The Impact of Radical Prostatectomy Versus Radiation Therapy on Cancer-Specific Mortality for Nonmetastatic Prostate Cancer: Analysis of an Other-Cause Mortality Matched Cohort

Previous population-based studies comparing radical prostatectomy to radiation therapy for prostate cancer were hindered by selection bias, with patients undergoing radiotherapy being at higher risk for any cause of death. Our report is the first to evaluate the impact of initial treatment on a cohort with no difference in othercause mortality risk, demonstrating that prostatectomy resulted in better cancer-specific mortality only for individuals with unfavorable intermediate-to-high risk disease. Introduction: Studies comparing radical prostatectomy (RP) to radiation therapy (RT) have consistently shown that patients undergoing RT have a higher risk of other-cause mortality (OCM) compared to RP, signifying poor health status of the former patients. We aimed to evaluate the impact of RP versus RT on cancer-specific mortality (CSM) over a cohort with equivalent OCM risk. Patients and Methods: The SEER database was queried to identify patients with nonmetastatic PCa between 2004 and 2009. Patients were matched based on their calculated 10-year OCM risk and further stratified for DA'mico Risk Score and Gleason Grade. A Cox-regression model was used to calculate the 10-year OCM risk. Propensity-score based on the calculated OCM risk were used to match RP and RT patients. Cumulative incidence curves and Competing-risk regression analyses were used to examine the impact of treatment on CSM in the matched cohort. Results: We identified 55,106 PCa patients treated with RP and 36,674 treated with RT. After match, 6,506 patients were equally distributed for RT versus RP, with no difference in OCM rates ( P = .2). The 10-year CSM rates were 8.8% versus 0.6% ( P = .01) for RT versus RP in patients with unfavorable-intermediate-risk (Gleason Score 4 + 3) and 7.9% versus 3.9% ( P = .003) for high-risk disease. There was no difference in CSM among RT and RP patients for favorable-intermediate-risk (Gleason Score 3 + 4) and low-risk disease. Conclusions: In a matched cohort of PCa patients with comparable OCM between the 2 arms, RP yielded a more favorable CSM rate compared to RT only for unfavorable-intermediate- and high-risk groups.