AI and Cancer Research ── Bladder ── 2026-11-05
Bladder cancer: treatment, trials and AI
- Abstract
Bladder cancer (BC) presents a significant clinical challenge, with a poor prognosis in its advanced stages. Although immunotherapy has improved outcomes, patient response rates remain limited due to complex immune evasion mechanisms. Robust biomarkers are urgently needed for prognostic stratification and to help identify patients who might benefit from immunotherapy. We developed an immune evasion-related signature (IRS) using transcriptomic data from The Cancer Genome Atlas (TCGA) and validated it across four independent Gene Expression Omnibus (GEO) cohorts. An integrative machine learning framework based on ten algorithms, fitted as 60 candidate model configurations, was used to construct the IRS from differentially expressed immune-related genes. The signature’s prognostic value, association with the tumor immune microenvironment (TME), and predictive power for immunotherapy response were assessed, and the key gene DOT1L was functionally validated in vitro. The IRS, comprising five genes (DOT1L, DVAC2, SRRT, PPP2R3C, and IRF9), effectively stratified BC patients into high-risk and low-risk groups. A high IRS score was an independent prognostic factor for poorer overall survival and was associated with an immunosuppressive TME, characterized by reduced CD8 + T cell infiltration and higher Tumor Immune Dysfunction and Exclusion (TIDE) scores. Conversely, low IRS scores correlated with enhanced response to anti-PD-1/L1 therapy in validation cohorts. Furthermore, DOT1L knockdown suppressed BC cell proliferation and migration. We developed and retrospectively validated a machine learning-derived IRS that showed consistent associations with prognosis and immunotherapy benefit across multiple independent bladder cancer cohorts. The IRS is a promising research biomarker, and prospective, multicenter validation is required before it can guide personalized treatment decisions.
Journal IF-equivalent: 2.4 (OpenAlex 2-year mean citedness, value as of 2026-10-12, retrieved 2026-10-13; not the official Clarivate IF)Reference: Huang B, Quan J, Xue W, Ma X. Development of a machine learning-based immune evasion signature for prognostic stratification and its association with immunotherapy response in bladder cancer. Discov Onc. 2026 Nov 2 [Epub ahead of print]. doi:10.1007/s12672-026-06185-5.Checked: Abstract only