Diagnostic impact of DWI absence on prostate lesion assessment using PI-RADS 2.1.

Abstract

Objective: To compare Prostate Imaging Reporting and Data System (PI-RADS) scores derived from a standard multiparametric prostate MRI (mpMRI) protocol with those from a protocol consisting only of T2-weighted and dynamic contrast-enhanced images (T2+DCE MRI).

Methods: In this retrospective, single-center, cross-sectional study approved by the IRB and compliant with HIPAA, 492 MRI exams performed in 2022 were analyzed. PI-RADS scores from mpMRIs were extracted from medical records, and new scores were generated for T2+DCE MRI following PI-RADS guidelines. Score differences were evaluated using Wilcoxon signed-rank and McNemar's tests, stratified by lesion location (peripheral zone, PZ, and transition zone, TZ). Diagnostic accuracies of the two methods were compared using ROC curves, and logistic regression was employed to identify predictors of score changes.

Results: Notable differences in PI-RADS scores were observed were observed in the PZ (P = 0.03) and TZ (P < 0.001). 4.8 % of PZ and 4.0 % of TZ PI-RADS 3-5 lesions were misclassified as PI-RADS 1-2 on T2W+DCE MRI (PZ vs TZ, P = 0.64). However, ROC curve analyses revealed no significant difference in diagnostic accuracy between mpMRI (Az = 0.77) and T2+DCE MRI (Az = 0.75, P = 0.07). PSA density was identified as a predictor of score changes from PI-RADS 3-5 to 1-2, although the effect size was modest.

Conclusions: Although T2+DCE MRI yields different PI-RADS scores compared to mpMRI, the clinical impact on diagnostic accuracy and decision-making is overall small. This supports the continued use of T2+DCE MRI, particularly when diffusion-weighted imaging is compromised.