Development of a novel five-gene immune-related risk model for the prognosis evaluation of prostate adenocarcinoma patients

The interaction between the immune cells and the host immune system with the tumor cells is significantly associated with the initiation and progression of prostate adenocarcinoma (PRAD), whereas the application of immune-related genes (IRGs) for the prognosis evaluation of PRAD patients is still lacking. In this study, we aimed to identify IRGs with prognostic values and to develop a clinically effective risk model. Wilcoxon rank-sum test and univariate Cox analysis were applied to identify the differentially expressed immune-related genes (DEIRGs) related to the survival of PRAD patients. The Least absolute shrinkage and selection operator (LASSO) analysis was performed to identify the independent prognostic DEIRGs and to establish an immune risk score prognostic model. The reliability and veracity of the prognostic model were validated in PRAD patients from the internal cohort (The Cancer Genome Atlas, TCGA dataset) and the external cohort (International Cancer Genome Consortium, ICGC dataset), respectively. Six of the 193 identified DEIRGs were survival-associated in PRAD patients. Five prognostic DEIRGs (SLPI, NOX1, DES, BIRC5 and AMH) were selected to construct the immune-related prognostic model with optimal robustness. In the 2 independent cohorts we chose, PRAD patients could be effectively stratified according to our risk model. Patients with high risk scores had worse survival. Clinical correlation analysis proved that the risk score was associated with advanced clinicopathologic features. Multivariate analysis indicated that the risk model was an independent prognostic indicator. We also established a nomogram based on the risk score model for clinical application. Additionally, the risk score model was correlated with immune cell infiltration and reflected the status of the immune microenvironment. The prognostic value of the five immune-related genes used in the prognostic model was also validated. Our immune-related prognostic model was an effective tool that could not only serve as a predictor for prognosis, but also provide potential prognostic and therapeutic molecular biomarkers for optimizing personalized therapies in clinical practice.