Negative magnetic resonance imaging cannot be used to omit an initial prostate biopsy - An ambispective study

Introduction: Up to 40% of patients with suspected prostate cancer (PCa) have a negative prebiopsy magnetic resonance imaging (nMRI), and up to 15% of them may have clinically significant PCa (csPCa). The ability to predict the presence of csPCa despite nMRI may help avoid unnecessary biopsies. We aimed to determine the negative predictive value (NPV) of mpMRI, the influence of MRI reporting patterns in clinical practice, and the factors that might predict csPCa among men with an nMRI. Methodology: In an IRB-approved, ambispective study, men who underwent prostate biopsy from 2016 to 2023 and had a prebiopsy MRI, were included to determine the presence of csPCa. The reporting patterns of institutional and noninstitutional MRI were evaluated. Age, digital rectal examination (DRE) findings, prostate specific antigen (PSA), PSA density (PSAD), and MRI reports were evaluated for their ability to predict csPCa in men with nMRI. Results: 1660 patients who underwent prostate biopsy were assessed for eligibility, and 685 patients were enrolled in the study. The median age, PSA and PSAD were 60 years, 11.63 ng/ml and 0.23 ng/ml/ cm3, respectively. 62 (9%) men had an nMRI, among which csPCa, non-csPCa, and negative biopsy were found in 34%, 5%, and 61% of men, respectively. 61% had an institutional MRI, while 39% had a noninstitutional MRI. The sensitivity and NPV of any MRI for csPCa were 93% and 66%, respectively, which improved to 96% and 81% for institutional MRI. Univariate and multivariate analyses showed abnormal DRE and PSAD >= 0.25 ng/ml/cc as predictive factors for csPCa in men with an nMRI. Conclusion: 34% of men with negative MRIs were found to harbor csPCa on prostate biopsy. The NPV of institutional MRI was higher than for noninstitutional MRI. Men with an abnormal DRE or PSAD >= 0.25 ng/ml/cc had a higher incidence of csPCa despite an nMRI. (c) 2024 The Asian Pacific Prostate Society. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).