Analysis of false-positive biopsy results of PIRADS 4 lesions in multiparametric magnetic resonance imaging of the prostate: Experience from a single nonacademic center using cognitive fusion.

Abstract

321 Background: Multiparametric magnetic resonance imaging (mpMRI) of the prostate has been increasingly used in the algorithm of the initial diagnosis of patients suspected to harbor prostate cancer (PCa) with the aim of improving the selection of patients who really need a biopsy. Yet patients with PI-RADS 5 lesions almost always have cancer found in their prostates, the same is not true for patients with PI-RADS 4. In this study we sought to describe pathological findings in target biopsies of PI-RADS4, and to identify clinical data that could predict those patients with benign findings. Methods: A retrospective study was conducted in a data bank collected prospectively from December 2015 through April 2022 in a single nonacademic center. The whole series consist of 547 patients, all of whom had a mpMRI of the prostate (reports as per PI-RADS version 2) followed by cognitive fusion prostate biopsy performed by a single, experienced radiologist. All biopsy specimens were read by a specialist in urological pathology. Out of these, 259 had PI-RADS 4 lesion, and 83 had PI-RADS 5. Results: We found a false positive rate of 29% and 3.7% for any cancer in PI-RADS 4 and 5 lesions, respectively. Different histologic patterns were observed among target biopsies. At multivariate analysis, PIRADS lesion ≤ 6 mm and previous negative biopsy were independent predictors of false positive PI-RADS4 lesions. The 76 cases showed normal histology (n = 22), a combination of chronic inflammation, reactive epithelium, and glandular atrophy (n = 18), glandular atrophy (n = 9), stromal proliferation consistent with nodular hyperplasia (n =7), a combination of chronic inflammation, reactive epithelium and glandular atrophy (n = 6), chronic inflammation (n = 4) and HGPIN (n = 10). The small number (n=3) of false PI-RADS5 lesions precluded further analyses. Conclusions: Benign findings are common in PI-RADS4 lesions and most of them did not show obvious glandular or stromal hypercellularity as expected in hyperplastic nodules. Size ≤ 6 mm and previous negative biopsy predict higher probability of false positive results in patients with PI-RADS 4 lesions.[Table: see text]