INTRODUCTION AND OBJECTIVE:
Despite the increased utilization of pre-biopsy multiparametric magnetic resonance imaging (mpMRI), the use of MRI as a staging tool remains limited. This is because mpMRI lacks sensitivity in detecting extracapsular extension (ECE). We aim to improve staging by assessing a novel MRI lesion/prostatic capsular relationship parameter, based on the standardized PI-RADS v2 report template, to predict the presence of locally advanced prostate cancer in men undergoing RP.
We identified men who underwent RP after 3-tesla multiparametric MRI (mpMRI) and subsequent concurrent systematic and MRI-targeted prostate biopsy between 2/2015-9/2019. Outcome of interest was ≥pT3a disease. The MRI lesion/prostatic capsular relationship was categorized using a standardized reporting format described in PI-RADS v2. A nomogram was developed including biopsy Gleason grade group (GGG), PSA density, PI-RADS score, and the MRI lesion/prostatic capsular relationship. Receiver operating characteristic curves (ROC) were calculated.
476 men were included in the study. Median age was 65 (IQR 60-70), median PSA was 6.09 (IQR 4.58-9.00). In univariate analysis, men were >5x more likely to harbor ≥pT3a disease if MRI demonstrated PI-RADS 5 lesion, if biopsy detected GGG 4/5, or if the capsular relationship demonstrated irregularity, or ECE (Table 1). Each remained significant predictors in the multivariate regression model. Use of the capsular relationship parameter improved the concordance index for the ROC curve from 0.7201 to 0.7859 (95%CI 0.737-0.822) (Figure 1).
We demonstrate that the MRI lesion/prostatic capsular relationship is a useful parameter in the prediction of locally advanced prostate cancer. Further, the incorporation of this parameter into a clinical nomogram improves the staging accuracy of mpMRI. The predictive model demonstrates more accurate staging as compared to widely used nomograms not inclusive of MRI parameters.
Source of Funding:
Joseph and Diane Steinberg Charitable Trust