JOURNAL OF MAGNETIC RESONANCE IMAGING 18:189 195 (2003)

Original Research

Hepatocellular Carcinoma of Diffuse Type: MR Imaging Findings and Clinical Manifestations

Masayuki Kanematsu, MD, Richard C. Semelka, MD,* Polytimi Leonardou, MD, Maria Mastropasqua, MD, and Joseph K.T. Lee, MD
Purpose: To assess MR imaging ndings and clinical manifestations of diffuse-type hepatocellular carcinoma (HCC). Materials and Methods: We retrospectively reviewed our experience with diffuse HCC from November 1994 to October 2001. MR imaging ndings and clinical features were assessed. Results: Twenty-two consecutive patients with diffuse-type HCC (19 men and three women, age range 16 80 years [mean, 52 years]) were identied in a review of liver MR studies. This represented 13% of all patients with HCC imaged during this time period. Diffuse HCC showed a permeative, inltrative pattern with ill-dened borders and no evidence of convex margination in all cases. At least 50% of the liver volume was involved with tumor. Diffuse-type HCC showed hypointensity in 15 patients, mixed intensity in three, and isointensity in four on T1-weighted images; heterogeneous hyperintensity in 16 patients; and homogeneous hyperintensity in six on T2-weighted MR images. Diffuse-type HCC showed patchy enhancement in 12 patients, miliary enhancement in nine, and minimal enhancement in one on postcontrast early-phase images, and showed heterogeneous washout in all patients on postcontrast late-phase images. Proximal portal venous tumor thrombosis was seen in all patients. Serum -fetoprotein (AFP) value was elevated ( 10 ng/mL) in 14 of 18 patients, and 13 showed a value greater than 500 ng/mL. The four patients who did not have elevated AFP had tumors which were indistinguishable from those in patients with elevated AFP; they also did not have a distinctive clinical history. Conclusion: Diffuse-type HCC was typically seen as an extensive, heterogeneous permeative hepatic tumor, with portal venous tumor thrombosis on MR images in all cases. Early enhancement, observed as patchy in 12 and miliary in nine of 22 patients, was a distinctive imaging feature. Elevated serum AFP value was a common nding; however, 22% had normal values. Key Words: Magnetic resonance; hepatocellular carcinoma; diffuse type; cirrhosis; contrast enhancement J. Magn. Reson. Imaging 2003;18:189 195. 2003 Wiley-Liss, Inc.

HEPATOCELLULAR CARCINOMA (HCC) is the most common primary malignant neoplasm arising in patients with chronic liver damage, and is most often related to hepatitis virus infection (1,2), alcohol abuse (3), or iron overload (4,5). A 1984 pathologic report (6) described that the common major gross patterns were expanding, spreading, and multifocal by means of aggregating a total of 529 HCC cases from Japan, the United States, and South Africa. Further modications have divided HCC into inltrative or expansive, single or multinodular, and mixed types, using observations on encapsulation and intrahepatic venous spread (7). Pathologically, diffuse-type HCC has been considered as a tumor that spreads throughout most of the liver and is not recognized as a focal tumor, typically accompanied by extensive portal venous tumor thrombosis and substantial elevation of serum -fetoprotein (AFP) value (8). Although many reports have described the MR appearance of focal forms of HCC, there is a relative paucity of descriptions of the MR appearance of diffusetype HCC in the literature. Diffuse-type HCC is often difcult to detect on imaging studies because of its permeative appearance and heterogeneity of background chronic liver disease. The purpose of this study was to describe the MR imaging ndings and clinical manifestations of diffuse-type HCC. MATERIALS AND METHODS We retrospectively searched the radiologic records of MR imaging of the liver performed at the Department of Radiology, University of North Carolina, from November 1994 to October 2001, and found records of 171 patients who had HCC and underwent MR imaging of the liver. Among this group, 22 patients were considered to have diffuse-type HCC based on the criterion that there was no distinct margination on any portion of the tumor on any sequence. Only patients with no prior treatment were included to avoid misinterpreting treatment changes as ndings of tumor. Fourteen patients, including four with normal serum AFP values and four in

Department of Radiology, University of North Carolina, Chapel Hill, North Carolina. Current address (M.K.): 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 27599-7510. E-mail: richsem@med.unc.edu Received February 4, 2003; Accepted April 8, 2003. DOI 10.1002/jmri.10336 Published online in Wiley InterScience (www.interscience.wiley.com).

2003 Wiley-Liss, Inc.

189

190

Kanematsu et al.

Figure 1. A 51-year-old man with diffuse-type HCC in alcoholic cirrhosis, showing serum alpha-fetoprotein value of 38000 ng/mL. A: Unenhanced T1-weighted spoiled gradient-echo (140/4.1) axial MR image shows the tumor (arrows) involving the entire right lobe of the liver as areas of homogeneous, mild hypointensity. Note a small volume of perihepatic and perisplenic ascites (curved arrow). B: Unenhanced T2-weighted fat-suppressed spin-echo (2400/90) axial MR image shows the tumor as areas of heterogeneous, moderate hyperintensity (arrows). C: Gadolinium-enhanced early-phase spoiled gradient-echo (140/ 4.1) axial MR image shows the tumor as areas of patchy enhancement (arrow). D: Gadolinium-enhanced fat-suppressed late-phase spoiled gradient-echo (140/4.1) axial MR image shows the tumor as areas of multiple foci of wash-out (arrow). Portal venous tumor thrombosis in the left main portal vein branch is shown as areas of discrete hypointensity (curved arrow).

whom serum AFP values were not available, had histopathologic conrmation (laparotomic wedge biopsy [N 1], percutaneous needle-core aspiration biopsy [N 13]). The remaining eight patients without histologic conrmation had MR imaging ndings consistent with extensive, malignant hepatic tumors associated with markedly high ( 2,000 ng/mL) serum AFP values. MR imaging was performed with a 1.5-T MR imager (Vision, Siemens Medical Systems, Iselin, NJ). MR imaging included a T1-weighted, in-phase, breath-hold, spoiled gradient-echo sequence (repetition time of 140 175 msec and echo time of 4.1 4.5 msec [140 175/ 4.1 4.5]; ip angle, 80; section thickness, 8 mm; intersection gap, 20%; one signal acquired; 21 sections in a 20-second breath hold), and a T2-weighted sequence performed on fat-suppressed spin-echo (2400/90; section thickness, 8 mm; intersection gap, 20%; two signals acquired) (one patient), short tau inversion recovery turbo spin-echo (5,110/76/170 [TR/TE/TI]; section thickness, 8 mm; intersection gap, 20%; two signals acquired) (four patients), or half-Fourier turbo spinecho sequence ( /90; section thickness, 8 mm; inter-

section gap, 20%; one signal; 20 sections) (17 patients). Transverse spoiled gradient-echo images were acquired prior to and after intravenous bolus injection of 0.1 mmol/kg gadolinium chelate (Magnevist, Berlex Laboratories, Wayne, NJ or Omniscan, Nycomed, New York, NY). Postcontrast spoiled gradient-echo sequences were initiated at 18 seconds and one minute, and fat-suppressed spoiled gradient-echo sequence was acquired two minutes postcontrast. Three investigators, Masa Kanematsu, Richard C. Semelka, and Polytimi Leonardou, who were blinded to the clinical and histopathologic information and to the original MR imaging reports, retrospectively evaluated, in consensus, the transverse unenhanced T1- and T2weighted MR and postcontrast early- and late-phase MR images in each patient. The MR images were reviewed for the following ndings: tumor extension in the liver expressed as number of Couinauds segments involved by tumors, brous capsules or septa, signal intensity characteristics of tumors on unenhanced T1and T2-weighted MR images, contrast enhancement characteristics of tumors on early- and late-phase post-

MRI of Diffuse HCC

191

Figure 2. A 44-year-old man with diffuse-type HCC in cirrhosis due to type-C viral hepatitis, showing serum alpha-fetoprotein value of 9000 ng/mL. A: Unenhanced T1-weighted spoiled gradient-echo (170/4.1) axial MR image shows the tumor (arrow) involving the right hepatic lobe and the medial segment of the left hepatic lobe as ill-demarcated areas of homogeneous, mild hypointensity, accompanied by extensive portal venous tumor thrombosis (curved arrow) shown as areas of mixed intensity. Note a moderate amount of perihepatic and perisplenic ascites (small arrows). B: Unenhanced T2-weighted turbo spin-echo ( /90) axial MR image shows the tumor (arrow) as areas of heterogeneous, moderate hyperintensity. Portal venous thrombosis (curved arrow) is shown as areas of mixed signal intensity. C: Gadolinium-enhanced early-phase spoiled gradient-echo (170/4.1) axial MR image shows the tumor as areas of miliary enhancement (arrow). Portal venous tumor thrombosis is shown as areas of heterogeneous hypointensity (curved arrow). Note areas of necrosis in the tumor (small arrow). D: Gadolinium-enhanced late-phase fat-suppressed spoiled gradient-echo (175/4.1) axial MR image shows the tumor as areas of heterogeneous isointensity with some areas of wash-out (arrow). Portal venous tumor thrombosis is shown as areas of discrete, heterogeneous, hypointensity (curved arrow).

contrast MR images, bile duct dilatation, portal venous tumor thrombosis, hepatic venous tumor thrombosis, ascites, and upper abdominal lymph-node metastases. The signal intensity of tumor on T1-weighted images was categorized as moderately hypointense when it was comparable to that of the spleen, and mildly hypointense when it was intermediate between liver and spleen. The signal intensity of tumor on T2-weighted images was categorized as moderately hyperintense when it was comparable to spleen, and mildly hyperintense when it was intermediate between liver and spleen. Enhancement characteristics of tumor were categorized into patchy and miliary: patchy enhancement referred to foci of enhancement that were irregular in size and shape and ill-dened, and miliary enhancement referred to foci that were small ( 5 mm) and relatively well dened. Other information on underlying hepatic disease (viral hepatitis, alcohol abuse, autoimmune hepatitis,

hemochromatosis, etc.), serological laboratory test results (total bilirubin, AFP, alkaline phosphatase, alanine aminotransferase, aspartate aminotransferase, -glutamyltranspeptidase), and distant metastases were obtained using the clinical information system of our institution. RESULTS From November 1994 to October 2001, we had 22 patients with diffuse-type HCC (Figs. 13), which represented 13% of all patients with HCC. The 22 patients included 19 men and three women, ranging in age from 16 80 years (mean age, 52 years). Sixteen patients had cirrhosis, and the remaining six had chronic hepatitis. The patient characteristics are summarized in Table 1 and the MR imaging ndings in Table 2. Serum total bilirubin value (normal range, 0 1.2 mg/ dL) was elevated in 19 of 20 patients in whom the test

192

Kanematsu et al.

one, and eight segments in 12 (mean, 6.7 1.6 segments). Fibrous capsules or septa were seen mildly but partly in seven, minimally in two, and none in 13. Unenhanced T1-weighted MR images showed the tumors as areas of homogeneous hypointensity in ve patients, heterogeneous hypointensity in ten, mixed intensity in three, and isointensity in four. T2-weighted MR images showed the tumors as areas of heterogeneous hyperintensity in 16 and homogeneous hyperintensity in six. Postcontrast early-phase MR images showed tumors as areas of patchy enhancement in 12 patients, miliary enhancement in nine, and minimal enhancement in one. Postcontrast late-phase MR images showed heterogeneous irregular areas of wash-out in all patients. Portal venous tumor thrombosis was seen in all patients: bilateral main portal vein branches were involved in 12 patients and either of the right or left main portal vein branch was involved in 10. Hepatic venous tumor thrombosis was not seen in any patient. Intrahepatic bile ducts were minimally to mildly dilated in three patients and of normal caliber in 19. Ascites was substantial in one patient, moderate in seven, minimal in eight, and absent in six. Upper abdominal lymphadenopathy was seen in three patients (porta-hepatis nodes in all three and one with additional porto-caval and peripancreatic nodes). Distant metastases were conrmed in three patients (bone in one, adrenal gland in one, and lung in one). DISCUSSION A 1981 report by Okuda et al (8) described the clinical and pathological ndings with six autopsy cases of diffuse-type HCC. In the report, the prominent clinical feature was the rapid deterioration of the patients general condition terminating in hepatic failure, and the liver size enlarged quickly at a perceptible speed, often accompanied by abdominal pain. They also described that the entire liver was studded with minute, uniformly sized tumor nodules evenly distributed throughout, with some of them grossly indistinguishable from cirrhotic nodules. The incidence of diffuse-type HCC in our institution was approximately 13% of all patients with HCC, which is slightly higher than previously reported (8). This may reect differences related to geography, as our study involved a North American population, variation related to the relative rarity of the disease, or that our study was imaging-based rather than autopsy-based. Portal venous tumor thrombosis is a common nding with diffuse-type HCC and may be a clue to the diagnosis. Accordingly, characterization of portal venous thrombosis is crucial in the diagnosis of diffuse-type HCC on contrast-enhanced CT or MR imaging. Tublin et al (9) evaluated 58 cirrhotic patients and concluded that identication of main portal venous thrombosis of 23 mm in diameter or greater or neovascularity of portal venous thrombosis resulted in a sensitivity and specicity for the CT characterization of malignant portal venous thrombosis of 86% and 100%, respectively. No previous reports have described the specic MR imaging ndings of diffuse-type HCC cases. We found that the unenhanced T1- and T2-weighted MR images

Figure 3. A 64-year-old man with diffuse-type HCC in hemochromatosis, showing negative serum alpha-fetoprotein value ( 10 ng/mL). A: Unenhanced T2-weighted short tau inversion recovery turbo spin-echo (5,110/76/170 [TR/TE/TI]) axial MR image shows the tumor involving most of the left hepatic lobe and a part of right hepatic lobe as areas of homogeneous, moderate hyperintensity (arrow). Note extensive portal venous tumor thrombosis (curved arrow). B: Gadolinium-enhanced early-phase spoiled gradient-echo (160/4.5) axial MR image shows the tumor as areas of intense, patchy enhancement (arrow). Note the very low signal intensity of the liver parenchyma (curved arrow), reecting underlying hemochromatosis.

result was available (0.723.5 mg/dL [mean, 6.5 8.1 mg/dL]), serum AFP value (0 10 ng/mL) was elevated in 14 of 18 patients (5242000 ng/mL, mean; 24822 56380 ng/mL), and 13 showed a value greater than 500 ng/mL. Serum alkaline phosphatase value (38 126 U/liter) was elevated in 17 of 20 patients (99 969 U/liter [356 267 U/liter]), serum alanine aminotransferase (15 48 U/liter) value was elevated in 14 of 20 patients (21190 U/liter [84 52 U/liter]), serum aspartate aminotransferase value (14 38 U/liter) was elevated in 18 of 20 patients (38 876 U/liter [243 226 U/liter]), serum -glutamyltranspeptidase value (11 48 U/liter) was elevated in all of 18 patients (54 1042 U/liter [361 270 U/liter]). Number of Couinauds segments involved by diffusetype HCC was four segments in four patients, ve segments in one, six segments in four, seven segments in

MRI of Diffuse HCC Table 1 Summary of Patient Characteristics in 22 Patients With Diffuse HCC Patients Age/sex/race 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
a

193

Etiology of underlying liver disease Alcohol abuse Alcohol abuse Alcohol abuse Alcohol abuse Alcohol abuse Hep. B Hep. B Hep. B Hep. C Hep. C Hep. C Hep. C Hep. C Hep. C Hep. C Alcohol abuse, Hep. C Hep. B, Hep. C Hemochromatosis Autoimmune hepatitis Cryptogenic Cryptogenic Cryptogenic

Cirrhosis Yes Yes Yes Yes Yes Noa Yes Noa Noa Yes Yes Yes Noa Yes Yes Yes Yes Noa Yes Yes Noa Yes

TB AFP ALP ALT AST -GTP (mg/dl) (ng/mL) (U/liter) (U/liter) (U/liter) (U/liter) 3.1 23.5 19.3 1.3 1.5 20.8 3.3 N.A. 8.7 2.2 1.8 2.7 2.3 2.0 2.0 2.2 5.2 0.7 N.A. 23.5 2.1 1.9 38,000 15,870 38,000 2,000 N.A. 564 5366 N.A. 42,000 1816 N.A. 136 5 9000 242,000 7000 38,000 10 N.A. 7000 10 10 242 114 99 245 551 623 851 N.A. 165 310 173 195 122 168 784 174 280 311 N.A. 193 969 548 136 28 53 66 28 155 147 N.A. 47 21 40 132 134 50 60 53 113 55 N.A. 190 134 46 395 65 118 105 38 700 404 N.A. 207 67 38 447 162 268 257 94 292 73 N.A. 876 168 86 257 350 84 274 666 177 1042 N.A. 184 445 139 135 254 N.A. N.A. 249 288 630 N.A. 54 822 451

No. of Capsules involved or septa segments 4 segs. 4 segs. All All All All All 6 segs. All 5 segs. All 7 segs. 4 segs. All All 6 segs. 4 segs. 6 segs. All All 6 segs. All

60/M/C 30/M/AA 51/M/C 48/M/AA 46/M/C 50/M/AA 64/M/C 49/M/A 49/M/AA 48/M/C 49/M/AA 51/M/AA 49/M/C 44/M/C 61/F/C 43/M/C 58/M/AA 64/M/C 16/F/C 64/M/C 71/M/C 80/F/C

Patients had chronic hepatitis. TB serum total bilirubin (normal range, 0 1.2 mg/dl), AFP serum alpha-fetoprotein (0 10 ng/mL), ALP serum alkaline phosphatase (38 126 U/liter), ALT serum alanine aminotransferase (15 48 U/liter), AST serum aspartate aminotransferase (14 38 U/liter), -GTP serum -glutamyltranspeptidase (11 48 U/liter), C Caucasian, AA African-American, A Asian, Hep. B type-B viral hepatitis, Hep. C type-C viral hepatitis, N.A. not available, mild, minimal, none.

Table 2 Summary of MRI Findings in 22 Patients With Diffuse HCC Patients 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 T1-weighted imaging ndings Iso Iso Homo, mild hypo Hetero, mild hypo Hetero, mild hypo Hetero, mild hypo Hetero, mod. hypo Hetero, mild hypo Homo, mod. hypo Mixed Hetero, mild hypo Hetero, mild hypo Homo, mild hypo Homo, mild hypo Hetero, mild hypo Hetero, mod. hypo Iso Hetero, mod. hypo Iso Mixed Mixed Homo, mod. hypo T2-weighted imaging ndings Hetero, mild hyper Hetero, mild hyper Hetero, mod. hyper Hetero, mild hyper Hetero, mod. hyper Hetero, mild hyper Hetero, mild hyper Hetero, mild hyper Homo, mod. hyper Homo, mild hyper Hetero, mild hyper Hetero, mild hyper Homo, mild hyper Hetero, mod. hyper Hetero, mild hyper Hetero, mild hyper Homo, mod. hyper Homo, mod. hyper Homo, mod. hyper Hetero, mild hyper Hetero, mild hyper Hetero, mod. hyper Enhancement pattern in early-phase postcontrast images Miliary Patchy Patchy Miliary Patchy Miliary Patchy Miliary Miliary Patchy Patchy Patchy Patchy Miliary Miliary Patchy Minimal Patchy Patchy Patchy Miliary Miliary Washed-out area % in late-phase postcontrast images 5075% 5075% 75100% 2550% 75100% 75100% 5075% 75100% 75100% 2550% 5075% 2550% 2550% 2550% 5075% 75100% 5075% 2550% 5075% 2550% 5075% 5075% homogeneous, Hetero Portal venous tumor thrombosis in MPV(s) Right Right Bilateral Left Bilateral Bilateral Bilateral Right Bilateral Right Right Bilateral Right Bilateral Bilateral Right Left Bilateral Bilateral Bilateral Right Bilateral Bile duct dilatation

MPV(s) main portal vein branch(es), moderate, hypo hypointensity, hyper

mild, minimal, none, Homo hyper intensity, iso iso intensity.

heterogeneous, mod.

194

Kanematsu et al.

showed the tumors as ill-demarcated areas of homogeneous or heterogeneous, abnormal signal intensities, typically mildly to moderately hypointense on T1weighted images and mildly to moderately hyperintense on T2-weighted images. We considered the ndings on noncontrast images somewhat nonspecic as these signal alterations may also be analogous to those in cirrhotic liver without tumor. However, multiple small areas of low signal intensity in the liver on T2-weighted images may suggest the presence of regenerating nodules in cirrhosis, which may help exclude the diagnosis of diffuse-type HCC (10). We observed patchy or miliary enhancement corresponding to the tumors on the early-phase postcontrast images. Miliary enhancement in particular, we believe, may be relatively specic for diffuse-type HCC, and may represent the enhancement of extensive micronodules of tumor as identied at histopathology (8). Heterogeneous wash-out of the tumors on the late-phase images is a feature typical for malignant tumor of all types in the liver. Increased hepatic parenchymal enhancement on early-phase postcontrast images is known to be caused by increased hepatic arterial supply in the setting of proximal portal vein obstruction (11), but this usually is observed as homogeneous, wedge-shaped enhancement in postcontrast early phase, with fading to homogeneous isointensity in postcontrast late phase. We could distinguish such non-neoplastic, transient early-phase enhancement, which was seen in some cases, from the enhancement of tumor. In our clinical experience diffuse HCC may be distinguished from other malignant diseases, specically cholangiocarcinoma and metastases, in the great majority of cases. Cholangiocarcinoma, in our experience, has better dened margins and we have not observed portal vein tumor thrombus with this entity, although portal vein comparison is common. Massive involvement of the liver in metastatic disease almost invariably shows a more clearly focal pattern of involvement. In our experience, breast cancer is the malignancy that most often may present with extensive inltration. In this setting the primary tumor is almost always known. Coexistence of breast cancer and cirrhosis is rare but may occur, so uncertainty whether the patient has breast cancer liver metastases or diffuse HCC is rare. Finally, while venous tumor thrombus was seen in all patients with diffuse HCC, it is rare in patients with metastatic disease. Lee et al (12) reported that the serum AFP level, which showed a positive predictive value of 95% for HCC, was 3200 ng/mL in livers with type-B viral hepatitis and 200 ng/mL in livers without type-B viral hepatitis. They concluded that the presence of underlying type-B viral hepatitis should be taken into consideration when serum AFP is measured to diagnose HCC. One of our patients had underlying type-B viral hepatitis and had a serum AFP value of 5366 ng/mL, which was considered diagnostic for HCC. Serum AFP values have been reported to be elevated in 43%72% of patients with HCC (1316). However, as more small HCCs are currently being detected, due to the advent of advanced radiologic imaging techniques

and increased imaging screening of patients with hepatitis or cirrhosis, this rate is likely much lower, as small well-differentiated focal HCCs rarely result in elevated AFP (17). We observed that 14 (78%) of 18 patients with diffuse HCC showed elevated serum AFP values in our study. Radiologists should be apprised of the fact that although the positive rate of AFP is high with diffuse HCC, a sizable number (22% in our study) have normal serum AFP value. There are some limitations to this study. We did not have a histopathologic diagnosis of HCC in eight patients. These patients were very ill and the combination of the MR ndings and markedly elevated serum AFP values ( 2000 ng/mL) were considered diagnostic for HCC, as also described in the literature (12). In conclusion, diffuse-type HCC represented 13% of the patients with HCC. Diffuse-type HCC was seen as extensive permeative hepatic tumor with ill-dened margins, with extensive portal venous tumor thrombosis in all patients, and markedly elevated serum AFP value in 78% of patients. Serum AFP value was normal in 22% of patients. Unenhanced MR images showed nonspecic ndings of abnormal signal intensities. Gadolinium-enhanced MR images showed patchy or miliary enhancement on immediate postcontrast images, with the miliary pattern possessing a distinctive appearance.

REFERENCES
1. Tsukuma H, Hiyama T, Tanaka S, et al. Risk factors for hepatocellular carcinoma among patients with chronic liver disease. N Engl J Med 1993;328:17971801. 2. National Institutes of Health Consensus Development Conference Panel statement: Management of hepatitis C. Hepatology 1997;26: 210. 3. Donato F, Tagger A, Chiesa R, et al. Hepatitis B and C virus infection, alcohol drinking, and hepatocellular carcinoma: a case-control study in Italy. Brescia HCC study. Hepatology 1997;26:579 584. 4. Niederau C, Fischer R, Sonnenberg A, Stremmel W, Trampisch HJ, Strohmeyer G. Survival and causes of death in cirrhotic and noncirrhotic patients with primary hemochromatosis. N Engl J Med 1985;313:1256 1262. 5. Bradbear RA, Bain C, Siskind V, et al. Cohort study of internal malignancy in genetic hemochromatosis and other non-alcoholic liver disease. J Natl Cancer Inst 1985;75:81 84. 6. Okuda K, Peters RL, Simson IW. Gross anatomic features of hepatocellular carcinoma from three disparate geographic areas. Proposal of new classication. Cancer 1984;54:21652173. 7. Anthony PP. Tumours and tumour-like lesions of the liver and biliary tract: aetiology, epidemiology, and pathology. In: MacSween RNM, Burt AD, Portmann BC, et al., eds. Pathology of the liver. 4th ed. London, England: Churchill Livingstone; 2002. p 711 825. 8. Okuda K, Noguchi T, Kubo Y, Shimokawa Y, Kojiro M, Nakashima T. A clinical and pathological study of diffuse type hepatocellular carcinoma. Liver 1981;1:280 289. 9. Tublin ME, Dodd 3rd GD, Baron RL. Benign and malignant portal vein thrombosis: differentiation by CT characteristics. AJR Am J Roentgenol 1997;168:719 723. 10. Itai Y, Ohnishi S, Ohtomo K, et al. Regenerating nodules of liver cirrhosis: MR imaging. Radiology 1987;165:419 423. 11. Itai Y, Moss AA, Goldberg HI. Transient hepatic attenuation difference of lobar or segmental distribution detected by dynamic computed tomography. Radiology 1982;144:835 839. 12. Lee HS, Chung YH, Kim CY. Specicities of serum alpha-fetoprotein in HBsAg and HBsAg- patients in the diagnosis of hepatocellular carcinoma. Hepatology 1991;14:68 72.

MRI of Diffuse HCC

13. Sherman M, Peltekian KM, Lee C. Screening for hepatocellular carcinoma in chronic carriers of hepatitis B virus: incidence and prevalence of hepatocellular carcinoma in a North American urban population. Hepatology 1995;22:432 438. 14. Aoyagi Y, Oguro M, Yanagi M, et al. Clinical signicance of simultaneous determinations of alpha-fetoprotein and desgamma-carboxy prothrombin in monitoring recurrence in patients with hepatocellular carcinoma. Cancer 1996;77:1781 1786. 15. Okuda H, Nakanishi T, Takatsu K, et al. Measurement of serum levels of des-gamma-carboxy prothrombin in patients with hepato-

195
cellular carcinoma by a revised enzyme immunoassay kit with increased sensitivity. Cancer 1999;85:812 818. 16. Nguyen MH, Garcia RT, Simpson PW, Wright TL, Keeffe EB. Racial differences in effectiveness of alpha-fetoprotein for diagnosis of hepatocellular carcinoma in hepatitis C virus cirrhosis. Hepatology 2002;36:410 417. 17. Peterson MS, Baron RL, Marsh Jr JW, Oliver 3rd JH, Confer SR, Hunt LE. Pretransplantation surveillance for possible hepatocellular carcinoma in patients with cirrhosis: epidemiology and CTbased tumor detection rate in 430 cases with surgical pathologic correlation. Radiology 2000;217:743749.