Introduction

The ProtecT trial reported excellent outcomes for patients with localized prostate cancer (PCa) treated with radical prostatectomy (RP). We aimed at assessing the generalizability of the ProtecT trial in men treated with RP at two high-volume institutions.

Methods

20,391 PCa patients aged 50-69 years treated with RP at two centers were included. We evaluated changes in disease characteristics over the study period and we identified patients who fulfilled the ProtecT criteria (cT1-cT2 and PSA <20ng/ml). Kaplan-Meier analyses assessed time to biochemical recurrence (BCR), clinical recurrence (CR), cancer-specific mortality (CSM), and overall mortality. A graphical illustration of the 10-year CSM and other-cause mortality (OCM) rates was generated using competing-risks Poisson regression analyses after stratifying patients according to eligibility in the ProtecT trial.

Results

The proportion of biopsy Gleason 8-10, PSA >20 ng/ml, and high-risk disease increased over time (5.1 vs. 18.8% and 6.9 vs. 9.1% and 13.3 vs. 24.6% for 1999-2005 vs. 2014-2016; all P<0.001). Overall, 18,555 (91.0%) patients fulfilled the ProtecT criteria. The proportion of eligible individuals decreased over time (91.2 vs. 89.7% for 1999-2005 vs. 2014-2016, respectively; P=0.02). The median PSA and the proportion of biopsy Gleason 8-10 were higher among men treated at referral centers compared to what reported in ProtecT (6.5 vs. 4.9 ng/ml and 9.0 vs. 2%; P<0.001). When considering patients eligible for the ProtecT, the proportion of biopsy Gleason score 8-10 and median PSA increased over time (3.8 vs. 15.7% and 6.3 vs. 7.1ng/ml for 1996-2005 vs. 2014-2016; all P<0.001). Median follow-up was 50 months. The 10-year BCR-, CR-, and CSM-free survival rates were higher in men eligible for the ProtecT trial as compared to those not eligible (74.2 vs. 42.6% and 95.4 vs. 79.0% and 98.8 vs. 91.4%; all P<0.001) and comparable to what observed in the ProtecT. The risk of dying from OCM was substantially higher than the risk of dying from PCa among men eligible for the ProtecT (6.9 vs. 2%). Conversely, in non-eligible patients the 10-year CSM and OCM rates were 9.7 vs. 7.8%.

Conclusions

A migration towards higher Gleason scores and PSA levels at diagnosis occurred over recent years. More than 10% of men treated at referral centers should have been excluded from the ProtecT trial due to less favourable characteristics. These patients are more likely to die from PCa than from OCM thus being the optimal candidate for testing the role of primary treatment in randomized trials.

Funding

none