MP13-03 FACTORS ASSOCIATED WITH FALSE-NEGATIVE MPMRI (PIRADS 1) OF THE PROSTATE IN PATIENTS UNDERGOING BIOPSY FOR SUSPICION OF PROSTATE CANCER

Abstract

INTRODUCTION AND OBJECTIVE: Multiparametric magnetic resonance imaging (mpMRI) of the prostate along subsequent grading with the Prostate Imaging Reporting and Data System version 2 (PI-RADSv2) score is one of the tools that aid in the decision making and performance of prostate biopsy. Although numerous studies have corroborated its value at predicting the presence of clinically significant prostate cancer (Gleason score ≥3+4), there still remains a large number of patients in which clinically significant disease is missed on mpMRI. This study’s objective was to investigate patient- and MRI-based factors that are associated with a false negative (FN) mpMRI of the prostate. METHODS: All patients with a negative mpMRI (PI-RADS 1) performed between 1/2016 and 7/2019 were identified. Patient charts were retrospectively reviewed for subsequent positive systematic prostate biopsy performed within one year, or negative biopsy at any later date. Patient and mpMRI specific factors were recorded. Two blinded radiologists evaluated the quality of different mpMRI-sequences (i.e. axial T2, DWI and ADC), measured the volume of rectal gas, bladder and prostate gland, and determined whether the peripheral zones (PZ) were ≥50% dark on T2-sequence. Interobserver agreement and multivariate logistic regression were performed. RESULTS: Out of a total of 550 patients with a negative mpMRI of the prostate, 150 patients (median age 64 years, interquartile range 58-68) underwent subsequent systematic prostate biopsy. Out of these, 12.7% (19/150) had Gleason score ≥3+4 prostate cancer, yielding a negative predictive value (NPV) of 87%. T2-dark PZs and PSA were statistically significantly associated with FN mpMRI (Odds ratios: 4.9, 95% CI:1.6-14.9, p=0.006 and 1.1, 1.0-1.2, p=0.030). Indication for mpMRI, radiologists’ experience and training, motion on axial T2, quality of DWI/ADC, prostate volume, bladder volume, amount of rectal gas and MRI vendor were not significant predictors. There was moderate interobserver agreement for T2-dark PZs (κ coefficient = 0.75). CONCLUSIONS: Even in select patients who undergo subsequent biopsy because of high clinical suspicion, a PI-RADS 1 prostate mpMRI has a high NPV for the absence of clinically significant prostate cancer. However, radiologists should be wary of assigning a PI-RADS 1 score when the PZs appear dark on mpMRI. Source of Funding: none