1. Federated Learning used for predicting outcomes in SARS-COV-2 patients
- Author
-
Matheus Ribeiro Furtado de Mendonça, Evan Leibovitz, Kristopher Kersten, Mona Flores, John Garrett, Baris Turkbey, Pablo F. Damasceno, Masoom A. Haider, Fred Kwon, Soo-Young Park, Chun-Nan Hsu, Keith J. Dreyer, Chien-Sung Tsai, Tatsuya Kodama, Daguang Xu, Min Kyu Kang, Tony Mazzulli, Andrew Feng, C. K. Lee, Isaac Yang, Deepi Bhatia, Marius George Linguraru, Byung Seok Kim, Aoxiao Zhong, Mohammad Adil, Pochuan Wang, Sheridan Reed, Peerapon Vateekul, Anas Z. Abidin, Sira Sriswa, J. D. Kaggie, Chia-Cheng Lee, Carlos Tor-Díez, Krishna Juluru, Xiang Li, Colin B. Compas, Xihong Lin, Jiahui Guan, Pierre Elnajjar, Yuhong Wen, Jung Gil Park, Hao-Hsin Shin, Amilcare Gentili, Weichung Wang, Colleen Ruan, Hui Ren, Hisashi Sasaki, Hitoshi Mori, Holger R. Roth, Felipe Kitamura, Chiu-Ling Lai, Jason C. Crane, Thomas M. Grist, Bradford J. Wood, Bernardo Bizzo, Dufan Wu, Jesse Tetreault, Andrew N. Priest, Mike Fralick, Anthony Costa, Andrew Liu, Benjamin S. Glicksberg, Griffin Lacey, Meena Abdelmaseeh, Thanyawee Puthanakit, Marcio Aloisio Bezerra Cavalcanti Rockenbach, Shelley McLeod, Pedro Mário Cruz e Silva, Chih-Hung Wang, Chia-Jung Hsu, Sarah E Hickman, Won Young Tak, Quanzheng Li, Yothin Rakvongthai, Watsamon Jantarabenjakul, Li-Chen Fu, Gustavo César de Antônio Corradi, Eric K. Oermann, Nicola Rieke, Varun Buch, Abood Quraini, Shuichi Kawano, Natalie Gangai, Yu Rim Lee, Krishna Nand Keshava Murthy, Christopher P. Hess, Stefan Gräf, Ittai Dayan, Stephanie Harmon, Jae Ho Sohn, Eddie Huang, Ahmed Harouni, Vitor de Lima Lavor, Sharmila Majumdar, Sheng Xu, Hirofumi Obinata, Fiona J. Gilbert, and Chin Lin
- Subjects
federated learning ,Computer science ,business.industry ,SARS-CoV-2 ,Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) ,Vital signs ,MEDLINE ,COVID-19 ,artificial intelligence ,Prognosis ,Data science ,Article ,Data sharing ,Machine Learning ,Data exchange ,Health care ,Outcome Assessment, Health Care ,Electronic Health Records ,Humans ,Set (psychology) ,business ,Anonymity - Abstract
Federated learning (FL) is a method used for training artificial intelligence models with data from multiple sources while maintaining data anonymity, thus removing many barriers to data sharing. Here we used data from 20 institutes across the globe to train a FL model, called EXAM (electronic medical record (EMR) chest X-ray AI model), that predicts the future oxygen requirements of symptomatic patients with COVID-19 using inputs of vital signs, laboratory data and chest X-rays. EXAM achieved an average area under the curve (AUC)0.92 for predicting outcomes at 24 and 72 h from the time of initial presentation to the emergency room, and it provided 16% improvement in average AUC measured across all participating sites and an average increase in generalizability of 38% when compared with models trained at a single site using that site's data. For prediction of mechanical ventilation treatment or death at 24 h at the largest independent test site, EXAM achieved a sensitivity of 0.950 and specificity of 0.882. In this study, FL facilitated rapid data science collaboration without data exchange and generated a model that generalized across heterogeneous, unharmonized datasets for prediction of clinical outcomes in patients with COVID-19, setting the stage for the broader use of FL in healthcare.
- Published
- 2021