{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,24]],"date-time":"2026-10-24T16:10:00Z","timestamp":1792858200409,"version":"4.5.0"},"reference-count":37,"publisher":"BMJ","content-domain":{"domain":["bmj.com"],"crossmark-restriction":true},"short-container-title":["Br J Ophthalmol"],"accepted":{"date-parts":[[2026,4,24]]},"abstract":"<jats:sec>\n                    <jats:title>Aims<\/jats:title>\n                    <jats:p>This study evaluated the use of ophthalmic foundation deep-learning models with cross-modal transfer learning to classify multiple diseases on optical coherence tomography angiography (OCTA) with limited sample size.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>The OCTA-500 dataset (n=500 subjects) was split into an 85% training\/validation set for fivefold cross-validation and a 15% held-out test set. Superficial and deep projections from OCTA were combined using intermediate fusion. Outcomes were multi-disease classification of normal, diabetic retinopathy, age-related macular degeneration and \u2018other\u2019. Transfer-learning from colour fundus photography was used to overcome the limitation of small sample sizes. Vision-Transformer-VisionFM and Vision-Transformer-RETFound domain-specific foundation models with cross-modal transfer learning were evaluated. Comparison was made with Vision-Transformer-ImageNet, a non-domain-specific model. Performance was evaluated using accuracy, F1-score, precision, recall and area under the receiver operating characteristic curve. Saliency maps were also computed.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>VisionFM with cross-modal transfer learning outperformed ImageNet in classifying different diseases on OCTA (accuracy: 0.8133\u00b10.0470\u2009vs 0.7600\u00b10.0502). RETFound with cross-modal transfer learning outperformed ImageNet in classifying different diseases on OCTA (accuracy: 0.8000\u00b10.0507\u2009vs 0.7600\u00b10.0521). Similar conclusions were reached with other performance metrics. Saliency maps from VisionFM and RETFound yielded attention patterns that localised pathology to relevant retinal structures on superficial and deep projections from OCTA, comparing favourably with those from ImageNet models.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusions<\/jats:title>\n                    <jats:p>Retinal foundation models with cross-modal transfer learning enable accurate multi-class classification using OCTA data, which had small sample size. Results from domain-specific foundation models compared favourably with a non-domain-specific model. Saliency analysis showed attention patterns of pathology localised to anatomically relevant retinal structures.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1136\/bjo-2025-329249","type":"journal-article","created":{"date-parts":[[2026,5,5]],"date-time":"2026-05-05T17:20:32Z","timestamp":1778001632000},"page":"bjo-2025-329249","update-policy":"https:\/\/doi.org\/10.1136\/crossmarkpolicy","source":"Crossref","is-referenced-by-count":3,"title":["Classification of retinal diseases based on optical coherence tomography angiography using cross-modal transfer learning of domain-specific foundation AI models"],"prefix":"10.1136","author":[{"given":"Parim","family":"Shah","sequence":"first","affiliation":[{"name":"Albert Einstein College of Medicine, Bronx, New York, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hamza Ahmed","family":"Farah","sequence":"additional","affiliation":[{"name":"Albert Einstein College of Medicine, Bronx, New York, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Daniel J","family":"Wisotsky","sequence":"additional","affiliation":[{"name":"Albert Einstein College of Medicine, Bronx, New York, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Paarth","family":"Nawani","sequence":"additional","affiliation":[{"name":"Albert Einstein College of Medicine, Bronx, New York, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Surya","family":"Hariharan","sequence":"additional","affiliation":[{"name":"Albert Einstein College of Medicine, Bronx, New York, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Arush","family":"Satasia","sequence":"additional","affiliation":[{"name":"Albert Einstein College of Medicine, Bronx, New York, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Eric R","family":"Muir","sequence":"additional","affiliation":[{"name":"Departments of Radiology and Ophthalmology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6314-8932","authenticated-orcid":false,"given":"Umar K","family":"Mian","sequence":"additional","affiliation":[{"name":"Ophthalmology, Montefiore Hospital and Medical Center, Bronx, New York, USA"},{"name":"Department of Neurology, Albert Einstein College of Medicine, Bronx, New York, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6403-2827","authenticated-orcid":false,"given":"Tim Q","family":"Duong","sequence":"additional","affiliation":[{"name":"Albert Einstein College of Medicine, Bronx, New York, USA"},{"name":"Ophthalmology, Montefiore Hospital and Medical Center, Bronx, New York, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"239","published-online":{"date-parts":[[2026,5,5]]},"reference":[{"key":"2026050510202632000_bjo-2025-329249v1.1","doi-asserted-by":"crossref","first-page":"2364","DOI":"10.1097\/IAE.0000000000000882","article-title":"Detection of microvascular changes in eyes of patients with diabetes but not clinical diabetic retinopathy using optical coherence tomography angiography","volume":"35","author":"de Carlo","year":"2015","journal-title":"Retina"},{"key":"2026050510202632000_bjo-2025-329249v1.2","doi-asserted-by":"crossref","first-page":"3000605231187933","DOI":"10.1177\/03000605231187933","article-title":"Optical coherence tomography angiography: a review of the current literature","volume":"51","author":"Javed","year":"2023","journal-title":"J Int Med Res"},{"key":"2026050510202632000_bjo-2025-329249v1.3","doi-asserted-by":"publisher","DOI":"10.1038\/srep42201"},{"key":"2026050510202632000_bjo-2025-329249v1.4","doi-asserted-by":"publisher","DOI":"10.1016\/j.ajo.2016.07.023"},{"key":"2026050510202632000_bjo-2025-329249v1.5","doi-asserted-by":"crossref","first-page":"1438739","DOI":"10.3389\/fendo.2025.1438739","article-title":"Recent advances and applications of optical coherence tomography angiography in diabetic retinopathy","volume":"16","author":"Zhang","year":"2025","journal-title":"Front Endocrinol"},{"key":"2026050510202632000_bjo-2025-329249v1.6","doi-asserted-by":"crossref","first-page":"1186098","DOI":"10.3389\/fmed.2023.1186098","article-title":"Retinal and choroidal microvascular characterization and density changes in different stages of diabetic retinopathy eyes","volume":"10","author":"Wang","year":"2023","journal-title":"Front Med"},{"key":"2026050510202632000_bjo-2025-329249v1.7","doi-asserted-by":"crossref","DOI":"10.1001\/jamanetworkopen.2019.19469","article-title":"Quantitative Microvascular Analysis With Wide-Field Optical Coherence Tomography Angiography in Eyes With Diabetic Retinopathy","volume":"3","author":"Tan","year":"2020","journal-title":"JAMA Netw Open"},{"key":"2026050510202632000_bjo-2025-329249v1.8","doi-asserted-by":"publisher","DOI":"10.1016\/j.ajo.2015.04.021"},{"key":"2026050510202632000_bjo-2025-329249v1.9","doi-asserted-by":"publisher","DOI":"10.1016\/j.preteyeres.2017.11.003"},{"key":"2026050510202632000_bjo-2025-329249v1.10","doi-asserted-by":"publisher","DOI":"10.1016\/j.ajo.2016.10.005"},{"key":"2026050510202632000_bjo-2025-329249v1.11","doi-asserted-by":"crossref","first-page":"724","DOI":"10.1097\/IAE.0000000000001576","article-title":"Optical coherence tomography angiography reveals blood flow in choroidal neovascular membrane in remission phase of neovascular age-related macular degeneration","volume":"37","author":"Ichiyama","year":"2017","journal-title":"Retina (Philadelphia, Pa)"},{"key":"2026050510202632000_bjo-2025-329249v1.12","doi-asserted-by":"crossref","first-page":"1428850","DOI":"10.3389\/fmed.2024.1428850","article-title":"Applications of optical coherence tomography angiography in glaucoma: current status and future directions","volume":"11","author":"Shen","year":"2024","journal-title":"Front Med"},{"key":"2026050510202632000_bjo-2025-329249v1.13","doi-asserted-by":"crossref","DOI":"10.3390\/diagnostics12061418","article-title":"Advances in OCT Imaging in Myopia and Pathologic Myopia","volume":"12","author":"Li","year":"2022","journal-title":"Diagnostics (Basel)"},{"key":"2026050510202632000_bjo-2025-329249v1.14","doi-asserted-by":"crossref","DOI":"10.3390\/diagnostics13193100","article-title":"An Analysis of Optical Coherence Tomography Angiography (OCT-A) Perfusion Density Maps in Patients Treated for Retinal Vein Occlusion with Intravitreal Aflibercept","volume":"13","author":"\u015apiewak","year":"2023","journal-title":"Diagnostics (Basel)"},{"key":"2026050510202632000_bjo-2025-329249v1.15","doi-asserted-by":"crossref","first-page":"173","DOI":"10.4103\/sjopt.sjopt_91_23","article-title":"Vision transformers: The next frontier for deep learning-based ophthalmic image analysis","volume":"37","author":"Wu","year":"2023","journal-title":"Saudi J Ophthalmol"},{"key":"2026050510202632000_bjo-2025-329249v1.16","doi-asserted-by":"publisher","DOI":"10.1016\/j.preteyeres.2024.101291"},{"key":"2026050510202632000_bjo-2025-329249v1.17","doi-asserted-by":"crossref","DOI":"10.1038\/s41746-023-00811-0","article-title":"Self-supervised learning for medical image classification: a systematic review and implementation guidelines","volume":"6","author":"Huang","year":"2023","journal-title":"NPJ Digit Med"},{"key":"2026050510202632000_bjo-2025-329249v1.18","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-023-06555-x"},{"key":"2026050510202632000_bjo-2025-329249v1.19","doi-asserted-by":"crossref","unstructured":"Qiu J , Wu J , Wei H , et al . Development and Validation of a Multimodal Multitask Vision Foundation Model for Generalist Ophthalmic Artificial Intelligence. NEJM AI 2024;1. doi:10.1056\/AIoa2300221","DOI":"10.1056\/AIoa2300221"},{"key":"2026050510202632000_bjo-2025-329249v1.20","doi-asserted-by":"crossref","DOI":"10.3390\/diagnostics14161751","article-title":"Cross-Modality Medical Image Segmentation via Enhanced Feature Alignment and Cross Pseudo Supervision Learning","volume":"14","author":"Yang","year":"2024","journal-title":"Diagnostics (Basel)"},{"key":"2026050510202632000_bjo-2025-329249v1.21","doi-asserted-by":"crossref","first-page":"2494","DOI":"10.1109\/TMI.2020.2972701","article-title":"Unsupervised Bidirectional Cross-Modality Adaptation via Deeply Synergistic Image and Feature Alignment for Medical Image Segmentation","volume":"39","author":"Chen","year":"2020","journal-title":"IEEE Trans Med Imaging"},{"key":"2026050510202632000_bjo-2025-329249v1.22","doi-asserted-by":"crossref","unstructured":"McNaughton J , Fernandez J , Holdsworth S , et al . Machine Learning for Medical Image Translation: A Systematic Review. Bioengineering (Basel) 2023; 10(9):1078. doi:10.3390\/bioengineering10091078","DOI":"10.3390\/bioengineering10091078"},{"key":"2026050510202632000_bjo-2025-329249v1.23","doi-asserted-by":"crossref","DOI":"10.1038\/s41746-025-01852-3","article-title":"Multimodal foundation model and benchmark for comprehensive retinal OCT image analysis","volume":"8","author":"Morano","year":"2025","journal-title":"NPJ Digit Med"},{"key":"2026050510202632000_bjo-2025-329249v1.24","doi-asserted-by":"crossref","first-page":"1711","DOI":"10.1109\/TMI.2024.3518067","article-title":"MultiEYE: Dataset and Benchmark for OCT-Enhanced Retinal Disease Recognition From Fundus Images","volume":"44","author":"Wang","year":"2025","journal-title":"IEEE Trans Med Imaging"},{"key":"2026050510202632000_bjo-2025-329249v1.25","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2024.103092"},{"key":"2026050510202632000_bjo-2025-329249v1.26","doi-asserted-by":"crossref","unstructured":"He K , Zhang X , Ren S , et al . Deep residual learning for image recognition. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR); Las Vegas, NV, USA. doi:10.1109\/CVPR.2016.90","DOI":"10.1109\/CVPR.2016.90"},{"key":"2026050510202632000_bjo-2025-329249v1.27","doi-asserted-by":"crossref","first-page":"101206","DOI":"10.1016\/j.preteyeres.2023.101206","article-title":"Optical coherence tomography angiography in diabetic retinopathy","volume":"97","author":"Waheed","year":"2023","journal-title":"Prog Retin Eye Res"},{"key":"2026050510202632000_bjo-2025-329249v1.28","doi-asserted-by":"crossref","DOI":"10.1371\/journal.pone.0217805","article-title":"Progression of subclinical choroidal neovascularization in age-related macular degeneration","volume":"14","author":"Heiferman","year":"2019","journal-title":"PLoS One"},{"key":"2026050510202632000_bjo-2025-329249v1.29","doi-asserted-by":"crossref","first-page":"4713","DOI":"10.1364\/BOE.495999","article-title":"Optimizing the OCTA layer fusion option for deep learning classification of diabetic retinopathy","volume":"14","author":"Ebrahimi","year":"2023","journal-title":"Biomed Opt Express"},{"key":"2026050510202632000_bjo-2025-329249v1.30","doi-asserted-by":"publisher","DOI":"10.1016\/S0140-6736(09)62124-3"},{"key":"2026050510202632000_bjo-2025-329249v1.31","doi-asserted-by":"publisher","DOI":"10.1016\/S0161-6420(03)00475-5"},{"key":"2026050510202632000_bjo-2025-329249v1.32","doi-asserted-by":"publisher","DOI":"10.1016\/S2214-109X(13)70145-1"},{"key":"2026050510202632000_bjo-2025-329249v1.33","doi-asserted-by":"publisher","DOI":"10.1038\/s41591-018-0300-7"},{"key":"2026050510202632000_bjo-2025-329249v1.34","doi-asserted-by":"crossref","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"},{"key":"2026050510202632000_bjo-2025-329249v1.35","doi-asserted-by":"crossref","first-page":"31","DOI":"10.1167\/tvst.13.12.31","article-title":"A Deep Learning Network for Accurate Retinal Multidisease Diagnosis Using Multiview Fusion of En Face and B-Scan Images: A Multicenter Study","volume":"13","author":"Ou","year":"2024","journal-title":"Trans Vis Sci Tech"},{"key":"2026050510202632000_bjo-2025-329249v1.36","doi-asserted-by":"crossref","DOI":"10.1016\/j.xops.2025.101030","article-title":"Federated Learning for Multi-Disease Ophthalmic Diagnostics Using OCT Angiography","volume":"6","author":"Nabil","year":"2026","journal-title":"Ophthalmol Sci"},{"key":"2026050510202632000_bjo-2025-329249v1.37","doi-asserted-by":"crossref","DOI":"10.3390\/bioengineering11070682","article-title":"Anomaly Detection in Optical Coherence Tomography Angiography (OCTA) with a Vector-Quantized Variational Auto-Encoder (VQ-VAE)","volume":"11","author":"Jebril","year":"2024","journal-title":"Bioengineering (Basel)"}],"container-title":["British Journal of Ophthalmology"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.1136\/bjo-2025-329249","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,5]],"date-time":"2026-05-05T17:20:44Z","timestamp":1778001644000},"score":1,"resource":{"primary":{"URL":"https:\/\/bjo.bmj.com\/lookup\/doi\/10.1136\/bjo-2025-329249"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,5,5]]},"references-count":37,"alternative-id":["10.1136\/bjo-2025-329249"],"URL":"https:\/\/doi.org\/10.1136\/bjo-2025-329249","relation":{},"ISSN":["0007-1161","1468-2079"],"issn-type":[{"value":"0007-1161","type":"print"},{"value":"1468-2079","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,5,5]]}}}