin (MAA), 105
Magnetic resonance cholangiopancreatography
(MRCP), 429
for cholangiocarcinoma, 10
Magnetic resonance imaging (MRI), 39, 225
3T, 190
of abdomen, 21
for cholangiocarcinoma, 10
contrast-enhanced dynamic, 226
diffuse-weighted, 189–190
dynamic, 39–40
hepatocellular carcinoma, appearance
of, 184–185
for hepatocellular carcinoma diagnosis,
267–268
hepatocellular carcinoma, features of, 229–230
of localized fibrous tumors of liver, 17–19
for metastatic liver disease, 7–8
perfusion-weighted, 190
preclinical liver metastases and, 371
scanning technique, 225–226
subtraction images and, 88
superparamagnetic iron oxide-enhanced,
213–214
of hepatocellular carcinoma, 230–231
technique of, 183–184
thermal injury and, 86–90
characteristics of, 86–88
extent of, 88–90
functional potential utility, 89–90
tissue-specific contrast agents, 187–188
tissue-specific contrast-enhanced, 226–227
vascular endothelial growth factor, Western
blotting measurement of, 202–209
Malignant duodenal obstruction, palliative stenting
of, 469–470
Mangafodipir trisodium (Mn-DPDP), 227
MCT. See Microwave coagulation therapy
MDCT. See Multi-detector computed tomography
Metastases, splenic
clinical features, 491–493
diagnostic procedures, 493–494
differential diagnosis, 494–495
epidemiology of, 489
pathogenesis of, 490–491
pathologic features, 489–490
Index 509

Metastatic liver disease duration of, 223

characterization of, 168–171 injection flow rate, 223
arterial phase, 170, 171 iodine concentration, 223
hyper-vascular, 171 image processing, 224–225
hypo-vascular, 170 optimal scanning delay, 223–224
parenchymal phase, 170, 171 scanning technique, 222
portal vein phase, 170, 171 scanning timing, 224
typical features, 170, 171 Multi-shot 2DFT-FSE, 22
colorectal cancer and, 66–67
2-fluoro-2-deoxy-D-glucose-positron emission Needle biopsy, 252–253
tomography for, 7–10 Nodular hepatocellular lesions, classification
liver dissection and, 73–74 of, 249–250
magnetic resonance imaging for, 7–8 Nodule-in-nodule hepatocellular carcinoma,
positron emission tomography assessment computed tomography of, 255
of, 104–105 Noncirrhotic liver, tumor arising in, 36
post-resectional control of, 73–74 Non Hodgkin’s lymphoma, 171
resection area definition in, 73 Nonlinear contrast enhancement, of focal liver
resection of, 73–74 lesions, 163–176
Microbubbles, 161 abscess, characterization of, 167
physics of, 146 arterial phase, 167
Microwave coagulation therapy (MCT), 85 parenchymal phase, 167
Mn-DPDP. See Mangafodipir trisodium portal vein phase, 167
Motion-probing gradient (MPG) pulses, 214 typical features, 167
MPG. See Motion-probing gradient pulses adenoma, characterization of, 166–167
MRCP. See Magnetic resonance arterial phase, 167
cholangiopancreatography atypical features, 167
MRI. See Magnetic resonance imaging parenchymal phase, 167
MUC1, 453–455 portal vein phase, 167
antibodies against, 455 typical features, 167
MUC4, 452–453 angiosarcoma, characterization of, 171
antibodies against, 453 cyst, characterization of, 166
Mucin characteristics, extrahepatic bile duct arterial phase, 166
carcinoma bile, 166
immunohistochemistry, 458 hydatid, 166
methods, 455–458 nonvital parasitic, 166
MUC1, 453–455 parenchymal phase, 166
MUC4, 452–453 portal vein phase, 166
Mucin hypersecretion, intraductal papillary typical features, 166
neoplasm of bile duct, 422 detection of, 171–174
Mucin phenotypes, intraductal papillary neoplasm hepatic, 172
of bile duct, 422 hepatocellular carcinoma surveillance,
Multicystic biliary hamartoma, 415–416 173–174
Multi-detector computed tomography (MDCT), oncologic patient follow-up, 172–173
33–34, 42, 429 staging, 172–173
limitations of, 233 dysplastic nodule, characterization of
prospects of, 233 arterial phase, 168
Multidetector-row helical computed tomography parenchymal phase, 168
(MDCT), 221 portal vein phase, 168
contrast administration for, 222–223 typical features, 168
dose, 222–223 fatty liver, characterization of, 165–166
510 Index

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
T2 fast spin echo imaging using,
216–218
PET. See Positron emission tomography
PGAA index, 139, 140
PGA index, 139, 140
Photodynamic therapy (PDT), 470–479
contraindications for, 471–472
for hilar cholangiocarcinoma palliation,
472–478
mechanism of action, 470–471
photosensitizers and, 472
Photosensitizers, photodynamic therapy and, 472
PIHI. See Pulse-inversion harmonic sonography
(PIHI)
Platelet counts, 139, 140
Portal branch
compression of, in hepatocellular carcinoma, 70
hooking of, in hepatocellular carcinoma, 70
puncture of, in hepatocellular carcinoma, 69–70
Positron emission tomography (PET)
for detection of extrahepatic disease, 50–51
for detection of intrahepatic disease, 51–52
in elevated carcinoembryonic antigen, 57–58
for follow-up, 57
in follow-up post hepatic resection, 58
liver metastases and
negative impact of, 54
positive impact of, 53–54
metastatic liver disease, assessment of, 104–105
preclinical liver metastases and, 372
as staging modality, to complement
conventional imaging, 50
for surveillance, 57
value of, prognostic clinical risk score, 52–53
in work-up of hepatic colorectal metastases,
58–59
Yttrium-90 microsphere radioembolization and,
109–112
Positron emission tomography (PET),
18
F-fluorodeoxyglucose (FDG)
colorectal liver metastases and, 389–404
abdominal cavity, 392–395
additional disease and, 398–399
anticancer systemic therapy, 401–402
benign increased activity of, 392, 397
change of management, 398–399
contribution of, 393, 397
cryosurgery ablation, 403
detection of, 389–391
distant lymph nodes metastases, 395
511
extra-abdominal organs, 395–397
extrahepatic metastases detection, 391–392
hepatic arterial chemotherapy, 403–404
on patient’s management, 397–398
pelvic recurrence of rectal cancer, 393–395
peritoneal metastases, 395
physiologic activity of, 392, 397
preoperative staging, 389
radiofrequency ablation, 403
regional therapy for, 401, 403
rising carcinoembryonic antigen, 400–401
survival impact, 399–400
systemic therapy for, 401
patient outcome of, 54–55
pitfalls of, 388
principle of, 388
Postoperative interferon alpha treatment,
of hepatocellular carcinoma
clinicopathological data in, 305–306
disease-free survival of, 306–307
immunohistochemistry, 304
materials in, 304–305
methods in, 304–305
overall survival of, 306–307
results of, 305–307
scoring of p48 immunohistochemistry, 304–305
statistical analysis, 305
survival of, 306
tissue microarray, 304
Power-Doppler ultrasound, for hepatocellular
carcinoma, 264–265
for tumor vascularity evaluation of, 279
Prebiliary cysts, non-neoplastic biliary cystic
lesions, 412–414
around extrahepatic bile ducts, 414
in hepatic hilum, 412–414
in large portal tracts, 412–414
Preclinical liver metastases
bioluminescent imaging, 372
fluorescent imaging, 372
high-frequency ultrasound, 373–374
high-frequency ultrasound imaging of, 376–379
magnetic resonance imaging and, 371
positron emission tomography and, 372
three-dimensional imaging and, 374–376
volume calculation, 374–376
Preoperative biliary drainage, 464
PROPELLER. See Periodically Rotated
Overlapping ParallEL Lines with
Enhanced Reconstruction
512
Prostate, splenic metastases to, 492
Proteomics, 343–344
for hepatitis C virus-related human
hepatocellular carcinoma, 345–348
amino acid sequencing by liquid
chromatography-tandem mass
spectrometry, 346
gel image analysis, 346
in-gel digestion, 346
isoelectric focusing, 345
mass spectrometry analysis, 346
proteins expression decreased in, 347–348
proteins expression increased in, 346–347
sodium dodecyl sulfate-polyacrylamide gel
electrophoresis, 345–346
tissue sample preparation, 345
two-dimensional gel electrophoresis, 345
for hepatocellular carcinoma tissues, 344
Prothrombin time, 139, 140
PTC. See Percutaneous transhepatic
cholangiography
PTK787/ZK 222584 (PTK/ZK), 193
Pulse-inversion harmonic sonography (PIHI),
contrast-enhanced, 148–150
Quantison, 161
Radial fast spin-echo, 27–30
Radioembolization, 290–291
Radiofrequency ablation (RFA)
endpoint monitoring of, 86
impedance, 86
temperature, 86
ultrasound and, 86
future perspective of, 128–129
intraoperative magnetic resonance imaging
and, 90–98
conduct of, 92–96
future developments of, 96–98
potential indications for, 92
rationale of, 90
technical developments in, 90–92
of liver, 14
for liver tumor treatment, 81–85
alternatives of, 85
indications for, 83–84
outcomes of, 84–85
principles of, 81–83
methodology of, 118–119
laparoscopic advantages of, 118
Index
laparoscopic disadvantages of, 119
open surgical advantages of, 119
open surgical disadvantages of, 119
percutaneous advantages of, 118
percutaneous disadvantages of, 118
operative (open), 83
percutaneous, 83
prognosis using, 120–124
intraoperative evaluation, 121
long-term results, 123–124
methods and, 120, 121–122
patients and, 120
postoperative follow-up, 122
preoperative evaluation, 121
short-term results, 123
statistical analyses, 122
tumor unresectability, 120–121
role of, 124–128
Radiologic diagnosis history, of hepatocellular
carcinoma, 194–196
vascular endothelial growth factor expression
and, 199
Receiver operating characteristic (ROC)
analyses, 231
Regenerative nodule, characterization of, 166
arterial phase, 166
parenchymal phase, 166
portal vein phase, 166
typical features, 166
Resection
of hepatocellular carcinoma (HCC), 69–73
limited, 70–71
parenchymal dissection, 71–72
portal branch, compression of, 70
portal branch, hooking of, 70
portal branch, puncture of, 69–70
post-resectional control, 72–73
systematic segmentectomy, 69
of liver metastases, 73–74
liver dissection, 73–74
post-resectional control, 73–74
resection area definition, 73
RFA. See Radiofrequency ablation
ROC. See Receiver operating characteristic analyses
Sarcomas, characterization of, 171
SDS-PAGE. See Sodium dodecyl sulfate-
polyacrylamide gel electrophoresis
Serum albumin, hepatitis C virus-related human
hepatocellular carcinoma and, 348
Index
Single-photon emission computed tomography
(SPECT), 106
with integrated computed tomography for
port perfusion scintigraphy, hepatic
intraarterial catheters, perfusion territory
visualization, 291–293
Single-shot echo planar (SSEPI), 22, 23
Single-shot fast spin-echo (SSFSE), 22, 23
image quality improvement in, 24–27
motion artifact reduction in, 24–27
Skin, splenic metastases to, 492
Smoothelin, hepatitis C virus-related human
hepatocellular carcinoma and, 348
Sodium dodecyl sulfate-polyacrylamide gel
electrophoresis (SDS-PAGE), for
hepatitis C virus-related human
hepatocellular carcinoma, 345–346
Solitary fibrous tumors. See Localized fibrous tumors
Sonazoid, 161
SonoVue®, 64, 161
SPECT. See Single photon emission computed
tomography
SPIO. See Superparamagnetic iron oxide
Spiral computed tomography, for hepatocellular
carcinoma, 265–266
Splenic metastases
clinical features, 491–493
diagnostic procedures, 493–494
differential diagnosis, 494–495
epidemiology of, 489
pathogenesis of, 490–491
pathologic features, 489–490
SSEPI. See Single-shot echo planar
SSFP. See Steady-state free precession
SSFSE. See Single-shot fast spin-echo
Steady-state free precession (SSFP), 22
Stomach, splenic metastases to, 492
Subtraction images, magnetic resonance imaging
and, 88
Superparamagnetic iron oxide (SPIO), 39, 40–42
enhanced magnetic resonance imaging, 213–214
of hepatocellular carcinoma, 230–231
gadolinium and, 187
Systematic segmentectomy, in hepatocellular
carcinoma, 69
T2-weighted images, of abdomen, 22
2DT-FSE, 22
lesion characterization with, 22–23
single-shot echo planar, 22, 23
513
single-shot fast spin-echo, 22, 23
steady-state free precession, 22
TACE. See Transcatheter arterial
chemoembolization
Thalidomide, hepatocellular carcinoma and,
vascular response evaluation, 279–282
Thermal injury, magnetic resonance imaging and
characteristics of, 86–88
extent of, 88–90
functional potential utility, 89–90
Thorotrast vinyl chloride, 171
3D-IDUS. See Three-dimensional intraductal
ultrasonography
Three-dimensional intraductal ultrasonography
(3D-IDUS)
cholangiocarcinoma, diagnostic accuracy for
extension, 433
cholangiocsarcoma
dual plane reconstruction images, 432
oblique reconstruction images, 432–433
tumor volume measurement, 433
3T magnetic resonance imaging, 190
Thyroid, splenic metastases to, 492
Transarterial radioembolization, 289
Transcatheter arterial chemoembolization (TACE),
hepatocellular carcinoma after, 232
Triosephosphate isomerase, hepatitis C virus-
related human hepatocellular carcinoma
and, 347
Tropomyosin beta chain, hepatitis C virus-related
human hepatocellular carcinoma and, 348
Tuberculosis, 167
TUF. See Tumor uptake fraction
Tumor(s)
cryoablation for, 85
ethanol injection for, 85
in liver cirrhosis, 35–36
in noncirrhotic liver, 36
in oncologic patients, 36
radiofrequency ablation for, 81–85
alternatives of, 85
indications for, 83–84
outcomes of, 84–85
principles of, 81–83
Tumor uptake fraction (TUF), 108
Tumor volume measurement, cholangiocsarcoma,
three-dimensional intraductal
ultrasonography, 433
2DFT-FSE. See Two-dimensional Fourier
transform-fast spin-echo
514
Two-dimensional Fourier transform (2DFT) fast
spin-echo (FSE), 22
image quality improvement in, 24–27
motion artifact reduction in, 24–27
Two-dimensional gel electrophoresis (2-DE),
for hepatitis C virus-related human
hepatocellular carcinoma, 345
2-DE. See Two-dimensional gel electrophoresis
Ultrasound
contrast-enhanced
arterial portal microfistulas, 164–165
arterial portal microfistulas and, 164–165
physical background of, 162–163
technological background of, 162–163
for hepatocellular carcinoma, 264–265
screening, 138–139
high-frequency, preclinical liver metastases,
373–374
intraductal, cholangiocarcioma
cancer extension precise diagnosis, 432
findings of, 432
indications, 431
insertion methods, 430–431
instruments, 429–430
methods, 429–431
percutaneous transhepatic approach, 430–431
three-dimensional, 432–433
transpapillary approach, 430
intraoperative
contrast-enhanced, 63–64
tumor vessel classification, 68–69
postoperative follow-up, colorectal cancer, liver
metastases and, 363
power-Doppler
for hepatocellular carcinoma, 264–265
for tumor vascularity evaluation of, 279
hepatocellular carcinoma, tumor vascularity
evaluation, 279
radiofrequency ablation and, endpoint
monitoring of, 86
radiofrequency ablation endpoint monitoring
and, 86
Index
three-dimensional intraductal
for cholangiocarcinoma, diagnostic accuracy
for extension, 433
for cholangiocarcinoma, dual plane
reconstruction images, 432
for cholangiocarcinoma, oblique
reconstruction images, 432–433
for cholangiocarcinoma, tumor volume
measurement, 433
Uveal tract, splenic metastases to, 492
Vascular endothelial growth factor
(VEGF)
expression, hepatocellular carcinoma,
196–199
radiologic diagnosis history, 199
Western blotting measurement of, 202–209
VCMs. See Von Meyenburg complexes
VEGF. See Vascular endothelial growth factor
VIBE. See Volumetric Interpolated Breath-hold
Examination
Volumetric Interpolated Breath-hold Examination
(VIBE), 226
Von Meyenburg complexes (VMCs), 415
Western blotting measurement
computed tomography, vascular endothelial
growth factor, 202–209
hepatocellular carcinoma, homeoprotein six1
overexpression, 315
X-ray computed tomography, preclinical liver
metastases and, 371–372
Y-90. See Yttrium-90 microsphere
radioembolization
Yttrium-90 (Y-90) microsphere radioembolization,
105–112
administration of, 105–106
clinical results of, 109–111
patient evaluation in, 106–108
positron emission tomography and, 109–112
treatment planning of, 105–106