Regression Discontinuity Analysis of Salvage Radiotherapy in Prostate Cancer

Abstract

There is a lack of randomized evidence comparing early (eSRT) to late (lSRT) salvage radiotherapy (SRT) after radical prostatectomy (RP) for prostate cancer (PCa). Moreover, the existing evidence is often affected by lead-time bias. We sought to address this gap in a cohort of 1458 PCa patients undergoing SRT for biochemical recurrence (BCR) after RP in two tertiary care centers between 1992 and 2013. Using a quasi-randomized study design known as regression discontinuity (RD) and adjusting for lead-time bias, we compared metastasis-free survival (MFS) at 5 and 10 years after surgery between eSRT (prostate-specific antigen [PSA] <0.5 ng/ml) and lSRT (PSA ≥ 0.5 ng/ml). Overall, 1049 patients (71.9%) underwent eSRT and 409 (28.1%) lSRT at a mean follow-up of 84 mo (interquartile range (IQR) 52-120.4). The MFS rate decreased nonsignificantly at the proposed cutoff by 0.04 (95% confidence interval [CI]: -0.06 to 0.19) at 5 years and by 0.07 (95% CI: - 0.12 to 0.32) at 10 years. Cox regression analysis revealed a hazard ratio for the cutoff examined of 1.3 (95% CI: 0.8-2.4; p = 0.2). In conclusion, in a quasirandomized study design accounting for lead-time bias, eSRT (PSA < 0.5 ng/ml) did not improve MFS. Our results underline the need for level-one evidence to compare eSRT and lSRT. PATIENT SUMMARY: We compared early versus late salvage radiotherapy (SRT) for biochemical recurrence after radical prostatectomy by simulating a randomized trial. We found that early SRT (initiated at prostate-specific antigen <0.5 ng/ml) compared to late SRT did not improve metastasis-free survival.

Keywords: Adverse pathological features; Prostate cancer; Regression discontinuity; Salvage radiation therapy.