Mp37-08 urinary microbiome as a biomarker for prostate cancer

You have accessJournal of UrologyCME1 May 2022MP37-08 URINARY MICROBIOME AS A BIOMARKER FOR PROSTATE CANCER Michael Liss, John Lee, James White, Teresa Johnson-Pais, Zhao Lai, Dean Troyer, Robin Leach, and Brian Wickes Michael LissMichael Liss More articles by this author , John LeeJohn Lee More articles by this author , James WhiteJames White More articles by this author , Teresa Johnson-PaisTeresa Johnson-Pais More articles by this author , Zhao LaiZhao Lai More articles by this author , Dean TroyerDean Troyer More articles by this author , Robin LeachRobin Leach More articles by this author , and Brian WickesBrian Wickes More articles by this author View All Author Informationhttps://doi.org/10.1097/JU.0000000000002591.08AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: To develop a diagnostic biomarker development strategy to investigate potential microbial biomarkers for prostate cancer. METHODS: We designed a prospective training cohort with a case-control design to create a microbiome-based biomarker. In the case group, we enrolled men that went to surgery. We collected urine before prostate biopsy, prostatectomy tissue, and urine after prostatectomy in the same patients to perform paired filtering to identify the most important microbiota (42 samples, n=14). In the control group, we collected urine from normal men with very low risk of prostate cancer (PSA<1.0, screened for more than 10 years) before and after prostate exam (44 samples, n=22). We used a shotgun metagenomic sequencing at 6M and single end 150 bp using Illumina HiSeq platform. We tested the biomarker in a prospectively enrolled validation cohort undergoing prostate biopsy (n=45). Our primary outcome clinically significant prostate cancer (ISUP Group 2). We additionally validated the 24 microbial taxa in TCGA prostate tissue and a proteomics cohort. RESULTS: Using our training cohort, we identified 24 candidate microbial taxa after performing paired comparisons and contaminant removal strategies. Using all microbes, the median area under the receiver operator curve (AUC) using 100 iterations of random subsampling was 0.87 (95% CI 0.81-0.92). We subsequently used a prostate biopsy validation cohort to test the best performing model. We identified eight microbial taxa that had an AUC of 0.84 (95% CI 0.74-0.85) that outperformed clinical risk factors including PSA (Figure 1). In the figure, we show the performance of the biomarker in the validation cohort (VAL) performance in high grade cancer (HG=ISUP>1), any cancer, and showed no improvement in AUC performance when clinical factors were added. We cross-referenced our 24 taxa to a published 187 microbiome taxa obtained from The Cancer Genome Atlas (TCGA) noting 19 (19/24, 76%) matches. We also investigated the 24 taxa within our existing proteomic data derived from mass spectrometry based proteomic analysis a previous cohort of 87 men on active surveillance confirming 8 of the 24 taxa. CONCLUSIONS: Using a unique cohort to filter bacterial signatures for prostate cancer, we identified eight bacterial taxa associated with cancer that performed well in a validation cohort. Source of Funding: Los Padres Foundation © 2022 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 207Issue Supplement 5May 2022Page: e611 Advertisement Copyright & Permissions© 2022 by American Urological Association Education and Research, Inc.MetricsAuthor Information Michael Liss More articles by this author John Lee More articles by this author James White More articles by this author Teresa Johnson-Pais More articles by this author Zhao Lai More articles by this author Dean Troyer More articles by this author Robin Leach More articles by this author Brian Wickes More articles by this author Expand All Advertisement PDF DownloadLoading ...