Changes of urinary immunity and microbiome after intravesical BCG therapy and their association with outcomes in NMIBC

Aim: Intravesical Bacillus Calmette–Guérin (BCG) is the standard therapy for non-muscle invasive bladder cancer (NMIBC); however, many patients experience recurrence or progression. We examined how urinary immune signals and the urinary microbiome change across BCG and are related to outcomes. Methods: In this single-center prospective cohort study, adults with NMIBC underwent transurethral resection of bladder tumor (TURBT), followed by BCG induction. Urine was collected before TURBT, before BCG, after BCG induction, and three months later. Urine sediment mRNA (PD-L1, PD-L2, CD33, and CD204) was quantified using TaqMan ΔCt. The urinary microbiome was profiled using 16S rRNA gene sequencing, and diversity, composition, and taxon balance were evaluated using nonparametric tests, PERMANOVA, repeated-measures correlations, and mixed-effects models. We analyzed the relationship between the urinary microbiome and prognosis. Results: Twenty-three patients were analyzed; ten recurrences, eight progressions, and three deaths were observed. Relative to baseline, CD33 increased after BCG and after three months; PD-L2 increased immediately after BCG and returned to baseline by three months; PD-L1 and CD204 increased after BCG. Shannon alpha-diversity was unchanged, but total read count was higher at three months, with stable beta-diversity and dispersion. Higher PD-L1 expression was associated with lower Actinobacteria abundance in the bladder cancer microenvironment. A higher post-BCG Firmicutes/Bacteroidetes ratio was associated with worse prognosis, with the clearest signal for progression-free survival (PFS) observed in the univariate Cox models. Higher post-BCG Corynebacterium and Enterobacteriaceae abundance was associated with better PFS. Conclusions: BCG was associated with higher urinary PD-L1/PD-L2 and myeloid marker transcripts, while overall community structure remained stable. These exploratory data support that pre-BCG microbial features may be related to early response, and post-BCG profiles may reflect durability and survival. Urine immune-microbiome profiling could be a feasible, noninvasive adjunct for monitoring and risk stratification in NMIBC.