TY - JOUR
T1 - A foundation model for generalizable disease detection from retinal images
AU - Zhou, Yukun
AU - Chia, Mark A.
AU - Wagner, Siegfried K.
AU - Ayhan, Murat S.
AU - Williamson, Dominic J.
AU - Struyven, Robbert R.
AU - Liu, Timing
AU - Xu, Moucheng
AU - Lozano, Mateo G.
AU - Woodward-Court, Peter
AU - Kihara, Yuka
AU - UK Biobank Eye and Vision Consortium
AU - Allen, Naomi
AU - Gallacher, John E.J.
AU - Littlejohns, Thomas
AU - Aslam, Tariq
AU - Bishop, Paul
AU - Black, Graeme
AU - Sergouniotis, Panagiotis
AU - Atan, Denize
AU - Dick, Andrew D.
AU - Williams, Cathy
AU - Barman, Sarah
AU - Barrett, J.
AU - Mackie, Sarah
AU - Braithwaite, Tasanee
AU - Carare, Roxana O.
AU - Ennis, Sarah
AU - Gibson, Jane
AU - Lotery, Andrew J.
AU - Self, Jay
AU - Chakravarthy, Usha
AU - Hogg, Ruth E.
AU - Paterson, Euan
AU - Woodside, Jayne
AU - Peto, Tunde
AU - Mckay, Gareth
AU - Mcguinness, Bernadette
AU - Foster, Paul J.
AU - Balaskas, Konstantinos
AU - Khawaja, Anthony P.
AU - Pontikos, Nikolas
AU - Rahi, Jugnoo S.
AU - Lascaratos, Gerassimos
AU - Patel, Praveen J.
AU - Chan, Michelle
AU - Chua, Sharon Y.L.
AU - Day, Alexander
AU - Desai, Parul
AU - Doney, Alexander
AU - Trucco, Emanuele
AU - Altmann, Andre
AU - Lee, Aaron Y.
AU - Topol, Eric J.
AU - Denniston, Alastair K.
AU - Alexander, Daniel C.
AU - Keane, Pearse A.
N1 - Funding Information:
We thank P. Rawlinson for project management, C. Green and L. Wickham for information governance expertise, and A. Wenban, S. St John-Green and M. Barnfield for information technology support. This work is supported by Engineering and Physical Sciences Research Council grant nos. EP/M020533/1, EP/R014019/1 and EP/V034537/1, as well as the NIHR UCLH Biomedical Research Centre. S.K.W. is supported by a Medical Research Council Clinical Research Training Fellowship (grant no. MR/TR000953/1). P.A.K. is supported by a Moorfields Eye Charity Career Development Award (grant no. R190028A) and a UK Research & Innovation Future Leaders Fellowship (grant no. MR/T019050/1). For the purpose of open access, the author has applied a Creative Commons Attribution (CC BY) licence to any Author Accepted Manuscript version arising.
Publisher Copyright:
© 2023, The Author(s).
PY - 2023/10/5
Y1 - 2023/10/5
N2 - Medical artificial intelligence (AI) offers great potential for recognizing signs of health conditions in retinal images and expediting the diagnosis of eye diseases and systemic disorders 1. However, the development of AI models requires substantial annotation and models are usually task-specific with limited generalizability to different clinical applications 2. Here, we present RETFound, a foundation model for retinal images that learns generalizable representations from unlabelled retinal images and provides a basis for label-efficient model adaptation in several applications. Specifically, RETFound is trained on 1.6 million unlabelled retinal images by means of self-supervised learning and then adapted to disease detection tasks with explicit labels. We show that adapted RETFound consistently outperforms several comparison models in the diagnosis and prognosis of sight-threatening eye diseases, as well as incident prediction of complex systemic disorders such as heart failure and myocardial infarction with fewer labelled data. RETFound provides a generalizable solution to improve model performance and alleviate the annotation workload of experts to enable broad clinical AI applications from retinal imaging.
AB - Medical artificial intelligence (AI) offers great potential for recognizing signs of health conditions in retinal images and expediting the diagnosis of eye diseases and systemic disorders 1. However, the development of AI models requires substantial annotation and models are usually task-specific with limited generalizability to different clinical applications 2. Here, we present RETFound, a foundation model for retinal images that learns generalizable representations from unlabelled retinal images and provides a basis for label-efficient model adaptation in several applications. Specifically, RETFound is trained on 1.6 million unlabelled retinal images by means of self-supervised learning and then adapted to disease detection tasks with explicit labels. We show that adapted RETFound consistently outperforms several comparison models in the diagnosis and prognosis of sight-threatening eye diseases, as well as incident prediction of complex systemic disorders such as heart failure and myocardial infarction with fewer labelled data. RETFound provides a generalizable solution to improve model performance and alleviate the annotation workload of experts to enable broad clinical AI applications from retinal imaging.
KW - Translational research
KW - Retinal diseases
KW - Prognosis
KW - Medical imaging
KW - Cardiovascular diseases
UR - http://www.scopus.com/inward/record.url?scp=85171296317&partnerID=8YFLogxK
U2 - 10.1038/s41586-023-06555-x
DO - 10.1038/s41586-023-06555-x
M3 - Article
C2 - 37704728
AN - SCOPUS:85171296317
SN - 0028-0836
VL - 622
SP - 156
EP - 163
JO - Nature
JF - Nature
IS - 7981
ER -
