Bacillus Calmette-Guérin-exposed Non-muscle-invasive Bladder Cancer: Survival Benchmarks, Bladder-sparing Strategies,...

Intravesical recurrence after induction bacillus Calmette-Guérin (BCG) therapy is common in non-muscle-invasive bladder cancer (NMIBC), but longitudinal, real-world data using contemporary definitions of the BCG-exposed (BCG-E) phenotype remain limited, which constrains trial design. We assessed long-term clinical outcomes and clinical trajectories for a large BCG-exposed cohort treated at a single tertiary care center to establish pragmatic benchmarks to inform patient counseling, surveillance strategies, and the design and interpretation of bladder-sparing studies.

We conducted a retrospective cohort study for adults with histologically confirmed high-grade NMIBC treated with intravesical BCG at Moffitt Cancer Center between 1988 and 2024. Relevant clinical data were extracted from electronic medical records into a prespecified database, including detailed BCG doses and timing and features of each recurrence episode. Initial management followed contemporary standards, with subsequent intravesical therapy or radical cystectomy (RC) after shared decision-making. Patients were classified as BCG-unresponsive (BCG-UR) or BCG-E, with BCG adequacy defined as previously published. Time-to-event endpoints were analyzed using Kaplan-Meier estimates and multivariable Cox proportional-hazards models; baseline characteristics were compared using Fisher's exact and Wilcoxon rank-sum tests.

Of 1076 NMIBC patients treated with BCG, 470 were classifiable: 245 (52.1%) were BCG-E, of whom 173 (70.6%) were resistant and 72 (29.4%) experienced delayed relapse, and 225 (47.9%) were BCG-UR. In the BCG-E group, 50.2% experienced recurrence and 15.5% experienced progression; median recurrence-free survival (RFS) was 27.2 mo. The 5-yr survival rate estimates were 32.25% for RFS, 79.3% for progression-free survival (PFS), 84.5% for metastasis-free survival (MFS), and 65.0% for overall survival (OS). Recurrence was associated with worse PFS (p < 0.001) and MFS (p = 0.03), but not OS (p = 0.2). RC was performed in 45 patients (18.4%). No consistent survival differences were observed across salvage therapies (62.0% BCG, 18.4% gemcitabine + docetaxel, 5.3% single-agent chemotherapy, 1.2% trials, 13% surveillance),. Multivariable OS estimates were computed. Addition of recurrence to the model did not improve discrimination; age and performance status were the strongest predictors of OS.

We present comprehensive outcomes for and an in-depth characterization of clinical trajectories in BCG-E NMIBC, for which salvage intravesical BCG is predominant and oncologic results are durable. These data provide pragmatic benchmarks for the interpretation and design of bladder-sparing trials in this setting.