A Novel mRNA Signature Related to Immunity to Predict Survival and Immunotherapy Response in Hepatocellular Carcinoma

BACKGROUND AND AIMS: Hepatocellular carcinoma (HCC) is the most common primary liver cancer and the incidence and mortality rates are increasing. Given the limited treatments of HCC and promising application of immunotherapy for cancer, we aimed to identify an immune-related prognostic signature that can predict overall survival (OS) rates and immunotherapy response in HCC. METHODS: The initial signature development was conducted using a training dataset from the Cancer Genome Atlas followed by independent internal and external validations from that resource and the Gene Expression Omnibus. A signature based nomogram was generated using multivariate Cox regression analysis. The associations of signature score with tumor immune phenotype and response to immunotherapy were analyzed using single-sample gene set enrichment analysis and tumor immune dysfunction and exclusion algorithm. A cohort from Zhongshan Hospital was employed to verify the predictive robustness of the signature regarding prognostic risk and immunotherapy response. RESULTS: The prognostic signature, IGS(HCC), consisting of 22 immune-related genes, had independent prognostic ability, with training and validation cohorts. Also, IGS(HCC) stratified HCC patients with different outcomes in subgroups. The prognostic accuracy of IGS(HCC) was better than three reported prognostic signatures. The IGS(HCC)-based nomogram had high accuracy and significant clinical benefits in predicting 3- and 5-year OS. IGS(HCC) reflected distinct immunosuppressive phenotypes in low- and high-score groups. Patients with low IGS(HCC) scores were more likely than those with high scores to benefit from immunotherapy. CONCLUSIONS: IGS(HCC) predicted HCC prognosis and response to immunotherapy, and contributed to individualized clinical management.