MP55-14 DEEP LEARNING ENABLED PREDICTION OF 5-YEAR SURVIVAL IN PEDIATRIC GENITOURINARY RHABDOMYOSARCOMA

Abstract

You have accessJournal of UrologyPediatric Urology V (MP55)1 Sep 2021MP55-14 DEEP LEARNING ENABLED PREDICTION OF 5-YEAR SURVIVAL IN PEDIATRIC GENITOURINARY RHABDOMYOSARCOMA Hriday Bhambhvani, Alvaro Zamora, Kyla Velaer, Daniel Greenberg, and Kunj Sheth Hriday BhambhvaniHriday Bhambhvani More articles by this author , Alvaro ZamoraAlvaro Zamora More articles by this author , Kyla VelaerKyla Velaer More articles by this author , Daniel GreenbergDaniel Greenberg More articles by this author , and Kunj ShethKunj Sheth More articles by this author View All Author Informationhttps://doi.org/10.1097/JU.0000000000002085.14AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: Genitourinary rhabdomyosarcoma (GU-RMS) is a rare, pediatric malignancy originating from embryonic mesenchyme. Current approaches to prognostication utilize nomograms that rely upon conventional statistical methods such as proportional hazards models and have suboptimal predictive ability. Given the success of deep learning approaches in other specialties, we sought to develop and compare deep learning models with Cox proportional hazards (CPH) models for the prediction of 5-year survival in pediatric GU-RMS patients. METHODS: Patients less than 20 years of age with GU were identified within the Surveillance, Epidemiology, and End Results (SEER) database (1998-2011). Deep neural networks (DNN) were trained and tested on an 80/20 split of the dataset in a 5-fold cross-validated fashion. Multivariable CPH models were developed in parallel. The primary outcomes were 5-year overall survival (OS) and disease-specific survival (DSS). Variables used for prediction were age, sex, race, primary site, histology, degree of tumor extension, tumor size, receipt of surgery, and receipt of radiation. Receiver operating characteristic curve analysis was conducted, and DNN models were tested for calibration. RESULTS: 277 patients were included. The area under the curve (AUC) for the DNN models was 0.93 for OS and 0.91 for DSS (Figure 1). AUC for the CPH models was 0.82 for OS and 0.84 for DSS. The DNN models were well-calibrated: OS model (slope=1.02, intercept=-0.06) and DSS model (slope=0.79, intercept=0.21) (Figure 2). CONCLUSIONS: In this study, a deep learning-based model demonstrated excellent performance, superior to that of CPH models, in the prediction of pediatric GU-RMS survival. Deep learning approaches may enable improved prognostication for patients with rare cancers. Source of Funding: None © 2021 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 206Issue Supplement 3September 2021Page: e964-e964 Advertisement Copyright & Permissions© 2021 by American Urological Association Education and Research, Inc.MetricsAuthor Information Hriday Bhambhvani More articles by this author Alvaro Zamora More articles by this author Kyla Velaer More articles by this author Daniel Greenberg More articles by this author Kunj Sheth More articles by this author Expand All Advertisement Loading ...

Ähnliche Arbeiten