Abstract
Neoadjuvant chemotherapy (NAC) is linked with clinical advantages in urothelial carcinoma for patients with muscle-invasive bladder cancer (MIBC). Despite comprehensive research into the influence of tumor mutation expression profiles and clinicopathologic factors on chemotherapy response, the role of the gut microbiome (GM) in bladder cancer chemotherapy response remains poorly understood. This study examines the variance in the GM of patients with bladder cancer compared with healthy adults, and investigates GM compositional differences between patients who respond to chemotherapy versus those who exhibit residual disease.Our study reveals distinct clustering, effectively separating the bladder cancer and healthy cohorts. However, no significant differences were observed between chemotherapy responders and nonresponders within community subgroups. Machine learning models based on responder status outperformed clinical variables in predicting complete response (AUC 0.88 vs. AUC 0.50), although no single microbial species emerged as a fully reliable biomarker.The evaluation of short chain fatty acid (SCFA) concentration in blood and stool revealed no correlation with responder status. Still, SCFA analysis showed a higher abundance of Akkermansia (rs = 0.51, P = 0.017) and Clostridia (rs = 0.52, P = 0.018), which correlated with increased levels of detectable fecal isobutyric acid. Higher levels of fecal Lactobacillus (rs = 0.49, P = 0.02) and Enterobacteriaceae (rs = 0.52, P < 0.03) correlated with increased fecal propionic acid.In conclusion, our study constitutes the first large-scale, multicenter assessment of GM composition, suggesting the potential for a complex microbial signature to predict patients more likely to respond to NAC based on multiple taxa. SIGNIFICANCE: Our study highlights results that link the composition of the GM to the efficacy of NAC in MIBC. We discovered that patients with higher levels of Bacteroides experienced a worse response to NAC. This microbial signature shows promise as a superior predictor of treatment response over traditional clinical variables. Although preliminary, our findings advocate for larger, more detailed studies to validate these associations.