MRI-Based Prediction of Positive Surgical Margins in Radical Prostatectomy: Insights from Clinical and Imaging Parameters.

INTRODUCTION Positive surgical margins (PSM) after radical prostatectomy are associated with adverse oncological outcomes. Preoperative identification of patients at risk may improve surgical decision-making.

Objective

To evaluate clinical and MRI-derived predictors of PSM and to explore their potential value for preoperative risk assessment in patients undergoing robot-assisted radical prostatectomy (RARP).

Methods

This retrospective, single-center study comprised 134 patients with localized prostate cancer who underwent preoperative multiparametric MRI (mpMRI) and RARP between July 2022 and April 2024. Clinical, histopathological, and MRI-derived parameters were analyzed using univariate and multivariate logistic regression. Odds ratios (ORs), 95% confidence intervals (CIs), and p-values were reported. Model performance was evaluated using receiver operating characteristic (ROC) curve analysis.

Results

PSMs were present in 20.1% (27/134) of patients. In multivariate analysis, the strongest independent predictors of PSM were suspicion of extraprostatic extension (EPE) on MRI (OR 12.62, 95% CI 3.24-49.15; p < 0.001), urethral involvement (distance < 2 mm) (OR 8.88, 95% CI 1.76-44.73; p = 0.008), and capsular contact length >12 mm (OR 8.84, 95% CI 2.44-32.03; p < 0.001). The exploratory model achieved excellent apparent discrimination (AUC 0.911). Infiltration of the perineural sheath was observed in 96% of R1 compared to 70% of R0 patients (p = 0.01). Nerve-sparing surgery was not significantly associated with a different rate of PSM (p = 0.385).

Conclusion

Several MRI-derived parameters, including capsular contact length, MRI-suspected extraprostatic extension, and urethral proximity, were associated with an increased likelihood of positive surgical margins. Incorporating these metrics alongside clinical factors may improve risk stratification. However, prospective studies with external validation are required before routine clinical implementation.