Pd19-06 Assessing the Clinical Utility of Published Prostate Cancer Polygenic Risk Scores in a Large Biobank Dataset

You have accessJournal of UrologyProstate Cancer: Detection & Screening II (PD19)1 May 2024PD19-06 ASSESSING THE CLINICAL UTILITY OF PUBLISHED PROSTATE CANCER POLYGENIC RISK SCORES IN A LARGE BIOBANK DATASET Helen H. Sun, Udit Singhal, Frederick R. Schumacher, Erika Trapl, Lars G. Fritsche, Todd M. Morgan, and Randy A. Vince Helen H. SunHelen H. Sun , Udit SinghalUdit Singhal , Frederick R. SchumacherFrederick R. Schumacher , Erika TraplErika Trapl , Lars G. FritscheLars G. Fritsche , Todd M. MorganTodd M. Morgan , and Randy A. VinceRandy A. Vince View All Author Informationhttps://doi.org/10.1097/01.JU.0001009448.41537.64.06AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: Prostate cancer (PCa) screening aims to maximize the detection of aggressive cancer while minimizing overdiagnosis of indolent disease. Many polygenic risk scores (PRSs) have been developed to identify patients at risk of developing PCa based on genetic variants. However, creating PRSs which correlate with high risk disease remains challenging. Our objective was to evaluate the performance of published PCa PRSs to differentiate PCa cases from controls, and to identify patients at risk for aggressive PCa. METHODS: We compared 16 published PRSs using a case-control study design using data from the Michigan Genomics Initiative. Patients diagnosed with PCa were cross-referenced in the Michigan Urologic Surgery Improvement Collaborative to stratify cases by risk group using biopsy and prostatectomy data. We assessed and ranked each PRS on accuracy for identifying PCa cases, using area under the covariate-adjusted receiver operating curve (AAUC), and association with disease aggressiveness, using the Cochran Armitage trend test. RESULTS: There were 15,310 controls and 2,740 cases meeting inclusion criteria. Overall there was a positive association between all PRSs and PCa diagnosis, but predictive performance using AAUC varied between 0.52 95% CI 0.51-0.53 to 0.67 0.66-0.68. Evaluation of the top three performing PRSs demonstrated a nearly identical PRS distribution across risk categories as determined by prostate biopsy (Figure 1). On sensitivity analysis the top performing PRS had nearly identical distributions across risk categories as determined from prostatectomy data (Figure 2). Collectively these findings indicate a lack of association between PRS and disease aggressiveness. CONCLUSIONS: Currently available PCa PRSs have modest performance for identifying patients at increased risk of PCa. Additionally, current PRS constructs do no aid in distinguishing patients who are at the highest risk of developing aggressive prostate cancer. Existing PRSs require further refinement based on disease risk to enable clinical applications. Download PPTDownload PPT Source of Funding: National Institutes of Health P30CA046592, University of Michigan Precision Health Investigators Award U063790, National Science Foundation grant # DMS-1712933 © 2024 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 211Issue 5SMay 2024Page: e441 Advertisement Copyright & Permissions© 2024 by American Urological Association Education and Research, Inc.Metrics Author Information Helen H. Sun More articles by this author Udit Singhal More articles by this author Frederick R. Schumacher More articles by this author Erika Trapl More articles by this author Lars G. Fritsche More articles by this author Todd M. Morgan More articles by this author Randy A. Vince More articles by this author Expand All Advertisement PDF downloadLoading ...