PD57-08 TRACKED BIOPSY DATA IMPROVES PREDICTION OF PROSTATE TUMOR SIZE AND STAGE

Abstract

You have accessJournal of UrologyProstate Cancer: Staging II (PD57)1 Apr 2020PD57-08 TRACKED BIOPSY DATA IMPROVES PREDICTION OF PROSTATE TUMOR SIZE AND STAGE Alan Priester*, Adam Kinnaird, Merdie Delfin, Danielle Barsa, Anthony Sisk, and Leonard Marks Alan Priester*Alan Priester* More articles by this author , Adam KinnairdAdam Kinnaird More articles by this author , Merdie DelfinMerdie Delfin More articles by this author , Danielle BarsaDanielle Barsa More articles by this author , Anthony SiskAnthony Sisk More articles by this author , and Leonard MarksLeonard Marks More articles by this author View All Author Informationhttps://doi.org/10.1097/JU.0000000000000967.08AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: Pre-operative assessment of prostate cancer (PCa) volume and pathologic stage are critical for risk stratification and disease management. Herein, we report on improved techniques for estimating tumor size and stage through analysis of the spatial distribution of tracked biopsy cores. METHODS: Subjects were 209 consecutive, treatment-naïve patients who received MRI-US fusion biopsy followed by radical prostatectomy and whole mount (WM) slide preparation at UCLA. Final Gleason Group was 1 in 7% of patients, 2 in 59%, 3 in 23%, and 4-5 in 11%. For each case, the minimum (Tm) and maximum (TM) tumor volume was estimated from the coordinates of positive and negative cores recorded by a MRI-US fusion system (Fig 1A). The diameter and volume of Tmand TMwere recorded along with other metrics: number of positive cores, prostate specific antigen (PSA), and maximum cancer core length (MCCL). Logistic regressions were performed to predict maximum tumor diameter and extraprostatic extension (EPE) on WM. Wilcoxon signed rank tests were used to assess statistical significance. RESULTS: Biopsy-defined minimum tumor (Tm) volume was the feature most strongly correlated with WM tumor diameter (Spearman R = 0.40). Tm diameter had a median absolute error of 7 mm, significantly lower than the median MCCL error of 15 mm (p < 0.01). For prediction of tumor diameter, the first logistic regression model (Fig 1B) incorporated PSA, positive core number, Tm/ TM measures, and prostate volume. The model had Spearman R = 0.63 and median absolute error = 4.0 mm. It was more accurate than any individual feature and a control model which excluded features derived from fusion biopsy (p = 0.01). For prediction of EPE (78/209), the second logistic regression model (Fig 1C) included PSA, Gleason Grade, MCCL, and PCa-positive MR target diameter. In the second model, Spearman R = 0.47 and AUC = 0.78, predicting EPE more accurately than any individual feature and the control model (p = 0.01). CONCLUSIONS: Tracked biopsy features (i.e. the spatial coordinates of positive and negative cores), when used alongside other clinical parameters, improved prediction of PCa volume and stage. These metrics, once validated in prospective trials, may improve PCa risk classification and disease management. Source of Funding: None © 2020 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 203Issue Supplement 4April 2020Page: e1196-e1197 Advertisement Copyright & Permissions© 2020 by American Urological Association Education and Research, Inc.MetricsAuthor Information Alan Priester* More articles by this author Adam Kinnaird More articles by this author Merdie Delfin More articles by this author Danielle Barsa More articles by this author Anthony Sisk More articles by this author Leonard Marks More articles by this author Expand All Advertisement PDF downloadLoading ...