Introduction

A potential risk of significant competing causes of mortality in men submitted to salvage radiation therapy (SRT) might be present. We aimed at reporting the natural history of patients treated with SRT from a large multi-institutional series.

Methods

The study included 715 patients who received SRT at six tertiary referral centres for either PSA rising after RP, or PSA persistence after surgery that was defined as PSA level ≥0.1 ng/ml at 1 month after surgery. The irradiation of the pelvic lymph nodes area was left at the discretion of the treating physician. The study outcomes were cancer-specific (CSM) and other cause mortality (OCM). Cox regression analyses were used to predict the risk of CSM and OCM. Predictors consisted of patient age, pT stage (≤pT3a vs. ≥pT3b), pathologic Gleason (≤7 vs. ≥8), surgical margins (negative vs. positive), PSA level at SRT, and concomitant hormonal therapy (HT) administration. Competing-risks Poisson regression methodologies were performed.

Results

Median patient age and PSA level at SRT were 66 years and 0.30 ng/ml, respectively. At a median follow-up of 102 months (inter-quartile range: 61, 135), local recurrence was detected in 14 (2.0%) patients, whereas 30 (4.3%), 16 (2.3%), 64 (9.2%), and 13 (1.9%) developed pelvic, retroperitoneal, skeletal, and visceral metastasis, respectively. Overall, 154 patients were died at last follow-up: 39 patients succumbed to prostate cancer, and 115 died from other cause. At 10 years follow-up, CSM and OCM rates were 13% and 37%, respectively. At multivariable competing-risks regression analyses, pathologic stage ≥pT3b (HR: 3.16, p=0.006) and Gleason score ?8 (HR: 3.56, p=0.003) were independent predictors of CSM, after accounting for the risk of dying from other cause. Conversely, age (HR: 1.08, p<0.0001) and concomitant HT (HR: 1.15, p=0.001) represented independent predictors of OCM, after accounting for the risk of dying from prostate cancer. When patients were stratified by age (≤65 vs. >65), the risk of OCM at 10 years was significantly higher in older patients (77% vs. 50%, p<0.0001).

Conclusions

This is the first study assessing the long-term natural history of SRT after accounting for the risk of OCM. We showed that roughly a third of men submitted to SRT died from other cause rather than from PCa and this rate increased in older men receiving concomitant HT. These results should be taken into account when deciding on the use of SRT after radical prostatectomy according to each patient profile in order to avoid potential overtreatment.

Funding

none