{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,25]],"date-time":"2026-05-25T14:14:44Z","timestamp":1779718484596,"version":"3.53.1"},"reference-count":23,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,5,25]],"date-time":"2026-05-25T00:00:00Z","timestamp":1779667200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Med."],"abstract":"<jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Automated disease screening systems face challenges when applied to multi-class medical image analysis, particularly under severe class imbalance inherent in clinical datasets. Retinal fundus imaging enables non-invasive screening for multiple ocular and systemic diseases simultaneously, yet current automated systems typically assess risk for only a single pathology or a limited disease range. We developed a comprehensive AI framework for simultaneous risk stratification of 15 distinct pathological conditions from retinal fundus images and report its preliminary evaluation in a real-world clinical setting.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>Our system combines fundus images from publicly available sources and proprietary clinical archives, addressing the significant class imbalance challenge inherent in rare condition risk assessment. We employed a hybrid deep learning architecture (CAFormer B36) with focal loss and targeted oversampling to ensure robust performance across common and rare conditions. The framework identifies retinal biomarkers associated with both primary ocular diseases (diabetic retinopathy, glaucoma, macular degeneration, cataracts, retinitis pigmentosa) and systemic conditions with retinal manifestations (hypertensive retinopathy, atherosclerotic changes, autoimmune manifestations). Dataset splitting was performed at the image level; all internal metrics should therefore be interpreted as exploratory upper-bound estimates pending patient-level replication.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>\n                      On the internal image-level test set, the system achieved ROC AUC of 0.9524\u20130.9971 across all 15 pathological classes, with F1 scores of 0.8968\u20130.9649. Notably, rare conditions with fewer than 100 training examples demonstrated robust risk stratification performance. In an exploratory single-site evaluation on 68 real-world cases, overall accuracy was 64.7% (95% CI: 52.9\u201376.5%). Due to the very limited number of sight-threatening cases in this small cohort (glaucoma\n                      <jats:italic>n<\/jats:italic>\n                      = 2, diabetic retinopathy\n                      <jats:italic>n<\/jats:italic>\n                      = 4), sensitivity estimates for these categories carry extremely wide confidence intervals and cannot be considered statistically reliable. These findings underscore the need for larger, prospective, multicenter studies.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusions<\/jats:title>\n                    <jats:p>\n                      This pilot proof-of-concept demonstrates the feasibility of multi-pathology risk stratification under severe class imbalance using a hybrid deep learning architecture. Internal image-level metrics show strong discriminative capability, while the preliminary single-site evaluation (\n                      <jats:italic>n<\/jats:italic>\n                      = 68, 64.7% accuracy) reveals the challenges of real-world translation and the substantial gap that remains before clinical deployment. This work provides a methodological foundation for future prospectively validated screening systems.\n                    <\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/fmed.2026.1778404","type":"journal-article","created":{"date-parts":[[2026,5,25]],"date-time":"2026-05-25T13:50:26Z","timestamp":1779717026000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Artificial intelligence framework for multi-pathology risk assessment from retinal fundus images: deep learning approach to 15-disease screening"],"prefix":"10.3389","volume":"13","author":[{"given":"Robert","family":"Vasilev","sequence":"first","affiliation":[{"name":"Z-union AI Technologies Consortium","place":["Moscow, Russia"]},{"name":"Department of the Problems of Physics and Power Engineering, Moscow Institute of Physics and Technology","place":["Dolgoprudny, Russia"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Andrey","family":"Savchenko","sequence":"additional","affiliation":[{"name":"Sber AI Lab","place":["Moscow, Russia"]},{"name":"HSE University, Laboratory for Theoretical Foundations of AI Models","place":["Moscow, Russia"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Pavel","family":"Blinov","sequence":"additional","affiliation":[{"name":"Sber AI Lab","place":["Moscow, Russia"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tadej","family":"Svetina","sequence":"additional","affiliation":[{"name":"Sber AI Lab","place":["Moscow, Russia"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Stepan","family":"Kudin","sequence":"additional","affiliation":[{"name":"Sber AI Lab","place":["Moscow, Russia"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Nikolay","family":"Romanenko","sequence":"additional","affiliation":[{"name":"Sber AI Lab","place":["Moscow, Russia"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yuliya","family":"Sarana","sequence":"additional","affiliation":[{"name":"Skolkovo Institute of Science and Technology, Center for Bio- and Medical Technologies","place":["Moscow, Russia"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Gleb","family":"Khizhnyak","sequence":"additional","affiliation":[{"name":"Z-union AI Technologies Consortium","place":["Moscow, Russia"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Andrey","family":"Demchinsky","sequence":"additional","affiliation":[{"name":"Association for Specialists in Progress, Engineering, and Clinical Technologies Unified by Medicine 'ASPECTUM', ELVIS Neuroimplants LLC","place":["Moscow, Russia"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Taisia","family":"Shcheglova","sequence":"additional","affiliation":[{"name":"Department of the Problems of Physics and Power Engineering, Moscow Institute of Physics and Technology","place":["Dolgoprudny, Russia"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,5,25]]},"reference":[{"key":"B1","volume-title":"World Health Statistics 2024: Monitoring Health for the SDGs, Sustainable Development Goals","year":"2024"},{"key":"B2","unstructured":"Blindness and Vision Impairment\n          \n          2024"},{"key":"B3","doi-asserted-by":"publisher","first-page":"1394","DOI":"10.3390\/biomedicines12071394","article-title":"Deep Learning evaluation of glaucoma detection using fundus photographs in highly myopic populations","volume":"12","author":"Chiang","year":"2024","journal-title":"Biomedicines"},{"key":"B4","doi-asserted-by":"publisher","first-page":"158","DOI":"10.22608\/APO.201976","article-title":"Artificial intelligence in diabetic eye disease screening","volume":"8","author":"Cheung","year":"2019","journal-title":"Asia-Pac J Ophthalmol"},{"key":"B5","doi-asserted-by":"publisher","first-page":"44","DOI":"10.1109\/JBHI.2021.3110593","article-title":"Explainable Diabetic Retinopathy Detection and Retinal Image Generation","volume":"26","author":"Niu","year":"2022","journal-title":"IEEE J Biomed Health Inform"},{"key":"B6","doi-asserted-by":"publisher","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":"B7","doi-asserted-by":"publisher","first-page":"2071","DOI":"10.1109\/JBHI.2020.3028180","article-title":"Accurate retinal vessel segmentation in color fundus images via fully attention-based networks","volume":"25","author":"Li","year":"2021","journal-title":"IEEE J Biomed Health Inform"},{"key":"B8","doi-asserted-by":"publisher","first-page":"411","DOI":"10.32604\/cmc.2021.013390","article-title":"Classification of Fundus images based on deep learning for detecting eye diseases","volume":"67","author":"Chea","year":"2021","journal-title":"Comput Mater Contin"},{"key":"B9","doi-asserted-by":"publisher","first-page":"4954","DOI":"10.1109\/EMBC.2012.6347104","article-title":"Automatic localization of retinal landmarks","author":"Cheng","year":"2012","journal-title":"Annual International Conference of the IEEE Engineering in Medicine and Biology Society"},{"key":"B10","doi-asserted-by":"publisher","first-page":"4828","DOI":"10.1038\/s41467-021-25138-w","article-title":"Automatic detection of 39 Fundus diseases and conditions in retinal photographs using deep neural networks","volume":"12","author":"Cen LP Ji","year":"2021","journal-title":"Nat Commun"},{"key":"B11","doi-asserted-by":"crossref","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":"B12","doi-asserted-by":"publisher","first-page":"2739","DOI":"10.1109\/JBHI.2022.3214086","article-title":"Multi-label retinal disease classification using transformers","volume":"27","author":"Rodr\u00edguez","year":"2023","journal-title":"IEEE J Biomed Health Inform"},{"key":"B13","doi-asserted-by":"publisher","first-page":"e13028","DOI":"10.1111\/exsy.13028","article-title":"Artificial intelligence with optimal deep learning enabled automated retinal Fundus image classification model","volume":"39","author":"Gupta","year":"2022","journal-title":"Expert Systems"},{"key":"B14","doi-asserted-by":"publisher","first-page":"e23472","DOI":"10.2196\/23472","article-title":"Deep learning-based detection of early renal function impairment using retinal fundus images: model development and validation","volume":"8","author":"Kang","year":"2020","journal-title":"JMIR Med Inform"},{"key":"B15","doi-asserted-by":"publisher","first-page":"5934","DOI":"10.1038\/s41598-023-32518-3","article-title":"An interpretable and interactive deep learning algorithm for a clinically applicable retinal Fundus diagnosis system by modelling finding-disease relationship","volume":"13","author":"Son","year":"2023","journal-title":"Sci Rep"},{"key":"B16","doi-asserted-by":"publisher","first-page":"944967","DOI":"10.3389\/fpubh.2022.944967","article-title":"Validating automated eye disease screening AI algorithm in community and in-hospital scenarios","volume":"10","author":"Han","year":"2022","journal-title":"Front Public Health"},{"key":"B17","doi-asserted-by":"publisher","first-page":"896","DOI":"10.1109\/TPAMI.2023.3329173","article-title":"Metaformer baselines for vision","volume":"46","author":"Yu","year":"2023","journal-title":"IEEE Trans Pattern Anal Mach Intell"},{"key":"B18","doi-asserted-by":"publisher","first-page":"770","DOI":"10.1109\/CVPR.2016.90","article-title":"Deep residual learning for image recognition","author":"He","year":"2016","journal-title":"Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)"},{"key":"B19","first-page":"6105","article-title":"EfficientNet: rethinking model scaling for convolutional neural networks","volume-title":"Proceedings of the International Conference on Machine Learning (ICML)","author":"Tan","year":"2019"},{"key":"B20","doi-asserted-by":"publisher","first-page":"1","DOI":"10.48550\/arXiv.2010.11929","article-title":"An image is worth 16x16 words: transformers for image recognition at scale","author":"Dosovitskiy","year":"2021","journal-title":"Proceedings of the International Conference on Learning Representations (ICLR)"},{"key":"B21","doi-asserted-by":"publisher","first-page":"10012","DOI":"10.1109\/ICCV48922.2021.00986","article-title":"Swin transformer: hierarchical vision transformer using shifted windows","author":"Liu","year":"2021","journal-title":"Proceedings of the IEEE\/CVF International Conference on Computer Vision (ICCV)"},{"key":"B22","doi-asserted-by":"publisher","first-page":"78125","DOI":"10.1007\/s11042-024-18627-9","article-title":"Towards improved fundus disease detection using Swin transformers","volume":"83","author":"Singh","year":"2024","journal-title":"Multimed Tools Appl"},{"key":"B23","doi-asserted-by":"publisher","first-page":"106299","DOI":"10.1016\/j.bspc.2024.106263","article-title":"Robust deep learning for eye fundus images: bridging real and synthetic data for enhancing generalization","volume":"94","author":"Oliveira","year":"2024","journal-title":"Biomed Signal Process Control"}],"container-title":["Frontiers in Medicine"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fmed.2026.1778404\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,25]],"date-time":"2026-05-25T13:50:29Z","timestamp":1779717029000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fmed.2026.1778404\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,5,25]]},"references-count":23,"alternative-id":["10.3389\/fmed.2026.1778404"],"URL":"https:\/\/doi.org\/10.3389\/fmed.2026.1778404","relation":{},"ISSN":["2296-858X"],"issn-type":[{"value":"2296-858X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,5,25]]},"article-number":"1778404"}}