Long-term Prediction of Metastatic Recurrence in Patients Candidate to Radical Prostatectomy Using MRI and Targeted...

Prostate cancer recurrence after radical prostatectomy (RP) is primarily driven by tumor aggressiveness, but accurate long-term prediction remains challenging. We evaluated long-term oncological outcomes using magnetic resonance imaging (MRI) and targeted biopsy data in 646 patients treated with RP between 2009 and 2018, and the median follow-up for survivors was 8.3 yr. The primary end point was metastatic recurrence and the secondary end point was biochemical recurrence (BCR). Preoperative variables included MRI-derived tumor volume and quantification of Gleason pattern 4/5 (GP4/5). Tumors were classified according to zonal origin as peripheral zone (PZ) or transition zone/anterior fibromuscular stroma (TZ/AFMS). At 10 yr, metastatic recurrence occurred in 11% (95% confidence interval [CI] = 9-15), including distant metastases in 7.3% (95% CI = 5.1-10). BCR occurred in 28% (95% CI = 24-32). MRI- and biopsy-derived GP4/5 burden strongly predicted metastatic recurrence (c-index 0.82, 95% CI = 0.76-0.88) and BCR (c-index 0.73, 95% CI = 0.69-0.78). Outcomes were consistently worse for PZ compared with TZ tumors, with higher 10-yr BCR rates (32% vs 14%, p = 0.004) and numerically higher metastatic recurrence (8.2% vs 3.7%), with overlapping CIs. These findings provide long-term validation of MRI-based quantification of tumor aggressiveness to predict metastatic recurrence after RP. Although PZ tumors showed consistently worse outcomes in unadjusted analyses, zonal origin was not an independent predictor after multivariable adjustment, suggesting mediation through higher GP4/5 burden.