Development and validation of a new tumor-based gene signature predicting prognosis of HBV/HCV-included resected hepatocellular carcinoma patients

Due to the phenotypic and molecular diversity of hepatocellular carcinomas (HCC), it is still a challenge to determine patients’ prognosis. We aim to identify new prognostic markers for resected HCC patients. 274 patients were retrospectively identified and samples collected from Zhongshan hospital, Fudan University. We analyzed the gene expression patterns of tumors and compared expression patterns with patient survival times. We identified a “9-gene signature” associated with survival by using the coefficient and regression formula of multivariate Cox model. This molecular signature was then validated in three patients cohorts from internal cohort (n = 69), TCGA (n = 369) and GEO dataset (n = 80). 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 cohort (concordance index, ci = 0.85), which was successfully validated (ci = 0.86 for internal cohort; ci = 0.78 for in silico cohort). The signature showed improved performance compared with clinical parameters alone (ci = 0.70). Furthermore, the signature predicted patient prognosis than previous gene signatures more accurately. It was also used to stratify early-stage, HBV or HCV-infected patients into low and high-risk groups, leading to significant differences in survival in training and validation (P < 0.001). The 9-gene signature, in which four were upregulated (ZC2HC1A, MARCKSL1, PTGS1, CDKN2B) and five (CLEC10A, PRDX3, PRKCH, MPEG1, LMO2) were downregulated in HCC with poor prognosis, stratified HCC patients into low and high risk group significantly in different clinical settings, including receiving adjuvant transarterial chemoembolization and especially in early stage disease. This new signature should be validated in prospective studies to stratify patients in clinical decisions.