Due to the phenotypic and molecular diversity of hepatocellular carcinomas (HCC), it is a challenge to determine a patient’s prognosis. We aimed to identify new prognostic markers of patients with HCC treated by liver resection.

We collected 314 HCC samples from patients at Bordeaux (1998−2007) and Créteil (2003−2007) hospitals in France. We analyzed the gene expression patterns of the tumors and compared expression patterns with patient survival times. Using the coefficient and regression formula of the multivariate Cox model, we identified a “5-gene score” associated with survival times. This molecular score was then validated in 2 groups of patients from Europe and the United States (n = 213) and China (n = 221).

The 5-gene score, based on combined expression level of HN1, RAN, RAMP3, KRT19, and TAF9, was associated with disease-specific survival times of 189 patients with resected HCC in Bordeaux (hazard ratio = 3.5; 95% confidence interval: 1.9−6.6; P < .0001). The association between the 5-gene score and disease-specific survival was validated in an independent cohort of 125 patients in Créteil (hazard ratio = 2.3; 95% confidence interval: 1.1−4.9; P < .0001). The 5-gene score more accurately predicted patient outcomes than gene expression signatures reported previously. In multivariate analyses, the 5-gene score was associated with disease-specific survival, independent of other clinical and pathology feature of HCC. Disease-specific survival was also predicted by combining data on microvascular invasion, the Barcelona Clinic Liver Cancer classification system, and the 5-gene score in a nomogram. The prognostic accuracy of the 5-gene score was further validated in European and US patients with hepatitis C, cirrhosis, and HCC (overall survival P = .002) and in Asian patients with HCC with hepatitis B (overall survival, P = .02). Combining the 5-gene score with the expression pattern of 186 genes in corresponding cirrhotic tissues increased its prognostic accuracy.

The molecular 5-gene score is associated with outcomes of patients with HCC treated by resection in different clinical settings worldwide. This new biomarker should be tested in clinical trials to stratify patients in therapeutic decisions.