Intravesical therapy and risk of cystectomy and bladder cancer death after BCG fails.

Abstract

e17026 Background: Management of non-muscle-invasive bladder cancer (NMIBC) when BCG fails engages complex decision-making that incorporates consideration of radical cystectomy (RC) and several intravesical options. We sought to characterize trends in practice variation and health outcomes for patients receiving intravesical therapy and radical cystectomy for presumption of recurrent or progressive high-risk NMIBC that has failed BCG. Methods: We identified patients with high-risk NMIBC from SEER-Medicare diagnosed from 2004-2015 who completed a full dose of BCG treatment. We evaluated receipt of intravesical therapy after BCG fails based on a code for a transurethral resection of bladder tumor (TURBT) within 6 months of BCG treatment followed by a switch to a different intravesical agent, systemic therapy, or radical cystectomy. We constructed risk estimates for disease-specific survival and subsequent need for radical cystectomy and using the Kaplan-Meier method and multivariate Cox proportional-hazards models adjusted for competing risks of death and patient clinical/demographic factors. Results: Among high-risk NMIBC patients who received BCG in their first year of diagnosis (n = 14,369), 9.7% (n = 1,273) went on to receive intravesical therapy, 4.2% (n = 607) proceeded directly to cystectomy, 15.8% (n = 2,272) underwent re-induction with BCG, 7.0% (n = 1,009) initiated systemic chemotherapy and 4.4% died without receiving treatment after BCG. Median follow-up after the post-BCG recurrence was 29 months (IQR 42, 14-56). Those receiving intravesical therapy after BCG fails were older, had more comorbidities, resided in rural areas, and had higher proportion of clinical Ta cancers (p < 0.05 for all) compared with immediate RC patients. Intravesical agents after BCG fails are listed in the Figure. Among those undergoing intravesical therapy, 12.9% (n = 478) eventually underwent RC a median 13 months (IQR 20, 7-27) after treatment post-BCG (2-year cystectomy free survival 89.4% [95% CI 88.3-90.5%]). Actuarial 2-year and 5-year bladder cancer-specific survival was 84.0% (95% CI 82.3-85.7%) and 71.0% (95% CI 68.3-73.7%), respectively, after intravesical treatment post-BCG. Conclusions: The heterogeneity in treatments used for NMIBC that fails BCG indicates potential provider uncertainty in the management of this challenging clinical situation. The paucity of effective bladder-sparing therapies, which is reflected in a high 2-year risk of RC and bladder cancer-specific death, highlights the urgent need for new therapies in this patient population.