Objectives: We studied the effect of conditional survival on 5-year cancer-specific mortality (CSM) probability in a contemporary North-American population-based nonmetastatic urothelial carcinoma of urinary bladder cohort treated with radical cystectomy.
Methods and materials: Within the SEER database (2004-2015), we identified pTa/pTis/pT1N0 high grade, pT2 to pT4N0 and pTanyN1-3 patients treated with radical cystectomy for nonmetastatic urothelial carcinoma of urinary bladder. Conditional 5-year CSM-free estimates were assessed after event-free follow-up duration. Multivariable Cox regression models predicted CSM according to event-free follow-up duration.
Results: According to T and N stages, 1,079 (7.9%) pTa/pTis/pT1N0, 5,058 (37.2%) pT2N0, 2,865 (21.1%) pT3N0, 1,211 pT4N0 (8.9%) and 3,382 (24.9%) pTanyN1-3 patients were included. Conditional CSM-free estimates increased from 90.1 to 91.8%, 80.6 to 92.5%, 62.5 to 90.7%, 53.1 to 84.5%, and 37.5 to 84.0% after 5 years of event-free follow-up, in respectively pTa/pTis/pT1N0, pT2N0, pT3N0, pT4N0, and pTanyN1-3 patients. Attrition due to mortality was highest in pTanyN1-3 cohort and lowest in pTa/pTis/pT1N0. In Multivariable Cox regression analyses, pT2N0 (hazard ratio [HR] 1.9 P< 0.001), pT3N0 (HR 4.3 P< 0.001), pT4N0 (HR 5.8 P< 0.001) and pTanyN1-3 (HR 9.1 P< 0.001) were independent predictors of higher CSM at baseline, relative to pTa/pTis/pT1N0. A decrease in all conditional HRs to nonsignificant levels was recorded at 60 months for pT4N0 and pTanyN1-3 and at 48 months for pT2N0 and pT3N0. CONCLUSIONS CONDITIONAL SURVIVAL: showed a direct relationship between event-free follow-up duration and survival probability. Conditional CSM-free estimates increased in proportion with event-free follow-up but also resulted in equally proportional increase in attrition rates.
Keywords: Cancer-specific mortality; Conditional survival; Nonmetastatic bladder cancer; Radical cystectomy.
Copyright © 2019 Elsevier Inc. All rights reserved.