Objective: To evaluate whether pathological factors are associated with differential effect of adjuvant chemotherapy (ACT).
Patients and methods: In this population-based retrospective cohort study, we linked electronic records of treatment and surgical pathology to the Ontario Cancer Registry. The study population included all patients with muscle-invasive bladder cancer undergoing cystectomy in Ontario 1994-2008. Factors associated with overall (OS) and cancer-specific survival (CSS) were evaluated using Cox proportional hazards. We tested for interaction between the following variables and ACT effect-size: N-stage, margin status, T-stage, and lymphovascular invasion (LVI).
Results: The study population included 2802 patients; 19% were treated with ACT. Interaction terms with ACT for OS/CSS are: N-stage (both P < 0.001); margin status (P = 0.054/P = 0.048); T-stage (P = 0.509/P = 0.286); and LVI (P = 0.361/P = 0.405). Magnitude of effect for ACT was greater for patients with node-positive disease [OS: hazard ratio (HR) 0.56, 95% confidence interval (CI) 0.47-0.67; CSS: HR 0.60, 95% CI 0.49-0.72] than for patients with node-negative disease (OS: HR 0.80, 95% CI 0.61-1.03; CSS: HR 0.79, 95% CI 0.59-1.07). ACT was also associated with greater effect among patients with involved margins (OS: HR 0.45, 95% CI 0.33-0.62; CSS: HR 0.40, 95% CI 0.28-0.57) compared with patients with negative margins (OS: HR 0.75, 95% CI 0.65-0.87; CSS: HR 0.79, 95% CI 0.67-0.93).
Conclusions: In this population-based cohort study we observe evidence of interaction between ACT effect and nodal stage and surgical margin status. Our results suggest that patients at highest risk of disease recurrence may derive greatest benefit from ACT.
Keywords: bladder cancer; chemotherapy; outcomes research; predictive factors.
© 2014 The Authors. BJU International © 2014 BJU International Published by John Wiley & Sons Ltd.