Identification and Validation of Prognostic Model for Tumor Microenvironment-Associated Genes in Bladder Cancer Based on Single-Cell RNA Sequencing Data Sets.

Abstract

The purpose of this study was to elucidate the relationship between the tumor microenvironment (TME) and cellular diversity in bladder cancer (BLCA) progression, leveraging single-cell RNA sequencing (scRNA-seq) data to identify potential prognostic biomarkers and construct a prognostic model for BLCA. We analyzed scRNA-seq data of normal and tumor bladder cells from the Gene Expression Omnibus (GEO) database to uncover crucial markers within the bladder TME. The study compared gene expression in normal versus tumor bladder cells, identifying differentially expressed genes. These genes were subsequently assessed for their prognostic significance using patient follow-up data from The Cancer Genome Atlas. Prognostic models were constructed using Least Absolute Shrinkage and Selection Operator and multivariate Cox regression analyses, focusing on eight genes of interest. The predictive performance of the model was also tested against additional GEO data sets (GSE31684, GSE13507, and GSE32894). The prognostic model demonstrated reliable prediction of patient outcomes. Validation through gene set enrichment analysis and immune cell infiltration assessment supported the model's efficacy. The results from both the univariate and multivariate analyses suggest that the risk score is an independent prognostic factor with a hazard ratio of 2.97 (95% CI, 2.28 to 3.9, <i>P</i> < .001). In the validation cohort, the AUC at 1, 2, and 3 years is 0.74, 0.74, and 0.72, respectively. Our findings proposed biomarkers with prognostic potential, laying the groundwork for future in vitro validation and therapeutic exploration. This contributes to a deeper understanding of the genes associated with bladder TME and may improve prognostic precision in BLCA management.