You have accessJournal of UrologyPediatrics: Imaging (Genital & Urinary Tract)/Infections and Vesicoureteral Reflux1 Apr 2012614 VUR INDEX (VURX): A NOVEL TOOL TO PREDICT VESICOURETERAL REFLUX (VUR) RESOLUTION BASED ON THE FIRST VCUG IN CHILDREN LESS THAN TWO YEARS OF AGE Arun Srinivasan, Lindsay Herrel, Edwin Smith, Jane Share, David Blews, Hal Scherz, and Andrew Kirsch Arun SrinivasanArun Srinivasan Atlanta, GA More articles by this author , Lindsay HerrelLindsay Herrel Atlanta, GA More articles by this author , Edwin SmithEdwin Smith Atlanta, GA More articles by this author , Jane ShareJane Share Atlanta, GA More articles by this author , David BlewsDavid Blews Atlanta, GA More articles by this author , Hal ScherzHal Scherz Atlanta, GA More articles by this author , and Andrew KirschAndrew Kirsch Atlanta, GA More articles by this author View All Author Informationhttps://doi.org/10.1016/j.juro.2012.02.692AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookTwitterLinked InEmail INTRODUCTION AND OBJECTIVES The usefulness of International Reflux Grading System is limited by evidence based on small sample sizes, poor inter-observer concordance, and poor predictive values for moderate grade VUR. We aim to provide an indexing tool to more reliably predict resolution rates by improving concordance rates among physicians and improve our ability to predict VUR resolution based on the first VCUG. METHODS We identified all patients aged < 2 years diagnosed with VUR between 2006-2010. We excluded patients with secondary VUR, ureteroceles, and megaureters. Patient age, indication for VCUG, ultrasound and renal scan findings as available, gender, anomalies, follow up data and eventual outcomes were collected. Two blinded radiologists reviewed the VCUGs and recorded the phase of VCUG when reflux (timing of reflux) was first noted. We categorized timing of reflux as early filling, end filling and voiding phase. Univariate and multivariate analysis was performed to identify factors that best predict VUR resolution. A Multilayer Perceptron neural network model was used to construct the VUR index (VURx) with normalized importance for each variable. The VURx was then validated by ROC and Kaplan-Meier curves. RESULTS 229 patients fit our patient criteria. 75% of the VCUGs were performed for UTI at a mean age of 0.42 years. By multivariate logistic regression analysis, absence of high grade reflux (OR 5.53), absence of anomalies (OR3.54) and timing of reflux (OR5.37) were found to be predictors of spontaneous resolution. The neural network model found the same variables to be predictive of reflux resolution with high normalized importance. A VURx was constructed using these three variables with a minimum value of 1 signifying a high resolution rate (92.3%) and maximum value of 5 signifying low rate of resolution (8.3%). This model was tested using resolution graphs (p<0.05), ROC curves (AUC 0.82) and Kaplan-Meier curves (p<0.05) all signifying high correlation. CONCLUSIONS VUR index – a score based on presence of high grade reflux, anomalies and timing of VUR is highly predictive of spontaneous resolution in children aged < 2 years. Patients with a low VURx may not require further studies by virtue of their high resolution rate while those with high VURx may be better served by earlier intervention. © 2012 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 187Issue 4SApril 2012Page: e250 Advertisement Copyright & Permissions© 2012 by American Urological Association Education and Research, Inc.MetricsAuthor Information Arun Srinivasan Atlanta, GA More articles by this author Lindsay Herrel Atlanta, GA More articles by this author Edwin Smith Atlanta, GA More articles by this author Jane Share Atlanta, GA More articles by this author David Blews Atlanta, GA More articles by this author Hal Scherz Atlanta, GA More articles by this author Andrew Kirsch Atlanta, GA More articles by this author Expand All Advertisement Advertisement PDF downloadLoading ...