Delayed Salvage Radiotherapy Guided by Contrast-Enhanced Local and Whole-Body Diffusion-Weighted Magnetic Resonance Imaging During Prostate-Specific-Antigen-Based Surveillance After Radical Prostatectomy: A Retrospective Cohort Study

Background Biochemical recurrence (BCR) after radical prostatectomy (RP) is common. Although early prostate-specific-antigen (PSA)-triggered salvage radiotherapy (SRT) improves outcomes, PSA elevation alone cannot reliably distinguish local recurrence from occult metastatic disease, leading to overtreatment. This study retrospectively evaluated long-term outcomes of a PSA-based surveillance strategy, wherein delayed SRT was guided by contrast-enhanced local and whole-body diffusion-weighted magnetic resonance imaging (MRI) and initiated only after radiologic confirmation of macroscopic local recurrence (MLR). Methods Between January 2015 and June 2023, 226 patients with MRI-detected MLR in the prostate bed were identified during PSA-based surveillance incorporating serial PSA measurements and repeat contrast-enhanced pelvic MRI combined with whole-body diffusion-weighted imaging after RP. SRT was intentionally deferred and initiated only after imaging-confirmed local recurrence with exclusion of distant metastases. Ninety patients who underwent SRT and had ≥6 months of follow-up were included in the final analysis. Radiotherapy was delivered at a total dose of 64–68 Gy. The primary endpoint was BCR-free survival (bRFS). Secondary endpoints included local-recurrence-free survival (lRFS), distant-metastasis-free survival (DMFS), overall survival (OS), patterns of failure, and late toxicity. Results The median follow-up was 67 months, and the median PSA level at SRT initiation was 0.98 ng/mL. The estimated 5-year bRFS and lRFS rates were 84.8% and 91.3%, respectively. The 5-year DMFS and OS rates were 94.1% and 97.3%, respectively. Disease recurrence occurred in 14 patients (15.6%), predominantly as local recurrence after SRT. Two patients exhibited late grade ≥3 gastrointestinal toxicity. However, no grade ≥3 genitourinary toxicity was reported. Conclusions The PSA-based surveillance strategy, incorporating serial MRI, facilitated the detection of localized recurrence and allowed delayed initiation of SRT in patients with imaging-confirmed MLR following RP. This approach achieved durable disease control with acceptable toxicity in the studied cohort. However, the broader clinical utility of this imaging-based strategy warrants further prospective evaluation.