Genomic and Transcriptomic Profiling of High-risk Bladder Cancer Reveals Diverse Molecular and Microenvironment Ecosystems

Abstract

Background and objective

High-risk bladder cancer recurs in 30% of cases and causes fatal outcomes in 10% within 2 yr despite surgical resection, endoscopic surveillance, and bacillus Calmette-Guérin (BCG) immunotherapy. The global shortage of BCG highlights the urgent need for alternative or complementary strategies. This study aimed to identify molecular subtypes and develop a precision framework to predict recurrence risk and guide treatment.

Methods

Transcriptomic profiling and targeted genomic sequencing were performed with validation by single-cell RNA sequencing and spatial transcriptomics. A machine learning model incorporating genomic and transcriptomic features was developed to predict recurrence risk.

Key findings and limitations

Four subtypes were identified, and an inflamed tumor subtype with high endogenous retroelement expression and increased commensal bacterial presence demonstrated the highest responsiveness to BCG therapy. The predictive model achieved high accuracy (area under the curve = 0.87, 95% confidence interval: 0.72–1.0) for recurrence risk. Findings are limited by sample size, necessitating validation in larger cohorts.

Conclusions and clinical implications

We describe a new conceptual framework for T1 tumors. Those with increased immune cells (subtype 2) have a better response to BCG, which may be secondary to enhanced baseline immune activity.