Conditional Survival of Patients with Prostate Cancer Undergoing External-Beam Radiotherapy on RTOG 0126 and 0415

Abstract

<h3>Purpose/Objective(s)</h3> Patients with low- and intermediate-risk prostate cancer (PCa) treated with radiation (RT) alone have excellent outcomes with anticipated long overall survival (OS). Although biochemical failure (BF) and OS rates are commonly reported from time of diagnosis, the relevance of such data diminishes over time for survivors. Conditional survival analyses can provide more relevant survival estimates for long-term survivors and help patients understand changing recurrence risks. The primary objective of this study was to report conditional OS and BF in patients with low- and intermediate-risk PCa treated with RT. A secondary objective was to determine if prognostic factors at initial treatment remain relevant later in survival. <h3>Materials/Methods</h3> We performed a pooled analysis of patients with low- or intermediate- risk PCa enrolled in RTOG 0126 or 0415 treated with RT alone. OS was calculated (Kaplan–Meier) at various survivorship timepoints. Cumulative incidence (Phoenix definition) was used to calculate BF rates. Risk factors (Gleason score [GS], T stage, prostate-specific antigen (PSA), and EQD2 of prescribed dose) were analyzed at different timepoints using multivariable Cox proportional hazards modeling (MVA). <h3>Results</h3> A total of 2,591 patients were included (median age, 69 y; median PSA, 6.4 ng/mL; 48.5% with GS 7; 34.2% with stage T2 disease). Median follow-up was 6.9 y. The 3-, 5-, and 8-y OS from time of treatment were 96.2%, 89.2%, and 41.0%, respectively. Patients with 1-, 3-, and 5y of survival, had an additional 5 y survival rate of 73.3%, 42.6%, and 20.8%, respectively. On MVA at diagnosis, GS 7 (HR 1.352, <i>P</i> = 0.003), increasing age (HR 1.054, <i>P</i> < 0.001), and increasing PSA (HR 1.054, <i>P</i> = 0.01), were associated with all-cause mortality. For patients who survived 5 y, only age (HR 1.071, <i>P</i> < 0.001) was predictive of mortality. The 3-, 5-, and 8-y rates of BF from time of treatment were 7.1%, 12.5%, and 22.3%, respectively. For patients surviving 1, 3, and 5 y without BF, the rates of BF in the next 5 y were 14.2%, 17.2%, and 18.8%, respectively. On MVA, GS 7 (HR 1.907, <i>P</i> < 0.001), T2 disease (HR 1.375, <i>P</i> < 0.001), and increasing PSA (HR 1.114, <i>P</i> < 0.001) were significantly associated with increased risk of BF, whereas increasing EQD2 was associated with decreased BF rates (HR 0.922, <i>P</i> < 0.001). For patients who survived 5 y without BF, all initial prognostic factors remained statistically significant: GS 7 (HR 2.401, <i>P</i> < 0.001), stage T2 (HR 1.344, <i>P</i> = 0.0396), PSA (HR 1.109, <i>P</i>< 0.001), and EQD2 (HR 0.887, <i>P</i> < 0.001). <h3>Conclusion</h3> Conditional risk of BF increases over time for patients with PCa initially treated with RT. Initial biologic and treatment variables remain prognostic at long-term follow-up, with dose-escalated RT continuing to be associated with decreased BF. This data can be used to counsel patients for their risk of BF in follow-up and to help decide on the frequency of PSA monitoring.