You have accessJournal of UrologyPlenary Session II – Late Breaking Abstracts1 Apr 2015PII-LBA8 USE OF LINEAR DISCRIMINANT ANALYSIS IN A URINE-BASED TEST FOR BLADDER CANCER DIAGNOSIS Ellen Wallace, Kathleen E. Mach, Leena McCann, Lai Yi Mandy Sin, Ruchika Mohan, Malini Satya, Huilin Wei, Jun Zhang, Chris Lykke, Russell Higuchi, and Joseph C. Liao Ellen WallaceEllen Wallace More articles by this author , Kathleen E. MachKathleen E. Mach More articles by this author , Leena McCannLeena McCann More articles by this author , Lai Yi Mandy SinLai Yi Mandy Sin More articles by this author , Ruchika MohanRuchika Mohan More articles by this author , Malini SatyaMalini Satya More articles by this author , Huilin WeiHuilin Wei More articles by this author , Jun ZhangJun Zhang More articles by this author , Chris LykkeChris Lykke More articles by this author , Russell HiguchiRussell Higuchi More articles by this author , and Joseph C. LiaoJoseph C. Liao More articles by this author View All Author Informationhttps://doi.org/10.1016/j.juro.2015.03.088AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookTwitterLinked InEmail INTRODUCTION AND OBJECTIVES We aim to develop a rapid, accurate and non-invasive molecular assay for bladder cancer diagnosis for use on the GeneXpert® (Cepheid, Sunnyvale CA) platform with sample-in-answer-out capability. Linear discriminant analysis (LDA) was used both to help choose the best markers for the assay, and to optimize the assay's ability to discriminate negative and positive samples. METHODS Appropriate local IRB approvals were obtained. The GeneXpert® Instrument Systems automate and integrate sample purification and multiplex RT-qPCR detection. The systems utilize single-use cartridges that hold the reagents and conduct the RT-qPCR process. The assay requires less than two minutes of hands-on time and results are available in approximately 90 minutes. A training set of 497 urine samples collected from 18 sites was tested for the expression levels of ABL1, CRH, IGF2, ANXA10, KRT20, AR, PIK3CA, UPK1B, UPK2 and MGEA5 mRNA. The ten markers were analyzed using stepwise logistic regression to choose the best 5 marker signature and develop an LDA equation to determine test outcome. An assay was developed for the mRNA detection of a 5 marker signature that includes a sample adequacy control and an internal RT-qPCR control. The assay with integrated LDA equation was tested with a set of 243 urine samples from 7 sites to validate the signature. RESULTS Backwards stepwise regression was used with the training set of samples and data with 10 markers to reduce the model and obtain the best 5-marker multivariable equation to determine test results. The resulting LDA equation using Ct values for ABL1, CRH, IGF2, ANXA10 and UPK1B was shown to yield sensitivities of 98% (79/81) for high grade cancer (papillary and CIS combined), 78% (31/40) for low grade and specificity of 83% (78/94) with patients undergoing hematuria workup. An assay for detection of ABL1, CRH, IGF2, ANXA10 and UPK1B mRNA was formulated and optimized for use on the GeneXpert®. A validation set of 243 urine samples was tested with the new assay to evaluate equivalence relative to the prior training set results. The new assay's sensitivity was 92% (12/13) for high-grade cancer (papillary and CIS combined) and 86% (12/14) for low-grade. The specificity was 89% (24/27) for patients undergoing workup for hematuria and 96% (144/150) with healthy volunteers. A prospective study with the newly developed test is underway. CONCLUSIONS We have developed a simple to use and accurate test for the detection of bladder cancer. The assay utilizes a proven diagnostic platform of GeneXpert® with automated sample preparation providing test results within 90 minutes. © 2015 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 193Issue 4SApril 2015Page: e498 Advertisement Copyright & Permissions© 2015 by American Urological Association Education and Research, Inc.MetricsAuthor Information Ellen Wallace More articles by this author Kathleen E. Mach More articles by this author Leena McCann More articles by this author Lai Yi Mandy Sin More articles by this author Ruchika Mohan More articles by this author Malini Satya More articles by this author Huilin Wei More articles by this author Jun Zhang More articles by this author Chris Lykke More articles by this author Russell Higuchi More articles by this author Joseph C. Liao More articles by this author Expand All Advertisement Advertisement PDF downloadLoading ...