Histological Features Predicting Malignant Transformation of Nonmalignant Hepatocellular Nodules: A Prospective Study

Transcription

1 GASTROENTEROLOGY 1998;115: LIVER, PANCREAS, AND BILIARY TRACT Histological Features Predicting Malignant Transformation of Nonmalignant Hepatocellular Nodules: A Prospective Study SHUICHI TERASAKI,*, SHUICHI KANEKO,* KENICHI KOBAYASHI,* AKITAKA NONOMURA, and YASUNI NAKANUMA *First Department of Internal Medicine, Pathology Section, and Second Department of Pathology, Kanazawa University School of Medicine, Kanazawa, Japan Background & Aims: Recent advances in imaging modalities allow the identification of borderline hepatocellular nodules that have the potential for malignant transformation. The aim of this study was to elucidate histological predictive features of borderline nodules by needle biopsy for the evolution to hepatocellular carcinoma (HCC). Methods: Thirty-four hepatocellular nodules diagnosed by needle biopsy were followed up for more than 6 months. Several histological parameters of these nodules that were related to malignant transformation were evaluated. Results: During the follow-up periods (median, 35 months), 5 of 34 nodules evolved to HCC during a follow-up of 6 15 months. Significant prognostic features of malignant transformation were an increased ratio of nuclear density of G1.5, clear cell change, small cell dysplasia, and fatty change of the hepatocytes. In multivariate analysis, an increased ratio of nuclear density of G1.5 and clear cell change were independent. Conclusions: A nodule with an increased ratio of nuclear density, clear cell change, small cell dysplasia, and fatty change should be recognized to be a high risk for evolution to HCC. Particularly, the former two were independent prognostic factors for malignant transformation. With advances in hepatic imaging modalities, many small nodular lesions are now detected in advanced chronic liver diseases, including those resulting in cirrhosis. The nodules range from 1 to 2 cm in diameter, and their imaging features differ from those of typical hepatocellular carcinoma (HCC) encountered during the few past decades. Specifically, the tumors do not usually enhance on angiography. Surgically resected small nodules are divided into several categories by pathology. Some nodules may be classified as overt HCC, whereas other nonmalignant hepatocellular nodules do not show enough atypia to be diagnosed as HCC. The latter lesions have been variably termed adenomatous hyperplasia (AH), 1,2 atypical AH, 2,3 and macroregenerative nodules. 4 7 Recently, it has been suggested that these nodules represent preneoplastic lesions, borderline tumors, or very early stage of HCCs. An International Working Party chaired by Wanless et al. has proposed terming such lesions dysplastic nodules. 8 These nodules are usually diagnosed histologically using tissue obtained by ultrasound-guided needle liver biopsy in a clinical setting. However, pathological diagnosis of needle biopsy specimens has some difficulties. First, differentiating nonmalignant hepatocellular nodules from HCCs is often difficult, even by experienced hepatopathologists. Second, sampling errors may occur because of the histopathologic heterogeneity of these nodules. Because the most atypical portions of a nodule are not always sampled by the biopsy, the histological grade of the nodule may be underestimated. Third, the ultimate fate of these hepatocellular nodules has not yet been clarified from the standpoint of histological changes. It is important to realize that not all nonmalignant hepatocellular nodules progress to overt HCC In clinical settings, the prediction of the outcome of the nodules may be very important. At present, it remains unsettled which radiological, laboratory, and histological findings predict the outcome of these nodules. Therefore, clinical follow-up studies for malignant transformation with an emphasis on histological features are needed to clarify these issues. In this prospective study, we analyzed the histological prognostic features of nonmalignant hepatocellular nodules by following up patients with nodules diagnosed by ultrasound-guided needle biopsy. Abbreviations used in this paper: AFP, -fetoprotein; AH, adenomatous hyperplasia; CT, computerized tomography by the American Gastroenterological Association /98/$3.00

2 November 1998 FATE OF NONMALIGNANT HEPATOCELLULAR NODULES 1217 Materials and Methods Subjects and Follow-up Ultrasound-guided needle biopsy specimens were obtained sequentially from 87 nodular lesions from 59 cirrhotic livers at Kanazawa University Hospital between 1991 and All nodules, detected by diagnostic ultrasound equipment (SSA-260A; Toshiba, Yokohama, Japan) using a MHz ultrasound transducer (PSF-37HT; Toshiba), were 3cm in diameter when evaluated by ultrasonography and did not enhance on angiography. A 21-gauge needle (Majima needle; Top, Tokyo, Japan) was used for needle biopsies. To compare histological changes, liver tissue from the inside and outside portions of the nodules was obtained in individual cases. If the nodules were not histologically diagnosed as HCC, a rebiopsy was performed similarly. Histologically, these 87 nodules were composed of 35 HCCs and 52 nonmalignant hepatocellular nodules. The HCCs were diagnosed according to available histological criteria. 8,13,14 Briefly, HCCs were characterized by nuclear atypia and structural changes according to Terada et al. 13 and Ferrell et al. 7,14 Invasion of tumor cells into portal areas was also recognized as a feature of malignancy. 15 Of the 52 nonmalignant hepatocellular nodules, 12 treated by percutaneous ethanol injection therapy were excluded. They were treated because 10 patients had coexistent HCC(s) in other parts of the liver, and 2 patients declined follow-up of the nodules and wanted treatment. Of the 40 untreated nodules, 4 nonmalignant nodules from 2 patients were excluded from the study because the patients discontinued visiting our hospital, and 2 nodules were excluded because they were masked by multiple nodules arising later in the livers of these patients. We followed up the remaining 34 nodules from 28 patients who did not undergo anticancerous treatment (age, years; 20 men and 8 women; 10 with hepatitis B virus infection and 18 with hepatitis C virus infection) for longer than 6 months with the patients informed consent. Four patients had multiple nodules. Two patients had 3 nodules, and the other 2 had 2 nodules. The median size of these 34 nodules by ultrasonography was 10 mm (range, 5 20 mm) in diameter. Echogenicity of the nodules was hyperechoic in 15, hypoechoic in 16, and mixed hyperechoic and hypoechoic in 3. Six cases had coexistent HCC apart from the nonmalignant nodules diagnosed during the follow-up period. These HCCs did not interfere with the evaluation of the nonmalignant lesions. Periodic imaging studies and evaluation of serum tumor markers for HCC, -fetoprotein (AFP) levels, and serum protein induced by vitamin K absence II levels were obtained. Specifically, ultrasonography was performed every 3 months, and computerized tomography (CT) and magnetic resonance imaging were performed every 6 months. Tumor markers were measured every 2 months. Angiography and CT during arterioportography were performed when the nodules enlarged, when the tumor marker levels became elevated, when HCC occurred in another part of the liver, when other hepatic complications developed, or when a patient desired further investigation 1 year after the first biopsy was performed. In addition, a follow-up biopsy of all nodules was performed when a nodule was detected by ultrasonography at readmission. An end point of the follow-up period was the specific time when the nodule was diagnosed as HCC or the time when the latest imaging was performed if the nodule had not been diagnosed as HCC. Histological Examination Sections were cut 3-µm thick from formalin-fixed, paraffin-embedded biopsy specimens. After deparaffinization, sections were stained with H&E and Gomori s silver impregnation. Histological specimens were examined blindly by a liver pathologist who was unaware of the outcome of the nodules. We examined the nuclear density of hepatocytes within the nodules and compared these measurements with those outside of the nodules. Nuclear density was calculated by counting the number of hepatocyte nuclei in a µm area inside and outside of the nodule. The ratio of hepatocellular nuclear density within the nodules to that outside of the nodules was also calculated. We surveyed the following histological features that are often encountered in AHs as well as well-differentiated HCCs 2,13,16 18 : liver cell dysplasia (small cell and large cell dysplasia), microacinar formation, clear cell change of the hepatocytes, nuclear deviation toward the sinusoids, fatty change of the hepatocytes, alteration of cytoplasmic stainability, and reduction or loss of reticulin fiber (Figure 1). Liver cell dysplasia was divided into small cell dysplasia and large cell dysplasia based on the size of dysplastic hepatocytes compared with that of hepatocytes without dysplasia. Small cell dysplasia was defined as minor changes in cell size, increases in the nucleocytoplasmic ratio, and pleomorphism 19 (Figure 1A). Large cell dysplasia was characterized by increasing size of both the nucleus and cytoplasm, nuclear pleomorphism, and multinucleation 20 (Figure 1B). Statistics All continuous data are expressed as median and range. Univariate analyses of prognostic histological factors on malignant transformation were performed with log rank tests for categorical variables. In multivariate analysis, Cox s proportional hazards regression model for survival analysis was used to assess the prognostic histological factors. A multivariate stepwise model using forward selection was applied to identify independent predictors. A P value of 0.05 was considered significant. Statistical calculations were performed using SPSS (SPSS, Inc., Chicago, IL) on a Macintosh computer (Apple Computer, Inc., Cupertino, CA). Results The 34 liver nodules were followed up for 6 64 months (median, 35 months). Among them, 22 nodules from 20 patients were examined by angiography and CT

3 1218 TERASAKI ET AL. GASTROENTEROLOGY Vol. 115, No. 5 Figure 1. Histological parameters shown in nodules. (A) Small cell dysplasia (arrows), (B) large cell dysplasia (arrows), (C) clear cell change of the hepatocytes (arrows), (D) nuclear deviation toward the sinusoids, (E ) fatty change of the hepatocytes, and (F ) reduction of reticulin fiber (arrows) (A E, H&E; F, reticulin stain; original magnification: A E, 190 ; and F, 120 ). during arterioportography on the second admission during the follow-up period. The patients with these nodules were readmitted because of occurrence of HCC that had arisen in another part of the liver away from the initial nodule (11 patients), enlargement of the nodules (3 patients), and massive ascites caused by hepatic failure (1 patient). Five patients wanted further investigation using angiography after 1 year of follow-up. The remaining 12 nodules did not change (3 nodules) or resolved (9 nodules) on periodic liver imaging studies between 15 and 54 months after the initial biopsy. In the reevaluation of 22 nodules, tumor enhancement was shown on angiography in 1 nodule (nodule no. 2), and 10 nodules (nodules no. 1 5, 6 [1], 9, 15, 22, and 26 [1]) were detected again by ultrasonography and examined histologically by needle biopsy. Five nodules were diagnosed as well differentiated HCC by needle biopsy between 6 and 15 months after the first biopsy (nodules no. 1 5; Figures 2 and 3). All 5 patients with the nodules had hepatitis C virus infection. The remaining 29 nodules did not show malignant transformation during follow-up of months (median, 47 months). Four of them did not change in either size or imaging features during long-term follow-up of months, and the 25 nodules became undetectable on imaging. During the follow-up period, HCCs distinct

4 November 1998 FATE OF NONMALIGNANT HEPATOCELLULAR NODULES 1219 Figure 2. Histological findings of nodule 1. (A) Initial biopsy specimen. The nodule was hypoechoic and 13 mm in diameter. A mild increase in nuclear density, small cell dysplasia, and clear cell change of the hepatocytes can be seen (H&E). (B) Second biopsy performed 6 months after the first biopsy. The nodule remained unchanged in echogenicity and size. It was diagnosed as a well-differentiated HCC. Nuclear hyperchromasia, marked increase of nuclear density, and trabecular pattern can be seen (H&E) (original magnification: A and B, 190 ). from the studied nodules developed in the livers of 3 patients. These HCCs did not interfere with the evaluation of the nonmalignant lesions. In patients who had no coexistent HCC in the liver at first biopsy, elevation of serum AFP level ( 100 ng/ml) was observed in only 2 cases (nodules 3 and 4). In the case with nodule 3, serum AFP level increased from 80 ng/ml at first biopsy to 107 ng/ml at diagnosis for HCC by second biopsy. In the case with nodule 4, serum AFP level increased from 111 ng/ml at first biopsy to 148 ng/ml at diagnosis for HCC. However, these changes were not thought to be relevant for progression to HCC because the serum AFP levels had fluctuated in both cases during the follow-up period. Serum protein induced by vitamin K absence II level did not increase in cases without coexistent HCC. The follow-up data, particularly those detailing malignant transformation and histological features and univariate analyses of them, are shown in Tables 1 and 2. Significant prognostic factors of malignant transformation were an increased ratio of the nuclear density of 1.5, clear cell change, small cell dysplasia, and fatty change of the hepatocytes. Other histological features, including large cell dysplasia, were not significant. An increased ratio of nuclear density of 1.3 was also a significant factor (P 0.05), but its probability value was less significant than that of an increased ratio of nuclear density of 1.5. Fatty change of the hepatocytes was observed in all specimens from nodules that transformed to HCC. Interestingly, all 4 nodules that showed both small cell dysplasia and fatty change of the hepatocytes in the first biopsy progressed to HCC (nodules no. 1, 2, 4, and 5). On the other hand, none of 9 nodules with only hepatocellular fatty change without any other atypical features excluding large cell dysplasia progressed to HCC. Figure 3. Histological findings of nodule 3. (A) Initial biopsy specimen. The nodule is hyperechoic and 10 mm in diameter. It was diagnosed as a borderline nodule. A mild increase in nuclear density and clear cell and macrovesicular fatty changes of some hepatocytes can be seen (H&E). (B) Second biopsy specimen taken 14 months after the first biopsy. The hyperechoic nodule is enlarged to 17 mm in diameter. It was diagnosed as a well-differentiated HCC. Marked nuclear crowding, nuclear atypia, and microacinar formation can be seen (H&E) (original magnification: A and B, 190 ).

5 1220 TERASAKI ET AL. GASTROENTEROLOGY Vol. 115, No. 5 Table 1. Histological Parameters of Needle Biopsy Specimens Nodule Progression to HCC Follow-up (mo) Nuclear density (/mm 2 ) Ratio of nuclear density SCD LCD Microacinar formation Clear cell change Nuclear deviation Fatty change Alteration of stainability Reduction of reticulin fiber (1) (2) (1) (2) (3) (1) (2) (3) (1) (2) SCD, small cell dysplasia; LCD, large cell dysplasia. Multivariate analysis showed that only increased ratio of nuclear density of 1.5 (P 0.028) and clear cell change (P 0.044) were independent prognostic factors toward HCC. Relative risk of the factors was 9.18 (95% confidence interval, ) and 7.05 (95% confidence interval, ), respectively. Small cell dysplasia and fatty change were not significant factors in multivariate analysis. Moreover, addition of fatty change Table 2. Result of Univariate Analysis: Prognostic Histological Factor of Malignant Transformation Variable Ratio of nuclear density ( 1.5) Small cell dysplasia Large cell dysplasia Microacinar formation Clear cell change Nuclear deviation toward the sinusoids Fatty change Alteration of staining Reduction of reticulin fiber P value as explanatory variable disturbed the multivariate analysis. Discussion This is the first prospective report about the association between the analytical histological features of nonmalignant hepatocellular nodules by needle biopsy specimens and their outcome. Because it is difficult to objectively define atypia of liver nodules, analysis of individual histological features is thought to provide more valuable data. There have been four follow-up studies of nonmalignant hepatocellular nodules diagnosed by needle biopsy that did not include detailed examination of prognostic histological factors In our study, the percentage of nonmalignant nodules that evolved toward HCC was 14.7%. Although this rate appears to differ from those reported previously, it is important to pay particular attention to differences in histological criteria used to classify nonmalignant hepatocellular nodules when com-

6 November 1998 FATE OF NONMALIGNANT HEPATOCELLULAR NODULES 1221 paring results. Kondo et al. 9 reported that none of 17 large regenerative nodules without cytological atypia developed into HCC after more than 1 year of follow-up. Takayama et al. 12 reported that 9 of 18 (50%) AHs with hypercellularity, but without cellular and structural atypia, developed into HCC after 1 5 years. Rapaccini et al. 11 and Lencioni et al. 10 reported high rates of transformation of regenerative nodules into HCC: 83% and 68%, respectively. The mean diameters of the nodules were 33.7 mm (range, mm) in the former study and 19.4 mm (range, 8 52 mm) in the latter study. These nodules were larger than those encountered in our study, which averaged 10 mm (range, 5 20 mm). Moreover, almost all nonmalignant hepatocellular nodules studied using surgically resected specimens, 17,18 autopsy material, 4,21 and cirrhotic liver explants 6,22 were 15 mm in diameter, and larger lesions tended to show more malignant features. Therefore, these investigators may have been following up nodules that were nearly malignant or HCC. The intervals between the first biopsy and the diagnosis of HCC of nodules 1 and 2 were short: 6 and 7 months, respectively. There are two possible explanations for this finding. First, the nodules did not contain cancerous foci at the first biopsy but transformed to HCC in a short time. Second, the first liver biopsy may have been obtained from an area within the nodule other than the cancerous foci. This possibility suggests that the histological diagnosis of malignancy may be underestimated at the time of the initial biopsy. Misdiagnosis is both possible and inevitable when diagnoses are made by needle biopsy. However, the histological features observed in needle biopsy specimens suggest either an existence of HCC in a nodule at this time or an occurrence of HCC from a nodule later. In any event, it is important to note that the histological features of needle biopsy specimens obtained from both nodules suggested that the lesions had malignant potential and may progress to overt HCC. In the univariate analysis, an increased ratio of nuclear density of 1.5, small cell dysplasia, clear cell change, and fatty change of the hepatocytes were prognostic histological features. However, the multivariate analysis showed that only increased ratio of nuclear density of 1.5 and clear cell change were independent prognostic factors toward HCC. The ratio of nuclear density may be the most objective histological parameter among the examined ones. Increase in nuclear density is thought to be associated with high proliferative activity. Sustained hepatocellular proliferation plays an important role in the development of HCC. 23 The International Working Party 8 included the ratio of nuclear density as a part of the diagnostic criteria to distinguish hepatocellular nodules. They proposed 1.3 as cutoff value. In this study, an increased ratio of nuclear density of 1.3 was statistically significant in univariate analysis but was not an independent prognostic factor. On the other hand, our results showed that an increased ratio of nuclear density of 1.5 was more significant and also independent. Because only 2 of 5 nodules that progressed to HCC showed an increased ratio of nuclear density of 1.5, further examination is needed about appropriate cutoff value. This study showed that small cell dysplasia was a significant factor in univariate analysis but not independent in multivariate analysis. However, these statistical results do not exclude the importance of small cell dysplasia. Because small cell dysplasia has high proliferative activity 24,25 and overexpression of p53 protein, 26 it is thought to be associated with hepatocarcinogenesis. Another important characteristic of small cell dysplasia is its clustering nature. It is possible that this characteristic may indicate monoclonal growth and also that its hepatocytes may have genetic alterations. Fatty change and clear cell change of the hepatocytes are often observed in both atypical AH and welldifferentiated HCC. 17,18 In this study, hepatocellular fatty change was a significant risk factor for the development of HCC. Although all nodules that showed both small cell dysplasia and fatty change of the hepatocytes in the biopsy progressed to HCC, none of 9 nodules with only fatty change as an atypical feature other than large cell dysplasia evolved toward HCC. The facts suggest that there are two types of hepatocellular fatty change of ultrasonically detected lesion that may disturb the multivariate analysis. One is associated with neoplastic nodule, and the other is observed in nonneoplastic lesion. Focal fatty change of noncirrhotic livers is thought to be associated with ischemia. 27 Similar fatty deposition of hepatocytes may be present in nonneoplastic regenerative nodules. Therefore, hepatocellular fatty change observed in biopsy specimens from nodular lesions should be interpreted carefully. The nodules showing fatty change combined with other atypical features, especially small cell dysplasia, were at high risk to progress to HCC. In the present study, large cell dysplasia inside of nonmalignant nodule was not a risk factor for progression to HCC. The relationship between large cell dysplasia and HCC has been disputed. Large cell dysplasia is thought to be a preneoplastic lesion because of its ploidy pattern and nuclear shape. 25,28 However, it also appears to be related to hepatitis B virus infection and may be a reactive change of hepatocytes. 29 Hytiroglou et al. 22 have proposed that liver cell dysplasia of the large cell type should not be used as a criterion for atypical nodules because of the lack of association with HCC.

7 1222 TERASAKI ET AL. GASTROENTEROLOGY Vol. 115, No. 5 Follow-up nodules examined in this study included multiple nodules from the same patients. There is a possibility that examined nodules did not have a completely independent nature. However, it is thought that multiple HCCs often arise in a multifocal fashon, especially premalignant lesions and well-differentiated HCC. 3,30 In conclusion, the histological features of nonmalignant hepatocellular nodules within needle biopsy specimens that predicted malignant transformation were an increased ratio of the nuclear density, clear cell change of the hepatocytes, small cell dysplasia, and fatty change of the hepatocytes. Hepatocellular nodules that show these histological features should be recognized as being at high risk for the development to HCC. Further examination is needed in the evaluation of the significance of small cell dysplasia and fatty change in the borderline nodules. References 1. Arakawa M, Sugihara S, Kenmochi K, Kage M, Nakashima T, Nakayama T, Tashiro S, Hiraoka T, Suenaga M, Okuda K. Small mass lesions in cirrhosis: transition from benign adenomatous hyperplasia to hepatocellular carcinoma? J Gastroenterol Hepatol 1986;1: Nakanuma Y, Terada T, Terasaki S, Ueda K, Nonomura A, Kawahara E, Matsui O. Atypical adenomatous hyperplasia in liver cirrhosis: low-grade hepatocellular carcinoma or borderline lesion? Histopathology 1990;17: Tsuda H, Hirohashi S, Shimosato Y, Terada M, Hasegawa H. Clonal origin of atypical adenomatous hyperplasia of the liver and clonal identity with hepatocellular carcinoma. Gastroenterology 1988;95: Furuya K, Nakamura M, Yamamoto Y, Togei K, Otsuka H. Macroregenerative nodule of the liver. A clinicopathologic study of 345 autopsy cases of chronic liver disease. Cancer 1988;61: Terada T, Nakanuma Y. Iron-negative foci in siderotic macroregenerative nodules in human cirrhotic liver. A marker of incipient neoplastic lesions. Arch Pathol Lab Med 1989;113: Theise ND, Schwartz M, Miller C, Thung SN. Macroregenerative nodules and hepatocellular carcinoma in forty-four sequential adult liver explants with cirrhosis. Hepatology 1992;16: Ferrell L, Wright T, Lake J, Roberts J, Ascher N. Incidence and diagnostic features of macroregenerative nodules vs. small hepatocellular carcinoma in cirrhotic livers. Hepatology 1992;16: International Working Party. Terminology of nodular hepatocellular lesions. Hepatology 1995;22: Kondo F, Ebara M, Sugiura N, Wada K, Kita K, Hirooka N, Nagato Y, Kondo Y, Ohto M, Okuda K. Histological features and clinical course of large regenerative nodules: evaluation of their precancerous potentiality. Hepatology 1990;12: Lencioni R, Caramella D, Bartolozzi C, Di Coscio G. Long-term follow-up study of adenomatous hyperplasia in liver cirrhosis. Ital J Gastroenterol 1994;26: Rapaccini GL, Pompili M, Caturelli E, Anti M, Aliotta A, Cedrone A, Amadei E, Grattagliano A, Squillante MM, Rabitti C, Gambassi G. Focal ultrasound lesions in liver cirrhosis diagnosed as regenerating nodules by fine-needle biopsy. Follow-up of 12 cases. Dig Dis Sci 1990;35: Takayama T, Makuuchi M, Hirohashi S, Sakamoto M, Okazaki N, Takayasu K, Kosuge T, Motoo Y, Yamazaki S, Hasegawa H. Malignant transformation of adenomatous hyperplasia to hepatocellular carcinoma. Lancet 1990;336: Terada T, Ueda K, Nakanuma Y. Histopathological and morphometric analysis of atypical adenomatous hyperplasia of human cirrhotic livers. Virchows Arch A Pathol Anat Histopathol 1993;422: Ferrell LD, Crawford JM, Dhillon AP, Scheuer PJ, Nakanuma Y. Proposal for standardized criteria for the diagnosis of benign, borderline, and malignant hepatocellular lesions arising in chronic advanced liver disease. Am J Surg Pathol 1993;17: Kondo F, Kondo Y, Nagato Y, Tomizawa M, Wada K. Interstitial tumour cell invasion in small hepatocellular carcinoma. Evaluation in microscopic and low magnification views. J Gastroenterol Hepatol 1994;9: Kondo F, Wada K, Nagato Y, Nakajima T, Kondo Y, Hirooka N, Ebara M, Ohto M, Okuda K. Biopsy diagnosis of well-differentiated hepatocellular carcinoma based on new morphologic criteria. Hepatology 1989;9: Sakamoto M, Hirohashi S, Shimosato Y. Early stages of multistep hepatocarcinogenesis: adenomatous hyperplasia and early hepatocellular carcinoma. Hum Pathol 1991;22: Eguchi A, Nakashima O, Okudaira S, Sugihara S, Kojiro M. Adenomatous hyperplasia in the vicinity of small hepatocellular carcinoma. Hepatology 1992;15: Watanabe S, Okita K, Harada T, Kodama T, Numa Y, Takemoto T, Takahashi T. Morphologic studies of the liver cell dysplasia. Cancer 1983;51: Anthony PP, Vogel CL, Barker LF. Liver cell dysplasia: a premalignant condition. J Clin Pathol 1973;26: Terada T, Terasaki S, Nakanuma Y. A clinicopathologic study of adenomatous hyperplasia of the liver in 209 consecutive cirrhotic livers examined by autopsy. Cancer 1993;72: Hytiroglou P, Theise ND, Schwartz M, Mor E, Miller C, Thung SN. Macroregenerative nodules in a series of adult cirrhotic liver explants: issues of classification and nomenclature. Hepatology 1995;21: Huang SN, Chisari FV. Strong, sustained hepatocellular proliferation precedes hepatocarcinogenesis in hepatitis B surface antigen transgenic mice. Hepatology 1995;21: Adachi E, Hashimoto H, Tsuneyoshi M. Proliferating cell nuclear antigen in hepatocellular carcinoma and small cell liver dysplasia. Cancer 1993;72: Zhao M, Zhang NX, Du ZY, Laissue JA, Zimmermann A. Three types of liver cell dysplasia (LCD) in small cirrhotic nodules are distinguishable by karyometry and PCNA labelling, and their features resemble distinct grades of hepatocellular carcinoma. Histol Histopathol 1994;9: Zhao M, Zhang NX, Laissue JA, Zimmermann A. Immunohistochemical analysis of p53 protein overexpression in liver cell dysplasia and in hepatocellular carcinoma. Virchows Arch 1994; 424: Brawer MK, Austin GE, Lewin KJ. Focal fatty change of the liver, a hitherto poorly recognized entity. Gastroenterology 1980;78: Thomas RM, Berman JJ, Yetter RA, Moore GW, Hutchins GM. Liver cell dysplasia: a DNA aneuploid lesion with distinct morphologic features. Hum Pathol 1992;23: Akagi G, Furuya K, Kanamura A, Chihara T, Otsuka H. Liver cell dysplasia and hepatitis B surface antigen in liver cirrhosis and hepatocellular carcinoma. Cancer 1984;54: Sheu JC, Huang GT, Chou HC, Lee PH, Wang JT, Lee HS, Chen DS. Multiple hepatocellular carcinomas at the early stage have different clonality. Gastroenterology 1993;105: Received July 17, Accepted July 16, Address requests for reprints to: Shuichi Terasaki, M.D., First Department of Internal Medicine, Kanazawa University School of Medicine, 13-1 Takaramachi, Kanazawa 920, Japan. Fax: (81)