NIH Research Festival
Purpose: To assess performance of bi-parametric MRI (bMRI) for detecting prostate cancer (PCa) in screening-positive patients. Materials and Methods: 6 readers with 3 different experience levels in prostate MRI (2 readers with high, moderate, low experience), used PI-RADSv2 to evaluate PCa on bMRI. bMRI was acquired at 3T with 32-channel surface coil (axial T2-weighted, diffusion weighted (DWI) (ADC, b1500-DWI)). All patients underwent prostate mpMRI with subsequent targeted+systematic 12-core TRUS guided biopsy (BX). bMRI results were correlated with BX. Mean sensitivity, positive predictive values (PPV) for all, clinically significant (CS) lesions (Gl≥3+4) were determined. Pairwise proportion of agreement between readers was calculated. Results: From 05/2015-01/2016, 49 consecutive patients were included (median age=65(52-77), median PSA=5.51ng/mL(0.63-37.2)). Histopathology revealed 100 PCa (39=CS PCa=37 patients). No cancer was detected in 12 patients. Mean sensitivity and PPV for all and CS lesions for all readers were 0.34 and 0.60, 0.48 and 0.35, respectively. For all lesions, mean sensitivity and PPV for low, moderate, high experience readers were 0.37 and 0.58, 0.37 and 0.51, 0.28 and 0.73, respectively. For CS lesions, mean sensitivity and PPV for low, moderate, high experience readers were 0.51 and 0.32, 0.49 and 0.28, 0.45 and 0.45, respectively. Overall percent agreement between readers was 0.59, and was 0.60, 0.49, and 0.78 between low, moderate, and high experience readers, respectively. Conclusion: Regardless of experience, readers had low-to-moderate sensitivity with reasonable PPV and percent agreement in detecting all and CS lesions. Currently, 10min bMRI cannot be recommended in isolation to detect PCa in screening-positive patients.
Scientific Focus Area: Biomedical Engineering and Biophysics
This page was last updated on Friday, March 26, 2021