Development and Validation of a New Tumor-Based Gene Signature Predicting Prognosis of HBV-Included Resected Hepatocellular Carcinoma Patients

Background: Due to the phenotypic and molecular diversity of hepatocellular carcinomas (HCC), it is a challenge to determine patients' prognosis. We aimed to identify new biomarkers of resected patients. Methods: We collected 274 HCC samples from patients at Zhongshan hospital, Fudan university. Gene expression analyses were performed using NanoString technology and compared with patient survival times. Using the coefficient and regression formula of the multivariate Cox model, we identified a "9-gene signature" associated with survival times. This molecular signature was then validated in three groups of patients from internal cohort (n=69), TCGA dataset (n=369) and HCC genomic consortium (n=80). Results: We identified 9-gene signature consisting of ZC2HC1A, MARCKSL1, PTGS1, CDKN2B, CLEC10A, PRDX3, PRKCH, MPEG1 and LMO2. The 9-gene signature was used, combined with clinical parameters, to fit a multivariable Cox model to the training data (concordance index, ci=0.85), which was successfully validated (ci=0.86 for internal cohort; ci=0.78 for external cohort). The signature showed improved performance compared with clinical parameters alone (ci=0.70). Furthermore, the signature more accurately predicted patient outcomes than gene signatures reported previously. It was also used to stratify early-stage, HBV or HCV-infected patients into groups with low and high risk of survival, leading to significant differences in survival in training and validation (P