Predictors of pathologic response after radical cystectomy for muscle invasive bladder cancer: A population-based analysis.

Abstract

503 Background: Neoadjuvant chemotherapy (NAC) improves overall survival (OS) for patients with muscle invasive bladder cancer (MIBC) undergoing radical cystectomy (RC), possibly through an increase in pathologic complete response (CR), defined as a pathologic stage T0 (pT0). We sought to identify predictors of CR for MIBC. Methods: The National Cancer Database from 2004 to 2013 was used to identify patients with cT2-4cN0cM0 urothelial cell carcinoma treated with RC. Patients were grouped based on pathology as CR (pT0), partial response (PR – pTa/Tis/T1) or no response (NR - pT2 or higher). Predictors of NAC and CR were identified with multivariable logistic regression. Cox proportional hazards model was used to compare OS based on cT stage, receipt of NAC and pathologic response. Results: The study population included 10,820 patients and NAC was administered in 16.4%. Use of NAC was associated with higher cT stage (p < 0.01) and increased over time (10% from 2003-2007 vs. 24% from 2011-2013, p < 0.01). Predictors of NAC use on multivariable analysis include younger age, lower comorbidity score, treatment at an academic center, and diagnosis from 2011-2013 (p < 0.01). Overall, CR was achieved in 3.3% without NAC and 16.3% with NAC (p < 0.01). NAC improved 5-year OS for all cT stages, however the survival benefit was only observed among those achieving CR (p < 0.01). Multivariable Cox regression demonstrates that both PR (HR 0. 58, p < 0.01) and CR (OR 0.26, p < 0.01) were independently associated with improved OS among those treated with NAC. Multivariable analysis identified age (OR 0.98, p < 0.01) and increased clinical stage (cT3: OR 0.47, p < 0.01; cT4 OR 0.54, p < 0.01) as negative predictors of CR. Utilization of NAC (OR 4.82 p < 0.01), academic institution, and diagnosis 2011-2013 (OR 1.92, p < 0.01) increased the odds of CR. Conclusions: Use of NAC increased over time and CR occurred in 16% of patients who received NAC. Treatment at an academic center, diagnosis from 2011-2013 and use of NAC were independently associated with CR, while increased age and clinical stage were negative predictors of CR. PR and CR are independently associated with improved OS relative to non-responders.