The aim of this study was to identify high-risk patients Hepatocellular carcinoma (HCC) is a frequent comfor hepatocellular carcinoma (HCC). Among 151 patients plication of cirrhosis, and its prevalence increases as with histologically proven cirrhosis hospitalized from treatment of other complications improves. Symptom-1987 to 1990 and prospectively followed-up until June atic HCC has a poor prognosis, and only early diagnosis 1994, 31 developed HCC. We assessed the predictive of isolated small tumors without vascular involvement value of 22 variables recorded at enrollment for HCC or metastases allows a chance for cure.1 Screening occurrence by the log rank test and the Cox proportional based on repeated liver ultrasonographic investigahazards model. Six clinical and biological variables sumtions and serum a-fetoprotein (AFP) determinations marized predictive information of HCC: age ¢50 years every 6 months was performed in patients with cirrho-(P Å .01), male (P Å .01), large esophageal varices (EV) (P Å .03), prothrombin activity õ70% (P Å .04), serum a-sis for the early detection of HCC. In Asia, and particufetoprotein (AFP) ¢15 ng/L (P Å .06), and anti-hepatitis C larly in Japan, the number of detected HCC is high, virus antibodies (P Å .08). A clinicobiological predictive with an annual incidence between 3% and 6%. More score identified two groups of patients at low (n Å 67; 3-than 50% of the detected tumors are õ 3 cm in diameyear cumulative incidence, 0%) and high risk for HCC (n ter 2-4 and most of them are resectable. 4 In Europe, the Å 84; 3-year cumulative incidence, 24%). The predictive annual incidence of HCC is as high as in Asia. Howvalue of this score was confirmed using an independent ever, only a minority of HCC are detected at an ''early'' population of 49 patients with cirrhosis. Furthermore, stage; the percentage of small isolated tumors is liver large-cell dysplasia (LCD) had an additional predictive value in high-risk patients (P Å 10 04 ), which thus approximately 30%, and very few of them are resecthelped to define a subgroup at very high risk for HCC able. [5][6][7][8] This discrepancy could be related to differences (n Å 12; 3-year cumulative incidence, 72%). In Western in the underlying liver disease and the growth pattern patients with cirrhosis, a limited number of usual vari-of HCC between Asian and Western patients. Whatables can identify a group of patients at high risk for ever the cause, different European studies have con-HCC. Among these patients, liver biopsy allows for the cluded that screening for HCC every 6 months should determination a subgroup of patients at very high risk not be recommended on a large scale in patients with for HCC requiring intensive screening or preventive cirrhosis.5-8 Nevertheless, the identification of a high-