Negative Predictive Value of Prostate Multiparametric Magnetic Resonance Imaging among Men with Negative Prostate Biopsy and Elevated Prostate Specific Antigen: A Clinical Outcome Retrospective Cohort Study

Purpose: We estimated the negative predictive value of prostate multiparametric magnetic resonance imaging to detect clinically significant (Gleason 7 or greater) prostate cancer at long-term followup (median 6.7 years, range 2.6 to 10.7), in men with negative biopsy findings before magnetic resonance imaging. We also assessed the diagnostic performance of multiparametric magnetic resonance imaging to detect clinically significant prostate cancer during this time. Materials and Methods: Following Institutional Research Ethics Board approval we retrospectively identified men who underwent prostate multiparametric magnetic resonance imaging after biopsy between 2004 and 2009 using a cancer registry database and magnetic resonance imaging reports. Multiparametric magnetic resonance imaging sequences comprised T2-weighted and dynamic contrast-enhanced series from 2004 to 2005 with diffusion-weighted imaging from 2006 and thereafter. Clinical outcomes were assessed up to July 2015 by reviewing subsequent pathology results, prostate specific antigen levels and electronic patient records. The primary outcome was clinically significant prostate cancer diagnosis during followup. We also estimated the sensitivity, specificity, and positive and negative predictive values of all prostate multiparametric magnetic resonance imaging during this period. Results: A total of 502 multiparametric magnetic resonance imaging scans with a prior biopsy were included in study. Of these scans 121 were done in men with a prior systematic biopsy negative for cancer. In these men median prostate specific antigen was 9.5 ng/dl and median age was 60 years. At a median followup of 6.7 years (95% CI 2.6 to 10.7) 70 of 73 (96%) men with negative multiparametric magnetic resonance imaging findings remained free of clinically significant prostate cancer. In this period the overall negative and positive predictive values of multiparametric magnetic resonance imaging were 86% (range 80% to 91%) and 54% (range 52% to 57%), respectively, in the entire cohort regardless of biopsy status before magnetic resonance imaging. Conclusions: Prostate multiparametric magnetic resonance imaging has high clinical negative predictive value. In men with a negative biopsy before magnetic resonance imaging and negative magnetic resonance imaging findings the risk of clinically significant prostate cancer was extremely low at a median of 6.7 years.