Predicting clinically significant prostate cancer in PI-RADS 3 lesions: an internally validated PSA density-based model with decision-curve analysis.

Authors

Federico Rovegno, Rajiv Pillai, Aniket Deshpande, John Corr

Published in

International urology and nephrology. Oct 02, 2026. Epub Oct 02, 2026.

Abstract

To identify clinical and imaging predictors of clinically significant prostate cancer (csPCa) in men with PI-RADS 3 index lesions on biparametric MRI (bpMRI), and to develop, internally validate, and evaluate the clinical utility of a risk-prediction model.
We retrospectively evaluated 450 consecutive, biopsy-naïve patients with a PI-RADS 3 index lesion on bpMRI who underwent transperineal targeted plus systematic biopsy at Colchester Hospital (2020-2025). Primary outcome was csPCa (ISUP Grade Group ≥2). Multivariable logistic regression identified independent predictors. Internal validation was performed using bootstrap resampling (1000 repetitions) to yield optimism-corrected discrimination (AUC) and calibration estimates. Decision curve analysis (DCA) assessed net clinical benefit.
csPCa was diagnosed in 98/450 patients (21.8%). On multivariable analysis, PSA density (PSAD) was the sole independent predictor of csPCa (OR 1.31 per 0.1 ng/mL/cc increase, 95% CI 1.11-1.55, p=0.002). Neither age, lesion size, location, DRE, nor family history reached significance. Bootstrap validation corrected the apparent AUC from 0.657 to 0.622, with excellent calibration (mean absolute error 0.014). On DCA, the model demonstrated superior net benefit over biopsy-all and biopsy-none strategies across the 10-30% threshold probability range. Stratification by PSAD demonstrated csPCa rates of 10.8% for PSAD <0.10, 18.5% for 0.10-0.15, and 34.1% for >0.15 ng/mL/cc.
PSAD was the only independent predictor of csPCa in this cohort of men with PI-RADS 3 lesions on bpMRI. Internal bootstrap validation demonstrated modest discrimination but good calibration, while decision-curve analysis showed net clinical benefit across clinically relevant threshold probabilities. These findings support further evaluation of PSAD-based risk stratification, although external validation is required before clinical implementation.

PMID:
42827186
Bibliographic data and abstract were imported from PubMed on 03 Oct 2026.

Read full publication at:
Please sign in to see all details.

Stats

1-terrible, 9-excellent. How would you rate this publication? Sign in in to submit your rating.

Post a comment

You need to be signed in to post comments. You can sign in here.

Comments

There are no comments yet.