AJR 2005; 184:832-841
© American Roentgen Ray Society
Expression of Vascular Endothelial Growth Factor in Hepatocellular Carcinoma and the Surrounding Liver and Correlation with MRI Findings
Masayuki Kanematsu1,2,
Shinji Osada3,
Nozomi Amaoka3,
Satoshi Goshima1,
Hiroshi Kondo1,
Hiroki Kato1,
Hironori Nishibori1,
Ryujiro Yokoyama2,
Hiroaki Hoshi1 and
Noriyuki Moriyama4
1 Department of Radiology Services, Gifu University School of Medicine, 1-1
Yanagido, Gifu 501–1193, Japan.
2 Department of Radiology Services, Gifu University Hospital, Gifu 501-1193,
Japan.
3 Department of Surgical Oncology, Gifu University School of Medicine, Gifu
501-1193, Japan.
4 Department of Diagnostic Radiology, National Cancer Center Hospital, Tsukiji,
Japan.
Received May 2, 2004;
accepted after revision July 23, 2004.
Supported in part by the Grant for Scientific Research Expenses for Health,
Labor and Welfare Programs; by the Foundation for the Promotion of Cancer
Research; and by the Research on Cancer Prevention and Health Services.
Address correspondence to M. Kanematsu.
Abstract
OBJECTIVE. The purpose of our study was to assess the correlation
between the quantitative and qualitative imaging findings on unenhanced and
gadolinium-enhanced MR images and the intensity of vascular endothelial growth
factor (VEGF) expression in hepatocellular carcinomas and in the surrounding
nontumorous liver.
MATERIALS AND METHODS. The intensities of VEGF expression in
hepatocellular carcinoma and in the surrounding liver by Western blot analysis
were converted to VEGF expression indexes (VEGFIND) in 22 surgical
specimens ranging in size from 14 to 126 mm (mean, 47.6 ± 29.5 mm) that
were resected in 22 patients (17 men and five women; age range, 41–85
years [mean, 64 years]) between April 2000 and October 2002. MR images were
retrospectively evaluated to determine contrast-to-noise ratios (CNRs), signal
intensity SD ratios, and phase-shift indexes. Signal intensity characteristics
of hepatocellular carcinomas were reviewed independently by two experienced
radiologists who were unaware of the pathologic diagnosis or the results of
immunoblotting. CNRs, SD ratios, and phase-shift indexes were correlated with
VEGFIND using a simple regression test, and signal intensity
characteristics were correlated with VEGFIND using the Spearman's
rank correlation test.
RESULTS. On opposed-phase T1-weighted spoiled gradient-recalled echo
(GRE) images, CNRs correlated inversely with the VEGFIND of
hepatocellular carcinomas (r = –0.46, p = 0.038). CNRs
on T2-weighted fast spin-echo images correlated directly with the
VEGFIND of hepatocellular carcinomas (r = 0.49, p
= 0.025), and on gadolinium-enhanced hepatic arterial phase GRE images
marginally and inversely correlated with VEGFIND (r =
–0.39, p = 0.081). On T2-weighted fast spin-echo images, SD
ratios correlated directly with the VEGFIND of hepatocellular
carcinomas (r = 0.44, p = 0.044). No correlation was found
between phase-shift indexes and VEGF expression. The qualitatively assessed
signal intensity heterogeneities of hepatocellular carcinomas correlated
directly with the VEGFIND of hepatocellular carcinomas on
opposed-phase T1-weighted GRE, T2-weighted fast spin-echo, hepatic arterial
phase GRE, and equilibrium phase GRE images.
CONCLUSION. Our results indicate that the signal intensity and
heterogeneity of hepatocellular carcinomas on MR images correlate with the
degree of VEGF expression in hepatocellular carcinomas.
Introduction
Angiogenesis is the process whereby new blood vessels develop from
preexisting vessels, which is known to occur physiologically during embryonic
development, normal tissue growth, and wound healing; during the female
reproductive cycle (i.e., ovulation, menstruation, and placental development);
and during the pathologic growth and metastasis of malignant neoplasms
[1]. A number of humoral agents
need to be activated to generate a neovascular blood supply or to initiate
angiogenesis in the human body; one of the most important humoral proteins
that must be activated to ensure the growth of the vascular endothelium is
vascular endothelial growth factor (VEGF). VEGF is an endothelial cell mitogen
that induces and promotes angiogenesis and endothelial cell proliferation,
which plays an important role in regulating angiogenesis
[2], and which was initially
identified as a vascular permeability factor
[3–5].
VEGF expression also plays an important role in the development of
hepatocellular carcinoma, and the degree of its expression is reported to be
associated with tumor size and histologic grade
[6–9].
Abundant evidence suggests that angiogenesis is preceded and accompanied by
enhanced microvascular permeability, although the mechanism remains enigmatic
[3–5].
Although MRI has been widely used as a tool for hepatic tumor detection,
evaluations of tumor vascularity and viability, differentiation of benign and
malignant tumors, and predictions of tumor growth, the relations between MRI
findings and biomolecular angiogenetic activities in hepatocellular carcinomas
and the surrounding liver have yet to be investigated. Moreover,
investigations of these relations may help radiologists understand radiologic
findings related to molecular biologic treatments.
The purpose of this study was to assess the correlation between the
quantitative and qualitative findings of MRI and the angiogenetic activities
as determined using the Western blot technique of VEGF in hepatocellular
carcinomas and in the surrounding liver.
Materials and Methods
Patients
From April 2000 to October 2002, 40 consecutive patients with
hepatocellular carcinoma underwent partial hepatectomy for tumor resection at
the Department of Surgical Oncology, Gifu University School of Medicine. Of
these 40 patients, 28 were selected whose tissue specimens were
histopathologically found not to have substantial degeneration or necrosis,
and 28 samples of hepatocellular carcinomas and of surrounding nontumorous
liver parenchyma were evaluated for the intensity of VEGF expression using
Western blot analysis. We retrospectively searched the radiologic records of
these 28 patients and found that 22, including 17 men and five women having an
age range of 41–85 years (mean, 64 years), had undergone preoperative
MRI of the liver within 2 weeks of surgery. All patients were informed that
the radiologic examinations were primarily for clinical diagnosis and
secondarily for radiologic research and that Western blot analysis of the
resected specimen was scheduled. Thereafter, all provided written consent in
accordance with the requirements of the institutional review board.
Of the 22 patients in this study, six had type B viral hepatitis and 16,
type C viral hepatitis. No patient had a history of alcohol abuse. The
clinical severity and progression of cirrhosis evaluated using the Child-Pugh
classification was grade A in 14 patients, grade B in seven, and grade C in
one. The technique of hepatectomy for tumor resection was left lobectomy in
two patients, trisegmentectomy in one, central bisegmentectomy in one,
segmentectomy in two, subsegmentectomy in seven, and partial resection of the
liver parenchyma harboring tumors surrounded by noncancerous margins in nine.
One tumor was resected in 14 patients, two in two, three in two, four in one,
and five in three. When a patient had multiple lesions resected, the largest
lesion and its surrounding liver were chosen to evaluate VEGF expression,
because multiple tumors in the same liver might influence each other in terms
of angiogenic activity and thus cause statistical bias. Eventually, 22
hepatocellular carcinomas ranging in size from 14 to 126 mm (mean, 47.6
± 29.5 mm) and samples of surrounding liver were evaluated for VEGF
expression. The 22 hepatocellular carcinomas comprised three
well-differentiated, 15 moderately differentiated, and four poorly
differentiated tumors. Underlying liver disease documented by histopathologic
study was chronic hepatitis to mild cirrhosis in five patients, moderate
cirrhosis in 11, and severe cirrhosis in six.
Immunoblotting Technique
All procedures of immunoblotting were conducted by two surgeons.
Immediately after surgical resection, the specimen obtained was sectioned
through the tumor center in the axial plane to ensure correlation with the
preoperative MRI. The hepatocellular carcinoma and surrounding liver samples
were obtained by slicing thin (5–10 mm) tissue sections so that the
samples were obtained evenly throughout the hepatocellular carcinoma and
surrounding liver in the section. The tissue samples were placed in liquid
nitrogen immediately after sampling and kept at –80°C until required
for the Western blot technique. Approximately 5-g samples were dissolved in 1
mL of radioimmunoprecipitation buffer (150 mmol/L NaCl, 50 mmol/L of
tris[hydroxymethyl] aminomethane hydroclhoride), pH 8.0, 0.1% sodium dodecyl
sulfate, 1% alkylaryl polyether alcohol, 1 mmol/L orthovanadate, 1 mmol/L
phenylmethylsulfonyl fluoride, 10 ng/mL of leupeptin, and 10 ng/mL of
aprotinin). Insoluble material was removed by microcentrifugation at 13,000
rpm for 15 min at 4°C. Cell lysates (20 µg of protein per lane) were
subjected to 10% sodium dodecyl (lauryl) sulfate–polyacrylamide gel
electrophoresis. Proteins were transferred to polyvinylidene difluoride
membranes. After blocking membranes with Tris-buffered saline containing
polysorbate 20 (10 mmol of tris[hydroxymethyl] aminomethane hydroclhoridel, pH
8.0, 150 mmol/L NaCl, 0.05% polysorbate 20 [Tween 20, Cayman Chemical]) and 5%
skim milk, membranes were incubated with anti-VEGF monoclonal antibody
(catalog no. LC-3350–10; Laboratory Vision Co.) and then with antimouse
IgG coupled to horseradish peroxidase. Detection was performed by enhanced
chemiluminescence (NEN Life Science).
We performed a control experiment to confirm linearity (r = 0.97,
p < 0.0001) between VEGF concentration and its corresponding
electrophoretic band intensity (Fig.
1). VEGF expression, observed as electrophoretic bands, was
quantified using image analysis software (Scion Image; Scion) that calculated
the area of histograms for the electrophoretic bands. Each pair of
hepatocellular carcinoma and surrounding liver samples was examined using
recombinant human VEGF solution (1.25 mg/mL) for calibration purposes (catalog
no. 2293; Genzyme-Techne). The VEGF expression index (VEGFIND) was
calculated by dividing the area of the histogram corresponding to the specimen
band by that of the calibration band.
View larger version (23K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1. —Schematic shows electrophoretic bands in Western blot technique and
corresponding electrophoretic band histograms with different concentrations of
vascular endothelial growth factor (VEGF) solution in control study. VEGF
solution concentrations and areas of electrophoretic band histogram were well
correlated (r = 0.97, p < 0.0001).
|
|
MRI Protocol
MRI was performed using a 1.5-T imager (Signa Horizon, GE Healthcare). All
images were obtained in the axial plane with a phased-array multicoil for the
body, a section thickness of 8 mm with a 2-mm intersection gap, and
field-of-view of 32 x 24–29 x 22 cm. The MRI protocol
included breath-hold, in-phase T1-weighted spoiled gradient-recalled echo
(GRE) imaging (TR/TE, 150/4.2; matrix, 512 x 224; flip angle, 90°;
one signal acquired; acquisition time, 20-sec for 10 sections; two data
acquisitions to cover the entire liver); breath-hold, opposed-phase
T1-weighted spoiled GRE imaging (150/1.6; matrix, 512 x 224 matrix; flip
angle, 90°; one signal averaged; 26-sec acquisition time for 18 sections);
and respiratory-triggered, chemical shift selective fat-suppressed T2-weighted
fast spin-echo imaging (effective TR range/effective TE, 3,333–8,500/80;
matrix, 512 x 256; echo-train length, 8–16; 3 or 4 signals
acquired; acquisition time, 3.5–5.2 min).
Gadolinium-enhanced spoiled GRE images (TR/TE, 150/1.6; matrix, 512 x
224; flip angle, 90; one signal averaged; 26-sec acquisition time for 18
sections) were obtained in all patients before and after an antecubital IV
bolus injection of 0.1 mmol of gadopentetate dimeglumine (Magnevist, Schering)
per kilogram of body weight followed by a 20-mL flush of sterile saline
solution at 2.5 mL/sec. The scanning delay for triphasic GRE imaging was
16–18 sec, 60 sec, and 3 min after initiating contrast injection,
predominantly representing the hepatic arterial, portal venous, and
equilibrium phases, respectively.
Quantitative Image Analysis
All procedures of quantitative analysis were conducted by one radiologist
who was blinded to the pathologic diagnosis or results of immunoblotting, on a
radiologic information system (Image VINS, Yokogawa Electric). To determine MR
signal intensity values and their SDs in hepatocellular carcinomas and in the
surrounding 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 surrounding liver devoid of large hepatic vessels and
prominent artifacts. The SD of the background noise, SDB, was measured in the
phase-encoding direction outside the anterior abdominal wall to calculate the
following:
 | (1) |
where SIlesion and SIliver are the signal intensities of
the hepatocellular carcinoma and of the surrounding liver, respectively. As a
quantitative parameter of signal intensity heterogeneity, the signal intensity
SD ratio was calculated as follows:
 | (2) |
where SDlesion is the signal intensity SD of the hepatocellular
carcinoma. As a quantitative parameter of fat deposition in hepatocellular
carcinomas, the phase-shift index was calculated as follows:
 | (3) |
where SIin-phase and SIopposed-phase are the signal
intensities of hepatocellular carcinoma on in-phase and opposed-phase
T1-weighted GRE images, respectively.
Qualitative Image Analysis
Two radiologists who had experience in abdominal MR image interpretation
for 14 and 6 years, respectively, independently reviewed MR images in a
retrospective manner. They subjectively evaluated the signal intensity in
hepatocellular carcinoma and in the surrounding liver using a 7-point scale:
–3, strong hypointensity; –2, moderate hypointensity; –1,
mild hypointensity; 0, isointensity; +1, mild hyperintensity; +2, moderate
hyperintensity; and + 3, strong hyperintensity. On T1-weighted images, a grade
of –3 was given when the signal intensity was as low as that of spinal
fluid, and +3 was given when the signal intensity was as high as that of
subcutaneous fat. On T2-weighted images, a grade of –3 was given when
the signal intensity was as low as that of air, and +3 was given when the
signal intensity was as high as that of spinal fluid. A grade of 0 was given
when the hepatocellular carcinoma was isointense to the surrounding liver. The
two radiologists further evaluated the degree of signal intensity
heterogeneity in hepatocellular carcinomas using a 4-point scale: 0, virtually
no heterogeneity; 1, mild heterogeneity; 2, moderate heterogeneity; and 3,
strong heterogeneity.
Statistical Analysis
We correlated the pathologic tumor sizes, CNRs, SD ratios, and phase-shift
indexes with the VEGFIND of hepatocellular carcinomas and of the
surrounding liver and with the VEGFIND difference (
VEGFIND), which was calculated by subtracting the
VEGFIND of the surrounding liver from that of the hepatocellular
carcinoma. We correlated the qualitative degrees of signal intensity and of
heterogeneity of hepatocellular carcinomas on MR images with the
VEGFIND of hepatocellular carcinomas and of the surrounding liver
and with VEGFIND. Statistical correlations were determined
using simple regression analysis for continuous data and the Spearman's rank
correlation test for categoric data. Interobserver variability was assessed
using the kappa test.
Results
The VEGFIND of hepatocellular carcinomas and of the surrounding
liver ranged from 0.46 to 9.3 (mean, 3.2 ± 2.5 [SD]) and from 0.44 to
5.6 (2.6 ± 1.5), respectively. The VEGFIND ranged
from –3.4 to 5.6 (mean, 0.6 ± 2.3). In 10 (45%) of 22
hepatocellular carcinomas, the VEGFIND of the surrounding liver was
greater than that of the corresponding hepatocellular carcinoma. The
quantitative and qualitative evaluations are summarized in
Table 1. The phase-shift
indexes of hepatocellular carcinomas ranged from –0.36 to 0.08 (mean,
–0.19 ± 0.11). Tumor size correlated directly with the
VEGFIND of hepatocellular carcinomas (r = 0.45, p
= 0.038) and with VEGFIND (r = 0.60, p =
0.003).
The CNRs correlated inversely with the VEGFIND of hepatocellular
carcinomas on opposed-phase T1-weighted GRE images (r = –0.46,
p = 0.038) and correlated directly with the VEGFIND of
hepatocellular carcinomas on T2-weighted fast spin-echo images (r =
0.49, p = 0.025). The CNRs on gadolinium-enhanced hepatic arterial
phase GRE images were marginally and inversely correlated with the
VEGFIND (r = –0.39, p = 0.081) and
inversely with VEGFIND (r = –0.49,
p = 0.024) (Table 2
and Figs. 2A,
2B,
2C, and
2D).
View larger version (23K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2A. —Correlations among contrast-to-noise ratios (CNRs), SD ratios, and
vascular endothelial growth factor expression indexes (VEGFIND) and
tumor-to-liver VEGFIND differences (VGEFIND).
Scattergram shows inverse correlation (r = –0.46, p =
0.038) between tumor-to-liver contrast-to-noise ratios (CNRs) on opposed-phase
T1-weighted gradient-recalled echo (GRE) images and VEGFIND of
hepatocellular carcinomas. Straight line and two curves in graph indicate
regression line and 95% confidence interval, respectively.
|
|
View larger version (25K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2B. —Correlations among contrast-to-noise ratios (CNRs), SD ratios, and
vascular endothelial growth factor expression indexes (VEGFIND) and
tumor-to-liver VEGFIND differences (VGEFIND).
Scattergram shows direct correlation (r = 0.49, p = 0.025)
between tumor-to-liver CNRs on T2-weighted fast spin-echo images and
VEGFIND of hepatocellular carcinomas.
|
|
View larger version (26K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2C. —Correlations among contrast-to-noise ratios (CNRs), SD ratios, and
vascular endothelial growth factor expression indexes (VEGFIND) and
tumor-to-liver VEGFIND differences (VGEFIND).
Scattergram shows inverse correlation (r = –0.49, p =
0.024) between tumor-to-liver CNRs on gadolinium-enhanced hepatic arterial
phase GRE images and VEGFIND difference.
|
|
View larger version (18K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2D. —Correlations among contrast-to-noise ratios (CNRs), SD ratios, and
vascular endothelial growth factor expression indexes (VEGFIND) and
tumor-to-liver VEGFIND differences (VGEFIND).
Scattergram shows direct correlation (r = 0.44, p = 0.044)
between SD ratios (SDR) of hepatocellular carcinomas on T2-weighted fast
spin-echo images and VEGFIND of hepatocellular carcinomas.
|
|
The SD ratios correlated directly with the VEGFIND of
hepatocellular carcinomas on T2-weighted fast spin-echo images (r =
0.44, p = 0.044) (Figs.
2A,
2B,
2C, and
2D) and with
VEGFIND on in-phase T1-weighted GRE (r = 0.48, p
= 0.027), opposed-phase T1-weighted GRE (r = 0.48, p =
0.029), T2-weighted fast spin-echo (r = 0.57, p = 0.007),
contrast-enhanced portal venous phase GRE (r = 0.50, p =
0.022), and equilibrium phase GRE (r = 0.58, p = 0.005)
images (Table 3). No
correlation was found between phase-shift index and VEGF expression
(Table 4).
The qualitative degrees of signal intensity on contrast-enhanced hepatic
arterial phase GRE images were inversely correlated with
VEGFIND (r = –0.43, p = 0.028)
(Table 5). The qualitative
degrees of signal intensity heterogeneity of hepatocellular carcinomas
correlated directly with the VEGFIND of hepatocellular carcinomas
on opposed-phase T1-weighted GRE (r = 0.64, p = 0.016),
T2-weighted fast spin-echo (r = 0.52, p = 0.038),
contrast-enhanced hepatic arterial phase GRE (r = 0.48, p =
0.045), and equilibrium phase GRE (r = 0.56, p = 0.018)
images, and correlated with VEGFIND on opposed-phase
T1-weighted GRE (r = 0.71, p = 0.004), T2-weighted fast
spin-echo (r = 0.58, p = 0.016, hepatic arterial phase GRE
(r = 0.55, p = 0.022), portal venous phase GRE (r =
0.66, p = 0.005), and equilibrium phase GRE (r = 0.63,
p = 0.008) images (Table
6 and Figs. 3A,
3B,
3C,
3D,
4A,
4B,
4C,
4D,
5A,
5B,
5C,
5D,
6A,
6B,
6C, and
6D).
View this table:
[in this window]
[in a new window]
|
TABLE 5 Statistical Results of Qualitative Degree of Signal Intensity Versus
Vascular Endothelial Growth Factor (VEGF) Expression
|
|
View this table:
[in this window]
[in a new window]
|
TABLE 6 Statistical Results of Qualitative Degree of Heterogeneity Versus
Vascular Endothelial Growth Factor (VEGF) Expression
|
|
View larger version (56K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3A. —57-year-old man with chronic type C viral hepatitis and poorly
differentiated 5.8-cm hepatocellular carcinoma showing high vascular
endothelial growth factor (VEGF) expression, discrete hypointensity on
in-phase T1-weighted, heterogeneous discrete hyperintensity on T2-weighted,
and weak enhancement on hepatic arterial dominant phase images. Child-Pugh
grade was A. Schematic shows electrophoretic bands and corresponding
histograms in this patient. VEGF solution (1.25 mg/mL) was used for
calibration. Area of histogram was 376 pixels for calibration band, 3,485
pixels for hepatocellular carcinoma band (HCC), and 1,395 pixels for
surrounding liver band. VEGF expression index (VEGFIND) was 9.27 in
hepatocellular carcinoma and 3.71 in surrounding liver, giving
VEGFIND difference of 5.56. Note that electrophoretic peaks
adjacent to those of hepatocellular carcinoma and liver are caused by
expression of irregular protein.
|
|
View larger version (131K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3B. —57-year-old man with chronic type C viral hepatitis and poorly
differentiated 5.8-cm hepatocellular carcinoma showing high vascular
endothelial growth factor (VEGF) expression, discrete hypointensity on
in-phase T1-weighted, heterogeneous discrete hyperintensity on T2-weighted,
and weak enhancement on hepatic arterial dominant phase images. Child-Pugh
grade was A. In-phase T1-weighted spoiled gradient-recalled echo (GRE) (TR/TE,
150/4.2) axial image shows hepatocellular carcinoma (arrow) as
discrete hypointense lesion with internal areas of lower signal intensity
(arrowhead). Likewise, opposed-phase T1-weighted spoiled GRE
(150/1.6) axial image (not shown here) shows hepatocellular carcinoma as
discrete hypointense lesion with internal areas of lower signal intensity.
|
|
View larger version (123K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3C. —57-year-old man with chronic type C viral hepatitis and poorly
differentiated 5.8-cm hepatocellular carcinoma showing high vascular
endothelial growth factor (VEGF) expression, discrete hypointensity on
in-phase T1-weighted, heterogeneous discrete hyperintensity on T2-weighted,
and weak enhancement on hepatic arterial dominant phase images. Child-Pugh
grade was A. Fat-suppressed T2-weighted fast spin-echo (4,286/80) axial image
shows hepatocellular carcinoma (arrow) as moderately hyperintense
lesion with internal areas of higher signal intensity (arrowheads)
presumably due to internal necrosis.
|
|
View larger version (143K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3D. —57-year-old man with chronic type C viral hepatitis and poorly
differentiated 5.8-cm hepatocellular carcinoma showing high vascular
endothelial growth factor (VEGF) expression, discrete hypointensity on
in-phase T1-weighted, heterogeneous discrete hyperintensity on T2-weighted,
and weak enhancement on hepatic arterial dominant phase images. Child-Pugh
grade was A. On gadolinium-enhanced hepatic arterial phase T1-weighted spoiled
GRE (150/1.6) axial image, hepatocellular carcinoma is slightly enhanced
peripherally (arrows). Central areas corresponding to hyperintense
internal areas on C remain unenhanced (arrowheads).
|
|
View larger version (55K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4A. —77-year-old woman with chronic type C viral hepatitis and moderately
differentiated 6.8-cm hepatocellular carcinoma showing moderate vascular
endothelial growth factor (VEGF) expression, isointensity on in-phase
T1-weighted, homogeneous moderate hyperintensity on T2-weighted, and
heterogeneous, moderate enhancement on hepatic arterial dominant phase images.
Child-Pugh grade was A. Schematic shows electrophoretic bands and
corresponding histograms in this patient. VEGF solution (1.25 mg/mL) was used
for calibration. Area of histogram was 312 pixels for calibration band, 1,654
pixels for hepatocellular carcinoma band (HCC), and 738 pixels for surrounding
liver band. VEGF expression index (VEGFIND) was 5.30 in
hepatocellular carcinoma and 2.37 in surrounding liver, giving
VEGFIND difference of 2.93.
|
|
View larger version (136K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4B. —77-year-old woman with chronic type C viral hepatitis and moderately
differentiated 6.8-cm hepatocellular carcinoma showing moderate vascular
endothelial growth factor (VEGF) expression, isointensity on in-phase
T1-weighted, homogeneous moderate hyperintensity on T2-weighted, and
heterogeneous, moderate enhancement on hepatic arterial dominant phase images.
Child-Pugh grade was A. In-phase T1-weighted spoiled gradient-recalled echo
(GRE) (TR/TE, 150/4.2) axial image shows hepatocellular carcinoma as virtually
isointense lesion (arrow) without internal heterogeneity.
Opposed-phase T1-weighted spoiled GRE (150/1.6) axial image (not shown here)
shows hepatocellular carcinoma as homogeneous, moderately hypointense lesion.
Signal intensity reduction was seen between in-phase and opposed-phase GRE
images, indicative of presence of intratumoral fat deposition.
|
|
View larger version (137K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4C. —77-year-old woman with chronic type C viral hepatitis and moderately
differentiated 6.8-cm hepatocellular carcinoma showing moderate vascular
endothelial growth factor (VEGF) expression, isointensity on in-phase
T1-weighted, homogeneous moderate hyperintensity on T2-weighted, and
heterogeneous, moderate enhancement on hepatic arterial dominant phase images.
Child-Pugh grade was A. Fat-suppressed T2-weighted fast spin-echo (3,750/80)
axial image shows hepatocellular carcinoma as slightly heterogeneous,
moderately hyperintense lesion (arrow).
|
|
View larger version (154K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4D. —77-year-old woman with chronic type C viral hepatitis and moderately
differentiated 6.8-cm hepatocellular carcinoma showing moderate vascular
endothelial growth factor (VEGF) expression, isointensity on in-phase
T1-weighted, homogeneous moderate hyperintensity on T2-weighted, and
heterogeneous, moderate enhancement on hepatic arterial dominant phase images.
Child-Pugh grade was A. On gadolinium-enhanced hepatic arterial phase
T1-weighted spoiled GRE (150/1.6) axial image, hepatocellular carcinoma
exhibits heterogeneous, moderate enhancement (arrow).
|
|
View larger version (53K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5A. —70-year-old man with cirrhosis due to chronic type C viral hepatitis
and moderately differentiated 7.4-cm hepatocellular carcinoma showing mild
vascular endothelial growth factor (VEGF) expression, moderate hypointensity
on in-phase T1-weighted, heterogeneous mild hyperintensity on T2-weighted, and
heterogeneous mild enhancement on hepatic arterial dominant phase images.
Child-Pugh grade was B. Schematic shows electrophoretic bands and
corresponding histograms in this patient. VEGF solution (1.25 mg/mL) was used
for calibration. Area of histogram was 723 pixels for calibration band, 1,559
pixels for hepatocellular carcinoma band (HCC), and 1,455 pixels for
surrounding liver band. VEGF expression index (VEGFIND) was 2.16 in
hepatocellular carcinoma and 2.01 in surrounding liver, giving
VEGFIND difference of 0.15.
|
|
View larger version (113K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5B. —70-year-old man with cirrhosis due to chronic type C viral hepatitis
and moderately differentiated 7.4-cm hepatocellular carcinoma showing mild
vascular endothelial growth factor (VEGF) expression, moderate hypointensity
on in-phase T1-weighted, heterogeneous mild hyperintensity on T2-weighted, and
heterogeneous mild enhancement on hepatic arterial dominant phase images.
Child-Pugh grade was B. In-phase T1-weighted spoiled gradient-recalled echo
(GRE) (TR/TE, 150/4.2) axial image shows hepatocellular carcinoma as
moderately hypointense lesion (arrow) without internal heterogeneity.
Likewise, opposed-phase T1-weighted spoiled GRE (150/1.6) axial image (not
shown here) showed hepatocellular carcinoma as moderately hypointense lesion
without internal heterogeneity.
|
|
View larger version (117K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5C. —70-year-old man with cirrhosis due to chronic type C viral hepatitis
and moderately differentiated 7.4-cm hepatocellular carcinoma showing mild
vascular endothelial growth factor (VEGF) expression, moderate hypointensity
on in-phase T1-weighted, heterogeneous mild hyperintensity on T2-weighted, and
heterogeneous mild enhancement on hepatic arterial dominant phase images.
Child-Pugh grade was B. Fat-suppressed T2-weighted fast spin-echo (4,286/80)
axial image shows hepatocellular carcinoma as mildly hyperintense lesion
(arrow) with internal areas of slightly higher signal intensity
(arrowhead).
|
|
View larger version (117K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5D. —70-year-old man with cirrhosis due to chronic type C viral hepatitis
and moderately differentiated 7.4-cm hepatocellular carcinoma showing mild
vascular endothelial growth factor (VEGF) expression, moderate hypointensity
on in-phase T1-weighted, heterogeneous mild hyperintensity on T2-weighted, and
heterogeneous mild enhancement on hepatic arterial dominant phase images.
Child-Pugh grade was B. On gadolinium-enhanced hepatic arterial phase
T1-weighted spoiled GRE (150/1.6) axial image, hepatocellular carcinoma
exhibits heterogeneous, moderate enhancement (arrow).
|
|
View larger version (54K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 6A. —59-year-old man with chronic type C viral hepatitis and moderately
differentiated 2.4-cm hepatocellular carcinoma showing weak vascular
endothelial growth factor (VEGF) expression, isointensity on in-phase
T1-weighted, subtle hyperintensity on T2-weighted, and moderate enhancement on
hepatic arterial dominant phase images. Child-Pugh grade was A. Schematic
shows electrophoretic bands and corresponding histograms in this patient. VEGF
solution (1.25 mg/mL) was used for calibration. Area of histogram was 902
pixels for calibration band, 1,174 pixels for hepatocellular carcinoma band
(HCC), and 930 pixels for surrounding liver band. VEGF expression index
(VEGFIND) was 1.30 in hepatocellular carcinoma and 1.03 in
surrounding liver, giving VEGFIND difference of 0.27.
|
|
View larger version (123K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 6B. —59-year-old man with chronic type C viral hepatitis and moderately
differentiated 2.4-cm hepatocellular carcinoma showing weak vascular
endothelial growth factor (VEGF) expression, isointensity on in-phase
T1-weighted, subtle hyperintensity on T2-weighted, and moderate enhancement on
hepatic arterial dominant phase images. Child-Pugh grade was A. In-phase
T1-weighted spoiled gradient-recalled echo (GRE) (TR/TE, 150/4.2) axial image
shows no abnormal imaging findings for hepatocellular carcinoma. Opposed-phase
T1-weighted spoiled GRE (150/1.6) axial image (not shown here) shows
hepatocellular carcinoma as area of partly decreased signal intensity, which
was not seen in B, indicating that this tumor harbors internal fat.
|
|
View larger version (99K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 6C. —59-year-old man with chronic type C viral hepatitis and moderately
differentiated 2.4-cm hepatocellular carcinoma showing weak vascular
endothelial growth factor (VEGF) expression, isointensity on in-phase
T1-weighted, subtle hyperintensity on T2-weighted, and moderate enhancement on
hepatic arterial dominant phase images. Child-Pugh grade was A. Fat-suppressed
T2-weighted fast spin-echo (6,000/80) axial image shows hepatocellular
carcinoma as homogeneous area of faintly increased signal intensity
(arrow).
|
|
View larger version (111K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 6D. —59-year-old man with chronic type C viral hepatitis and moderately
differentiated 2.4-cm hepatocellular carcinoma showing weak vascular
endothelial growth factor (VEGF) expression, isointensity on in-phase
T1-weighted, subtle hyperintensity on T2-weighted, and moderate enhancement on
hepatic arterial dominant phase images. Child-Pugh grade was A. On
gadolinium-enhanced hepatic arterial phase T1-weighted spoiled GRE (150/1.6)
axial image, hepatocellular carcinoma exhibits slightly heterogeneous, mild
enhancement (arrow).
|
|
The kappa values of the two observers ranged from 0.62 to 0.89 (mean, 0.73)
in terms of rating images independently, indicating a substantial to almost
perfect agreement.
Discussion
In our study, the tumor-to-liver CNRs on opposed-phase T1-weighted GRE
images were found to be inversely correlated with the VEGFIND of
hepatocellular carcinomas, and CNRs on T2-weighted fast spin-echo images were
found to be directly correlated with the VEGFIND of hepatocellular
carcinomas, which indicates that the stronger the VEGF expression is in
hepatocellular carcinomas, the more prolonged are the T1 and T2 relaxation
times of hepatocellular carcinomas. T1 and T2 relaxation time prolongation
commonly occurs in tissues containing increased amounts of free water in
extracellular or interstitial spaces, typically as seen in malignant tumors.
Moreover, VEGF, also referred to as "vascular permeability
factor," raises the permeability of blood vessels
[3–5].
We suspect that increased vascular permeability regulated by VEGF peptides can
predispose increased free water in the extracellular or interstitial spaces of
hepatocellular carcinomas and thus cause prolongation of T1 and T2 relaxation
times in hepatocellular carcinomas.
The SD ratios showed a direct correlation with the VEGFIND of
hepatocellular carcinomas on T2-weighted fast spin-echo images, and the
qualitative degree of heterogeneity showed a direct correlation with
VEGFIND of hepatocellular carcinomas on opposed-phase T1-weighted
GRE, T2-weighted fast spin-echo, contrast-enhanced hepatic arterial phase GRE,
and equilibrium phase GRE images. These observations indicate more
heterogeneous signal intensity of hepatocellular carcinoma corresponds to
stronger VEGF expression of hepatocellular carcinoma. Heterogeneity of
hepatocellular carcinoma on unenhanced MR images may be explained by an uneven
distribution of extracellular free water or, more commonly, by the presence of
intratumoral necrosis. In an experimental model using rat livers, VEGF
peptides were produced by nonparenchymal and parenchymal cells after necrosis
[10]. Heterogeneity on
gadolinium-enhanced MR images may indicate unevenness of vascularity and of
concomitant oxygenation in hepatocellular carcinoma. Moreover, the hypoxic
regions of solid tumors are known to produce powerful and directly acting
angiogenic proteins such as VEGF
[11].
We found a marginal inverse correlation between CNRs on contrast-enhanced
hepatic arterial phase GRE images and VEGFIND of hepatocellular
carcinomas, and a significant inverse correlation between the CNR and
VEGFIND, which suggests that there might be an inverse correlation
between arterial vascularity and VEGF expression in hepatocellular carcinomas.
Some researchers have reported that VEGF activity is not correlated with the
hepatocellular carcinoma vascularity as determined by conventional angiography
[12,
13], whereas others have found
that VEGF activity is correlated directly with the intensity of tumor staining
on angiography [7].
Kwak et al. [14], who
correlated tumor attenuation on contrast-enhanced CT with the intensity of
VEGF expression using immunohistochemical staining, concluded that the degree
of VEGF expression in hepatocellular carcinomas is directly correlated with
the degree of contrast enhancement during the hepatic arterial phase. Although
why their results differ from ours is not known, in view of our previous
results, which showed an inverse correlation between hepatic arterial
enhancement on CT during hepatic arteriography and VEGF expression in
hepatocellular carcinomas
[15], and the results of the
present study, we suspect that there is an inverse correlation between hepatic
arterial vascularity and VEGF expression in hepatocellular carcinomas.
Furthermore, in the study by Kwak et al.
[14], 18 (82%) of 22 patients
had type B hepatitis and one (5%) had type C hepatitis, whereas in our study
only six (27%) of 22 patients had type B hepatitis and 16 (73%) had type C
hepatitis. This substantial difference in the patient populations and in the
underlying hepatic disease might reflect the contradiction in results. In
addition, we need to consider the optimal correlation methodology: for
example, the use of single-detector helical CT without a bolus tracking device
can lead to inconstant acquisitions of optimal hepatic arterial phase images,
the use of a fixed amount of iodine load per body may cause varying iodine
concentrations in hepatic arterial blood in individual patients, and
subjective ratings of contrast enhancement on CT images and of VEGF expression
may obscure statistical correlations.
An inefficient vascular supply and the resultant reduction in tissue oxygen
tension lead to neovascularization to satisfy the needs of tissue
[11,
16]. Moreover, in
hepatocellular carcinoma in humans, it has been suggested that hypoxia induces
the upregulation of VEGF gene expression
[17]—that is,
hypoxia-inducible factor-1
and hypoxia-inducible factor-2, which
are upregulated by hypoxia, induce proangiogenic peptides such as VEGF
[18].
Fat deposition frequently occurs in various types of hepatocellular
carcinomas [19]. Kutami et al.
[20] reported that fatty
changes in small hepatocellular carcinomas are closely related to tumor size,
histologic grade, and insufficient development of arterial tumor vessels.
Previous studies have shown that lipid bodies in endothelial cells are induced
during hypoxia in any cell type
[21]. On the basis of these
previous results, we suspect that the VEGF expression, hepatic arterial
perfusion, hypoxia, and fat deposition in hepatocellular carcinomas are
closely related. However, in our study, no correlation was found between
phase-shift indexes, a parameter of intratumoral fat deposition as determined
by phase-shift GRE sequences, and the degree of VEGF expression.
Some limitations of our study should be mentioned. First, our study
population was small because the study was performed at a single institution.
Furthermore, the distribution of different types of histologic tumor grades
was uneven. Multiinstitutional studies will be needed to confirm our results
with greater statistical power. Finally, although we used the Western blot
technique to semiquantify VEGF peptides, this technique is limited in terms of
its ability to differentiate VEGF peptides in cell membranes, the cytoplasm,
and interstitial spaces. Although we used phase-shift indexes to represent the
degree of fat deposition in hepatocellular carcinomas, that calculation has
not been commonly used in the clinical setting to date. Although we performed
preliminary correlations between VEGF activity and MRI findings and found some
significant correlations, the real effects of our results on radiology
practice are still debatable. However, we believe that our results may help
the future practice of radiology in connection with biomolecular or genetic
treatments for hepatocellular carcinoma.
In conclusion, the tumor-to-liver CNRs and VEGF expression in
hepatocellular carcinomas were found to be inversely correlated on
opposed-phase T1-weighted GRE images, directly correlated on T2-weighted fast
spin-echo images, and marginally inversely correlated on gadolinium-enhanced
hepatic arterial phase GRE images. The SD ratios and VEGF expression in
hepatocellular carcinomas were found to be directly related on T2-weighted
fast spin-echo images, and the qualitative heterogeneity of hepatocellular
carcinomas and VEGF expression to be directly related on opposed-phase
T1-weighted GRE, T2-weighted fast spin-echo, contrast-enhanced hepatic
arterial phase GRE, and equilibrium phase GRE images. Our results indicate
that the MRI findings of hepatocellular carcinoma are indeed correlated with
the degree of VEGF expression in hepatocellular carcinoma.
References
- Folkman J. Tumor angiogenesis: therapeutic implications.
N Engl J Med1971; 285:1182
-1186
- Ferrara N. Vascular endothelial growth factor: molecular and
biological aspects. Curr Top Microbiol Immunol1999; 237:1
-30[Medline]
- Senger DR, Galli SJ, Dvorak AM, et al. Tumor cells secrete a
vascular permeability factor that promotes accumulation of ascites fluid.
Science 1983;219:983
-985[Abstract/Free Full Text]
- Keck PJ, Hauser SD, Krivi G, et al. Vascular permeability factor:
an endothelial cell mitogen related to PDGF. Science1989; 246:1309
-1312[Abstract/Free Full Text]
- Connolly DT. Vascular permeability factor: a unique regulator of
blood vessel function. J Cell Biochem1991; 47:219
-223[Medline]
- Yamane A, Seetharam L, Yamaguchi S, et al. 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). Oncogene1994; 9:2683
-2690[Medline]
- Mise M, Arii S, Higashituji H, et al. Clinical significance of
vascular endothelial growth factor and basic fibroblast growth factor gene
expression in liver tumors. Hepatology1996; 23:455
-464[Medline]
- Yamaguchi R, Yano H, Iemura A, Ogasawara S, Haramaki M, Kojiro M.
Expression of vascular endothelial growth factor in human hepatocellular
carcinoma. Hepatology1998; 28:68
-77[Medline]
- Ohmori S, Shiraki K, Sugimoto K, et al. High expression of
CD34-positive sinusoidal endothelial cells is a risk factor for hepatocellular
carcinoma in patients with HCV-associated chronic liver diseases.
Hum Pathol2001; 32:1363
-1370[Medline]
- Ishikawa K, Mochida S, Mashiba S, et al. Expressions of vascular
endothelial growth factor in non-parenchymal as well as parenchymal cells in
rat liver after necrosis. Biochem Biophys Res Commun1999; 254:587
-593[Medline]
- Mukhopadhyay D, Tsiokas L, Zhou XM, Foster D, Brugge JS, Sukhatme
VP. Hypoxic induction of human vascular endothelial growth factor expression
through c-Src activation. Nature1995; 375:577
-581[Medline]
- El-Assal ON, Yamanoi A, Soda Y, et al. Clinical significance of
microvessel density and vascular endothelial growth factor expression in
hepatocellular carcinoma and surrounding liver: possible involvement of
vascular endothelial growth factor in the angiogenesis of cirrhotic liver.
Hepatology1998; 27:1554
-1562[Medline]
- Suzuki K, Hayashi N, Miyamoto Y, et al. Expression of vascular
permeability factor/vascular endothelial growth factor in human hepatocellular
carcinoma. Cancer Res1996; 56:3004
-3009[Abstract/Free Full Text]
- Kwak BK, Shim HJ, Park ES, et al. Hepatocellular carcinoma:
correlation between vascular endothelial growth factor level and degree of
enhancement by multiphase contrast-enhanced computed tomography.
Invest Radiol2001; 36:487
-492[Medline]
- Kanematsu M, Osada S, Amaoka N, et al. Expression of vascular
endothelial growth factor in hepatocellular carcinoma and the surrounding
liver: correlation with angiographically assisted CT.
AJR 2004;183:1585
-1593[Abstract/Free Full Text]
- Shweiki D, Itin A, Soffer D, Keshet E. Vascular endothelial growth
factor induced by hypoxia may mediate hypoxia-initiated angiogenesis.
Nature 1992;359:843
-845[Medline]
- von Marschall Z, Cramer T, Hocker M, Finkenzeller G, Wiedenmann B,
Rosewicz S. Dual mechanism of vascular endothelial growth factor upregulation
by hypoxia in human hepatocellular carcinoma. Gut2001; 48:87
-96[Abstract/Free Full Text]
- Blancher C, Moore JW, Talks KL, Houlbrook S, Harris AL.
Relationship of hypoxia-inducible factor (HIF)-1alpha and HIF-2alpha
expression to vascular endothelial growth factor induction and hypoxia
survival in human breast cancer cell lines. Cancer Res2000; 60:7106
-7113[Abstract/Free Full Text]
- Yoshikawa J, Matsui O, Takashima T, et al. Fatty metamorphosis in
hepatocellular carcinoma: radiologic features in 10 cases.
AJR 1988;151:717
-720[Abstract/Free Full Text]
- Kutami R, Nakashima Y, Nakashima O, Shiota K, Kojiro M.
Pathomorphologic study on the mechanism of fatty change in small
hepatocellular carcinoma of humans. J Hepatol2000; 33:282
-289[Medline]
- Scarfo LM, Weller PF, Farber HW. Induction of endothelial cell
cytoplasmic lipid bodies during hypoxia. Am J Physiol Heart Circ
Physiol 2001; 280:294
-301
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Top
Abstract
Introduction
