Early Heterogeneous Enhancement of the Liver: Magnetic Resonance Imaging Findings and Clinical Significance

Transcription

1 JOURNAL OF MAGNETIC RESONANCE IMAGING 20: (2004) Original Research Early Heterogeneous Enhancement of the Liver: Magnetic Resonance Imaging Findings and Clinical Significance Masayuki Kanematsu, MD, 1 Maria I. Danet, MD, 1 Polytimi Leonardou, MD, 1 Maria Mastropasqua, MD, 1 Maria A. Mosetti, MD, 1 Larissa Braga, MD, 1 John T. Woosley, MD, 2 and Richard C. Semelka, MD 1 * Purpose: To investigate and assess the radiologic, serological, and histopathologic findings in patients who presented with early heterogeneous enhancement (EHE) on gadolinium-enhanced early-phase magnetic resonance imaging (MRI) of the liver. Materials and Methods: We searched our radiologic records of MRI of the liver from July 1999 to April 2002 to identify patients with EHE. Three investigators retrospectively evaluated in consensus the MR images in each patient for intensity and characteristic of EHE blinded to clinical information. Serological laboratory values and clinical information were obtained in all patients, and histologic findings were available in 19. Results: We identified 67 patients with EHE. Of them, 62 patients (93%) had underlying chronic liver disease. Twenty-seven patients had viral hepatitis, 13 had alcohol abuse, 6 had primary sclerosing cholangitis, and the others had miscellaneous etiologies. The five patients without chronic liver disease had the following clinical histories: concurrent chemotherapy for extrahepatic malignancy (two patients), concurrent intraabdominal infection (one), and no known associated disease (two). Intensity of EHE was intense in 6 (9%), moderate in 22 (33%), and mild in 39 (58%). Pattern of EHE was geographic in 15 patients (22%), patchy in 37 (55%), and miliary in 15 (22%). All EHE showed rapid fading on postcontrast late-phase images. EHE showed mild to moderate hyperintensity on T2-weighted images in 30 patients (45%). In the 19 patients with histological correlation, 19 (100%) had hepatocellular necrosis, 19 (100%) had fibrosis, 18 (95%) had inflammatory cell infiltration, 17 (89%) had capillary-size vessels within fibrous septa, and 1 Department of Radiology, University of North Carolina, Chapel Hill, North Carolina. 2 Department of Pathology and Laboratory Medicine, University of North Carolina, Chapel Hill, North Carolina. The current address of Masayuki Kanem is Department of Radiology, Gifu University School of Medicine, Gifu, Japan. *Address reprint requests to: R.C.S., Department of Radiology, CB 7510, University of North Carolina, Chapel Hill, NC richsem@med.unc.edu Received October 23, 2003; Accepted February 24, DOI /jmri Published online in Wiley InterScience ( 16 (84%) had ductal proliferation. No statistical correlation was found between the intensity or pattern of EHE on MR images and the extent of elevation of serological laboratory values or severity of histologic findings. Conclusion: Several different types of underlying chronic liver disease were observed in most of the patients with EHE. Hepatocyte necrosis, fibrosis, or inflammatory processes were found in all patients with EHE who had histopathological correlation. Key Words: liver; MR; gadolinium; contrast enhancement; chronic hepatitis J. Magn. Reson. Imaging 2004;20: Wiley-Liss, Inc. VARIATIONS OF HEPATIC PARENCHYMAL enhancement occurring in the absence of focal liver lesions have been observed in a number of settings, including portal vein thrombosis or compression (1,2), Budd-Chiari syndrome (3), hepatic arterial occlusion (4), hemolytic anemia, elevated liver function tests, and low platelets (HELLP) syndrome (5), and acute on chronic hepatitis (6,7). Many of the underlying causes are vascular in nature and are readily explained by vascular phenomena related to the hepatic dual blood supply, blood delivery autoregulation, and hepatic tissue pressure. However, the causes for early heterogeneous enhancement (EHE) not related to major vessel compression or occlusion are less well described. Acute on chronic hepatitis (7) and acute cholecystitis (8,9) are recognized as resulting in EHE, with the former condition having an association with cellular infiltration and the latter resulting from increased arterial perfusion stimulated by inflammation of the adjacent gallbladder. We undertook this study to ascertain the underlying disease processes that result in EHE, that are not related to major vessel compression or occlusion or acute cholecystitis, in a relatively large, consecutive group of patients. We identified patients with EHE and assessed these patients for possible clinical, serological, and histopathological correlations Wiley-Liss, Inc. 242

2 Early Heterogeneous Enhancement in Liver MRI 243 MATERIALS AND METHODS Patients We retrospectively searched the magnetic resonance imaging (MRI) reports of the liver in 2700 patients documented by one radiologist (R.C.S) at the Department of Radiology, University of North Carolina, from July 1999 to April 2002 for patients with EHE seen on gadolinium-enhanced early-phase MR images. We included a total of 67 patients with EHE with strict exclusion criteria for the following patients: patients with any type of hypervascular benign or malignant neoplasm, patients with enhancing dysplastic nodules or regenerative nodules in cirrhosis, patients with arterioportal shunting showing early portal venous filling and the surrounding parenchymal enhancement with a straight-line sign, patients with portal venous obstruction or compression or hepatic venous or arterial occlusion, patients with early, focal enhancement in the vicinity of the gallbladder or in the posterior aspect of the quadrate lobe of the liver, presumably due to aberrant venous return from cystic or pancreatico-pyloro-duodenal veins, respectively, and patients with acute cholecystitis. Evaluation for EHE was made on images obtained 18 seconds after initiating gadolinium administration, with contrast in hepatic arteries and portal veins and no contrast in hepatic veins. Information on underlying liver diseases (viral hepatitis, alcohol abuse, etc.) and serological laboratory values (alanine aminotransferase (ALT, normal range U/liter), aspartate aminotransferase (AST, U/liter), alkaline phosphatase (ALP, U/liter), -glutamyltranspeptidase ( -GTP, U/liter), within two weeks (mean days) of the MRI) were obtained from the clinical information system of our institution. Liver Pathology Nineteen patients underwent histopathologic evaluation (liver explant (N 10), percutaneous fine-needle aspiration biopsy (N 7), and transjugular needle biopsy (N 2)) within one month of the MRI. Histologic evaluation of hematoxylin-eosin specimens for the degree of hepatocyte necrosis, fibrosis, inflammatory cell infiltration, capillary-size vessels within fibrous septa, and ductal proliferation was made by an experienced liver pathologist (J.T.W.) at our institution, who was blinded to the MRI and clinical findings. The pathologist evaluated the degree of each histologic finding, as evaluated in a semiquantitative fashion (prominent, moderate, mild, minimal, and absent). MRI MRI was performed with a 1.5-T MR imager (Vision, Siemens Medical Systems, Iselin, NJ). MRI included breath-hold in-phase T1-weighted spoiled gradientecho (repetition time msec, echo time msec, flip angle 80, section thickness 8 mm, 20% intersection gap, one signal acquired, sections in a 20-second breath hold) and breath-hold T2-weighted half-fourier single-shot turbo spin-echo (HASTE; TR/TE 8/90, section thickness 8 mm, 20% intersection gap, one signal acquired, 20 sections in a 20-second breath hold) imaging. Transverse T1- weighted spoiled gradient-echo images were acquired prior to and after intravenous bolus injection of 0.1 mmol/kg of gadolinium chelate (Magnevist, Berlex Laboratories, Wayne, NJ, or Omniscan, Amersham Health, New York) at 2 ml/second, using a power injector. Postcontrast spoiled gradient-echo imaging was initiated at 18 seconds and 1 minute, and fat-suppressed spoiled gradient-echo images were acquired 2 minutes postcontrast. All patients were examined using the same technical protocol during the study period. Image Analysis Three investigators (M.I.D., M.K., P.L.), who were blinded to the clinical and histopathologic information and to the original MRI reports, retrospectively evaluated in consensus the unenhanced T1- and T2- weighted and early-phase (18 seconds) and late-phase (2 minutes) postcontrast images in the 67 patients. The MR images were reviewed for the following findings: intensity (intense, moderate, or mild) and pattern (geographic, patchy, or miliary) of EHE on early-phase postcontrast images; signal intensity (hyperintense, isointense, or hypointense) and heterogeneity (homogeneous or heterogeneous) of EHE on late-phase postcontrast images; signal intensity (hyperintense, isointense, or hypointense) on T1- and T2-weighted images corresponding to EHE; and presence and degree of intrahepatic bile duct dilatation. Enhancement was considered intense when it was comparable to normal pancreas or renal cortex, moderate when it was intermediate between the pancreas or renal cortex and the liver, and mild when it was slightly higher than the liver. The pattern of enhancement was considered miliary when it was observed predominantly as small, multiple foci smaller than 1 cm in size, patchy as multiple areas 1 5 cm in size, and geographic as multiple areas greater than 5 cm in size. Signal intensity on T1-weighted images was considered moderately hypointense when it was comparable to the spleen, and mildly hypointense when it was intermediate between the liver and spleen. Signal intensity on T2-weighted images was considered moderately hyperintense when it was comparable to the spleen, and mildly hyperintense when it was intermediate between the liver and spleen. The Spearman s rank correlation coefficient was calculated to compare the intensity and pattern of EHE determined on MR images and the extent of elevation of serological laboratory values in all patients and severity of histologic changes in 19 patients. RESULTS During the study period from July 1999 to August 2002, we identified 67 patients (34 men and 33 women, age range years, mean 46 years) with EHE. Sixty-two patients (93%) had underlying chronic liver disease (Figs. 1 4). The etiologies of underlying liver disease in the 62 patients were type C viral hepatitis in 15 patients (22%), alcohol abuse in 8 (12%), type B viral

3 244 Kanematsu et al. Figure 1. MR images in a 62-year-old man with cirrhosis of the liver associated with chronic type C viral hepatitis. Serum ALT (73 U/liter), AST (58 U/liter), and -GTP (90 U/liter) values, obtained 12 days before MRI, were elevated. a: Transverse postcontrast early-phase MR image (140/4.5) shows predominantly mild, miliary EHE in the entire liver. b: Transverse fat-suppressed postcontrast late-phase MR image (147/4.1) shows abnormal enhancement with a tiny nodular and fine reticular pattern typical of cirrhosis. c and d: Transverse unenhanced T1-weighted MR image (140/4.5) (c) and transverse unenhanced T2-weighted MR image (8/90) (d) show no abnormal imaging findings corresponding to EHE seen on postcontrast early-phase image. e: Photomicrograph (hematoxylin-eosin stain, 100) of explanted liver specimen shows macronodular cirrhosis. Moderate lymphocytic infiltrate (arrows) is seen within thick fibrous bands surrounding regenerative nodules in cirrhosis. This infiltrate also enters and damages more peripheral portions of regenerative nodules, resulting in a piecemeal necrosis-like pattern. hepatitis in 6 (9%), primary sclerosing cholangitis in 6 (9%), alcohol abuse and type C viral hepatitis in 5 (7%), Wilson s disease in 3 (4%), diffuse fatty liver in 2 (3%), autoimmune hepatitis in 2 (3%), allograft rejection in 2 (3%), mixed types B and C viral hepatitis in 1 (1%), primary biliary cirrhosis in 1 (1%), secondary biliary cirrhosis in 1 (1%), alpha-1-antitrypsin deficiency in 1 (1%), and cryptogenic hepatitis in 9 (13%). Twentyseven patients (40%) had viral hepatitis (Figs. 1 and 2), 13 (19%) had alcohol abuse (Fig. 2), 6 (9%) had primary

4 Early Heterogeneous Enhancement in Liver MRI 245 Figure 2. MR images in a 36-year-old man with cirrhosis of the liver associated with alcohol abuse and chronic type C viral hepatitis. Serum AST (59 U/liter), -GTP (133 U/liter), and ALP (477 U/liter) values, obtained seven days before MRI, were elevated. a: Transverse postcontrast early-phase MR image (140/4.5) shows intense, patchy EHE in the entire liver. b: Transverse fat-suppressed postcontrast late-phase MR image (147/4.1) shows homogeneous parenchymal enhancement sparing perihepatic venous areas (curved arrows). c: Transverse unenhanced T1-weighted MR image (140/4.5) shows no abnormal imaging findings corresponding to EHE seen on postcontrast early-phase image. d: Transverse unenhanced T2-weighted MR image (8/90) shows mildly to moderately increased signal intensity in the entire liver parenchyma, presumably reflecting diffuse inflammation throughout the liver. sclerosing cholangitis, and the other had miscellaneous etiologies (Figs. 3 and 4). Forty-five patients (67%) had cirrhosis of the liver, and the remaining 22 did not. Five patients (7%) had no underlying chronic liver disease but had the following clinical histories: concurrent chemotherapy for extrahepatic malignant disease (two patients), concurrent intraabdominal infection (one patient), and no known associated disease (two patients). Intrahepatic biliary dilatation was moderate in 2 patients, mild in 5, and absent in 60. Serum ALT (range U/liter, mean U/liter) was elevated in 39 (58%) of the 67 patients, serum AST ( U/liter, U/liter) was elevated in 53 (79%), serum ALP ( U/liter, IU/liter) was elevated in 35 (52%), and serum -GTP ( IU/liter, IU/liter) was elevated in 62 (93%). In the 19 patients with histologic correlation of the liver, hepatocyte necrosis was present in all patients (100%): prominent in 1, moderate in 0, mild in 3, and minimal in 15. Fibrosis was present in all patients (100%): prominent in 4, moderate in 3, mild in 7, and minimal in 5. Inflammatory cell infiltration was present in 18 (95%): prominent in 3, moderate in 7, mild in 5, and minimal in 3. Capillary-size vessels within fibrous septa were present in 17 patients (89%): prominent in 3, moderate in 2, mild in 8, and minimal in 4. Ductal proliferation was present in 16 (84%): prominent in 4, moderate in 2, mild in 2, and minimal in 8. The intensity of EHE was intense in 6 patients (9%), moderate in 22 (33%), and mild in 39 (58%). The pattern of EHE was geographic in 15 patients (22%), patchy in 37 (55%), and miliary in 15 (22%). The EHE showed rapid fading to homogeneous isointensity in 37 patients (55%), heterogeneous isointensity in 27 (40%), or homogeneous hyperintensity in 3 (4%) on postcontrast late-phase images. For the EHE, unenhanced T1- weighted images showed mild hypointensity in 2 patients (3%), isointensity in 63 (94%), and mild hyperintensity in 2 (3%). T2-weighted images showed moderate hyperintensity in 2 patients (3%), mild hyperintensity in 28 (42%), and isointensity in 37 (55%). There was no statistical correlation between the intensity or pattern of EHE determined on MR images and the extent of elevation of serological laboratory values or severity of histologic findings.

5 246 Kanematsu et al. Figure 3. MR images in a 24-year-old woman with chronic liver damage due to Wilson s disease. Serum ALT (1822 U/liter), AST (3167 U/liter), and -GTP (68 U/liter) values, obtained five days after MRI, were elevated. a: Transverse postcontrast early-phase MR image (140/4.1) shows predominantly moderate, geographic EHE throughout liver. b: Transverse fat-suppressed postcontrast late-phase MR image (175/4.1) shows abnormal enhancement with a tiny nodular and fine reticular pattern, suggesting the presence of cirrhosis. c: Transverse unenhanced T1-weighted MR image (140/4.1) shows no abnormal imaging findings. d: Transverse unenhanced T2-weighted MR image (8/90) shows mildly increased signal intensity in the entire liver parenchyma, presumably reflecting diffuse inflammation in the liver. e: Photomicrograph (hematoxylin-eosin stain, 100) of explanted liver specimen shows mild to moderate macro- and microvesicular steatosis with lymphocytic infiltrate within the lobule (arrow) as well as within the zones of bridging fibrosis (curved arrow). Numerous, capillary-size vascular profiles (C) and proliferating ductules (D) are noted within septa. There is marked hepatocanalicular cholestasis with ballooning of hepatocyte cytoplasm. DISCUSSION Prior researchers have described gadolinium-enhanced MRI findings of the liver with active hepatitis (6,7). Heterogeneous hepatic enhancement on postcontrast MR images was associated with a higher likelihood of liver disease and biochemical evidence of hepatic dysfunction than homogeneous enhancement (6). In livers with chronic hepatitis, early patchy enhancement was shown to indicate either concurrent or recent hepatocellular damage (7). In our current study, 93% of the

6 Early Heterogeneous Enhancement in Liver MRI 247 Figure 4. MR images in a 25-year-old man with allograft rejection 10 years after orthotopic liver transplantation for autoimmune hepatitis. Serum ALT (160 U/liter), AST (237 U/liter), -GTP (87 U/liter), and ALP (144 U/liter) values, obtained four days after MRI, were elevated. a: Transverse postcontrast early-phase MR image (140/4.5) shows intense, geographic EHE in the right hepatic lobe and moderate, geographic EHE in the left hepatic lobe. b: Transverse fat-suppressed postcontrast late-phase MR image (147/4.1) shows abnormal enhancement with a tiny nodular and fine reticular pattern, suggesting the presence of cirrhosis. c and d: Transverse unenhanced T1-weighted MR image (140/4.4) (c) and transverse unenhanced T2-weighted MR image (8/90) (d) show virtually no abnormal imaging findings corresponding to EHE on postcontrast early-phase image. e: Photomicrograph (hematoxylin-eosin stain, 100) of biopsy specimen subsequently obtained from the right lobe of the liver shows chronic hepatitic pattern with prominent piecemeal necrosis. In addition, there are mixed portal lymphocytic and eosinophilic infiltrates with endothelialitis and decreased numbers of interlobular ducts, which is consistent with a low level of superimposed acute and chronic rejection. Nodular lymphocytic infiltrates (arrow) are seen focally in fibrous septa. patients with EHE had underlying chronic liver disease. Although we did not achieve direct correlation between the EHE and its causative histopathologic findings, the histologic specimens obtained from various locations in the liver showed underlying hepatocyte necrosis and fibrosis in all, and inflammatory cell infiltration, capillary-size vessels within fibrous septa, and ductal proliferation in the majority. Biopsies were generally directed to the most abnormal enhancing regions appreciated on MRI.

7 248 Kanematsu et al. Prior reports that described unenhanced MRI findings in patients with chronic liver disease showed that livers with higher histologic activity of periportal and lobular necrosis and portal inflammation showed a decreased signal on T1-weighted images and an increased signal on T2-weighted images (10). In our study, T2- weighted images showed mild to moderate hyperintensity in 30 of the 67 patients (45%) with EHE, suggesting that although increased T2 relaxation time does occur with EHE, it is less sensitive than findings on postcontrast early-phase images (7). We did not find any statistical correlation between the intensity or pattern of EHE on MR images and the extent of elevation of serological laboratory values. The lack of correlation between the severity of radiologic, serologic, and histological findings may be partly due to time differences between temporal changes between acquisition of these tests, and the lack of strict control of biopsy location imaging, blood, and tissue sampling. It is known that serological laboratory values do not necessarily correlate with the activity of inflammation in the liver (11), and that histologic examination is commonly required to establish the diagnosis of activity of hepatitis. Although the number of histologically proven patients was few, and there was no control group of patients without EHE in our study, hepatocyte necrosis, fibrosis, inflammatory cell infiltration, capillarysize vessels within fibrous septa, and ductal proliferation were seen quite frequently in patients with EHE who underwent histologic examination of the liver, which suggested a relationship between EHE and activity of hepatic inflammation. We speculate some possible causes of EHE: First, EHE in patients with chronic liver damage most likely reflects rapid extravasation of gadolinium chelate into liver regions with increased extracellular spaces due to hepatocyte necrosis and damage to the capillaries (7). Second, hepatic arterial perfusion increases by the hyperemia due to active inflammation of the liver. Supporting this, increased hepatic arterial blood flow was shown in patients with chronic type C viral hepatitis on duplex Doppler color sonography (12). Third, decreased or uneven distribution of portal perfusion due to varying tissue pressure in varying extent of acute injury superimposed on chronic liver damage results in a compensatory increase of hepatic arterial perfusion. Previous researchers reported that portal perfusion was affected by multiple factors associated with chronic liver damage (13 15). Heterogeneous enhancement of liver parenchyma on CT during arterial portography, reflecting heterogeneous portal venous supply, is often observed in patients with cirrhosis (15). Vollmar et al (16) performed an animal experimental study to investigate the relationship between the hepatic microcirculation and the development of hepatic fibrosis in cirrhosis. They observed an increase in vascular luminal area caused by the formation of new shunting vessels bypassing the sinusoids as one of the most significant microvascular changes progressively observed in vivo. We also found capillary-size vessels within fibrous septa in most of the pathologic specimens. The EHE on MR images may be associated with such an altered hepatic microcirculation in the development of cirrhosis, and perhaps in its early stages or when acute inflammation develops on chronic hepatitis, more active alteration in microcirculation is present. One of the most important clinical aspects of EHE is to avoid confusion with diffuse-type hepatocellular carcinoma, which can also have EHE. Although not the subject of this study, in our clinical experience EHE tends to rapidly become more homogeneous on postcontrast late-phase images, while diffuse-type hepatocellular carcinoma tends to show heterogeneous washout and is generally associated with portal venous tumor thrombosis and a highly elevated serum alphafetoprotein value, neither of which are features associated with EHE. There are some limitations to this study: First, a small proportion of the total number of patients had histologic correlation; these patients generally had chronic liver damage of a moderate to severe degree, which might predispose them to a high frequency of histopathologically abnormal findings when histopathologically evaluated. Second, the biopsy specimens were not necessarily obtained from the region in the liver with the most extensive enhancement abnormalities, although we directed biopsy to the most abnormal region of the liver in our clinical practice. In conclusion, several different types of underlying chronic liver disease were observed in most of the patients with EHE. Histologic correlation suggested the relationship with hepatocyte necrosis, fibrosis, inflammatory cell infiltration, capillary-size vessels within fibrous septa, and ductal proliferation in the liver; this should be further clarified by a histologic correlation between patients with and without EHE. At present, however, radiologists should be aware that EHE on postcontrast early-phase images may indicate the presence of active hepatic inflammation. REFERENCES 1. Itai Y, Moss AA, Goldberg HI. Transient hepatic attenuation difference of lobar or segmental distribution detected by dynamic computed tomography. Radiology 1982;144: Schlund JF, Semelka RC, Kettritz U, Eisenberg LB, Lee JK. Transient increased segmental hepatic enhancement distal to portal vein obstruction on dynamic gadolinium-enhanced gradient echo MR images. J Magn Reson Imaging 1995;5: Noone TC, Semelka RC, Siegelman ES, et al. Budd-Chiari syndrome: spectrum of appearances of acute, subacute, and chronic disease with magnetic resonance imaging. J Magn Reson Imaging 2000;11: Semelka RC, Braga L, Armao D, et al. Liver. In: Semelka RC, editor. Abdominal-pelvic MRI, 1st edition. New York: Wiley-Liss; p Chiang KS, Athey PA, Lamki N. Massive hepatic necrosis in the HELLP syndrome: CT correlation. J Comput Assist Tomogr 1991; 15: Brown JJ, Borrello JA, Raza HS, et al. Dynamic contrast-enhanced MR imaging of the liver: parenchymal enhancement patterns. Magn Reson Imaging 1995;13: Semelka RC, Chung JJ, Hussain SM, Marcos HB, Woosley JT. Chronic hepatitis: correlation of early patchy and late linear enhancement patterns on gadolinium-enhanced MR images with histopathology: initial experience. J Magn Reson Imaging 2001;13: Ito K, Awaya H, Mitchell DG, et al. Gallbladder disease: appearance of associated transient increased attenuation in the liver at biphasic, contrast-enhanced dynamic CT. Radiology 1997;204:

8 Early Heterogeneous Enhancement in Liver MRI Loud PA, Semelka RC, Kettritz U, Brown JJ, Reinhold C. MRI of acute cholecystitis: comparison with the normal gallbladder and other entities. Magn Reson Imaging 1996;14: Marti-Bonmati L, Talens A, del Olmo J, et al. Chronic hepatitis and cirrhosis: evaluation by means of MR imaging with histologic correlation. Radiology 1993;188: Jamal MM, Soni A, Quinn PG, Wheeler DE, Arora S, Johnston DE. Clinical features of hepatitis C-infected patients with persistently normal alanine transaminase levels in the southwestern United States. Hepatology 1999;30: Walsh KM, Leen E, MacSween RN, Morris AJ. Hepatic blood flow changes in chronic hepatitis C measured by duplex Doppler color sonography: relationship to histological features. Dig Dis Sci 1998; 43: Silva G, Fluxa F, Bresky G, et al. Splanchnic and systemic hemodynamics in early abstinence and after ethanol administration in non-cirrhotic alcoholic patients. J Hepatol 1994;20: Lawrance JA, McDermott VG, Paulson EK, et al. Zebra pattern: a diagnostically challenging hepatic parenchymal enhancement pattern at CT arterial portography. Radiology 1997;203: Taura T, Nakamura K, Takashima S, et al. Heterogeneity of hepatic parenchymal enhancement on computed tomography during arterial portography: quantitative analysis of correlation with severity of hepatic fibrosis. Hepatol Res 2001;20: Vollmar B, Siegmund S, Menger MD. An intravital fluorescence microscopic study of hepatic microvascular and cellular derangements in developing cirrhosis in rats. Hepatology 1998;27: