An apoptosis-related gene signature for the prognosis of hepatocellular carcinoma

Abstract Hepatocellular carcinoma (HCC) is a major public health burden worldwide owning to high incidence and poor prognosis. Although a mushrooming number of apoptosis-related genes had been disclosed in HCC, the prognostic value and clinical utility of them remain to be illustrated. Here, we defined the data from Gene Expression Omnibus (GEO) as a training cohort and data from The Cancer Genome Atlas-Liver Hepatocellular Carcinoma data set (TCGA-LIHC) as a validation cohort. The apoptosis-related differentially expressed genes (AR-DEGs) were identified with the two cohorts and the Gene Set Enrichment Analysis. Then, we constructed a Lasso-penalized Cox regression model using AR-DEGs and conducted a signature including 14 apoptotic genes to calculate the risk score. Patients with a high risk score indicated worse overall survival than those with low risk. Besides, the 3-year and 5-year area under curve (AUC) values of the signature were above 0.7 in both training and validation cohorts (0.762, 0.818, 0.717, 0.745, respectively). Moreover, a nomogram containing the signature and clinical characteristics presented reliable net benefits for the survival prediction. And the nomogram was tested by probability calibration curves and Decision Curve Analysis (DCA). Furthermore, protein-protein interaction (PPI) and Gene Ontology (GO) enrichment analysis disclosed several noticeable pathways that might clarify the hidden mechanism. Collectively, the present study formed a novel signature based on the 14 apoptotic genes and this possibly predicted prognosis and strengthened the communication with HCC patients about the likely treatment.