Introduction

The effectiveness of salvage radiation therapy (SRT) may be limited to specific categories of patients. We aimed at identifying the optimal candidate for SRT.

Methods

The study included 693 node-negative 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 RP. The study outcome consisted of distant metastasis after SRT: retroperitoneal (M1a), skeletal (M1b), and visceral metastasis (M1c). Regression tree analysis was used to develop a distant metastasis risk-stratification tool. Covariates consisted of pT stage (≤pT3a vs. ≥pT3b), pathologic Gleason (≤7 vs. ≥8), surgical margins (negative vs. positive), post-operative undetectable PSA (no vs. yes), and PSA level at SRT.

Results

At a median follow-up of 96 months, 82 (12%) patients developed distant metastasis. The metastasis location was retroperitoneal, skeletal, and visceral in 18 (2.6%), 33 (4.8%), and 9 (1.3%) patients, respectively. Using regression tree analysis, 5 risk groups for distant metastasis were identified: 1) very low-risk: undetectable PSA after RP, Gleason sum ≤7, and ≤pT3a; 2) low-risk: undetectable PSA after RP, Gleason sum ≤7, and ≥pT3b; 3) intermediate-risk: undetectable PSA after RP, and Gleason sum ≥8; 4) high-risk: PSA persistence after RP, and Gleason sum ≤7); 5) very high-risk: PSA persistence after RP, and Gleason sum ≥8. Frequencies and proportions of the five groups were 294 (42%), 211 (30%), 110 (16%), 58 (8%), 20 (3%), respectively. Metastasis-free survival at 8 years of the five groups was 86%, 75%, 74%, 72%, and 60%, respectively (p=0.003). The PSA level at SRT was significantly associated with the risk of distant metastasis in low-, intermediate-, and high-risk patients, where an early SRT was associated with better cancer control (all p<0.01). Conversely, this effect was not evident in the very low- and very high-risk patients, where PSA level at SRT was not significantly associated with the risk of metastasis (all p>0.05)

Conclusions

We developed a risk stratification tool that identified five prognostic risk groups. The early SRT administration provides better cancer control in low-, intermediate-, and high-risk patients. On the other hand, very low- (undetectable PSA after RP, Gleason score ≤7, and ≤pT3a) and very high-risk patients (PSA persistence after RP and Gleason score ≥8) do not benefit from an early treatment administration.

Funding

none