Development and internal validation of a biparametric MRI-based nomogram for predicting extraprostatic extension in...

This study aimed to identify predictors of extraprostatic extension (EPE) in prostate cancer and develop a nomogram to guide nerve-sparing decision-making in robot-assisted radical prostatectomy (RARP). The study included 402 patients who underwent RARP between January 2020 and December 2025. Fifty five patients were excluded as they received neoadjuvant therapy and did not undergo magnetic resonance imaging (MRI) based on the Prostate Imaging Reporting and Data System (PI-RADS). For side-specific analysis, data were analyzed on a lobe-by-lobe basis; among a total of 694 lobes, 73 were excluded due to being difficult to evaluate for EPE or lacking PI-RADS evaluation. Clinical, imaging, and biopsy factors associated with EPE per side were retrospectively analyzed, and a nomogram was developed to predict EPE. Among the 621 lobes analyzed, 101 (16.3%) were EPE-positive, while 520 (83.7%) were EPE-negative. Multivariate analysis identified prostate-specific antigen (PSA) (odds ratio [OR]: 1.041, 95% confidence interval [CI]: 1.003-1.081, p = 0.034), PIRADS category (OR: 2.630, 95%CI: 1.049-6.593, p = 0.039), suspected EPE on MRI (OR: 2.887, 95%CI: 1.461-5.702, p = 0.002), biopsy Grade group (OR: 2.502, 95%CI: 1.239-5.052, p = 0.011), and biopsy positivity in the peripheral zone (OR: 4.080, 95%CI: 1.744-9.548, p = 0.001) as significant factors associated with EPE. Subsequently, a nomogram was created based on these factors. The area under the curve of the nomogram was 0.868, indicating high diagnostic accuracy. In summary, we developed an accurate nomogram for predicting EPE. This nomogram may contribute to the appropriate selection of nerve-sparing grades for RARP.