{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,2,27]],"date-time":"2026-02-27T22:42:48Z","timestamp":1772232168655,"version":"3.50.1"},"reference-count":36,"publisher":"Oxford University Press (OUP)","issue":"11","license":[{"start":{"date-parts":[[2024,9,3]],"date-time":"2024-09-03T00:00:00Z","timestamp":1725321600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"Lions Sight Centre"},{"DOI":"10.13039\/501100001804","name":"Canada Research Chairs Program","doi-asserted-by":"crossref","id":[{"id":"10.13039\/501100001804","id-type":"DOI","asserted-by":"crossref"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2024,11,1]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:sec>\n                  <jats:title>Objectives<\/jats:title>\n                  <jats:p>The retinal age gap (RAG) is emerging as a potential biomarker for various diseases of the human body, yet its utility depends on machine learning models capable of accurately predicting biological retinal age from fundus images. However, training generalizable models is hindered by potential shortages of diverse training data. To overcome these obstacles, this work develops a novel and computationally efficient distributed learning framework for retinal age prediction.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Materials and Methods<\/jats:title>\n                  <jats:p>The proposed framework employs a memory-efficient 8-bit quantized version of RETFound, a cutting-edge foundation model for retinal image analysis, to extract features from fundus images. These features are then used to train an efficient linear regression head model for predicting retinal age. The framework explores federated learning (FL) as well as traveling model (TM) approaches for distributed training of the linear regression head. To evaluate this framework, we simulate a client network using fundus image data from the UK Biobank. Additionally, data from patients with type 1 diabetes from the UK Biobank and the Brazilian Multilabel Ophthalmological Dataset (BRSET) were utilized to explore the clinical utility of the developed methods.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Results<\/jats:title>\n                  <jats:p>Our findings reveal that the developed distributed learning framework achieves retinal age prediction performance on par with centralized methods, with FL and TM providing similar performance (mean absolute error of 3.57\u2009\u00b1\u20090.18 years for centralized learning, 3.60\u2009\u00b1\u20090.16 years for TM, and 3.63\u2009\u00b1\u20090.19 years for FL). Notably, the TM was found to converge with fewer local updates than FL. Moreover, patients with type 1 diabetes exhibited significantly higher RAG values than healthy controls in all models, for both the UK Biobank and BRSET datasets (P\u2009&amp;lt;\u2009.001).<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Discussion<\/jats:title>\n                  <jats:p>The high computational and memory efficiency of the developed distributed learning framework makes it well suited for resource-constrained environments.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Conclusion<\/jats:title>\n                  <jats:p>The capacity of this framework to integrate data from underrepresented populations for training of retinal age prediction models could significantly enhance the accessibility of the RAG as an important disease biomarker.<\/jats:p>\n               <\/jats:sec>","DOI":"10.1093\/jamia\/ocae220","type":"journal-article","created":{"date-parts":[[2024,9,3]],"date-time":"2024-09-03T21:56:14Z","timestamp":1725400574000},"page":"2550-2559","source":"Crossref","is-referenced-by-count":7,"title":["Foundation model-driven distributed learning for enhanced retinal age prediction"],"prefix":"10.1093","volume":"31","author":[{"given":"Christopher","family":"Nielsen","sequence":"first","affiliation":[{"name":"Department of Radiology, University of Calgary , Calgary, AB T2N 4N1,","place":["Canada"]},{"name":"Biomedical Engineering Graduate Program, University of Calgary , Calgary, AB T2N 4N1,","place":["Canada"]}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-7455-3383","authenticated-orcid":false,"given":"Raissa","family":"Souza","sequence":"additional","affiliation":[{"name":"Department of Radiology, University of Calgary , Calgary, AB T2N 4N1,","place":["Canada"]},{"name":"Biomedical Engineering Graduate Program, University of Calgary , Calgary, AB T2N 4N1,","place":["Canada"]},{"name":"Hotchkiss Brain Institute, University of Calgary , Calgary, AB T2N 4N1,","place":["Canada"]}]},{"given":"Matthias","family":"Wilms","sequence":"additional","affiliation":[{"name":"Department of Radiology, University of Calgary , Calgary, AB T2N 4N1,","place":["Canada"]},{"name":"Hotchkiss Brain Institute, University of Calgary , Calgary, AB T2N 4N1,","place":["Canada"]},{"name":"Alberta Children\u2019s Hospital Research Institute, University of Calgary , Calgary, AB T2N 4N1,","place":["Canada"]},{"name":"Department of Pediatrics, University of Calgary , Calgary, AB T2N 4N1,","place":["Canada"]},{"name":"Department of Community Health Sciences, University of Calgary , Calgary, AB T2N 4N1,","place":["Canada"]}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2556-3224","authenticated-orcid":false,"given":"Nils D","family":"Forkert","sequence":"additional","affiliation":[{"name":"Department of Radiology, University of Calgary , Calgary, AB T2N 4N1,","place":["Canada"]},{"name":"Hotchkiss Brain Institute, University of Calgary , Calgary, AB T2N 4N1,","place":["Canada"]},{"name":"Alberta Children\u2019s Hospital Research Institute, University of Calgary , Calgary, AB T2N 4N1,","place":["Canada"]},{"name":"Department of Clinical Neurosciences, University of Calgary , Calgary, AB T2N 4N1,","place":["Canada"]}]}],"member":"286","published-online":{"date-parts":[[2024,9,3]]},"reference":[{"issue":"3","key":"2024102107520798400_ocae220-B1","doi-asserted-by":"publisher","first-page":"237","DOI":"10.1111\/1753-0407.13364","article-title":"The association of retinal age gap with metabolic syndrome and inflammation","volume":"15","author":"Zhu","year":"2023","journal-title":"J Diabetes"},{"issue":"4","key":"2024102107520798400_ocae220-B2","doi-asserted-by":"publisher","first-page":"547","DOI":"10.1136\/bjophthalmol-2021-319807","article-title":"Retinal age gap as a predictive biomarker for mortality risk","volume":"107","author":"Zhu","year":"2023","journal-title":"Br J Ophthalmol"},{"issue":"11","key":"2024102107520798400_ocae220-B3","doi-asserted-by":"publisher","first-page":"3320","DOI":"10.1161\/STROKEAHA.122.038809","article-title":"Association of retinal age gap with arterial stiffness and incident cardiovascular disease","volume":"53","author":"Zhu","year":"2022","journal-title":"Stroke"},{"issue":"5","key":"2024102107520798400_ocae220-B4","doi-asserted-by":"publisher","first-page":"537.e1","DOI":"10.1053\/j.ajkd.2022.09.018","article-title":"Association of retinal age gap and risk of kidney failure: a UK Biobank study","volume":"81","author":"Zhang","year":"2023","journal-title":"Am J Kidney Dis"},{"issue":"3","key":"2024102107520798400_ocae220-B5","doi-asserted-by":"publisher","DOI":"10.1093\/ageing\/afac062","article-title":"Retinal age gap as a predictive biomarker of future risk of Parkinson\u2019s disease","volume":"51","author":"Hu","year":"2022","journal-title":"Age Ageing"},{"issue":"2","key":"2024102107520798400_ocae220-B6","doi-asserted-by":"publisher","first-page":"274","DOI":"10.1038\/s41433-021-01926-y","article-title":"Feasibility and clinical utility of handheld fundus cameras for retinal imaging","volume":"37","author":"Das","year":"2023","journal-title":"Eye"},{"issue":"S4","key":"2024102107520798400_ocae220-B7","doi-asserted-by":"publisher","first-page":"259","DOI":"10.1186\/s12859-020-03587-2","article-title":"Comparison of smartphone-based retinal imaging systems for diabetic retinopathy detection using deep learning","volume":"21","author":"Karakaya","year":"2020","journal-title":"BMC Bioinformatics"},{"issue":"1","key":"2024102107520798400_ocae220-B8","doi-asserted-by":"publisher","first-page":"195","DOI":"10.1186\/s12916-019-1426-2","article-title":"Key challenges for delivering clinical impact with artificial intelligence","volume":"17","author":"Kelly","year":"2019","journal-title":"BMC Med"},{"issue":"4","key":"2024102107520798400_ocae220-B9","doi-asserted-by":"publisher","first-page":"e003424","DOI":"10.1136\/bmjdrc-2023-003424","article-title":"Artificial intelligence for diabetic retinopathy in low-income and middle-income countries: a scoping review","volume":"11","author":"Cleland","year":"2023","journal-title":"BMJ Open Diab Res Care"},{"issue":"1","key":"2024102107520798400_ocae220-B10","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1038\/s43856-021-00028-w","article-title":"Mitigating bias in machine learning for medicine","volume":"1","author":"Vokinger","year":"2021","journal-title":"Commun Med"},{"issue":"7","key":"2024102107520798400_ocae220-B11","doi-asserted-by":"publisher","first-page":"1932","DOI":"10.1148\/rg.2020200040","article-title":"One algorithm may not fit all: how selection bias affects machine learning performance","volume":"40","author":"Yu","year":"2020","journal-title":"RadioGraphics"},{"issue":"4","key":"2024102107520798400_ocae220-B12","doi-asserted-by":"publisher","first-page":"184","DOI":"10.1200\/CCI.19.00047","article-title":"Systematic review of privacy-preserving distributed machine learning from federated databases in health care","volume":"4","author":"Zerka","year":"2020","journal-title":"JCO Clin Cancer Inform"},{"key":"2024102107520798400_ocae220-B13","doi-asserted-by":"publisher","author":"McMahan","year":"2023","DOI":"10.48550\/arXiv.1602.05629"},{"issue":"8","key":"2024102107520798400_ocae220-B14","doi-asserted-by":"publisher","first-page":"945","DOI":"10.1093\/jamia\/ocy017","article-title":"Distributed deep learning networks among institutions for medical imaging","volume":"25","author":"Chang","year":"2018","journal-title":"J Am Med Inform Assoc"},{"key":"2024102107520798400_ocae220-B15","doi-asserted-by":"publisher","author":"Souza","DOI":"10.1117\/12.2612728"},{"issue":"1","key":"2024102107520798400_ocae220-B16","doi-asserted-by":"publisher","first-page":"112","DOI":"10.1093\/jamia\/ocac204","article-title":"An analysis of the effects of limited training data in distributed learning scenarios for brain age prediction","volume":"30","author":"Souza","year":"2023","journal-title":"J Am Med Inform Assoc"},{"issue":"11","key":"2024102107520798400_ocae220-B17","doi-asserted-by":"publisher","first-page":"2835","DOI":"10.3390\/diagnostics12112835","article-title":"Federated learning in ocular imaging: current progress and future direction","volume":"12","author":"Nguyen","year":"2022","journal-title":"Diagnostics"},{"key":"2024102107520798400_ocae220-B18","doi-asserted-by":"publisher","first-page":"1259017","DOI":"10.3389\/fmed.2023.1259017","article-title":"Federated learning for diagnosis of age-related macular degeneration","volume":"10","author":"Gholami","year":"2023","journal-title":"Front Med"},{"issue":"2","key":"2024102107520798400_ocae220-B19","doi-asserted-by":"publisher","first-page":"100928","DOI":"10.1016\/j.patter.2024.100928","article-title":"FedEYE: a scalable and flexible end-to-end federated learning platform for ophthalmology","volume":"5","author":"Yan","year":"2024","journal-title":"Patterns"},{"issue":"7981","key":"2024102107520798400_ocae220-B20","doi-asserted-by":"publisher","first-page":"156","DOI":"10.1038\/s41586-023-06555-x","article-title":"A foundation model for generalizable disease detection from retinal images","volume":"622","author":"Zhou","year":"2023","journal-title":"Nature"},{"key":"2024102107520798400_ocae220-B21","author":"Li","year":"2022"},{"key":"2024102107520798400_ocae220-B22","author":"Alam","year":"2023"},{"key":"2024102107520798400_ocae220-B23","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1145\/3508396.3512869","author":"Wang","year":"2022"},{"issue":"7726","key":"2024102107520798400_ocae220-B24","doi-asserted-by":"publisher","first-page":"203","DOI":"10.1038\/s41586-018-0579-z","article-title":"The UK Biobank resource with deep phenotyping and genomic data","volume":"562","author":"Bycroft","year":"2018","journal-title":"Nature"},{"issue":"23","key":"2024102107520798400_ocae220-B25","doi-asserted-by":"publisher","first-page":"E215","DOI":"10.1161\/01.cir.101.23.e215","article-title":"PhysioBank, PhysioToolkit, and PhysioNet: components of a new research resource for complex physiologic signals","volume":"101","author":"Goldberger","year":"2000","journal-title":"Circulation"},{"key":"2024102107520798400_ocae220-B26","doi-asserted-by":"publisher","author":"Nakayama","year":"2024","DOI":"10.1101\/2024.01.23.24301660"},{"key":"2024102107520798400_ocae220-B27","doi-asserted-by":"publisher","first-page":"48","DOI":"10.1007\/978-3-030-32239-7_6","author":"Fu","year":"2019"},{"key":"2024102107520798400_ocae220-B28","year":"2024"},{"key":"2024102107520798400_ocae220-B29","doi-asserted-by":"publisher","author":"Sun","year":"2021","DOI":"10.48550\/arXiv.2104.11375"},{"key":"2024102107520798400_ocae220-B30","doi-asserted-by":"publisher","DOI":"10.3389\/frai.2024.1301997","article-title":"A multi-center distributed learning approach for Parkinson\u2019s disease classification using the traveling model paradigm","volume":"7","author":"Souza","year":"2024","journal-title":"Front Artif Intell"},{"key":"2024102107520798400_ocae220-B31","doi-asserted-by":"publisher","author":"Abay","year":"2020","DOI":"10.48550\/arXiv.2012.02447"},{"key":"2024102107520798400_ocae220-B32","author":"Ye","year":"2024"},{"issue":"1","key":"2024102107520798400_ocae220-B33","doi-asserted-by":"publisher","first-page":"70","DOI":"10.1186\/1477-7525-2-70","article-title":"The use of bootstrap methods for analysing health-related quality of life outcomes (particularly the SF-36)","volume":"2","author":"Walters","year":"2004","journal-title":"Health Qual Life Outcomes"},{"key":"2024102107520798400_ocae220-B34","doi-asserted-by":"publisher","first-page":"1061","DOI":"10.1162\/tacl_a_00413","article-title":"Compressing large-scale transformer-based models: a case study on BERT","volume":"9","author":"Ganesh","year":"2021","journal-title":"Trans Assoc Comput Linguist"},{"issue":"12","key":"2024102107520798400_ocae220-B35","doi-asserted-by":"publisher","first-page":"1330","DOI":"10.1038\/s41551-022-00898-y","article-title":"Shifting machine learning for healthcare from development to deployment and from models to data","volume":"6","author":"Zhang","year":"2022","journal-title":"Nat Biomed Eng"},{"issue":"5","key":"2024102107520798400_ocae220-B36","doi-asserted-by":"publisher","first-page":"e0127914","DOI":"10.1371\/journal.pone.0127914","article-title":"Suitability of UK Biobank retinal images for automatic analysis of morphometric properties of the vasculature","volume":"10","author":"MacGillivray","year":"2015","journal-title":"PLOS One"}],"container-title":["Journal of the American Medical Informatics Association"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/jamia\/article-pdf\/31\/11\/2550\/59813503\/ocae220.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/jamia\/article-pdf\/31\/11\/2550\/59813503\/ocae220.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,10,21]],"date-time":"2024-10-21T07:52:22Z","timestamp":1729497142000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/jamia\/article\/31\/11\/2550\/7748301"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,9,3]]},"references-count":36,"journal-issue":{"issue":"11","published-online":{"date-parts":[[2024,9,3]]},"published-print":{"date-parts":[[2024,11,1]]}},"URL":"https:\/\/doi.org\/10.1093\/jamia\/ocae220","relation":{},"ISSN":["1067-5027","1527-974X"],"issn-type":[{"value":"1067-5027","type":"print"},{"value":"1527-974X","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2024,11]]},"published":{"date-parts":[[2024,9,3]]}}}