{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,26]],"date-time":"2026-03-26T12:43:29Z","timestamp":1774529009258,"version":"3.50.1"},"reference-count":24,"publisher":"Oxford University Press (OUP)","issue":"8","license":[{"start":{"date-parts":[[2018,3,29]],"date-time":"2018-03-29T00:00:00Z","timestamp":1522281600000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by-nc\/4.0\/"}],"funder":[{"name":"National Institutes of Health Blueprint for Neuroscience Research","award":["T90DA022759\/R90DA023427"],"award-info":[{"award-number":["T90DA022759\/R90DA023427"]}]},{"DOI":"10.13039\/100000002","name":"NIH","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100000002","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000070","name":"National Institute of Biomedical Imaging and Bioengineering","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100000070","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000002","name":"National Institutes of Health","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100000002","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000002","name":"National Institutes of Health","doi-asserted-by":"publisher","award":["U01CA154601"],"award-info":[{"award-number":["U01CA154601"]}],"id":[{"id":"10.13039\/100000002","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000002","name":"National Institutes of Health","doi-asserted-by":"publisher","award":["U24CA180927"],"award-info":[{"award-number":["U24CA180927"]}],"id":[{"id":"10.13039\/100000002","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000002","name":"National Institutes of Health","doi-asserted-by":"publisher","award":["U24CA180918"],"award-info":[{"award-number":["U24CA180918"]}],"id":[{"id":"10.13039\/100000002","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000002","name":"National Institutes of Health","doi-asserted-by":"publisher","award":["U01CA190214"],"award-info":[{"award-number":["U01CA190214"]}],"id":[{"id":"10.13039\/100000002","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000002","name":"National Institutes of Health","doi-asserted-by":"publisher","award":["U01CA187947"],"award-info":[{"award-number":["U01CA187947"]}],"id":[{"id":"10.13039\/100000002","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2018,8,1]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:sec>\n                  <jats:title>Objective<\/jats:title>\n                  <jats:p>Deep learning has become a promising approach for automated support for clinical diagnosis. When medical data samples are limited, collaboration among multiple institutions is necessary to achieve high algorithm performance. However, sharing patient data often has limitations due to technical, legal, or ethical concerns. In this study, we propose methods of distributing deep learning models as an attractive alternative to sharing patient data.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Methods<\/jats:title>\n                  <jats:p>We simulate the distribution of deep learning models across 4 institutions using various training heuristics and compare the results with a deep learning model trained on centrally hosted patient data. The training heuristics investigated include ensembling single institution models, single weight transfer, and cyclical weight transfer. We evaluated these approaches for image classification in 3 independent image collections (retinal fundus photos, mammography, and ImageNet).<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Results<\/jats:title>\n                  <jats:p>We find that cyclical weight transfer resulted in a performance that was comparable to that of centrally hosted patient data. We also found that there is an improvement in the performance of cyclical weight transfer heuristic with a high frequency of weight transfer.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Conclusions<\/jats:title>\n                  <jats:p>We show that distributing deep learning models is an effective alternative to sharing patient data. This finding has implications for any collaborative deep learning study.<\/jats:p>\n               <\/jats:sec>","DOI":"10.1093\/jamia\/ocy017","type":"journal-article","created":{"date-parts":[[2018,2,15]],"date-time":"2018-02-15T20:10:52Z","timestamp":1518725452000},"page":"945-954","source":"Crossref","is-referenced-by-count":282,"title":["Distributed deep learning networks among institutions for medical imaging"],"prefix":"10.1093","volume":"25","author":[{"given":"Ken","family":"Chang","sequence":"first","affiliation":[{"name":"Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts General Hospital, Charlestown, MA, 02129, USA"}]},{"given":"Niranjan","family":"Balachandar","sequence":"additional","affiliation":[{"name":"Department of Radiology and Biomedical Data Science, Stanford University, Palo Alto, CA, 94305, USA"}]},{"given":"Carson","family":"Lam","sequence":"additional","affiliation":[{"name":"Department of Radiology and Biomedical Data Science, Stanford University, Palo Alto, CA, 94305, USA"}]},{"given":"Darvin","family":"Yi","sequence":"additional","affiliation":[{"name":"Department of Radiology and Biomedical Data Science, Stanford University, Palo Alto, CA, 94305, USA"}]},{"given":"James","family":"Brown","sequence":"additional","affiliation":[{"name":"Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts General Hospital, Charlestown, MA, 02129, USA"}]},{"given":"Andrew","family":"Beers","sequence":"additional","affiliation":[{"name":"Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts General Hospital, Charlestown, MA, 02129, USA"}]},{"given":"Bruce","family":"Rosen","sequence":"additional","affiliation":[{"name":"Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts General Hospital, Charlestown, MA, 02129, USA"}]},{"given":"Daniel L","family":"Rubin","sequence":"additional","affiliation":[{"name":"Department of Radiology and Biomedical Data Science, Stanford University, Palo Alto, CA, 94305, USA"}]},{"given":"Jayashree","family":"Kalpathy-Cramer","sequence":"additional","affiliation":[{"name":"Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts General Hospital, Charlestown, MA, 02129, USA"},{"name":"MGH and BWH Center for Clinical Data Science, Massachusetts General Hospital, Boston, MA, 02114, USA"}]}],"member":"286","published-online":{"date-parts":[[2018,3,29]]},"reference":[{"key":"2020110612240817900_ocy017-B1","first-page":"1","article-title":"ImageNet classification with deep convolutional neural networks","author":"Krizhevsky","year":"2012","journal-title":"Adv Neural Inf Process Syst"},{"key":"2020110612240817900_ocy017-B2","doi-asserted-by":"crossref","first-page":"82","DOI":"10.1109\/MSP.2012.2205597","article-title":"Deep neural networks for acoustic modeling in speech recognition: The shared views of four research groups","volume":"29","author":"Hinton","year":"2012","journal-title":"IEEE Signal Process Mag"},{"key":"2020110612240817900_ocy017-B3","first-page":"160","article-title":"A unified architecture for natural language processing: deep neural networks with multitask learning","author":"Collobert","year":"2008","journal-title":"Proc 25th Int Conf Mach Learn"},{"key":"2020110612240817900_ocy017-B4","doi-asserted-by":"crossref","first-page":"2402","DOI":"10.1001\/jama.2016.17216","article-title":"Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal fundus photographs","volume":"316","author":"Gulshan","year":"2016","journal-title":"JAMA"},{"key":"2020110612240817900_ocy017-B5","doi-asserted-by":"crossref","first-page":"115","DOI":"10.1038\/nature21056","article-title":"Dermatologist-level classification of skin cancer with deep neural networks","volume":"542","author":"Esteva","year":"2017","journal-title":"Nature"},{"key":"2020110612240817900_ocy017-B6","doi-asserted-by":"crossref","DOI":"10.1158\/1078-0432.CCR-17-2236","article-title":"Residual Convolutional Neural Network for the Determination of IDH Status in Low- and High-Grade Gliomas from MR Imaging","author":"Chang","year":"2018","journal-title":"Clin Cancer Res"},{"key":"2020110612240817900_ocy017-B7","doi-asserted-by":"crossref","first-page":"26094","DOI":"10.1038\/srep26094","article-title":"Deep patient: an unsupervised representation to predict the future of patients from the electronic health records","volume":"6","author":"Miotto","year":"2016","journal-title":"Sci Rep"},{"key":"2020110612240817900_ocy017-B8","doi-asserted-by":"crossref","first-page":"436","DOI":"10.1038\/nature14539","article-title":"Deep learning","volume":"521","author":"LeCun","year":"2015","journal-title":"Nature"},{"key":"2020110612240817900_ocy017-B9","doi-asserted-by":"crossref","first-page":"10","DOI":"10.1016\/j.neuroimage.2017.03.027","article-title":"Multi-center machine learning in imaging psychiatry: a meta-model approach","volume":"155","author":"Dluho\u0161","year":"2017","journal-title":"Neuroimage"},{"key":"2020110612240817900_ocy017-B10","article-title":"It's all in the timing: calibrating temporal penalties for biomedical data sharing","author":"Xia","journal-title":"J Am Med Informatics Assoc"},{"key":"2020110612240817900_ocy017-B11","first-page":"1","article-title":"Experiments on parallel training of deep neural network using model averaging","author":"Su","year":"2015","journal-title":"ArXiv"},{"key":"2020110612240817900_ocy017-B12","first-page":"1","article-title":"Large scale distributed deep networks","author":"Dean","year":"2012","journal-title":"NIPS 2012 Neural Inf Process Syst"},{"key":"2020110612240817900_ocy017-B13","volume-title":"Diabetic Retinopathy Detection","author":"Kaggle","year":"2015"},{"key":"2020110612240817900_ocy017-B14","doi-asserted-by":"crossref","first-page":"178","DOI":"10.1016\/j.media.2017.04.012","article-title":"Deep image mining for diabetic retinopathy screening","volume":"39","author":"Quellec","year":"2017","journal-title":"Med Image Anal"},{"key":"2020110612240817900_ocy017-B15","volume-title":"Kaggle Diabetic Retinopathy Detection Competition Report","author":"Graham","year":"2015"},{"key":"2020110612240817900_ocy017-B16","doi-asserted-by":"crossref","first-page":"770","DOI":"10.1109\/CVPR.2016.90","article-title":"Deep residual learning for image recognition","author":"He","year":"2016","journal-title":"2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). IEEE"},{"key":"2020110612240817900_ocy017-B17","first-page":"1","article-title":"Keras: Deep Learning library for Theano and TensorFlow","author":"Chollet","year":"2015","journal-title":"GitHub Repos"},{"key":"2020110612240817900_ocy017-B18","first-page":"19","article-title":"Theano: a Python framework for fast computation of mathematical expressions","author":"Theano Development Team","year":"2016","journal-title":"arXiv e-prints"},{"key":"2020110612240817900_ocy017-B19","article-title":"Understanding the difficulty of training deep feedforward neural networks","author":"Glorot","journal-title":"Proc Int Conf Artif Intell Stat (AISTATS\u201910) Soc Artif Intell Stat"},{"key":"2020110612240817900_ocy017-B20","volume-title":"DDSM: Digital Database for Screening Mammography","author":"USF Digital Mammography"},{"key":"2020110612240817900_ocy017-B21","doi-asserted-by":"crossref","first-page":"211","DOI":"10.1007\/s11263-015-0816-y","article-title":"ImageNet large scale visual recognition challenge","volume":"115","author":"Russakovsky","year":"2015","journal-title":"Int J Comput Vis"},{"key":"2020110612240817900_ocy017-B22","doi-asserted-by":"crossref","first-page":"993","DOI":"10.1109\/34.58871","article-title":"Neural network ensembles","volume":"12","author":"Hansen","year":"1990","journal-title":"IEEE Trans Pattern Anal Mach Intell"},{"key":"2020110612240817900_ocy017-B23","doi-asserted-by":"crossref","first-page":"1345","DOI":"10.1109\/TKDE.2009.191","article-title":"A survey on transfer learning","volume":"22","author":"Pan","year":"2010","journal-title":"IEEE Trans Knowl Data Eng"},{"key":"2020110612240817900_ocy017-B24","doi-asserted-by":"crossref","DOI":"10.1088\/1361-6560\/aa93d4","article-title":"Multi-task transfer learning deep convolutional neural network: application to computer-aided diagnosis of breast cancer on mammograms","author":"Samala","year":"2017","journal-title":"Phys Med Biol"}],"container-title":["Journal of the American Medical Informatics Association"],"original-title":[],"language":"en","link":[{"URL":"http:\/\/academic.oup.com\/jamia\/article-pdf\/25\/8\/945\/34150259\/ocy017.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"http:\/\/academic.oup.com\/jamia\/article-pdf\/25\/8\/945\/34150259\/ocy017.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2020,11,6]],"date-time":"2020-11-06T17:31:55Z","timestamp":1604683915000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/jamia\/article\/25\/8\/945\/4956468"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2018,3,29]]},"references-count":24,"journal-issue":{"issue":"8","published-online":{"date-parts":[[2018,3,29]]},"published-print":{"date-parts":[[2018,8,1]]}},"URL":"https:\/\/doi.org\/10.1093\/jamia\/ocy017","relation":{},"ISSN":["1067-5027","1527-974X"],"issn-type":[{"value":"1067-5027","type":"print"},{"value":"1527-974X","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2018,8]]},"published":{"date-parts":[[2018,3,29]]}}}