Limited progress has been made in the treatment of gastric adenocarcinoma (GAC) in recent years, but the potential of immunotherapy in GAC is worthy of consideration. The purpose of this study was to develop a reliable, personalized signature based on immune genes to predict the prognosis of GAC. Here, we identified two groups of patients with significantly different prognoses by performing unsupervised clustering analysis of The Cancer Genome Atlas (TCGA) database based on 881 immune genes. The immune signature was constructed with a training set composed of 350 GAC samples from the TCGA and subsequently validated with 431 samples from GSE84437, 432 samples from GSE26253, and 145 GAC samples from real-time quantitative reverse transcription polymerase chain reaction data. This classification system can also be used to predict prognosis in different clinical subgroups. Further analysis suggested that high-risk patients were characterized by low immune scores, distinctive immune cell proportions, different immune checkpoint profiles, and a low tumor mutational burden. Ultimately, the signature was identified as an independent prognostic factor. In general, the signature can accurately predict recurrence and overall survival in patients with GAC and may serve as a powerful prognostic tool to further optimize cancer immunotherapy.
Keywords: gastric adenocarcinoma; immune signature; prognosis; qRT-PCR; tumor microenvironment.