| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, Kurt Amplatz Center, Innsbruck University Hospital,
Anichstr. 35, 6020 Innsbruck, Austria.
2 Institute of Pathology, Feldkirch Academic Teaching Hospital, 6800 Feldkirch,
Austria.
3 Department of Transplantation-Surgery, Innsbruck University Hospital, 6020
Innsbruck, Austria.
4 Department of Gastroenterology, Innsbruck University Hospital, 6020 Innsbruck,
Austria.
Received July 24, 2002;
accepted after revision December 24, 2002.
Address correspondence to A. Mallouhi.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Fifty-nine patients with liver cirrhosis were examined with CT arterioportography and digital subtraction angiography as a part of preoperative diagnostic workup for liver transplantation. Before liver explantation, CT arterioportograms and digital subtraction angiograms were prospectively evaluated in a blinded manner, separately by two CT radiologists and two angiographers, respectively, and combined by two reviewer teams, each including a CT radiologist and an angiographer. In addition, each examination was retrospectively evaluated using direct comparison with the corresponding thin-section liver explant specimens
RESULTS. There were 39 histologically confirmed hepatocellular carcinomas. In both prospective and retrospective assessments, the reviewers achieved the best performance with CT arterioportography and digital subtraction angiography combined (area under the curve [Az] 0.82). The diagnostic confidence in the detection of hepatocellular carcinoma was higher with digital subtraction angiography (Az, 0.81) than that with CT arterioportography (Az, 0.68). Prospectively, sensitivity and specificity were 75% and 60% for CT arterioportography, 77% and 80% for digital subtraction angiography, and 84% and 81% for CT arterioportography and digital subtraction angiography combined, respectively. Retrospectively, sensitivity and specificity were 80% and 62% for CT arterioportography; 82% and 79% for digital subtraction angiography; 87% and 81% for CT arterioportography and digital subtraction angiography combined, respectively. Five hepatocellular carcinomas, one poorly and four well differentiated, with a mean size of 1.4 cm were not detectable on the CT arterioportography and digital subtraction angiography combination. False-positive findings were 20, 11, and 10 on CT arterioportography, digital subtraction angiography, and the CT arterioportography and digital subtraction angiography combination.
CONCLUSION. Combining CT arterioportography with digital subtraction angiography enabled reliable detectability of moderately and poorly differentiated hepatocellular carcinomas in cirrhotic livers but was less sensitive for the detection of well-differentiated hepatocellular carcinomas and resulted in a relatively high rate of false-positive findings.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Despite its invasive nature, CT arterioportography was recommended by several investigators for the preoperative evaluation of cirrhotic livers [3, 20, 21]. The accuracy of CT arterioportography was mostly assessed on the basis of pathologic evidence acquired from percutaneous biopsy or partial liver resection rather than from transplantation correlation, the reference that enables a more accurate evaluation of sensitivity and specificity in the detection of hepatocellular carcinoma.
The purpose of our study was to assess the clinical reliability of CT arterioportography and digital subtraction hepatic angiography, separately and combined, for the detection of hepatocellular carcinoma in the cirrhotic liver, by using histopathologic findings of thin-section liver explants as the standard of reference.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Imaging Technique
After placement of a 5-French sidewinder I catheter in the proximal
superior mesenteric artery in the angiography suite, CT was performed on a
HiSpeed Advantage scanner (General Electric Medical Systems, Milwaukee, WI).
The liver was first examined with helical scanning in 5-mm sections without
contrast material. For CT arterioportography, continuous scanning commenced at
approximately 2 cm cranial to the liver and proceeded in a caudal direction
during portal and venous enhancement with a scan delay of 25 sec between both
phases. Acquisition parameters were 5-mm collimation and 10-mm/sec table
speed. Data were always acquired during breath-hold. A total volume of 90 mL
of nonionic contrast material was administered at a rate of 3 mL/sec by using
a power injector connected to the catheter. Scan delay was determined by using
the Smart Prep technique (General Electric Medical Systems) with a region of
interest placed in the portal vein. When the attenuation value increased up to
100–120 H after contrast injection, the helical scanning was manually
initiated under the supervision of the radiologist who performed the
catheterization.
The radiologic study was concluded by intraarterial digital subtraction angiography including celiac and mesenteric angiography and selective angiography of the common hepatic artery. Separate selective angiography of the right and left hepatic arteries was performed in seven patients with the right hepatic artery originating from the superior mesenteric artery. Angiograms were obtained by using a 4-French cobra or 4-French sidewinder I catheter with injection volumes (16–25 mL) and rates (4–5 mL/sec) varying with vessel size.
Image Analysis
Prospectively, CT studies were interpreted by two independent radiologists
experienced in abdominal CT and unaware of digital subtraction angiography
findings; digital subtraction angiograms were interpreted, likewise, in
blinded fashion by two independent senior interventional radiologists. After
reviewers completed their independent assessment, CT and digital subtraction
angiograms were collectively evaluated by two independent teams, each
including a CT radiologist and an angiographer. Within each team, final
decisions on the findings were determined by consensus. Retrospectively, at
the time of specimen sectioning, one radiologist experienced in abdominal
imaging and one senior pathologist performed a direct correlation between
radiologically detected lesions and pathologically confirmed lesions. In
comparison with the corresponding transverse sections of explanted livers, CT
arterioportograms were initially reviewed, then digital subtraction
angiograms, and finally CT arterioportograms and digital subtraction
angiograms combined. In both prospective and retrospective analyses, CT
arterioportograms and digital subtraction angiograms were evaluated on PACS
(picture archiving and communication system), and multiplanar reformations of
CT arterioportography data sets were performed when necessary to assist in the
exact correlation of intrahepatic lesions among CT arterioportography, digital
subtraction angiography, and explant liver slices. The reviewers assessed the
number, site, and enhancement characteristics of all detected lesions and
pseudolesions.
The criterion for hepatocellular carcinoma on CT arterioportography was a round perfusion defect larger than 1 cm showing a soft-tissue attenuation. Round perfusion defects smaller than 1 cm were considered to be dysplastic nodules. Wedge-shaped or irregular perfusion defects or those found around the gallbladder fossa or intersegmental fissure were interpreted as pseudolesions. The criterion for hepatocellular carcinoma at hepatic digital subtraction angiography was a round enhancing lesion irrespective of size. Hypoattenuating round lesions or irregular enhanced areas were interpreted as dysplastic nodules or pseudolesions, respectively. The criterion for hepatocellular carcinoma on CT arterioportography and digital subtraction angiography combined was a round hypervascular area on digital subtraction angiography that coincided in localization with a round perfusion defect on CT arterioportography or a round hypervascular area on digital subtraction angiography that coincided with no significant findings on CT arterioportography.
The diagnostic confidence in the presence of a hepatocellular carcinoma was scored at the prospective assessment by using a 5-point ordinal scale, in which 1 represented definitely present; 2, probably present; 3, uncertain; 4, probably absent; 5, definitely absent. At the retrospective assessment, the presence of a hepatocellular carcinoma was scored as 1, present, or 2, absent. Because a nodule smaller than 1 cm detected on CT arterioportography may not allow the characterization of that nodule as a hepatocellular carcinoma and follow-up imaging or biopsy is performed to confirm the diagnosis, the 1-cm threshold was applied at the prospective analysis. At the retrospective analysis, all hepatocellular carcinomas were included.
Imaging—Pathologic Correlation
Immediately after the explantation, the liver specimens were conserved for
2 weeks in formalin (4%). To enable an exact radiologic—pathologic
correlation, we cut the specimens into 5-mm sections that corresponded to the
axial CT planes. Every nodule differing from the cirrhotic liver parenchyma in
color, texture, or size was analyzed by one pathologist. The histologic
criteria included cell density, the nuclear—cytoplasmic ratio, the
heterogeneity of nuclear size, the eosinophilic affinity, and the cellular
organization in the nodule. All cross-sections of each liver were
photographed, and all lesions were sampled. The specimens were also examined
to identify any lesion categorized on CT arterioportography or digital
subtraction angiography as a hepatocellular carcinoma or as suspicious for
hepatocellular carcinoma. If a lesion was indicated on preoperative CT or
digital subtraction angiography and could not be detected on the liver
specimen, the suspected area was histopathologically examined.
Statistical Analysis
Data entry procedures and statistical analysis were performed with a
statistical software system (SPSS for Windows, version 10.0.0, Statistical
Package for the Social Sciences, Chicago, IL). In the first step, receiver
operating characteristic (ROC) curves were generated to show the relative
accuracy of CT arterioportography and digital subtraction angiography
separately and combined for the detection of hepatocellular carcinoma by
comparing the area under the curve of each reviewer of CT arterioportography
and digital subtraction angiography and of each reviewing team of the CT
arterioportography and digital subtraction angiography combination. From the
observed data points, sensitivity and specificity were calculated on a
per-hepatocellular carcinoma (i.e., the ability to correctly identify all
hepatocellular carcinomas) basis with the use of "definitely
present" and "probably present" as positive and all other
categories as negative. In the second step of the analysis, the kappa
statistic was used for comparison of observer performance. Kappa values were
calculated on the basis of each reviewer's and team's confidence level for the
ROC analysis. Interobserver agreement was considered as slight, (
≤
0.2), fair ( = 0.21–0.40), moderate ( = 0.41–0.60),
substantial ( = 0.61–0.80), or almost perfect ( =
0.81–1.00).
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Prospective Assessment
The parametric ROC models performed for the detection of any hepatocellular
carcinoma and the results of calculating the Az values for
each imaging modality separately and combined are shown in Figures
1A,
1B and
2A,
2B. The reviewers achieved the
best performance in the detection of hepatocellular carcinoma with CT
arterioportography and digital subtraction angiography combined
(Az = 0.82 and Az = 0.84 for teams 1
and 2, respectively). When all patients were included, the area under the ROC
curve for digital subtraction angiography (Az = 0.78 and 0.84 for
reviewers 1 and 2, respectively) was substantially higher than that for CT
arterioportography (Az = 0.66 and 0.70 for reviewers 1 and
2, respectively). No statistically significant difference was found between
these areas. However, considering only those patients with positive findings
on either imaging modality or at pathologic analysis caused a dramatic
decrease in the area under the curve of CT arterioportography by both
reviewers with loss of statistical significance (Az =
0.44, p = 0.432; and Az = 0.50, p =
0.976, respectively). On the contrary, a significant Az
value was maintained on digital subtraction angiography alone
(Az = 0.67, p = 0.027; and Az
= 0.73, p = 0.003 for reviewers 1 and 2, respectively) and on the CT
arterioportography and digital subtraction angiography combination
(Az = 0.68, p = 0.015; and Az
= 0.73, p = 0 .003 for teams 1 and 2, respectively).
|
|
|
|
At CT arterioportography, reviewer 1 misclassified 31 findings (20 incorrectly identified as hepatocellular carcinoma and 11 not diagnosed as hepatocellular carcinomas), whereas reviewer 2 misclassified 30 findings (21 incorrectly identified as hepatocellular carcinoma and nine not diagnosed as hepatocellular carcinomas). The number of miscategorized lesions decreased considerably on digital subtraction angiography to 21 lesions (12 incorrectly identified as hepatocellular carcinomas and nine not diagnosed as hepatocellular carcinomas) by reviewer 1 and 18 lesions (nine incorrectly identified as hepatocellular carcinomas and nine not diagnosed as hepatocellular carcinomas) by reviewer 2. These results yielded an almost similar sensitivity on digital subtraction angiography and CT arterioportography associated with a substantial increase in specificity and positive predictive value on digital subtraction angiography in comparison with CT arterioportography (Figs. 3A, 3B, 3C). Combining the findings of CT arterioportography with digital subtraction angiography showed a further slight advantage over digital subtraction angiography and a substantial advantage over CT arterioportography (Table 1). From five hepatocellular carcinomas (four poorly and one well differentiated) smaller than 1 cm, three poorly differentiated hepatocellular carcinomas were detected on the combination of CT arterioportography and digital subtraction angiography (Figs. 4A, 4B, 4C). The smallest well-differentiated hepatocellular carcinoma that was detected on CT arterioportography and digital subtraction angiography measured 1.5 cm (Figs. 5A, 5B, 5C).
|
|
|
|
|
|
|
|
|
|
Some disagreement as to the confidence levels of hepatocellular carcinoma
detection was observed between the reviewers of each modality. Interobserver
agreement was moderate for CT arterioportography ( = 0.59) and
substantial for digital subtraction angiography ( = 0.63) and the CT
arterioportography and digital subtraction angiography combination ( =
0.71).
Retrospective Assessment
With the direct correlation to pathologic findings, the detection of
hepatocellular carcinoma was considerably improved, compared with the blinded
prospective evaluation (Table
2). By evaluating the ability to identify the presence of a
hepatocellular carcinoma, we determined that there were 20 false-positive and
eight false-negative findings on CT arterioportography, 11 false-positive and
seven false-negative findings on digital subtraction angiography, and 10
false-positive (Figs. 6A,
6B,
6C) and five false-negative
findings on the CT arterioportography and digital subtraction angiography
combination. These results yielded a considerably better specificity and
positive predictive value on digital subtraction angiography in comparison
with CT arterioportography and a substantial advantage for the CT
arterioportography and digital subtraction angiography combination. Two
hepatocellular carcinomas were detected only on CT arterioportography and
three only on digital subtraction angiography. The former were well
differentiated and the latter were moderately and poorly differentiated
hepatocellular carcinomas.
|
|
|
|
On CT arterioportography, digital subtraction angiography, and the CT arterioportography and digital subtraction angiography combination, the false-positive interpretations were most commonly attributed to the presence of regenerative nodules (13/20, 6/11, and 6/10, respectively). One cavernous hemangioma and one focal area of fibrosis were incorrectly interpreted as hepatocellular carcinomas. For the other false-positive findings (five, on CT arterioportography; three, on digital subtraction angiography; and two, on the CT arterioportography and digital subtraction angiography combination), a correlating histopathologic factor could not be identified.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CT arterioportography has been considered the most sensitive imaging technique for the detection of hepatocellular carcinoma in the cirrhotic liver. A number of authors have reported a sensitivity that ranged from 73% to 95% [18–22], and several studies have addressed its limitations, particularly the high false-positive rates (≤ 15%) [29] and perfusion abnormalities [30, 31] that might obscure a tumor. The accuracy of CT arterioportography has been assessed using pathologic proof acquired from percutaneous biopsy or partial liver resection rather than from transplantation correlation, the reference that enables the more accurate evaluation in the detection of hepatocellular carcinoma. To our knowledge, only one study [22] evaluated the sensitivity of CT arterioportography and two studies [22, 23] evaluated the sensitivity of hepatic digital subtraction angiography by using whole explant correlation. Both studies, however, included only patients with hepatocellular carcinoma. To enable the estimation of the sensitivity and specificity (i.e., accuracy) of CT arterioportography and hepatic digital subtraction angiography, we recruited all patients with cirrhosis who were examined with both techniques and ultimately underwent orthotopic liver transplantation, even those who were found not to have hepatocellular carcinoma.
For this study, all conserved livers were cut in total into 5-mm anatomic slices and precisely photo-documented. Every nodule differing from the cirrhotic liver parenchyma in appearance or size was analyzed by the pathologist. As opposed to partial liver resection and biopsy used as the standard of reference in several studies [4, 18–20], the explanted livers offered the opportunity to study the true extension of hepatocellular carcinoma throughout the entire parenchyma and to assess the diagnostic accuracy of CT arterioportography and digital subtraction angiography in the detection of hepatocellular carcinoma. In fact, the accuracy of songraphically guided needle biopsy in the detection of hepatocellular carcinoma is presumed not to be high [32, 33]. One reason is believed to be the sampling error at needle biopsy because the target lesion is small. Another reason is that the exact histopathologic diagnosis may not be obtained because hepatocellular carcinoma usually contains different types of tissues, such as necrosis, fibrosis, and more differentiated and more dedifferentiated components. These conditions influence the accurate histopathologic diagnosis.
Our study consisted of a prospective component based on data acquired from the blinded evaluation of CT arterioportography and hepatic digital subtraction angiography by two independent reviewers before liver explantation and a retrospective component based on data acquired from direct comparison between CT arterioportograms and hepatic digital subtraction angiograms and the corresponding liver explant specimens. The specific objective of the prospective component of the study was to determine the reviewer-dependent accuracy of CT arterioportography and hepatic digital subtraction angiography in the detection of hepatocellular carcinoma. For the retrospective portion of the study, the objective was to assess the ultimate accuracy of both techniques separately and combined by reducing the effect of reviewer-dependent diagnostic decisions and therefore enhancing the diagnostic confidence level.
The results of the prospective analysis showed that digital subtraction angiography allows a considerably improved diagnostic confidence in comparison with CT arterioportography. By combining hepatic digital subtraction angiography with CT arterioportography, we found that sensitivity improved slightly, whereas specificity and positive predictive value increased significantly. Because of the lack of a corresponding hypervascular lesion at digital subtraction angiography, the CT arterioportography and digital subtraction angiography combination improved the characterization of many low-attenuation nodules depicted on CT arterioportography (nine by team 1 and 12 by team 2) that were confirmed not to be hepatocellular carcinoma. Nevertheless, two well-differentiated hepatocellular carcinomas in our patient cohort were depicted as hypoattenuating nodules on CT arterioportography and as isoattenuating on digital subtraction angiography and therefore were prospectively underestimated as non—hepatocellular carcinoma lesions. Furthermore, six and seven hepatocellular carcinomas were not detected on combined CT arterioportography and digital subtraction angiography by the reviewing teams 1 and 2, respectively. The result was assigning a stage that was too low to six patients before transplantation. On the contrary, 11 and nine lesions were incorrectly interpreted as hepatocellular carcinomas by the reviewing teams 1 and 2, respectively, whereas the pathologic analysis confirmed them as nonmalignant findings.
The retrospective analysis using direct comparison with liver specimens revealed a relatively good sensitivity on CT arterioportography and digital subtraction angiography, separately and combined. All moderately to poorly differentiated 0.8-cm hepatocellular carcinomas (except one) and 12 of 16 well-differentiated hepatocellular carcinomas were detectable by the combined CT arterioportography and digital subtraction angiography. The specificity and positive predictive value were significantly better on digital subtraction angiography alone and combined with CT arterioportography than on CT arterioportography alone. However, four (25%) of the 16 well-differentiated hepatocellular carcinomas in four patients were isoattenuating on digital subtraction angiography and CT arterioportography and thus were missed by both modalities. Furthermore, in well-differentiated nodules, the combination of hepatic artery degeneration and preserved portal veins results in isoattenuation on CT arterioportography and hypoattenuation on digital subtraction angiography [34]. In fact, the two well-differentiated hepatocellular carcinomas found in our study were depicted only on CT arterioportography as hypoattenuating nodules. These results outline the main drawback of digital subtraction angiography, namely the lack of conspicuity of hypoattenuating particularly well-differentiated hepatocellular carcinomas, the perception of which may be improved on CT hepatic arteriography [34].
The main disadvantage of CT arterioportography is its low specificity relative to the presence of pseudolesions and perfusion defects. The diagnostic pitfalls of CT arterioportography can be classified in three categories: first, nonuniform enhancement pattern of the liver parenchyma due to diffuse morphologic changes characteristic for cirrhosis, including arteriovenous or arterioportal shunts, parenchymal necrosis, fatty infiltration, and parenchymal fibrosis. Several authors [29, 30] remarked that the presence of portosystemic shunts, portal thrombosis, or reversed portal blood flow diverts the contrast agent away from the liver and leads to an inadequate parenchymal enhancement. In our study, two hepatocellular carcinomas measuring 1.5 and 1.9 cm could not be visualized on CT arterioportography because of existing significant portosystemic shunts. Both lesions were depicted on digital subtraction angiography. In addition, the presence of small arterioportal shunts may mimic a hepatocellular carcinoma on CT arterioportography [4]. However, because of the specific appearance of arterioportal shunts on digital subtraction angiography, false-positive results can be reduced [35]. Furthermore, inhomogeneous segmental enhancement has also been reported [29, 31]. In seven cases in our study, the left hepatic segments, including the fourth segment in three cases, were not perfused with the contrast agent. In such cases, digital subtraction angiography could help detect hepatocellular carcinoma.
Second, nontumorous perfusion defects were usually located immediately anterior to the porta hepatis adjacent to the gallbladder in the medial segment of the left lobe; the posterior peripheral segment of the left lobe; and close to the eighth, ninth, and 10th ribs [36, 37]. In this regard, digital subtraction angiography ruled out the presence of hepatocellular carcinoma in two of our patients.
Third, regarding perfusion defects due to regenerative and dysplastic nodules, both of which pose particular diagnostic dilemmas, Chezmar et al. [38] described a regenerative nodule as a high-attenuation area on CT arterioportography; Oliver et al. [30] described it as a low attenuation area; and Peterson et al. [37] described it as an intermediate attenuation area. Lim et al. [6, 39] concluded that the enhancement of dysplastic nodules was so variable that no consistent pattern could be found. Those different appearances of regenerative and dysplastic nodules may be explained by the multistep development of hepatocellular carcinoma in liver cirrhosis, in which regenerative liver nodules are an integral part of the process. Regenerative nodules may grow large and appear tumorlike. Over time, they can undergo malignant transformation and then are described as "atypical" or "early malignant." Within these steps, vascularization changes. As the grade of malignancy increases, arterial blood flow deteriorates before the decrease in portal blood flow, followed by initiation of neovascularized arteries [34, 40]. Because of this behavior, regenerative nodules may have portal vein and hepatic artery supply, which results in different appearances on CT arterioportography.
Noninvasive cross-sectional imaging techniques such as dynamic CT and MR imaging have been implemented for the detection of hepatocellular carcinoma in the cirrhotic liver. Numerous studies showed the high sensitivity of these imaging techniques for hepatocellular carcinoma detectability. Sensitivity range was 53–94% for monoslice CT [2–4, 7, 21], 76–90% for MDCT [8–10], 77–90% for gadolinium-enhanced MR imaging [12–14, 16], and 57–93% for ferumoxides-enhanced MR imaging [12–14, 17, 41]. Review of these studies revealed that the standard of reference was mainly based on pathologic proof acquired from percutaneous biopsy or partial liver resection rather than on whole liver explant correlation. Choi et al. [16, 17] concluded that MR imaging could be used successfully in place of combined CT arterioportography and CT hepatic arteriography for the preoperative evaluation of patients with hepatocellular carcinoma.
Studies that evaluated hepatocellular carcinoma detectability in correlation with explant pathologic analysis revealed, on the contrary, moderate results. Sensitivity ranged from 37% to 76% for dynamic monoslice CT [1, 5, 6, 15, 41] and from 55% to 77% for dynamic MR imaging [11, 15]. To our knowledge, the accuracy of MDCT in the detection of hepatocellular carcinoma has not yet been reported with explant correlation. Murakami et al. [9] found that a double arterial phase on MDCT improves hepatocellular carcinoma detectability. On the contrary, Ichikawa et al. [10] showed that the acquisition technique described by Murakami et al. added no significant yield to hepatocellular carcinoma detectability in comparison with single late arterial phase MDCT. In addition to cross-sectional imaging, sonography with explant correlation was found by Bennett et al. [42] to yield a low sensitivity (20.5%) in revealing hepatocellular carcinomas. Not only a relatively large number of small hepatocellular carcinomas remain isodense or isointense relative to the background and go undetected on CT and MR, respectively, despite optimal arterial phase imaging, but also these noninvasive techniques are not free from false-positive results that have been reported up to 57% on MR imaging [11], up to 28% on CT [6, 41, 43], and up to 42% on sonography [42].
A main limitation of our investigation was the relatively long time between CT arterioportography and digital subtraction angiography and transplantation. Of 26 patients with hepatocellular carcinoma, six underwent transplantation within more than 5 months (two patients within 6 months; one within 7, two within 8, and one within 11). Two of the undetected hepatocellular carcinomas, an 8-mm poorly differentiated carcinoma and a 6-mm well-differentiated carcinoma, were found in these patients. In light of this fact, we were unable to confirm the presence of these two hepatocellular carcinomas at the time of the diagnostic examination. Another limitation of our study was the relatively small patient cohort, particularly those with hepatocellular carcinoma. Finally, because of the lack of depth on posteroanterior digital subtraction angiograms, the correlation between them and axial CT arterioportograms and explant slices could have introduced some bias. However, the ability to interactively perform coronal multiplanar reformations of CT arterioportography data sets on the reviewing console substantially assisted tumor localization and correlation.
In summary, digital subtraction hepatic angiography in combination with CT arterioportography enables a significant accuracy for the detection of hepatocellular carcinoma in the cirrhotic liver. However, despite an excellent detection rate of advanced hepatocellular carcinoma, digital subtraction angiography and CT arterioportography combined are less sensitive for the detection of early hepatocellular carcinomas and may lead to assigning the patient a stage that is too low before transplantation. Hepatic digital subtraction angiography and CT arterioportography, separately and combined, have a relatively high false-positive rate. Recognizing the difficulty in the differentiation between early hepatocellular carcinoma and cirrhosis-relevant lesions, we believe that a percutaneous liver biopsy may be warranted in cases of suspected intrahepatic nodules.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  M. Kanematsu, A. Mallouhi, E. M. Gassner, I. Steingruber, and W. R. Jaschke How Should We Evaluate CT During Arterial Portography in Patients with Cirrhosis? Am. J. Roentgenol., February 1, 2004; 182(2): 529 - 530. [Full Text] [PDF]
|
|
| ||||||||||||
