Immediate radiotherapy versus observation in patients with node-positive prostate cancer after radical prostatectomy.

BACKGROUND

The optimal management of node-positive (pN1) prostate cancer following radical prostatectomy (RP) remains uncertain. Despite randomized evidence, utilization 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 adjuvant radiotherapy versus observation in men with pN1 prostate cancer.

METHODS

Using the RADICALS-RT trial to inform the design of a hypothetical 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 to 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

In total, 3510 patients were included in the study, of whom 587 (17%) received aRT (73% with concurrent ADT). Median follow-up was 40.0 -months, during which 333 deaths occurred. After sIPW re-weighting, baseline characteristics were well-balanced. Adjusted overall survival (OS) was 93% versus 89% at 5-years and 82% versus 79% at 7-years for aRT versus observation (p = 0.11). In IPW-reweighted Cox regression, aRT was associated with a lower risk of all-cause mortality (ACM) than observation, but this did not reach statistical significance (HR 0.70 p = 0.06). In analyses examining heterogeneity of treatment effects, aRT was associated with improved ACM only for men with Gleason 8-10 disease (HR 0.59, p = 0.01), ≥2 positive LNs (HR 0.49, p = 0.04 for 2 positive LNs; HR 0.42, p = 0.01 for ≥3 positive LNs), or negative surgical margins (HR 0.50, p = 0.02).

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, ≥2 positive LNs, or negative surgical margins.