Impact of Molecular Androgen and Estrogen Response Scores on Outcomes Following Neoadjuvant Pembrolizumab in Muscle-Invasive Bladder Cancer

Abstract Purpose: Androgen and estrogen influence bladder cancer pathogenesis, but their role in modulating treatment response, especially immunotherapy, remains unclear. This post hoc analysis of the PURE-01 trial evaluated the impact of androgen and estrogen signatures on outcomes in patients with muscle-invasive bladder carcinoma treated with neoadjuvant pembrolizumab followed by radical cystectomy. Experimental Design: Pretreatment transurethral resection of bladder tumor samples from 102 PURE-01 patients was analyzed using the Decipher Bladder whole-transcriptome assay to classify genomic subtypes. Hallmark androgen response (ARS), early estrogen response (ERS-early), and late estrogen response (ERS-late) scores were derived and visualized with boxplots. Subgroup differences were assessed using Mann–Whitney U, χ2, and Kruskal–Wallis tests. Event-free survival (EFS) was the primary endpoint, whereas pathologic response was the secondary endpoint. Cox proportional hazard models and Kaplan–Meier analyses evaluated the association of molecular scores with outcomes. Immune190 and FGFR3 signatures were also analyzed to explore their interactions with ARS and ERS. Results: Luminal subtypes had the highest ARS and ERS scores across molecular classifications. Negative correlations were observed between Immune190 scores and ERS and ARS scores. ERS-late scores were highest in nonresponders and significantly associated with worse EFS. Conversely, ARS was not significantly linked to EFS. FGFR3 signature positively correlated with ARS and ERS and negatively with Immune190 scores. Conclusions: The strong association of ERS-late scores with both pathologic response and EFS suggests that this signature could guide patient stratification and treatment selection. These findings support future trials combining immunotherapy with antiestrogens in both neoadjuvant and advanced urothelial carcinoma settings.