Identification and validation of a six immune-related gene signature for prediction of biochemical recurrence in localized prostate cancer following radical prostatectomy.

BACKGROUND

Prostate cancer (PCa) is the second lethal heterogeneous cancer among males worldwide, and approximately 20% of PCa patients following radical prostatectomy (RP) will undergo biochemical recurrence (BCR). This study is aimed to identify the immune-related gene signature that can predict BCR in localized PCa following RP.

METHODS

Expression profile of genes together with clinical parameters from The Cancer Genome Atlas (TCGA) and the Gene Expression Omnibus database (GEO) and the immune-related genes from the Molecular Signatures Database v4.0 were applied to construct and validate the gene signature. The Cox regression analyses were conducted to identify the candidate genes and establish the gene signature. To estimate the prognostic power of the risk score, the time-dependent receiver operating characteristic (ROC) analysis and Harrell's index of concordance (C-index) were utilized. We also established a nomogram to forecast the probability of patients' survival.

RESULTS

A total of 268 patients from the TCGA and 77 patients from GSE70770 and six immune-related genes () were eventually selected. The Kaplan-Meier analysis demonstrated that patients in the low-risk group had a significantly longer recurrence-free survival (RFS) compared to those in the high-risk group. In the multivariate Cox model, the signature was identified as an independent prognostic factor, which was significantly associated with RFS (TCGA: HR =5.232, 95% CI: 1.762-15.538, P=0.003; GSE70770: HR =2.158, 95% CI: 1.051-4.432, P=0.036). Moreover, the C-index got improved after incorporating the risk score into original clinicopathological parameters. In addition, the novel nomogram was constructed to better predict the 1-, 3- and 5-year RFS.

CONCLUSIONS

This signature could serve as an independent prognostic factor for BCR. Incorporation of our signature into traditional risk classification might further stratify patients with different prognosis, which could assist practitioners in developing clinical decision-making.