Predictors of prostate bed recurrence on magnetic resonance imaging in patients with rising prostate-specific antigen after radical prostatectomy

Abstract

Introduction: Radical prostatectomy (RP) is a standard treatment modality for localized prostate cancer. Biochemical failure after RP is usually evaluated with whole body imaging to exclude distant metastatic disease, and pelvic magnetic resonance imaging (MRI) to detect local recurrence in the prostatectomy bed. The goal of this study is to correlate disease characteristics and demographic data in patients with rising prostate-specific antigen (PSA) after RP to determine association with MRI-detected cancer recurrence. Methods: Sixty-four patients who underwent pelvic MRI for rising PSA after RP and had complete clinical and pathological data available were included. Using Chi-squared testing, we analyzed PSA levels, pathological disease characteristics (prostate cancer risk group, Gleason score, extracapsular extension, positive surgical margin, seminal vesicle involvement, perineural invasion, lymphovascular invasion, and PSA level before MRI), time from surgery to biochemical failure, and patient demographic characteristics as potential predictors of MR-detected local recurrence. Results: Definite MRI-detected local recurrence was observed in 17/64 patients (27%). Eleven (17%) patients had a suspicious lesion with the differential of scarring, retained seminal vesicle, or recurrent cancer. Thirty-six (56%) patients had no evidence of tumor in the prostate bed or pelvis on MRI. Patient race was associated with likelihood of detecting a prostate nodule on MRI (p=0.04), with African-American patients having 82% lower odds of MRI-detected tumor recurrence compared with White patients (p=0.045). No other tumor or patient characteristic was significantly associated with MRI-detected recurrence. Conclusions: African-American patients with biochemical failure after RP are less likely to have MRI-detectable recurrence in the prostate bed compared with White patients.