MP42-03 HUMAN VS MACHINE: COMPARISON OF MANUAL VERSUS AUTOMATED SOFTWARE CT MEASUREMENTS OF STONE PHANTOMS

Abstract

You have accessJournal of UrologyImaging/Radiology: Uroradiology II (MP42)1 Apr 2020MP42-03 HUMAN VS MACHINE: COMPARISON OF MANUAL VERSUS AUTOMATED SOFTWARE CT MEASUREMENTS OF STONE PHANTOMS Andrew M. Harris, Morgan N. Cash, Leslie M. Peard*, Mohamed M. Issa, James T. Lee, Jason R. Bylund, Amul A. Bhalodi, and John R. Bell Andrew M. HarrisAndrew M. Harris More articles by this author , Morgan N. CashMorgan N. Cash More articles by this author , Leslie M. Peard*Leslie M. Peard* More articles by this author , Mohamed M. IssaMohamed M. Issa More articles by this author , James T. LeeJames T. Lee More articles by this author , Jason R. BylundJason R. Bylund More articles by this author , Amul A. BhalodiAmul A. Bhalodi More articles by this author , and John R. BellJohn R. Bell More articles by this author View All Author Informationhttps://doi.org/10.1097/JU.0000000000000891.03AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: Given the significance of stone size management of urolithiasis, obtaining accurate and precise measurements is paramount. We compared software and manual stone length and volume measurements to actual measurements of stone phantoms. METHODS: Seventeen unique stone phantoms were created in a variety of shapes and sizes using BegoStone Plus. Stone length, width, and height were measured with electronic calipers. Stone volume was measured using volume of displacement in water. Stone phantoms were placed in saline and imaged with computed tomography (CT). The CT images were uploaded into post-processing software, which analyzed the stones for maximum length and volume. Two urologists assessed the imaged stones for length, width and height and were blinded to the software measurements. Stone volume was calculated using the ellipsoid formula 0.52 x length x width x height. RESULTS: The stone phantoms ranged in length from 5.5mm to 63mm with volumes of 40mm3 to 16,000mm3. The software had the closest approximation to the actual stone length with a median percent error of 6.2% (IQR 3.1%-9.4%). Readers 1 and 2 had median percent errors of 5.7% (IQR 3.7%-14.1%) and 7.8% (IQR 2.1%-17.2%) respectively. Reader 1 and 2 both differed from the software with median percent differences of 16% (IQR 8.6%-22.4%) and 17.1% (IQR 5.2%-28.4%). The readers had a median percent difference from each other of 4.1% (IQR 3%-6.6%). None of the length measurements by the software or the human readers differed significantly from the actual stone phantoms (P>0.610). The percent errors with regards to volume for the software, reader 1 and reader 2 compared to actual stone volume were 20% (IQR 14.2%-30%), 18% (IQR 5.2%-39.8%), 24.1% (IQR 10.2%-33.2%), respectively. Reader 1 and 2 differed from the software with median percent differences of 33.1% (IQR 21.7%-59.5%) and 26.3% (IQR 15%-48%), respectively. The readers had a median percent difference from each other of 17.3% (IQR 7.5%-32.5%). None of the measurements by the software or the human readers differed significantly from the actual stone phantom volume (P>0.697). CONCLUSIONS: Software measurements approximated the actual measurements for the stone phantoms while eliminating interobserver variability. Using post-processing software to measure urinary stones appears to be as accurate as human readers, but with greater precision. Source of Funding: None © 2020 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 203Issue Supplement 4April 2020Page: e603-e604 Advertisement Copyright & Permissions© 2020 by American Urological Association Education and Research, Inc.MetricsAuthor Information Andrew M. Harris More articles by this author Morgan N. Cash More articles by this author Leslie M. Peard* More articles by this author Mohamed M. Issa More articles by this author James T. Lee More articles by this author Jason R. Bylund More articles by this author Amul A. Bhalodi More articles by this author John R. Bell More articles by this author Expand All Advertisement PDF downloadLoading ...