AI and Cancer Research ── Bladder ── 2026-10-26
Bladder cancer: treatment, trials and AI
- Abstract
Bladder cancer (BCa) displays marked molecular heterogeneity and carries substantial risks of recurrence, progression and treatment resistance, highlighting the need for new biomarkers defined at the level of cellular states. Here, we integrated single-cell RNA sequencing (scRNA-seq), spatial transcriptomics (ST), bulk RNA sequencing (bulk RNA-seq), machine-learning (ML) models and immunotherapy cohorts to characterize highly stem-like malignant epithelial cells in BCa. Based on scRNA-seq data, malignant epithelial cells were identified using CopyKAT and then divided into LowStem, IMStem and HighStem groups according to within-patient CytoTRACE scores. Cell–cell communication and pathway analyses showed extensive inferred interactions between HighStem malignant epithelial cells and immune, fibroblast and endothelial cells. HighStem cells also showed higher scores for DNA repair, oxidative phosphorylation, MYC and E2F pathways, but lower scores for transforming growth factor-β (TGF-β), Notch and WNT/β-catenin pathways. By integrating high-dimensional weighted gene co-expression network analysis (hdWGCNA), differential-expression analysis, survival analysis, ST validation and ML modelling, we identified TXNL4A, APEX1, TIMM8B, MDH2, TRMT112, C1QBP, CYC1 and BANF1 as core risk genes. The Gaussian naïve Bayes (NB) model showed favorable overall classification performance in both cohorts. Shapley additive explanations (SHAP) indicated that all eight genes contributed to the identification of HighStem cells. Immunotherapy-cohort analyses suggested that TXNL4A, BANF1 and MDH2 may be associated with immunotherapy response. Virtual knockout analysis suggested that the BANF1-associated network involves protein synthesis, mitochondrial function and cell adhesion. These findings provide a potential molecular basis for understanding stemness heterogeneity in BCa and may inform the development of individualized treatment strategies.
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: Liu X, Xiao Z, Wang Y, Huang L. Integrative single-cell and spatial transcriptomics combined with machine learning reveals molecular features and key risk genes of high-stemness malignant epithelial cells in bladder cancer. Discov Onc. 2026 Oct 23 [Epub ahead of print]. doi:10.1007/s12672-026-06196-2.Checked: Abstract only