An Application Based Nomogram to Predict Clinically Significant Diseases in Patients With an Elevated PSA and MRI Prior to Prostate Biopsy

Introduction and Objective: A nomogram-based approach can accurately predict the outcome of prostate biopsy (PBx) in biopsy naïve men with an elevated PSA using known traditional risk factors. However, MRI of the prostate has been shown to be of value before performing a PBx. This study aimed to predict clinically significant prostate cancer (csPC) on PBx by combining traditional clinical risk factors with the addition of MRI imaging findings in biopsy naïve men before PBx. An application was then created to improve accessibility to predicted outcomes and aid in decision-making. Methods: A cohort of 448 biopsy-naive men with elevated PSA, prostate MRI, and who underwent transrectal ultrasound-guided prostate biopsy were included. Age, race, PSA, prostate volume, PSA density, Prostate Imaging Reporting & Data System, digital rectal examination, and Gleason score were evaluated. CsPC was defined as Gleason Grade Group 2 or higher. Logistical models were created with the aforementioned factors. A nomogram with model performance using 10-fold cross-validation was then developed with subsequent application development. The calibration plot, ROC, and decision curve were developed to evaluate nomogram performance. Results: Of the 448 men, 36% were diagnosed with csPC. The mean age of the cohort was 65.97 ± 7.74 years, and 81% of the cohort were White. The mean PSA of the men was 12.60 ± 13.04 ng/mL. An interaction with age was found, so stratified models were developed above and below 67 years of age. The model achieved 82% AUC, with 61% sensitivity, 82% specificity, and 74% overall accuracy. Conclusions: The application-based nomogram combines the significant traditional risk factors for prostate cancer with findings on MRI to predict csPC in biopsy naïve men. By combining these variables, the nomogram provides a powerful predictive tool for guiding decision-making in the discovery of csPC. Our model outperformed traditional risk factors in predicting clinically significant cancer before PBx (82% AUC). The nomogram is especially specific (82%) and effective in ruling out the need for biopsy in low-risk patients. In addition, the development of an accurate and user-friendly application increases usefulness to urologists and patients alike.