Federated learning for predicting clinical outcomes in patients with COVID-19

被引：406

作者：

Dayan, Ittai ^{[1
,2
]}

Roth, Holger R. ^{[3
]}

Zhong, Aoxiao ^{[4
,5
]}

Harouni, Ahmed ^{[3
]}

Gentili, Amilcare ^{[6
]}

Abidin, Anas Z. ^{[3
]}

Liu, Andrew ^{[3
]}

Costa, Anthony Beardsworth ^{[7
]}

Wood, Bradford J. ^{[8
,9
]}

Tsai, Chien-Sung ^{[10
]}

Wang, Chih-Hung ^{[11
,12
]}

Hsu, Chun-Nan ^{[13
]}

Lee, C. K. ^{[3
]}

Ruan, Peiying ^{[3
]}

Xu, Daguang ^{[3
]}

Wu, Dufan ^{[4
]}

Huang, Eddie ^{[3
]}

Kitamura, Felipe Campos ^{[14
]}

Lacey, Griffin ^{[3
]}

de Antonio Corradi, Gustavo Cesar ^{[14
]}

Nino, Gustavo ^{[15
]}

Shin, Hao-Hsin ^{[16
]}

Obinata, Hirofumi ^{[17
]}

Ren, Hui ^{[4
]}

Crane, Jason C. ^{[18
]}

Tetreault, Jesse ^{[3
]}

Guan, Jiahui ^{[3
]}

Garrett, John W. ^{[19
,20
]}

Kaggie, Joshua D. ^{[21
]}

Park, Jung Gil ^{[22
]}

Dreyer, Keith ^{[1
,2
,23
]}

Juluru, Krishna ^{[16
]}

Kersten, Kristopher ^{[3
]}

Rockenbach, Marcio Aloisio Bezerra Cavalcanti ^{[23
]}

Linguraru, Marius George ^{[24
,25
]}

Haider, Masoom A. ^{[26
,27
]}

AbdelMaseeh, Meena ^{[27
]}

Rieke, Nicola ^{[3
]}

Damasceno, Pablo F. ^{[18
]}

Silva, Pedro Mario Cruz E. ^{[3
]}

Wang, Pochuan ^{[28
,29
]}

Xu, Sheng ^{[8
,9
]}

Kawano, Shuichi ^{[17
]}

Sriswasdi, Sira ^{[30
,31
]}

Park, Soo Young ^{[32
]}

Grist, Thomas M. ^{[19
,20
,33
]}

Buch, Varun ^{[23
]}

Jantarabenjakul, Watsamon ^{[34
,35
]}

Wang, Weichung ^{[28
,29
]}

Tak, Won Young ^{[32
]}

机构：

[1] MGH Radiol, Boston, MA USA

[2] Harvard Med Sch, Boston, MA 02115 USA

[3] NVIDIA, Santa Clara, CA 95051 USA

[4] Harvard Med Sch, Massachusetts Gen Hosp, Dept Radiol, Ctr Adv Med Comp & Anal, Boston, MA 02115 USA

[5] Harvard Univ, Sch Engn & Appl Sci, Boston, MA 02115 USA

[6] San Diego VA Hlth Care Syst, San Diego, CA USA

[7] Icahn Sch Med Mt Sinai, Dept Neurosurg, New York, NY 10029 USA

[8] NIH, Radiol & Imaging Sci Clin Ctr, Bldg 10, Bethesda, MD 20892 USA

[9] NCI, NIH, Bethesda, MD 20892 USA

[10] Triserv Gen Hosp, Natl Def Med Ctr, Dept Surg, Div Cardiovasc Surg, Taipei, Taiwan

[11] Triserv Gen Hosp, Natl Def Med Ctr, Dept Otolaryngol Head & Neck Surg, Taipei, Taiwan

[12] Natl Def Med Ctr, Grad Inst Med Sci, Taipei, Taiwan

[13] Univ Calif San Diego, Ctr Res Biol Syst, San Diego, CA 92103 USA

[14] DasaInova, Diagnost Amer SA, Barueri, Brazil

[15] Childrens Natl Hosp, Div Pediat Pulm & Sleep Med, Washington, DC USA

[16] Mem Sloan Kettering Canc Ctr, 1275 York Ave, New York, NY 10021 USA

[17] Self Def Forces Cent Hosp, Tokyo, Japan

[18] Univ Calif San Francisco, Ctr Intelligent Imaging, Dept Radiol & Biomed Imaging, San Francisco, CA 94143 USA

[19] Univ Wisconsin, Dept Radiol, Sch Med & Publ Hlth, Madison, WI 53706 USA

[20] Univ Wisconsin, Dept Med Phys, Sch Med & Publ Hlth, 1530 Med Sci Ctr, Madison, WI 53706 USA

[21] Univ Cambridge, NIHR Cambridge Biomed Resource Ctr, Dept Radiol, Cambridge, England

[22] Yeungnam Univ, Dept Internal Med, Coll Med, Daegu, South Korea

[23] Massachusetts Gen Brigham, Ctr Clin Data Sci, Boston, MA USA

[24] Childrens Natl Hosp, Sheikh Zayed Inst Pediat Surg Innovat, Washington, DC USA

[25] George Washington Univ, Dept Radiol, Sch Med & Hlth Sci, Washington, DC USA

[26] Univ Toronto, Joint Dept Med Imaging, Sinai Hlth Syst, Toronto, ON, Canada

[27] Lunenfeld Tanenbaum Res Inst, Toronto, ON, Canada

[28] Natl Taiwan Univ, MeDA Lab Inst Appl Math Sci, Taipei, Taiwan

[29] Natl Taiwan Univ, Dept Comp Sci & Informat Engn, Taipei, Taiwan

[30] Chulalongkorn Univ, Fac Med, Res Affairs, Bangkok, Thailand

[31] Chulalongkorn Univ, Fac Med, Ctr Artificial Intelligence Med, Bangkok, Thailand

[32] Kyungpook Natl Univ, Sch Med, Dept Internal Med, Daegu, South Korea

[33] Univ Wisconsin, Dept Biomed Engn, Sch Med & Publ Hlth, Madison, WI USA

[34] Chulalongkorn Univ, Fac Med, Dept Pediat, Bangkok, Thailand

[35] King Chulalongkorn Mem Hosp, Thai Red Cross Emerging Infect Dis Clin Ctr, Bangkok, Thailand

[36] Harvard Univ, Harvard TH Chan Sch Publ Hlth, Boston, MA 02115 USA

[37] Cambridge Univ Hosp, NIHR Cambridge Biomed Resource Ctr, Dept Radiol, Cambridge, England

[38] NIH, Dept Radiol & Imaging Sci, Bldg 10, Bethesda, MD 20892 USA

[39] Icahn Sch Med Mt Sinai, Hass Plattner Inst Digital Hlth Mt Sinai, New York, NY 10029 USA

[40] Icahn Sch Med Mt Sinai, Dept Genet & Genom Sci, New York, NY 10029 USA

[41] Catholic Univ Daegu, Dept Internal Med, Sch Med, Daegu, South Korea

[42] Triserv Gen Hosp, Natl Def Med Ctr, Planning & Management Off, Taipei, Taiwan

[43] Natl Def Med Ctr, Sch Med, Taipei, Taiwan

[44] Natl Def Med Ctr, Sch Publ Hlth, Taipei, Taiwan

[45] Natl Def Med Ctr, Grad Inst Life Sci, Taipei, Taiwan

[46] Natl Hlth Insurance Adm, Med Review & Pharmaceut Benefits Div, Taipei, Taiwan

[47] NYU Grossman Sch Med, Dept Neurosurg, New York, NY USA

[48] Natl Taiwan Univ, MOST NTU All Vista Healthcare Ctr, Ctr Artificial Intelligence & Adv Robot, Taipei, Taiwan

[49] Sinai Hlth Syst, Div Gen Internal Med & Geriatr Fralick, Toronto, ON, Canada

[50] Chulalongkorn Univ, Fac Engn, Dept Comp Engn, Bangkok, Thailand

来源：

NATURE MEDICINE | 2021年 / 27卷 / 10期

基金：

英国工程与自然科学研究理事会;

关键词：

PRIVACY; CHALLENGES; MODELS;

D O I：

10.1038/s41591-021-01506-3

中图分类号：

Q5 [生物化学]; Q7 [分子生物学];

学科分类号：

071010 ; 081704 ;

摘要：

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.

引用

页码：1735 / +

页数：25

共 73 条

[41] Data sharing for novel coronavirus (COVID-19) [J].

Moorthy, Vasee ;

Restrepo, Ana Maria Henao ;

Preziosi, Marie-Pierre ;

Swaminathan, Soumya .

BULLETIN OF THE WORLD HEALTH ORGANIZATION, 2020, 98 (03) :150-150

[42] What the radiologist should know about artificial intelligence - an ESR white paper [J].

Neri, Emanuele ;

deSouza, Nandita ;

Brady, Adrian .

INSIGHTS INTO IMAGING, 2019, 10 (01)

[43] Pediatric lung imaging features of COVID-19: A systematic review and meta-analysis [J].

Nino, Gustavo ;

Zember, Jonathan ;

Sanchez-Jacob, Ramon ;

Gutierrez, Maria J. ;

Sharma, Karun ;

Linguraru, Marius George .

PEDIATRIC PULMONOLOGY, 2021, 56 (01) :252-263

[44] Artificial intelligence in medical imaging: threat or opportunity? Radiologists again at the forefront of innovation in medicine [J].

Pesapane F. ;

Codari M. ;

Sardanelli F. .

European Radiology Experimental, 2 (1)

[45] The state of OA: a large-scale analysis of the prevalence and impact of Open Access articles [J].

Piwowar, Heather ;

Priem, Jason ;

Lariviere, Vincent ;

Alperin, Juan Pablo ;

Matthias, Lisa ;

Norlander, Bree ;

Farley, Ashley ;

West, Jevin ;

Haustein, Stefanie .

PEERJ, 2018, 6

[46] Privacy in the age of medical big data [J].

Price, W. Nicholson ;

Cohen, I. Glenn .

NATURE MEDICINE, 2019, 25 (01) :37-43

[47] SCOR: A secure international informatics infrastructure to investigate COVID-19 [J].

Raisaro, J. L. ;

Marino, Francesco ;

Troncoso-Pastoriza, Juan ;

Beau-Lejdstrom, Raphaelle ;

Bellazzi, Riccardo ;

Murphy, Robert ;

Bernstam, Elmer, V ;

Wang, Henry ;

Bucalo, Mauro ;

Chen, Yong ;

Gottlieb, Assaf ;

Harmanci, Arif ;

Kim, Miran ;

Kim, Yejin ;

Klann, Jeffrey ;

Klersy, Catherine ;

Malin, Bradley A. ;

Mean, Marie ;

Prasser, Fabian ;

Scudeller, Luigia ;

Torkamani, Ali ;

Vaucher, Julien ;

Puppala, Mamta ;

Wong, Stephen T. C. ;

Frenkel-Morgenstern, Milana ;

Xu, Hua ;

Musa, Maiyaki ;

Habib, Abdulrazaq G. ;

Cohen, Trevor ;

Wilcox, Adam ;

Salihu, Hamisu M. ;

Sofia, Heidi ;

Jiang, Xiaoqian ;

Hubaux, J. P. .

JOURNAL OF THE AMERICAN MEDICAL INFORMATICS ASSOCIATION, 2020, 27 (11) :1721-1726

[48] Emergency Radiology During the COVID-19 Pandemic: The Canadian Association of Radiologists Recommendations for Practice [J].

Redmond, Ciaran E. ;

Nicolaou, Savvas ;

Berger, Ferco H. ;

Sheikh, Adnan M. ;

Patlas, Michael N. .

CANADIAN ASSOCIATION OF RADIOLOGISTS JOURNAL-JOURNAL DE L ASSOCIATION CANADIENNE DES RADIOLOGISTES, 2020, 71 (04) :425-430

[49]

Remedios S. W., 2020, FEDERATED GRADIENT A

[50] The future of digital health with federated learning [J].

Rieke, Nicola ;

Hancox, Jonny ;

Li, Wenqi ;

Milletari, Fausto ;

Roth, Holger R. ;

Albarqouni, Shadi ;

Bakas, Spyridon ;

Galtier, Mathieu N. ;

Landman, Bennett A. ;

Maier-Hein, Klaus ;

Ourselin, Sebastien ;

Sheller, Micah ;

Summers, Ronald M. ;

Trask, Andrew ;

Xu, Daguang ;

Baust, Maximilian ;

Cardoso, M. Jorge .

NPJ DIGITAL MEDICINE, 2020, 3 (01)

← 1 2 3 4 5 6 7 8 →