A novel biomarker associated with EBV infection improves response prediction of immunotherapy in gastric cancer

Novel biomarkers are required in gastric cancer (GC) treated by immunotherapy. Epstein-Barr virus (EBV) infection induces an immune-active tumor microenvironment, while its association with immunotherapy response is still controversial. Genes underlying EBV infection may determine the response heterogeneity of EBV + GC. Thus, we screened hub genes associated with EBV infection to predict the response to immunotherapy in GC. Prognostic hub genes associated with EBV infection were screened using multi-omic data of GC. EBV + GC cells were established and confirmed by EBV-encoded small RNA in situ hybridization (EBER-ISH). Immunohistochemistry (IHC) staining of the hub genes was conducted in GC samples with EBER-ISH assay. Infiltrating immune cells were stained using immunofluorescence. CHAF1A was identified as a hub gene in EBV + GC, and its expression was an independent predictor of overall survival (OS). EBV infection up-regulated CHAF1A expression which also predicted EBV infection well. CHAF1A expression also predicted microsatellite instability (MSI) and a high tumor mutation burden (TMB). The combined score (CS) of CHAF1A expression with MSI or TMB further improved prognostic stratification. CHAF1A IHC score positively correlated with the infiltration of NK cells and macrophages M1. CHAF1A expression alone could predict the immunotherapy response, but its CS with EBV infection, MSI, TMB, or PD-L1 expression showed better effects and improved response stratification based on current biomarkers. CHAF1A could be a novel biomarker for immunotherapy of GC, with the potential to improve the efficacy of existing biomarkers.