Long-term incidence of secondary bladder and rectal cancer in patients treated with brachytherapy for localized prostate cancer: a large-scale population-based analysis.

OBJECTIVE: To examine the incidence and time trends of secondary bladder cancer (BCa) and rectal cancer (RCa) after brachytherapy (BT) relative to radical prostatectomy (RP). MATERIALS AND METHODS: Within the Surveillance, Epidemiology and End Results (SEER) database (1988-2015), we identified patients with localized PCa as an only or first primary cancer, who underwent BT or RP. Cumulative incidence plots and multivariable competing-risks regression (CRR) models were used. Sensitivity analyses focused on patients' age and year of diagnosis intervals and tested the effect of an unmeasured confounder. RESULTS: Of 318 058 patients with localized prostate cancer (PCa), 55 566 (18.4%) underwent BT. After propensity score-matching, 20-year secondary BCa incidence was 6.0% in patients who had undergone BT vs 2.4% in those who had undergone RP (P < 0.001) and the respective 20-year secondary RCa incidence was 1.1% vs 0.5% (P < 0.001). In multivariable CRR models, BT predicted higher secondary BCa (hazard ratio [HR] 1.58; P < 0.001) and RCa rates (HR 1.59; P < 0.001) vs RP. Sensitivity analyses replicated the same results after stratification according to age and showed HRs of decreasing magnitude for historical, intermediate and contemporary years of diagnosis. An unmeasured confounder with an HR of 2 would render the effect of BT statistically insignificant if it affected patients in the RP group with a ratio of 2 relative to those in the BT group. Finally, temporal trends showed a decrease of secondary 5-year BCa and RCa rates.> CONCLUSIONS: Brachytherapy predominantly increases the risk of secondary BCa and, to a lesser extent, that of RCa. Follow-up of such patients is therefore required. It is encouraging that both secondary BCa, and RCa rates, in particular, have recently decreased, RCa.