Prostate cancer-specific survival differences in patients treated by radical prostatectomy versus curative radiotherapy

Objective: We compared the cause-specific survival of patients who received radiotherapy to those who received surgery for cure of their prostate cancer using a number of design and analytic steps to mitigate confounding by indication. Methods: This was a case-cohort study of 2213 patients in the Ontario Cancer Registry diagnosed between 1990 and 1998 who were either treatment candidates or received curative radiotherapy or surgery. Cases included patients who died of prostate cancer within 10 years. The study population was restricted to those who were candidates for either treatment (radiotherapy or surgery) based on disease severity (low and intermediate risk using the Genitourinary Radiation Oncologists of Canada risk groups). The median follow-up was 51 months. Cause-specific survival was analyzed using Cox-proportional hazards regression with case-cohort variance adjustment. Results from intent-to-treat analyses were compared to results by treatment received. Results: Adjusted hazard ratios for risk of prostate cancer death for radiotherapy compared to surgery for the entire study population were 1.62 (95% CI 1.00-2.61) and 2.02 (1.19-3.43) analyzing by intent-to-treat and treatment received, respectively. Intent-to-treat hazard ratios for the low-and intermediate-risk groups were 0.87 (0.28-2.76) and 1.57 (0.95-2.61), respectively. Conclusion: Overall results were driven by the finding in the intermediate-risk group, which indicated that radiotherapy was not as effective as surgery in this group. Confirmation was needed with special attention paid to risk stratification and the impact of more contemporary delivery of these treatment options.