Identification Of An 11-Autophagy-Related-Gene Signature As Promising Prognostic Biomarker For Bladder Cancer Patients

Simple SummaryHuman bladder cancer, one of the most common cancers worldwide, is a molecularly heterogenous and complex disease. Identifying novel prognostic biomarkers and establishing new predictive signatures are important for personalized medicine and effective treatment of bladder cancer patients. Autophagy, a cell self-maintenance process that removes damaged organelles and misfolded proteins, displays both tumor promotion and suppression activities. The aim of our study is to investigate the function of autophagy-related genes in bladder cancer with the main focus on their contribution to prognostic outcome. By analyzing data obtained from The Cancer Genome Atlas (TCGA), we identified 32 autophagy-related genes that were highly associated with overall survival of bladder cancer patients. Further statistical assessment established an 11-autophagy-related-gene signature as an effective prognostic biomarker to predict the survival outcomes of bladder cancer patients.Background: Survival rates for highly invasive bladder cancer (BC) patients have been very low, with a 5-year survival rate of 6%. Accurate prediction of tumor progression and survival is important for diagnosis and therapeutic decisions for BC patients. Our study aims to develop an autophagy-related-gene (ARG) signature that helps to predict the survival of BC patients. Methods: RNA-seq data of 403 BC patients were retrieved from The Cancer Genome Atlas Urothelial Bladder Carcinoma (TCGA-BLCA) database. Univariate Cox regression analysis was performed to identify overall survival (OS)-related ARGs. The Lasso Cox regression model was applied to establish an ARG signature in the TCGA training cohort (N = 203). The performance of the 11-gene ARG signature was further evaluated in a training cohort and an independent validation cohort (N = 200) using Kaplan-Meier OS curve analysis, receiver operating characteristic (ROC) analysis, as well as univariate and multivariate Cox regression analysis. Results: Our study identified an 11-gene ARG signature that is significantly associated with OS, including APOL1, ATG4B, BAG1, CASP3, DRAM1, ITGA3, KLHL24, P4HB, PRKCD, ULK2, and WDR45. The ARGs-derived high-risk bladder cancer patients exhibited significantly poor OS in both training and validation cohorts. The prognostic model showed good predictive efficacy, with the area under the ROC curve (AUCs) for 1-year, 3-year, and 5-year overall survival of 0.702 (0.695), 0.744 (0.640), and 0.794 (0.658) in the training and validation cohorts, respectively. A prognostic nomogram, which included the ARGs-derived risk factor, age and stage for eventual clinical translation, was established. Conclusion: We identified a novel ARG signature for risk-stratification and robust prediction of overall survival for BC patients.