Prediction of clinically significant prostate cancer after negative prostate biopsy: The current value of microscopic...

To assess the current ability of atypical small acinar proliferation (ASAP), multifocal high-grade prostatic intraepithelial neoplasia (mHGPIN), HGPIN with atypia (PINATYP) and other non-malignant lesions to predict clinically significant prostate cancer (csPCa) in repeat prostate biopsies.

This retrospective study analyzed 377 repeat prostate biopsies, carried out between 2.014 and 2.017, and excluding those with previous PCa or 5-alpha reductase inhibitors treatment. ASAP, mHGPIN, PINATYP, prostatic atrophy, prostatic hyperplastic atrophy, proliferative inflammatory atrophy (PIA), chronic prostatitis, acute prostatitis, or granulomatous prostatitis, were prospectively reported after 12-core transrectal ultrasound (TRUS) systematic negative previous biopsies. 3T-multiparametric magnetic resonance imaging (mpMRI) was performed previous repeat biopsies. At least 2-core TRUS targeted biopsies of Prostate Imaging-Reporting and Data Systemv2 lesions ≥3, and/or 12-core TRUS systematic biopsy were performed in repeat prostate biopsies. The main outcome measurements were csPCa detection, which was defined when the International Society of Uro-Pathology group grade >1 and avoided biopsies. After logistic regression analysis the most efficient model was selected, nomogram was designed with internal validation, and clinical utility was analyzed.

Normal benign tissue alone was present in less than 2% of previous negative biopsies. mHGPIN (39.7%), ASAP (4.3%) and PINATYP (3.7%) failed to predict csPCa risk in repeat biopsies. The finding of PIA (38.2%) associated with a decreased the risk of csPCa with an Odd ratio of 0.54 (95% confidence interval: 0.31-0.95), P= 0.031. The area under the curve, to predict csPCa, of mpMRI was 0.736, increasing up to 0.860 (95% confidence internal:0.82-0.90) when PSA density, age, digital rectal examination, and differential PSA between biopsies and PIA finding were integrated in a predictive model. At 6% threshold, more than 20% of repeat prostate biopsies were saved without missing csPCa.

Currently, mHGPIN in negative prostate biopsy seems not able to predict the risk of future csPCa. The low incidence of ASAP and PINATYP, in our series, did not allow us to draw conclusions. PIA finding associated with a reduced risk of csPCa, and it could be integrated in a useful based-mpMRI predictive nomogram.