1. Federated learning for predicting clinical outcomes in patients with COVID-19
- Author
-
Jiahui Guan, Krishna Juluru, Yothin Rakvongthai, Benjamin S. Glicksberg, Watsamon Jantarabenjakul, Li-Chen Fu, Mike Fralick, Anthony Costa, Quanzheng Li, Andrew Feng, Eric K. Oermann, Joshua D. Kaggie, Xihong Lin, Pedro Mário Cruz e Silva, Deepeksha Bhatia, Byung Seok Kim, Hitoshi Mori, Pablo F. Damasceno, Peiying Ruan, Yuhong Wen, Hao-Hsin Shin, Amilcare Gentili, Weichung Wang, Chiu-Ling Lai, Jason C. Crane, Andrew N. Priest, Soo-Young Park, Peerapon Vateekul, Matheus Ribeiro Furtado de Mendonça, Gustavo César de Antônio Corradi, Griffin Lacey, Meena AbdelMaseeh, Yu Rim Lee, Tatsuya Kodama, Pierre Elnajjar, Krishna Nand Keshava Murthy, Xiang Li, Evan Leibovitz, Vitor Lavor, Christopher P. Hess, Colin B. Compas, Stefan Gräf, Masoom A. Haider, Daguang Xu, Nicola Rieke, Thanyawee Puthanakit, Sarah E Hickman, Hui Ren, Marcio Aloisio Bezerra Cavalcanti Rockenbach, Jung Gil Park, Jesse Tetreault, Hisashi Sasaki, Min Kyu Kang, Won Young Tak, Chun-Nan Hsu, Fiona J. Gilbert, Chin Lin, Varun Buch, Felipe Kitamura, Tony Mazzulli, Eddie Huang, Abood Quraini, Shelley McLeod, Young Joon Kwon, Gustavo Nino, Dufan Wu, Chien-Sung Tsai, Mona Flores, Baris Turkbey, Sira Sriswasdi, Pochuan Wang, Mohammad Adil, Aoxiao Zhong, Chih-Hung Wang, Sheng Xu, C. K. Lee, Isaac Yang, Marius George Linguraru, Holger R. Roth, Chia-Jung Hsu, Anas Z. Abidin, Thomas M. Grist, Hirofumi Obinata, Sheridan Reed, Andrew Liu, Ahmed Harouni, Natalie Gangai, Ittai Dayan, Kristopher Kersten, Stephanie Harmon, Jae Ho Sohn, John Garrett, Bradford J. Wood, Sharmila Majumdar, Bernardo Bizzo, Shuichi Kawano, Keith J. Dreyer, Carlos Tor-Díez, and Chia-Cheng Lee
- Subjects
medicine.medical_specialty ,Information privacy ,Coronavirus disease 2019 (COVID-19) ,Computer science ,business.industry ,Vital signs ,General Medicine ,General Biochemistry, Genetics and Molecular Biology ,Data sharing ,Data exchange ,Health care ,medicine ,Generalizability theory ,Medical physics ,In patient ,business - 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. Federated learning, a method for training artificial intelligence algorithms that protects data privacy, was used to predict future oxygen requirements of symptomatic patients with COVID-19 using data from 20 different institutes across the globe.
- Published
- 2021
- Full Text
- View/download PDF