Utility of Multiparametric Magnetic Resonance Imaging With PI-RADS, Version 2, in Patients With Prostate Cancer Eligible for Active Surveillance: Which Radiologic Characteristics Can Predict Unfavorable Disease?

Abstract

We investigated the utility of multiparametric magnetic resonance imaging (mpMRI) using Prostate Imaging Reporting and Data System, version 2 (PI-RADSv2), scoring in patients with prostate cancer eligible for active surveillance (AS). The medical records of the patients who had undergone mpMRI before radical prostatectomy from 2014 to 2018 were reviewed. All the patients had met the Prostate Cancer Research International AS criteria. PI-RADSv2 scores were assigned to 12 prostate regions. Unfavorable disease was stratified using the American Joint Committee on Cancer (AJCC) prognostic scale as stage IIB (Gleason score [GS], 3+4 and pathologic stage T2) and IIC-III (GS,≥ 4+3 or pathologic stage T3). Of 376 eligible patients, 184 (48.9%), 129 (34.3%), and 63 (16.8%) had AJCC stage I, IIB, and IIC-III disease, respectively. The patients with IIC-III disease were older and had a higher prostate-specific antigen density than those with stage I or IIB disease. PI-RADS 5 lesions were more frequent in patients with stage IIC-III than in patients with stage I or IIB disease. Multivariable analysis revealed that≥ 2 lesions with a PI-RADS 5 score was an independent predictor for unfavorable disease (hazard ratio [HR], 3.612; P< .001 for IIB; HR, 6.562; P< .001 for IIC-III), and PI-RADS score of≥ 4 was limited for predicting AJCC stage IIB disease (HR, 2.387; P= .01). mpMRI with PI-RADSv2 showed high negative predictive value for patients with prostate cancer eligible for AS. Multiple PI-RADS 4-5 lesions were associated with unfavorable disease compared with solitary lesions. Multiple PI-RADS 5 lesions were strongly associated with GS≥ 4+3 or pathologic T3 disease. Targeted biopsy or radical treatment should be considered for these patients.