Development of a biochemical recurrence prediction model based on pyroptosis related genes in prostate cancer

Abstract

Abstract Background: Pyroptosis plays important roles in the development and progression of cancer. However, the role of pyroptosis-related genes (PRGs) in biochemical recurrence (BCR) of prostate cancer (PCa) remain unclear.Methods: Expression data and clinical information of PCa patients used in the current study were downloaded from the Cancer Genome Atlas (TCGA) and the Gene Expression Omnibus (GEO) database. Differentially expressed PRGs (DEPRGs) were identified between PCa and control samples and between BCR and BCR-free samples. Univariate and LASSO Cox regressions were performed to identify BCR-related DEPRGs in PCa, followed by the construction of the risk score model. Receiver operating characteristic (ROC) curves were plotted to assess the performance of the risk score model. Univariate and multivariate Cox regressions were carried out to determine independent BCR factors and to establish the nomogram in predicting BCR of PCa patients. The microenvironment of low- and high-risk groups were evaluated by GSVA and ssGSEA.Results: By overlapping 29 DEPRGs between PCa and control samples and 10 DEPRGs between BCR and BCR-free samples, TP63, CHMP4C, CHMP7, GSDMB, CASP8, PLCG1 and TP53 were obtained. By univariate and LASSO Cox regressions, CHMP4C, GSDMB, PLCG1 and TP53 were identified as BCR biomarkers in PCa. ROC curves revealed the good performance of the risk score model based on BCR biomarkers in both TCGA and GEO cohorts. Univariate and multivariate Cox regressions showed that the risk score was an independent BCR factor in PCa. A nomogram with good performance to predict the BCR of PCa patients were established based on risk score and other independent prognostic factors. After GSVA and ssGSEA, we observed that the immune and metabolic microenvironment of two groups were much different.Conclusion: Our study revealed the role of PRGs in the BCR of PCa, and constructed reliable models in predicting the BCR of PCa patients.