Predictive Value of Preoperative PSA Levels Combined with MRI Features for BCR after Radical Prostatectomy: A Retrospective Study.

Abstract

Existing models for predicting that biochemical recurrence (BCR) will occur in patients after radical prostatectomy (RP) vary in their predictive results from magnetic resonance imaging (MRI). This study aimed to assess the predictive value of preoperative prostate-specific antigen (PSA) levels combined with MRI features in determining BCR following radical prostatectomy. A retrospective analysis was conducted on a cohort comprising 102 patients who underwent radical prostatectomy at our hospital between January 2019 and December 2019. On the basis of the outcomes observed during a 4-year follow-up after surgery, the patients were categorised into BCR group (n = 52) and non-BCR group (n = 50). Differences in preoperative PSA levels and MRI characteristics between the two groups were compared, and factors influencing postoperative BCR were analysed. The receiver operating characteristic curve was drawn, and the sensitivity, specificity, area under the curve (AUC) and Youden index were calculated to observe the predictive value of the combination of preoperative PSA level and MRI features for BCR following radical prostatectomy. Logistic regression analysis showed that preoperative PSA level, postoperative Gleason score, data system (Prostate Imaging-Reporting and Data System (PI-RADS)) score and clinical T stage were independent risk factors for BCR in patients following radical prostatectomy, with odds ratio (OR) greater than 1. The AUC value of preoperative PSA level combined with PI-RADS score was 0.921, surpassing the AUC values of 0.783, 0.822, 0.617 and 0.608 predicted by preoperative PSA level, postoperative Gleason score, PI-RADS score and clinical T stage alone, respectively. Postoperative BCR in patients with prostate cancer undergoing radical prostatectomy is associated with preoperative PSA level, postoperative Gleason score, PI-RADS score and clinical T stage. The combination of preoperative PSA level and MRI features can improve the predictive efficiency for postoperative BCR.