The aim of this work was to develop an individualized score for predicting hepatocellular carcinoma (HCC) in patients with hepatitis C (HCV)-compensated cirrhosis. Among 1,323 patients with HCV cirrhosis enrolled in the French prospective ANRS CO12 CirVir cohort, 720 and 360 were randomly assigned to training and validation sets, respectively. Cox's multivariate model was used to predict HCC, after which a nomogram was computed to assess individualized risk. During follow-up (median, 51.0 months), 103 and 39 patients developed HCC in the training and validation sets, respectively. Five variables were independently associated with occurrence of HCC: age > 50 years (hazard ratio [HR], 1.94; 95% confidence interval [CI], 1.16; 3.25; P = 0.012); past excessive alcohol intake (HR, 1.55; 95% CI, 1.02; 2.36; P = 0.041); low platelet count (<100 Giga/mm(3) : HR, 2.70; 95% CI, 1.62; 4.51; P < 0.001; [100; 150] Giga/mm(3) : HR, 1.87; 95% CI, 1.10; 3.18; P = 0.021); gamma-glutamyl transpeptidase above the upper limit of normal (HR, 1.96; 95% CI, 1.11; 3.47; P = 0.021); and absence of a sustained virological response during follow-up (HR, 3.02; 95% CI, 1.67; 5.48; P < 0.001). An 11-point risk score was derived from the training cohort and validated in the validation set. Based on this score, the population was stratified into three groups, in which HCC development gradually increased, from 0% to 30.1% at 5 years for patients with the lowest (≤3) and highest (≥8) scores (P < 0.001). Using this score, a nomogram was built enabling individualized prediction of HCC occurrence at 1, 3, and 5 years.
Conclusion: This HCC score can accurately predict HCC at an individual level in French patients with HCV cirrhosis. (Hepatology 2016;64:1136-1147).
© 2016 by the American Association for the Study of Liver Diseases.