Basal/squamous and mixed subtype bladder cancers present poor outcomes after neoadjuvant chemotherapy in the VESPER trial

Background: Neoadjuvant chemotherapy (NAC) is the standard treatment for muscle-invasive bladder cancer (MIBC), yet 40% of patients progress, emphasizing the need for biomarkers predictive for response or chemoresistance. Gene expression-based subtypes may serve as biomarkers, though which subtypes will respond, notably when it comes to the basal subtype, remains contentious. Patients and methods: This post hoc study analyzed 300 NAC-treated patients enrolled in the GETUG/AFU VESPER trial, with transurethral diagnostic formalin-fixed paraffin-embedded which underwent pathological review before being sequenced. ‘Mixed’ subtype was defined for tumors displaying at least two different Consensus molecular subtypes in separate regions. We evaluated the association between molecular subtypes and outcome after NAC. Tumors with remaining tissue at cystectomy (n ¼ 83) were compared with pre-treatment tumors. Results: Cases were classified basal/squamous (Ba/Sq) (n ¼ 84), luminal unstable (n = 57), stroma-rich (n = 53), mixed (n = 48), luminal papillary (n ¼ 39), luminal non-specific (n = 18), and neuroendocrine-like (n ¼ 1), with 30/48 mixed cases including a Ba/Sq component. Compared with other molecular subtypes in a multivariate Cox model, Ba/Sq (pure or mixed) patients had an increased hazard ratio (HR) of progression-free survival [HR 2.0, 95% confidence interval (CI) 1.36-3.0]. Mixed tumors were associated with decreased metabolic activity that could account for chemoresistance. Ba/Sq and mixed non-responders mostly maintained their subtype at cystectomy and have fewer myeloid dendritic cells after NAC. Tumors classified luminal papillary at transurethral resection of the urinary bladder tumor exhibited an increase in T CD4þ and macrophage signatures after NAC. Other subtypes did not show significant immune changes after NAC. Our study design relied on detailed pathological review, which precluded evaluating the mixed subtype in published datasets. Furthermore, the sample size for post-NAC analyses constrained the statistical power of these findings. Conclusions: Our findings underscore the importance of recognizing intra-tumor heterogeneity in MIBC and its role in chemoresistance associated with Ba/Sq subtype, and provide valuable insights that could help future treatment development and improve patient outcomes.