Machine Learning and Mendelian Randomization Reveal a Tumor Immune Cell Profile for Predicting Bladder Cancer Risk and Immunotherapy Outcomes.

This study's objective was to develop predictive models for bladder cancer (BLCA) using tumor infiltrated immune cell (TIIC)-related genes. Multiple RNA expression data and scRNA-seq were downloaded from the TCGA and GEO databases. A tissue specificity index was calculated and a computational framework developed to identify TIIC signature scores based on three algorithms. Univariate Cox analysis was performed, and the TIIC-related model was generated by 20 machine learning algorithms. A significant correlation between TIIC signature score and survival status, tumor stage, and TNM staging system was found. Patients in the high-score BLCA group had more favorable survival outcomes and enhanced response to PD-L1 immunotherapy as compared to those in the low-score group. This TIIC model showed better performance in prognosing BLCA. Diverse frequencies of mutations were observed in human chromosomes across groups categorized by TIIC score. No statistically significant correlation was observed between noncancerous bladder conditions and BLCA when examining the single nucleotide polymorphisms (SNPs) associated with the genes in the prognostic model. However, a statistically significant association was found at the SNP sites of rs3763840. There was no significant association between bladder stones and BLCA, but there was a significant association on the SNP sites of rs3763840. A novel TIIC signature score was constructed for the prognosis and immunotherapy for BLCA, which offers direction for predicting overall survival of patients with BLCA.