Login to Access Video or Poster Abstract: MP38-14
Sources of Funding: This work was carried out with support from the St Vincent&[prime]s Hospital (Melbourne) Research Endowment Fund.

Introduction

Multiparametric MRI (mpMRI), with enhanced visualization of tissue within the prostate, has been shown to improve the detection of prostate cancer with biopsy. Our study aims to determine the accuracy of mpMRI by precisely mapping the location and description of mpMRI lesions with the grade and location of positive biopsy cores.

Methods

Retrospective analysis across two Australian tertiary referral centres between 2012 and 2016 was performed. Patients underwent 3-Tesla mpMRI followed by prostate biopsy ≤3 months later, by cognitive or software fusion, with either transrectal or transperineal ultrasound guided techniques. _x000D_ Lesions on mpMRI were correlated to the location of positive biopsy cores after targeted biopsy, with a positive match being defined as both in the same or neighbouring location as per standard extended biopsy protocol. A suspicious mpMRI lesion was defined as having a Prostate Imaging Reporting and Data System (PI-RADS) score ≥4, whereas clinically significant cancer was defined as cancer with Gleason grade ≥3+4. Descriptive statistics and two-way t-tests were performed with STATA® 14. _x000D_

Results

Of 355 patients, 49.3% (175/355) had a suspicious mpMRI lesion and 47.9% (170/355) of the cohort had significant cancer on biopsy overall. There was moderate sensitivity (67.1%), negative predictive value (71.3%), specificity (69.7%) and positive predictive value (60.9%) for the mpMRI detection of significant cancer in the same or neighbouring location at biopsy. Of the 56 patients with significant cancer but non-suspicious mpMRI (PI-RADS <4), 66.1% (37/56) were Gleason 3+4 and 32.1% (18/56) were matched to a PI-RADS 3 lesion._x000D_ For Gleason grade ≥4+4 cancers, 90.4% (47/52) had a corresponding mpMRI lesion with PI-RADS score ≥4. Examining all patients with significant cancer at biopsy matching to a suspicious mpMRI lesion, 84.2% (96/114) were the highest Gleason grade found on biopsy._x000D_ _x000D_ The mean size of suspicious mpMRI lesions with significant cancer on biopsy was significantly larger than those with Gleason 3+3 cancer or no cancer at biopsy (16.7 mm vs.13.9 mm respectively; p=0.03). _x000D_

Conclusions

When precisely mapping the location of mpMRI lesions to positive prostate biopsy cores for significant cancer, a moderate sensitivity and positive predictive value was found suggesting that concomitant systematic biopsy cannot be abandoned. PI-RADS ≥4 lesions had excellent prediction of high-grade cancers at biopsy. Larger mpMRI lesions may yield more clinically significant cancers on biopsy.

Funding

This work was carried out with support from the St Vincent&[prime]s Hospital (Melbourne) Research Endowment Fund.