PD22-04: Immediate Radiotherapy versus Observation in Patients with Node-Positive Prostate Cancer after Radical Prostatectomy

Christian Schaufler, Sumedh Kaul, Aaron Fleishman, Ruslan Korets, Peter Chang, Andrew Wagner, Boston, MA, Simon Kim, Aurora, CO, Joaquim Bellmunt, Irving Kaplan, Aria F. Olumi, Boris Gershman*, Boston, MA

Introduction: The optimal management of node-positive (pN1) prostate cancer following radical prostatectomy (RP) remains uncertain. Despite randomized evidence, use of immediate, life-long androgen deprivation therapy (ADT) remains poor, and recent trials of early salvage radiotherapy included only a minority of pN1 patients. We therefore emulated a hypothetical pragmatic trial of immediate (adjuvant) radiotherapy versus observation in men with pN1 prostate cancer.

Methods: Using RADICALS-RT to inform the design of a hypothetical target trial, we identified men aged 50-69 years with pT2-3 Rany pN1 M0, pre-treatment PSA <50 ng/mL prostate cancer in the NCDB from 2006-2015 treated with 60-72 Gy of adjuvant RT (aRT) ± ADT within 26 weeks of RP or observation. After estimating a propensity score for receipt of aRT, we estimated absolute and relative treatment effects using stabilized inverse probability of treatment (sIPW) re-weighting.

Results: A total of 3,510 patients were included in the study, of whom 587 (17%) received aRT (73% with concurrent ADT). Median follow-up was 42.3 months, during which time 333 deaths occurred. After sIPW re-weighting, baseline characteristics were well-balanced. Adjusted OS was 94% versus 89% at 5-years and 82% versus 80% at 8-years for aRT versus observation (Figure 1; p=0.12). In IPW-reweighted Cox regression modelling, aRT was associated with a lower risk of all-cause mortality than observation, but this did not reach statistical significance (HR 0.71, p=0.06). In analyses examining heterogeneity of treatment effects, aRT was associated with improved OS only for men with Gleason 8-10 disease (HR 0.59, p=0.01), ³2 positive LNs (HR 0.50, p=0.04 for 2 positive LNs; HR 0.43, p=0.01 for ³3 positive LNs), or negative surgical margins (HR 0.50, p=0.03).

Conclusions: In observational analyses designed to emulate a hypothetical target trial of aRT versus observation in pN1 prostate cancer, aRT was associated with improved OS only for men with Gleason 8-10 disease, ³ 2 positive LNs, or negative surgical margins.

Source of Funding: The authors have no funding disclosures.