Novel clinical risk calculator for improving cancer predictability of mpMRI fusion biopsy in prostates

• Published in: International urology and nephrology Vol. 56; no. 9; pp. 2851 - 2860
• Main Authors: Bruccoliere, Anthony, Tran, Vivie, Helo, Naseem, Awal, Abdul, Stroever, Stephanie, de Riese, Werner T. W.
• Format: Journal Article
• Language: English
• Published: Dordrecht Springer Netherlands 01.09.2024; Springer Nature B.V
• Subjects: Aged; Biopsy; Decision making; Humans; Image-Guided Biopsy - methods; Information processing; Lesions; Magnetic resonance imaging; Male; Medicine; Medicine & Public Health; Middle Aged; Multiparametric Magnetic Resonance Imaging; Nephrology; Pilot Projects; Prediction models; Predictive Value of Tests; Prostate cancer; Prostatic Neoplasms - pathology; Regression analysis; Retrospective Studies; Risk Assessment; Urology; Urology - Original Paper; Prostate cancer; Risk calculator; MRI-fusion biopsy
• ISSN: 1573-2584; 0301-1623; 1573-2584
• DOI: 10.1007/s11255-024-04037-1

Purpose Prostate Imaging-Reporting and Data System (PI-RADS) assists in evaluating lesions on multiparametric magnetic resonance imaging (mpMRI), but there are still ongoing efforts in improving the predictive value for the presence of clinically significant PCa (csPCa) with a Gleason grade group ≥ 2 on Fusion-Biopsy. This pilot study intends to propose an easily implementable method for augmenting predictability of csPCa for PI-RADS. Methods A cohort of 151 consecutive patients underwent mpMRI Fusion and random US Biopsy as a result of having at least one PI-RADS lesion grade 3–5 between January 1, 2019 and December 31, 2022. A single radiologist reads all films in this study applying PI-RADS V2. Results Of the 151 consecutive patients, 49 had a highest lesion of PI-RADS 3, 82 had a highest lesion of PI-RADS 4, and 20 had a highest lesion of PI-RADS 5. For each respective group, 12, 42, and 18 patients had proven csPCa. Two predictive models for csPCa were created by employing a logistical regression with parameters readily available to providers. The models had an AUC of 0.8133 and 0.8206, indicating promising effective models. Conclusion PI-RADS classification has relevant predictability problems for grades 3 and 4. By applying the presented risk calculators, patients with PI-RADS 3 and 4 are better stratified, and thus, a significant number of patients can be spared biopsies with potential complications, such as infection and bleeding. The presented predictive models may be a valuable diagnostic tool, adding additional information in the clinical decision-making process for biopsies.