| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
Department of Radiology, Faculty of Medicine, Kyushu University, 3-1-1
Maidashi, Higashi-Ku, Fukuoka, 812-8582, Japan.
2
Department of Pathology II, Faculty of Medicine, Kyushu University, Fukuoka,
812-8582, Japan.
3
Department of Surgery II, Faculty of Medicine, Kyushu University, Fukuoka,
812-8582, Japan.
Received June 27, 2001;
accepted after revision August 14, 2001.
Supported in part by a grant-in-aid for scientific research (C) of the
Ministry of Education, Science, Sports, and Culture of Japan.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Eighty-six hepatocellular lesions were confirmed at pathology in 49 patients who underwent CT with both hepatic arteriography and arterioportography. These images were compared with lesion-to-liver vascular ratios of cumulative arteries, preexisting hepatic arteries, and portal veins in resected specimens. Lesions were classified in five groups according to intranodular hemodynamics determined by CT hepatic arteriography and CT during arterioportography: group 1, isoattenuating on both procedures; group 2, hypoattenuating on CT hepatic arteriography and isoattenuating on CT during arterioportography; group 3, hypoattenuating on both procedures; group 4, isoattenuating on CT hepatic arteriography and hypoattenuating on CT during arterioportography; and group 5, hyperattenuating on CT hepatic arteriography and hypoattenuating on CT during arterioportography.
RESULTS. Among 86 lesions, we identified seven low-grade dysplastic nodules, eight high-grade dysplastic nodules, 14 well-differentiated hepatocellular carcinomas, 45 moderately differentiated hepatocellular carcinomas, and 12 poorly differentiated hepatocellular carcinomas. The lesions were classified as group 1 (n = 5), group 2 (n = 13), group 3 (n = 6), group 4 (n = 2), or group 5 (n = 60). Intranodular hemodynamics was significantly correlated with pathologic grading (p < 0.001). For correlations between combinations of the groups and pathologic gradings, the order "groups 1-2-3-4-5" was the most significant (p < 0.001).
CONCLUSION. During hepatocarcinogenesis, most hepatocellular nodules show deterioration of arterial blood flow before loss of portal blood flow. Vascular imaging of hepatic nodules may predict malignant abnormality via the early loss of hepatic arterial flow seen before portal flow changes.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Dysplastic nodule is defined as a well-circumscribed hepatic parenchymal nodule that occurs following acute or chronic hepatic disorders [7] and is characterized by the presence of a portal tract. In cirrhotic livers, a dysplastic nodule can undergo malignant transition to advanced hepatocellular carcinoma [8]; malignant foci are often detected within dysplastic nodules [9,10,11].
Early hepatocellular carcinoma is defined as a lesion that does not substantially destroy the existing architecture of the liver lobule or pseudolobules and that frequently includes portal tract structures such as the portal vein and bile ducts [1]. Histopathologically, cancer cells in early hepatocellular carcinoma are extremely well-differentiated, corresponding to grade I in the Edmondson-Steiner classification [12]. Early hepatocellular carcinoma is considered an intermediate step between dysplastic nodule and early advanced hepatocellular carcinoma [1].
Early advanced hepatocellular carcinoma is a nodular lesion that contains a component of advanced hepatocellular carcinoma within early hepatocellular carcinoma. Pathologically, it is defined as a nodule that simultaneously has both macroscopically identifiable advanced hepatocellular carcinoma (Edmondson-Steiner grades II or III) and early hepatocellular carcinoma (Edmondson-Steiner grade I) [4, 5].
Neovascular growth (angiogenesis) can be seen in a nodule irrespective of whether it is malignant [13]. Advanced hepatocellular carcinoma is characterized by abundant arterial neovascularization. By contrast, dysplastic nodule and early hepatocellular carcinomas often contain residual portal tracts, indicating that they are fed not only by arteries but also by portal veins [7, 9,10,11]. Neovascularized arteries are seen even in these early-stage hepatic nodules, and the total number of intranodular cumulative arteries (preexisting hepatic arteries and neovascularized arteries) is often greater in such nodules than in the surrounding normal hepatic parenchyma [14].
Many of the early-stage hepatic nodules are not enhanced at the arterial phase of angiography, dynamic CT, or dynamic MR imaging. This inconsistency between the histopathologically identified increase in intranodular cumulative arteries and the hypovascular imaging characteristics has not been explained. Recent advances in imaging have enabled researchers to identify not only advanced hepatocellular carcinoma but also dysplastic nodules or early hepatocellular carcinomas [2, 15, 16]. Researchers can now obtain images of CT hepatic arteriography and CT during arterial portography, enabling them to evaluate the hemodynamics of a lesion concurrently and pre-operatively [16, 17].
The purpose of this study was to evaluate intratumoral hemodynamic changes during the transition from dysplastic nodule to advanced hepatocellular carcinoma and to compare enhancement patterns of CT hepatic arteriography and CT during arterioportography with various pathologic gradings.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Hepatitis virus markers were positive in 43 of 49 patients (five patients tested positive for hepatitis B surface antigen, and 40 patients tested positive for hepatitis C virus antibody; two patients tested positive for both). Histopathology of the liver parenchyma showed that 32 patients had cirrhosis, eight had hepatic fibrosis, six had chronic hepatitis, and three had normal livers according to the classification of the Japanese Liver Cancer Study Group [6].
Imaging Techniques
CT hepatic arteriography and CT during arterioportography were part of the
preoperative angiographic examination. One hour before CT hepatic
arteriography and CT during arterioportography, the Seldinger technique was
used to place 5-French catheters through both femoral arteries. The hepatic
arterial catheter for CT hepatic arteriography was placed in the proper
hepatic artery in 25 patients, through the common hepatic artery in six
patients, through the right hepatic artery in 10 patients, through the left
hepatic artery in five patients, through the middle hepatic artery in one
patient, through the replaced right hepatic artery in one patient, and through
the replaced left hepatic artery in one patient. Anatomic anomalies, hepatic
arterial supply, and portal venous patency were assessed with digital
subtraction angiography before CT hepatic arteriography and CT during
arterioportography. The volume of contrast material during angiography was
minimized before CT hepatic arteriography and CT during
arterioportography.
Helical CT imaging (X-Vigor; Toshiba Medical Systems, Tokyo, Japan) was performed in a craniocaudal direction during one-breath-hold helical acquisitions of 32-40 sec (depending on liver size), using the following parameters: collimation, 5 mm; table speed, 5 mm/sec; kilovoltage, 130 kVp; and amperage, 150 mAs. For CT during arterioportography, data acquisition was started 25 sec after initiation of a transcatheter arterial injection of 90-100 mL of nonionic contrast material (Iopamidol, Iopamiron 150 [iodine, 150 mg I/mL]; Schering, Osaka, Japan) at 2.5 mL/sec using the automated power injector (Medrad, Pittsburgh, PA). Before the contrast material was injected, a transarterial infusion of prostaglandin E1 (10 mg) as a vasodilator was performed. The catheter was placed in the root of the superior mesenteric artery in 41 patients and in a more distal portion of the branching site in five patients with a replaced or accessory right hepatic artery and in three patients who had hepatopetal blood flow originating from the pancreatic arterial arcade.
CT hepatic arteriography was performed approximately 5 min after CT during arterioportography. For CT hepatic arteriography, scan parameters were identical to CT during arterioportography. Data acquisition began 5 sec after initiation of a transcatheter hepatic arterial injection of 20-80 mL of nonionic contrast material at a speed of 1-4 mL/sec using the automated power injector. The appropriate injection rate for CT hepatic arteriography was determined as the maximum injection rate (basically depending on the vessel caliber) that would not cause backward flow of contrast material on the hepatic arteriography. The duration of arterial injection was 20 sec.
Imaging Analysis
The image review was performed prospectively. The radiologists were unaware
of all clinical data. Attenuation of lesions on CT hepatic arteriography and
CT during arterioportography was visually judged by consensus as
hyperattenuation, isoattenuation, or hypoattenuation by comparing each lesion
to the surrounding hepatic parenchyma. All lesions were classified in the
following groups based on the enhancement patterns observed on CT hepatic
arteriography and CT during arterioportography: group 1, nodules that appeared
isoattenuating on both CT hepatic arteriography and CT during
arterioportography; group 2, hypoattenuating on CT hepatic arteriography and
isoattenuating on CT during arterioportography; group 3, hypoattenuating both
on CT hepatic arteriography and CT during arterioportography; group 4,
isoattenuating on CT hepatic arteriography and hypoattenuating on CT during
arterioportography; and group 5, hyperattenuating on CT hepatic arteriography
and hypoattenuating on CT during arterioportography.
Pathologic Analysis
After the resected liver tissue was fixed in 10% formalin,
paraffin-embedded specimens were grossly examined by a pathologist to
determine the maximum tumor diameter and whether fibrous capsules were
present. All specimens were examined microscopically using H and E staining.
The degree of cancer cell differentiation was classified as Edmondson-Steiner
grades I—IV, and the histologic pattern was classified according to the
criteria of the World Health Organization
[18]. In an attempt to
standardize the terminology applied to diseases of the liver, the organizers
of the World Congress of Gastroenterology sponsored an International Working
Party [19] on the terminology
of nodular hepatocellular lesions, and they suggested inclusion of low-grade
and high-grade dysplastic nodules. In our study, low-grade and high-grade
dysplastic nodules correspond to adenomatous hyperplasia and atypical
adenomatous hyperplasia, respectively.
Growth patterns at the tumor margins were categorized as replacing growth or as expansive growth according to Nakashima et al. [20].
Blood Vessel Counts
All specimens were examined by two pathologists to count the blood vessels
in the lesion and surrounding hepatic parenchyma. The total numbers of
cumulative arteries, preexisting hepatic arteries, and portal veins per unit
area were counted. Ratios of the numbers of cumulative arteries, pre-existing
hepatic arteries, and portal veins in the lesion versus the parenchyma
(lesion-to-liver ratios) were calculated and were termed cumulative artery
ratio, hepatic artery ratio, and portal vein ratio, respectively. Blood
vessels were counted in five low-power fields, and the mean value per square
millimeter was obtained. The number of blood vessels included those in the
septum of lesions but not in the fibrous capsule. For preexisting hepatic
arteries, only arterial vessels with normal architecture were counted.
Arterial vessels that were abnormal, including fragile or ill-demarcated
vessels or vessels whose lumina were narrowed because of arterial wall
thickening, were counted for the total number of cumulative arteries.
Statistical Analysis
Statistical analysis was performed with Statview J 4.02 software, Universal
version (Abacus Concepts, Berkeley, CA). Tests for association between
categorical variables were performed using a chisquare statistic. The relation
between the enhancement patterns on CT hepatic arteriography and CT during
arterioportography was analyzed with the Spearman's rank correlation test. The
enhancement pattern on CT hepatic arteriography was placed in the order of
hypoattenuation, isoattenuation, and hyperattenuation, and the enhancement
pattern on CT during arterioportography was placed in the order of
hyperattenuation, isoattenuation, and hypoattenuation. Similarly, the
relationship between enhancement on CT hepatic arteriography or CT during
arterioportography and pathologic grading was analyzed with Spearman's rank
correlation test. Pathologic gradings were ordered as low-grade dysplastic
nodule, high-grade dysplastic nodule, well-differentiated hepatocellular
carcinoma, moderately differentiated hepatocellular carcinoma, and poorly
differentiated hepatocellular carcinoma, and poorly differentiated
hepatocellular carcinoma. To find between-group differences in continuous
variables (lesion-to-liver ratios), the two-tailed Student's t test
was used. To assess hemodynamic changes in arterial and portal blood flow
during hepatocarcinogenesis, combinations of CT hepatic arteriography and CT
during arterioportography were correlated with pathologic gradings and
analyzed by the nonparametric Spearman's and Kendall's rank correlation tests.
A p value less than 0.05 was considered statistically
significant.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
|
|
|
|
|
|
|
Histopathology of the Lesions
The 86 lesions were classified as low-grade dysplastic nodule (n =
7), high-grade dysplastic nodule (n = 8), early well-differentiated
hepatocellular carcinoma (n = 0), well-differentiated hepatocellular
carcinoma (n = 14), moderately differentiated hepatocellular
carcinoma (n = 45), and poorly differentiated hepatocellular
carcinoma (n = 12). Pathologic gradings were compared with
enhancement patterns on CT hepatic arteriography, CT during
arterioportography, and both. The enhancement pattern was significantly
different in the five pathologic gradings in all instances (chi-square test,
p <0.001) (Tables 1
and 2). A significant
correlation was shown between the enhancement pattern of CT hepatic
arteriography and the pathologic gradings when the CT hepatic arteriography
patterns were placed in the order of hypoattenuation, isoattenuation, and
hyperattenuation (r = 0.848, p <0.001)
(Table 1). A significant
correlation was also shown between the CT during arterioportography patterns
and the pathologic gradings when the CT during arterioportography patterns
were placed in the order of hyperattenuation, isoattenuation, and
hypoattenuation (r = 0.564, p <0.001)
(Table 1). These findings
suggest that intranodular arterial flow increases and intranodular portal flow
decreases as the disease advances.
|
|
Pathologic gradings were also compared among the five groups according to hemodynamic status (Table 2).
Lesion Classification According to Intranodular Arterial and Portal
Blood Supply Correlation with Lesion-to-Liver Vascular Ratios
All lesions in group 1 were either low-grade or high-grade dysplastic
nodules; all lesions in groups 2 and 3 were either low-grade dysplastic
nodules, high-grade dysplastic nodules, or well-differentiated hepatocellular
carcinomas (Table 3). Lesions
in group 4 were well-differentiated hepatocellular carcinomas, whereas all
lesions in group 5 were either moderately differentiated hepatocellular
carcinomas or poorly differentiated hepatocellular carcinomas. A significant
correlation was shown between pathologic grading and the five hemodynamic
groups (r = 0.568, p = 0.0027) when the enhancement patterns
were placed in order 1-2-3-4-5 (Table
2).
|
|
For tumor growth pattern, the lesions in groups 1-4 were classified as replacing growth and lesions in group 5 were classified as expansive growth.
Blood Vessel Counts
Lesion-to-liver ratios of blood vessels are shown in
Figure 4. The number of
cumulative artery ratios was slightly increased in groups 1 (mean ± SD,
1.15 ± 0.08) through 4 (1.40 ± 0.85) and was markedly increased
in group 5 (2.29 ± 1.68). Hepatic artery ratios decreased in the
following order: group 4 (1.23 ± 0.61), group 1 (1.02 ± 0.12),
group 3 (0.84 ± 0.24), group 2 (0.73 ± 0.21), and group 5 (0.16
± 0.30). In group 5, preexisting hepatic arteries were rare, indicating
that the increase in intranodular cumulative arteries was primarily caused by
neovascularized arteries. By contrast, in the lesions of group 4, the number
of preexisting hepatic arteries was increased compared with the surrounding
hepatic parenchyma. Portal vein ratios decreased in the following order: group
1 (0.81 ± 0.18), group 2 (0.69 ± 0.24), group 3 (0.55 ±
0.17), group 4 (0.35 ± 0.14), and group 5 (0.08 ± 0.14).
The disappearance, obliteration, or weakening of the wall of preexisting hepatic arteries and luminal narrowing of the arteries caused by a remarkably thickened wall appeared in low-grade dysplastic nodules, high-grade dysplastic nodules, and well-differentiated hepatocellular carcinomas (Fig. 3A,3B,3C). In lesions of group 5, nearly all of the preexisting hepatic arteries disappeared. Portal veins were relatively preserved in lesions of low-grade dysplastic nodules or early hepatocellular carcinomas (groups 1-3), but these veins disappeared in group 5 (Figs. 1A,1B,1C,2A,2B,2C,2D,3A,3B,3C).
Figure 5 indicates the relationship between pathologic grades and lesion-to-liver vascular ratios. Cumulative artery ratios were greater than 1 in every histologic grading. Moderately differentiated hepatocellular carcinomas and poorly differentiated hepatocellular carcinomas showed higher cumulative artery ratios compared with well-differentiated hepatocellular carcinomas, and moderately differentiated hepatocellular carcinomas showed the highest cumulative artery ratio among all histologies. On the other hand, hepatic artery ratios decreased in the following order: low-grade dysplastic nodules (0.94 ± 0.27), high-grade dysplastic nodules (0.84 ± 0.18), well-differentiated hepatocellular carcinomas (0.76 ± 0.35), moderately differentiated hepatocellular carcinomas (0.17 ± 0.30), and poorly differentiated hepatocellular carcinomas (0.02 ± 0.04). Portal vein ratios also decreased in the following order: low-grade dysplastic nodule (0.72 ± 0.18), high-grade dysplastic nodule (0.77 ± 0.22), well-differentiated hepatocellular carcinomas (0.52 ± 0.23), moderately differentiated hepatocellular carcinomas (0.06 ± 0.11), and poorly differentiated hepatocellular carcinomas (0.03 ± 0.07); these ratios were significantly different from each other. Analysis of the lesion-to-liver vascular ratios (cumulative artery, hepatic artery, and portal vein) showed no statistically significant difference between low-grade dysplastic nodules and high-grade dysplastic nodules or between moderately differentiated hepatocellular carcinomas and poorly differentiated hepatocellular carcinomas.
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Although hepatocellular carcinoma is generally a hypervascular tumor, some patients have hypovascular nodules that are not enhanced on imaging [16, 22,23,24]; such nodules do not retain iodized oil [23]. These lesions are classified as well-differentiated hepatocellular carcinoma, grade 1 in the Edmondson-Steiner classification, and they correspond to early hepatocellular carcinoma [22, 23].
Poor enhancement may be caused by the tumor itself: replacing growth by the tumor with preservation of the intratumor portal branch [22]; intratumor necrosis, fibrosis, or diffuse fatty metamorphosis [23, 25, 26]; immature development of tumor vessels [16, 26]; or poor contrast access to the tumor vessels compared with arteries in the surrounding hepatic parenchyma [27]. Causes not attributable to the tumor include obliteration of segmental staining as a result of occlusion of the intrahepatic portal branch [27] and loss of the tumor stain because of poor parenchymal uptake with severe cirrhosis [25]. Why some hepatic nodules appear relatively hypovascular compared with the surrounding hepatic parenchyma has not been fully explained [17].
Increased numbers of tumor vessels are seen in hepatocellular carcinoma, and even in early hepatocellular carcinoma [2, 3]. The total number of intranodular cumulative arteries may increase as the lesion becomes more malignant—that is, from low-grade dysplastic nodule to high-grade dysplastic nodule and then to advanced hepatocellular carcinoma [14]. This finding is inconsistent with tumor hypovascularity by imaging, and this discrepancy has not been clearly explained.
In our preliminary study involving fewer cases [28], we noted in early hepatocellular carcinomas degeneration of preexisting hepatic arteries in portal tracts (fragility or obliteration of the wall, narrowing of the lumen, or decreased numbers). The status of degenerated hepatic arteries was associated with lesser arterial blood flow in early hepatocellular carcinoma based on CT hepatic arteriography. Sequential changes in intranodular hemodynamics during evolution to hepatocellular carcinoma and the relationship between intranodular arterial blood flow and pathologic gradings have not been extensively examined.
Dynamic imaging generally provides more accurate information on the blood supply of hepatic nodular lesions than does histopathology. Researchers can perform CT scanning during angiography, thereby acquiring data on lesion location and intranodular blood flow simultaneously [2]. In this study, we performed an in vivo analysis of tumor hemodynamics based on CT hepatic arteriography and CT during arterioportography at various pathologic stages.
Our data suggest that intranodular arterial flow increases and intranodular portal flow decreases as the histopathologic grading advances. Significant correlations were also shown between the individual enhancement patterns of CT hepatic arteriography and CT during arterioportography and pathologic grading. The correlation with CT hepatic arteriography was more significant, perhaps because retrograde inflow of arterial blood into the portal blood flow via intranodular sinusoids affected portal blood flow on CT during arterioportography; alternatively, there may be a bias in the hemodynamic distribution of patients by CT during arterioportography, because few patients had findings classified as hyperattenuation and isoattenuation. Therefore, the stepwise transition in histopathologic differentiation will be discussed mainly in association with intranodular arterial blood flow.
It is generally accepted that hepatocellular carcinoma results from multistep carcinogenesis [1]; a dysplastic nodule produced in a cirrhotic liver eventually advances to advanced hepatocellular carcinoma through the repeated substitution of malignant and poorly differentiated tissue in the nodule (dedifferentiation). This may explain the nodule-in-nodule pattern of an outer component of early hepatocellular carcinoma and an inner component of advanced hepatocellular carcinoma at a certain growth stage during multistep development. In our study, five lesions appeared as early advanced hepatocellular carcinoma on imaging. Vascular ratios of the outer and inner components in the lesions corresponded to early hepatocellular carcinoma and advanced hepatocellular carcinoma, respectively.
The decrease in preexisting hepatic arteries was associated with hypovascularity on imaging. In groups 1-4 (well-differentiated lesions), the cumulative artery ratio was slightly increased, but this does not explain the hypoattenuation of groups 2 and 3 on CT hepatic arteriography. An increase in the total number of intranodular cumulative arteries may not be depicted by CT hepatic arteriography, whereas a decrease in the number of preexisting hepatic arteries is clearly shown on CT hepatic arteriography. Therefore, we assume that the degeneration or disappearance of preexisting hepatic arteries causes a decline in intranodular arterial blood flow in dysplastic nodules or well-differentiated hepatocellular carcinomas, even though intranodular angiogenesis has already begun. In group 5, preexisting hepatic arteries were rare, and most of the intranodular arteries were neovascularized, causing the hyperattenuation seen on CT hepatic arteriography.
Cumulative artery ratios were markedly increased in group 5 compared with other groups, and most lesions in group 5 showed a cumulative artery ratio greater than 1.5. Compared with groups 1-4, the border between hyperattenuation and hypoattenuation on CT hepatic arteriography was between 1.2 and 1.5. When the number of intranodular cumulative arteries increases by 20-50%, arterial inflow into the tumor is substantially increased, and the tumor begins showing hypervascular characteristics on CT hepatic arteriography.
Portal vein ratios were preserved in groups 1 and 2, but were substantially decreased in groups 3, 4, and 5. In group 5, portal veins were rarely seen. We found that a portal vein ratio between 0.5 and 0.7 is the border between isoattenuation and hypoattenuation on CT during arterioportography. When the number of intranodular portal veins decreases by 30-50%, intranodular portal venous flow decreases enough to be identified on CT during arterioportography.
With respect to the correlation between vascular ratios and pathologic gradings, the cumulative artery ratio was greater than 1 in all gradings, and in moderately differentiated hepatocellular carcinoma and poorly differentiated hepatocellular carcinoma, the cumulative artery ratio was greater than 2. No significant difference was seen in vascular ratios between low-grade dysplastic nodule and high-grade dysplastic nodule or between moderately differentiated hepatocellular carcinoma and poorly differentiated hepatocellular carcinoma; these findings agree with those of other researchers [14].
In this study, lesions were classified hemodynamically into five groups. The order of groups 1-2-3-4-5 had the highest p value, suggesting a highly probable growth process from group 1 (dysplastic nodule) through groups 2 and 3 (dysplastic nodule or early hepatocellular carcinoma), to group 4 (well-differentiated hepatocellular carcinoma), and finally to group 5 (advanced hepatocellular carcinoma).
Five nodules had isoattenuation on both CT hepatic arteriography and CT during arterioportography; all were well-differentiated low-grade or high-grade dysplastic nodules. The smaller tumors may have shown isoattenuation because of a partial volume-averaging effect. The lesion-to-liver vascular ratios were equivalent to the surrounding liver parenchyma, consistent with the results of CT hepatic arteriography and CT during arterioportography. This pattern can be explained by the preexisting hepatic arteries and portal veins not yet having decreased or by the presence of neovascularized arteries before a decrease in portal blood flow. The former is thought to be at a stage closer to low-grade dysplastic nodule, whereas the latter is more related to moderately differentiated hepatocellular carcinoma. Despite similar imaging features and hemodynamics, they are histopathologic opposites.
Thirteen nodules showed hypoattenuation on CT hepatic arteriography and isoattenuation on CT during arterioportography; these were low-grade dysplastic nodules, high-grade dysplastic nodules, and well-differentiated hepatocellular carcinomas. This hemodynamic state suggests that preexisting hepatic arteries have decreased, portal veins are preserved, and neovascularized arteries are insufficient.
Six nodules showed hypoattenuation on both CT hepatic arteriography and CT during arterioportography, corresponding to mainly highgrade dysplastic nodule and well-differentiated hepatocellular carcinoma. In the hemodynamic state, preexisting hepatic arteries and portal veins have decreased, but development of neovascularized arteries is still insufficient.
Two nodules of well-differentiated hepatocellular carcinoma showed isoattenuation on CT hepatic arteriography and hypoattenuation on CT during arterioportography; Takayasu et al. [16] reported three early hepatocellular carcinoma nodules with similar findings, representing early disappearance of portal veins before preexisting hepatic arteries, or disappearance of preexisting hepatic arteries with delayed increase of neovascularized arteries.
Sixty nodules showed hyperattenuation on CT hepatic arteriography and hypoattenuation on CT during arterioportography, representing mainly moderately differentiated hepatocellular carcinoma and poorly differentiated hepatocellular carcinoma, or so-called advanced hepatocellular carcinoma. This hemodynamic state shows loss of preexisting hepatic arteries and portal veins with neovascularized arteries markedly increased.
No nodules showed hyperattenuation with CT during arterioportography, indicating increased portal blood flow in the lesion. One group of researchers [29] reported a case of low-grade dysplastic nodule with increased intranodular portal blood flow and multiple dilated portal lumina inside the nodule. The portal tract may have been preserved in the growth stage, or this condition could have resulted from fibrosis of the surrounding hepatic parenchyma.
In summary, in early-stage nodules (dysplastic nodule or early hepatocellular carcinoma), insufficient growth of neovascularized arteries coupled with the disappearance of preexisting hepatic arteries results in hypovascularity compared with the surrounding hepatic parenchyma. Subsequently, as a result of a marked increase in neovascularized arteries, neovascular blood flow becomes dominant. The portal blood supply decreases with advancement of the tumor, and eventually, the tumor is fed mainly by arterial flow. Kudo et al. [3] suggested that portal blood flow in nodules decreases before the increase in arterial blood flow. However, to our knowledge, there have been no reports about a relationship between deterioration of arterial flow and portal flow.
Our data suggest several possible patterns of changes in arterial and portal blood flow during evolution of hepatocellular nodules (Figs. 6 and 7A,7B,7C,7D). In the most probable scenarios of hemodynamic change during evolution of hepatocellular carcinoma (based on the analysis of the permutations of all five groups), the deterioration in arterial blood flow precedes the decrease in portal blood flow during advancement to more aggressive disease (Fig. 7B). Our study of hepatocellular carcinoma and angiogenesis suggests that full-scale inflow via neovascularized arteries does not occur in the early-stage nodule, even though angiogenesis has already begun. The tumor is fed by portal flow or flow through the hepatic sinusoid. Gradually, preexisting hepatic arteries degenerate, and intratumor arterial blood flow decreases. Loss of portal veins results in a reduction in portal blood flow, followed by initiation of full-scale inflow into neovascularized arteries. In very specific cases, both the initial decrease and subsequent increase of intranodular arterial blood flow may precede the decrease in intranodular portal blood flow (Fig. 7A), or both intranodular arterial and portal blood flow may decrease simultaneously (Fig. 7C). Also it is possible in rare cases that a deterioration in intranodular portal blood flow may precede a decrease in arterial blood flow (Fig. 7D).
|
|
|
|
|
In our study, CT hepatic arteriography and CT during arterioportography have clearly shown the hemodynamic properties of various hepatocellular lesions. The next step in the evaluation of these lesions may be a quantitative measure on CT that can be compared with histologic parameters. However, individual differences in degree of liver fibrosis can greatly affect attenuation of the liver tumor and liver parenchyma. For quantitative measuring of attenuation values of various nodules in cirrhosis, this difficulty in evaluation of these hepatocellular lesions on the basis of blood supply may add limitations to CT hepatic arteriography and CT during arterioportography in clinical practice.
Several studies have evaluated whether CT angiography could be replaced by other less invasive techniques such as helical CT and MR imaging. Jang et al. [30] have concluded that triple-phase helical CT should replace the combination of CT during arterioportography and CT hepatic arteriography for the preoperative evaluation of hepatocellular carcinoma. On the contrary, some investigators have reported that the combination of CT hepatic arteriography and CT during arterioportography is recommended to improve the detection sensitivity of small hypervascular nodules of hepatocellular carcinoma [31].
In our study, we conclude from a correlation between the findings of qualifying dual-phase helical CT and the combination of CT hepatic arteriography and CT during arterioportography that the simultaneous use of CT hepatic arteriography and CT during arterioportography is superior to dual-phase helical CT in detecting both hypervascular and hypovascular lesions, especially well-differentiated lesions, and that dual-phase helical CT is insufficient in evaluating subtle differences in arterial and portal blood supply of these lesions and the surrounding hepatic parenchyma (unpublished data). For the most accurate determination of lesions concealed in the liver and precise evaluation of intranodular arterial and portal blood supply, additional CT hepatic arteriography and CT during arterioportography are recommended. However, because CT hepatic arteriography and CT during arterioportography are invasive techniques, the concept of imaging findings shown here should be applied to dual-phase or triple-phase helical CT, dynamic MR imaging, and Doppler sonography.
Angiogenesis maintains tumor growth through timing in the process of tumor formation. Folkman [32] suggested avascular and vascular phases in the growth of solid tumors. Folkman et al. [33] also found that angiogenic activity first appears in a subset of hyperplastic cells before the onset of tumor formation. It is generally accepted that in hepatocarcinogenesis the increased number of abnormal arteries is part of the tumor process [2]. It is not known at what stage angiogenesis is initiated in the process of tumor formation.
Studies of angiogenic factors in the carcinogenetic process have been popular recently among researchers. Especially in hepatocellular carcinoma, it is suggested that intranodular expression of vascular endothelial growth factors is an important factor for intratumor angiogenesis and tumor growth and advancement. Vascular endothelial growth factor is assumed to be a factor that acts specifically on the endothelial cells and induces angiogenesis and in vivo vascular permeability. Holash et al. [34] reported that a subset of tumors initially grows by coopting preexisting host vessels and that growth factors such as vascular endothelial growth factors and angiopoietins may be critical regulators of the balance between regression and growth in tumors. On the other hand, in human hepatocellular carcinoma, hypoxia is suggested as a central stimulus of angiogenesis through upregulation of vascular endothelial growth factor gene expression by means of activation of vascular endothelial growth factor gene transcription and increase in vascular endothelial growth factor mRNA stability [35].
Recently, noninvasive techniques of nuclear medicine using 18F-fluoromisonidazole uptake and positron emission tomography have enabled detection of hypoxic tissues or regional hypoxia within tumors [36, 37]. Studies of hypoxia of liver tissue and hepatic arterial flow have shown that decreased hepatic arterial flow in the cirrhotic liver can cause hypoxic status in regional liver tissues [36].
Therefore, a hypoxic state is probably created by decreased intranodular arterial blood flow in early-stage nodules. Hepatocarcinogenesis may be initiated or advanced by this hypoxic state. We conclude that vascular imaging of hepatic nodules may predict malignant pathology via the early loss of hepatic arterial flow before portal flow changes. It will be necessary to determine whether hypoxia occurs before angiogenesis or vice versa in the process of hepatocellular carcinoma.
Acknowledgments
We greatly appreciate the encouragement and helpful advice of Keizou
Sugimachi, Professor and Chairman, Department of Surgery and Science; and
Masazumi Tsuneyoshi, Professor and Chairman, Department of Anatomic Pathology,
Graduate School of Medical Sciences, Kyushu University, for providing
pathology data.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
