{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,17]],"date-time":"2026-10-17T21:51:26Z","timestamp":1792273886221,"version":"4.4.0"},"reference-count":41,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2025,7,17]],"date-time":"2025-07-17T00:00:00Z","timestamp":1752710400000},"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. Cell Dev. Biol."],"abstract":"<jats:sec><jats:title>Purpose<\/jats:title><jats:p>This study aimed to develop a hybrid deep learning model for classifying multiple fundus diseases using ultra-widefield (UWF) images, thereby improving diagnostic efficiency and accuracy while providing an auxiliary tool for clinical decision-making.<\/jats:p><\/jats:sec><jats:sec><jats:title>Methods<\/jats:title><jats:p>In this retrospective study, 10,612 UWF fundus images were collected from the JiuJiang No. 1 People\u2019s Hospital and the Seventh Affiliated Hospital, Sun Yat-sen University between 2020 and 2025, covering 16 fundus diseases, including normal fundus, nine common eye diseases, and six rare retinal conditions. The model employed DenseNet121 as a feature extractor combined with an XGBoost classifier. Gradient-weighted Class Activation Mapping (Grad-CAM) was used to visualize the model\u2019s decision-making process. Performance was evaluated on validation and external test sets using accuracy, recall, precision, F1 score, and AUC-ROC. The model\u2019s diagnostic accuracy was also compared with that of junior and intermediate ophthalmologists.<\/jats:p><\/jats:sec><jats:sec><jats:title>Results<\/jats:title><jats:p>The model demonstrated exceptional diagnostic performance. For common diseases such as retinal vein occlusion, age-related macular degeneration, and diabetic retinopathy, AUC values exceeded 0.975, with accuracy rates above 0.980. For rare diseases, AUC values were above 0.970, and accuracy rates surpassed 0.998. Grad-CAM visualizations confirmed that the model\u2019s focus areas aligned with clinical pathological features. Compared to ophthalmologists, the model achieved significantly higher accuracy across all diagnostic tasks.<\/jats:p><\/jats:sec><jats:sec><jats:title>Conclusion<\/jats:title><jats:p>The proposed deep learning model can automatically identify and classify multiple ophthalmic diseases using UWF images. It holds promise for enhancing clinical diagnostic efficiency, assisting ophthalmologists in optimizing workflows, and improving patient care quality.<\/jats:p><\/jats:sec>","DOI":"10.3389\/fcell.2025.1630667","type":"journal-article","created":{"date-parts":[[2025,7,17]],"date-time":"2025-07-17T05:25:37Z","timestamp":1752729937000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":14,"title":["Deep learning-based classification of multiple fundus diseases using ultra-widefield images"],"prefix":"10.3389","volume":"13","author":[{"given":"Ming-Ming","family":"Duan","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xiang","family":"Tu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2025,7,17]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"123","DOI":"10.4103\/sjopt.sjopt_151_23","article-title":"Peripheral retinal lesions in diabetic retinopathy on ultra-widefield imaging","volume":"38","author":"Ashrafkhorasani","year":"2024","journal-title":"Saudi J. Ophthalmol. Off. J. Saudi Ophthalmol. Soc."},{"key":"B2","doi-asserted-by":"publisher","first-page":"187","DOI":"10.1007\/s10772-024-10089-6","article-title":"Continuous feature learning representation to XGBoost classifier on the aggregation of discriminative features using DenseNet-121 architecture and ResNet 18 architectures towards apraxia recognition in the child speech therapy","volume":"27","author":"Ashwini","year":"2024","journal-title":"Int. J. Speech Technol."},{"key":"B3","doi-asserted-by":"publisher","first-page":"948","DOI":"10.1007\/s10278-021-00479-6","article-title":"Artificial intelligence\u2013assisted early detection of retinitis pigmentosa \u2014 the Most common inherited retinal degeneration","volume":"34","author":"Chen","year":"2021","journal-title":"J. Digit. Imaging"},{"key":"B4","doi-asserted-by":"publisher","first-page":"e1010778","DOI":"10.1371\/journal.pcbi.1010778","article-title":"Ten quick tips for computational analysis of medical images","volume":"19","author":"Chicco","year":"2023","journal-title":"PLOS Comput. Biol."},{"key":"B5","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1167\/tvst.13.4.1","article-title":"Deep learning-based automated detection of retinal breaks and detachments on fundus photography","volume":"13","author":"Christ","year":"2024","journal-title":"Transl. Vis. Sci. Technol."},{"key":"B6","doi-asserted-by":"publisher","first-page":"611","DOI":"10.3390\/s18020611","article-title":"Gaofen-3 PolSAR image classification via XGBoost and polarimetric spatial information","volume":"18","author":"Dong","year":"2018","journal-title":"Sensors"},{"key":"B7","doi-asserted-by":"publisher","first-page":"e1221","DOI":"10.1016\/S2214-109X(17)30393-5","article-title":"Global causes of blindness and distance vision impairment 1990-2020: a systematic review and meta-analysis","volume":"5","author":"Flaxman","year":"2017","journal-title":"Lancet Glob. Health"},{"key":"B8","doi-asserted-by":"publisher","first-page":"2774","DOI":"10.21037\/qims-24-1644","article-title":"Development and validation of the artificial intelligence-proliferative vitreoretinopathy (AI-PVR) insight system for deep learning-based diagnosis and postoperative risk prediction in proliferative vitreoretinopathy using multimodal fundus imaging","volume":"15","author":"Gan","year":"","journal-title":"Quant. Imaging Med. Surg."},{"key":"B9","doi-asserted-by":"publisher","first-page":"1084118","DOI":"10.3389\/fnins.2022.1084118","article-title":"Application of artificial intelligence models for detecting the pterygium that requires surgical treatment based on anterior segment images","volume":"16","author":"Gan","year":"2022","journal-title":"Front. Neurosci."},{"key":"B10","doi-asserted-by":"publisher","first-page":"1182388","DOI":"10.3389\/fnins.2023.1182388","article-title":"Application of artificial intelligence for automatic cataract staging based on anterior segment images: comparing automatic segmentation approaches to manual segmentation","volume":"17","author":"Gan","year":"","journal-title":"Front. Neurosci."},{"key":"B11","doi-asserted-by":"publisher","first-page":"94","DOI":"10.1016\/j.csbj.2025.02.039","article-title":"Deep learning-enabled transformation of anterior segment images to corneal fluorescein staining images for enhanced corneal disease screening","volume":"28","author":"Gan","year":"","journal-title":"Comput. Struct. Biotechnol. J."},{"key":"B12","doi-asserted-by":"publisher","first-page":"1097291","DOI":"10.3389\/fnins.2023.1097291","article-title":"Artificial intelligence method based on multi-feature fusion for automatic macular edema (ME) classification on spectral-domain optical coherence tomography (SD-OCT) images","volume":"17","author":"Gan","year":"","journal-title":"Front. Neurosci."},{"key":"B13","doi-asserted-by":"publisher","first-page":"187","DOI":"10.1016\/j.visres.2017.02.009","article-title":"Ultra-wide-field imaging in diabetic retinopathy","volume":"139","author":"Ghasemi Falavarjani","year":"2017","journal-title":"Vis. Res."},{"key":"B14","doi-asserted-by":"publisher","first-page":"1410","DOI":"10.1016\/j.ophtha.2018.02.037","article-title":"A deep learning algorithm for prediction of age-related eye disease study severity scale for age-related macular degeneration from color fundus photography","volume":"125","author":"Grassmann","year":"2018","journal-title":"Ophthalmology"},{"key":"B15","doi-asserted-by":"publisher","first-page":"1684","DOI":"10.3390\/diagnostics12071684","article-title":"Progress of imaging in diabetic retinopathy-from the past to the present","volume":"12","author":"Horie","year":"2022","journal-title":"Diagn. Basel Switz."},{"key":"B16","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1038\/s41746-023-00991-9","article-title":"The performance of a deep learning system in assisting junior ophthalmologists in diagnosing 13 major fundus diseases: a prospective multi-center clinical trial: 1","volume":"7","author":"Li","year":"2024","journal-title":"Npj Digit. Med."},{"key":"B17","doi-asserted-by":"publisher","first-page":"101095","DOI":"10.1016\/j.xcrm.2023.101095","article-title":"Artificial intelligence in ophthalmology: the path to the real-world clinic","volume":"4","author":"Li","year":"2023","journal-title":"Cell. Rep. Med."},{"key":"B18","doi-asserted-by":"publisher","first-page":"1311","DOI":"10.2147\/DMSO.S506494","article-title":"EfficientNetB0-Based end-to-end diagnostic system for diabetic retinopathy grading and macular edema detection","volume":"18","author":"Long","year":"2025","journal-title":"Diabetes Metab. Syndr. Obes. Targets Ther."},{"key":"B19","doi-asserted-by":"publisher","first-page":"e6900","DOI":"10.7717\/peerj.6900","article-title":"Accuracy of a deep convolutional neural network in detection of retinitis pigmentosa on ultrawide-field images","volume":"7","author":"Masumoto","year":"2019","journal-title":"PeerJ"},{"key":"B20","doi-asserted-by":"publisher","first-page":"e5696","DOI":"10.7717\/peerj.5696","article-title":"Accuracy of deep learning, a machine learning technology, using ultra-wide-field fundus ophthalmoscopy for detecting idiopathic macular holes","volume":"6","author":"Nagasawa","year":"2018","journal-title":"PeerJ"},{"key":"B21","doi-asserted-by":"publisher","first-page":"660","DOI":"10.1097\/IAE.0000000000000937","article-title":"Ultra-widefield fundus imaging: a review of clinical applications and future trends","volume":"36","author":"Nagiel","year":"2016","journal-title":"Retina Phila. Pa"},{"key":"B22","doi-asserted-by":"publisher","first-page":"105","DOI":"10.3390\/diagnostics14010105","article-title":"Retinal disease diagnosis using deep learning on ultra-wide-field fundus images: 1","volume":"14","author":"Nguyen","year":"2024","journal-title":"Diagnostics"},{"key":"B23","doi-asserted-by":"publisher","first-page":"9425","DOI":"10.1038\/s41598-017-09891-x","article-title":"Accuracy of deep learning, a machine-learning technology, using ultra-wide-field fundus ophthalmoscopy for detecting rhegmatogenous retinal detachment","volume":"7","author":"Ohsugi","year":"2017","journal-title":"Sci. Rep."},{"key":"B24","doi-asserted-by":"publisher","first-page":"1138","DOI":"10.1038\/s41433-018-0064-9","article-title":"Automated diabetic retinopathy detection in smartphone-based fundus photography using artificial intelligence","volume":"32","author":"Rajalakshmi","year":"2018","journal-title":"Eye Lond. Engl."},{"key":"B25","article-title":"Grad-CAM: visual explanations from deep networks via gradient-based localization","volume-title":"2017 IEEE int. conf. comput. vis. ICCV","author":"Selvaraju","year":"2024"},{"key":"B26","doi-asserted-by":"publisher","first-page":"8330833","DOI":"10.1155\/2022\/8330833","article-title":"Early diagnosis of brain tumour MRI images using hybrid techniques between deep and machine learning","volume":"2022","author":"Senan","year":"2022","journal-title":"Comput. Math. Methods Med."},{"key":"B27","doi-asserted-by":"publisher","first-page":"275","DOI":"10.3390\/life12020275","article-title":"Potential applications of artificial intelligence in clinical trials for alzheimer\u2019s disease","volume":"12","author":"Seo","year":"2022","journal-title":"Life"},{"key":"B28","doi-asserted-by":"publisher","first-page":"575","DOI":"10.4103\/0301-4738.167122","article-title":"Update on wide- and ultra-widefield retinal imaging","volume":"63","author":"Shoughy","year":"2015","journal-title":"Indian J. Ophthalmol."},{"key":"B29","doi-asserted-by":"publisher","first-page":"549","DOI":"10.1016\/j.ajo.2012.03.019","article-title":"Nonmydriatic ultrawide field retinal imaging compared with dilated standard 7-field 35-mm photography and retinal specialist examination for evaluation of diabetic retinopathy","volume":"154","author":"Silva","year":"2012","journal-title":"Am. J. Ophthalmol."},{"key":"B30","doi-asserted-by":"publisher","first-page":"2587","DOI":"10.1016\/j.ophtha.2013.05.004","article-title":"Peripheral lesions identified by mydriatic ultrawide field imaging: distribution and potential impact on diabetic retinopathy severity","volume":"120","author":"Silva","year":"2013","journal-title":"Ophthalmology"},{"key":"B31","doi-asserted-by":"publisher","first-page":"18","DOI":"10.1186\/s40942-019-0163-1","article-title":"Wide-field angiography in retinal vein occlusions","volume":"5","author":"Tan","year":"2019","journal-title":"Int. J. Retina Vitr."},{"key":"B32","doi-asserted-by":"publisher","first-page":"1173","DOI":"10.2147\/IJGM.S507100","article-title":"Practical applications of artificial intelligence diagnostic systems in fundus retinal disease screening","volume":"18","author":"Wei","year":"2025","journal-title":"Int. J. Gen. Med."},{"key":"B33","doi-asserted-by":"publisher","first-page":"485","DOI":"10.1080\/14737159.2023.2208751","article-title":"Detection of macular atrophy in age-related macular degeneration aided by artificial intelligence","volume":"23","author":"Wei","year":"2023","journal-title":"Expert Rev. Mol. diagn."},{"key":"B34","doi-asserted-by":"publisher","first-page":"e106","DOI":"10.1016\/S2214-109X(13)70145-1","article-title":"Global prevalence of age-related macular degeneration and disease burden projection for 2020 and 2040: a systematic review and meta-analysis","volume":"2","author":"Wong","year":"2014","journal-title":"Lancet Glob. Health"},{"key":"B35","doi-asserted-by":"publisher","first-page":"1079155","DOI":"10.3389\/fncom.2022.1079155","article-title":"Classification of dry and wet macular degeneration based on the ConvNeXT model","volume":"16","author":"Wu","year":"2022","journal-title":"Front. Comput. Neurosci."},{"key":"B36","doi-asserted-by":"publisher","first-page":"1365350","DOI":"10.3389\/fendo.2024.1365350","article-title":"An ensemble deep learning diagnostic system for determining clinical activity scores in thyroid-associated ophthalmopathy: integrating multi-view multimodal images from anterior segment slit-lamp photographs and facial images","volume":"15","author":"Yan","year":"2024","journal-title":"Front. Endocrinol."},{"key":"B37","doi-asserted-by":"publisher","first-page":"840011","DOI":"10.3389\/fphys.2022.840011","article-title":"Automatic multichannel electrocardiogram record classification using XGBoost fusion model","volume":"13","author":"Ye","year":"2022","journal-title":"Front. Physiol."},{"key":"B38","doi-asserted-by":"publisher","first-page":"2225","DOI":"10.1007\/s00417-021-05105-3","article-title":"Development of a deep-learning system for detection of lattice degeneration, retinal breaks, and retinal detachment in tessellated eyes using ultra-wide-field fundus images: a pilot study","volume":"259","author":"Zhang","year":"2021","journal-title":"Graefes Arch. Clin. Exp. Ophthalmol. Albr. Von. Graefes Arch. Klin. Exp. Ophthalmol."},{"key":"B39","doi-asserted-by":"publisher","first-page":"3240","DOI":"10.1038\/s41598-025-87953-1","article-title":"An automatic cervical cell classification model based on improved DenseNet121","volume":"15","author":"Zhang","year":"2025","journal-title":"Sci. Rep."},{"key":"B40","doi-asserted-by":"publisher","first-page":"808402","DOI":"10.3389\/fmed.2022.808402","article-title":"Screening of common retinal diseases using six-category models based on EfficientNet","volume":"9","author":"Zhu","year":"2022","journal-title":"Front. Med."},{"key":"B41","doi-asserted-by":"publisher","first-page":"39","DOI":"10.1038\/s41377-022-00714-x","article-title":"Deep learning in optical metrology: a review","volume":"11","author":"Zuo","year":"2022","journal-title":"Light Sci. Appl."}],"container-title":["Frontiers in Cell and Developmental Biology"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fcell.2025.1630667\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,7,17]],"date-time":"2025-07-17T05:25:38Z","timestamp":1752729938000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fcell.2025.1630667\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,7,17]]},"references-count":41,"alternative-id":["10.3389\/fcell.2025.1630667"],"URL":"https:\/\/doi.org\/10.3389\/fcell.2025.1630667","relation":{},"ISSN":["2296-634X"],"issn-type":[{"value":"2296-634X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,7,17]]},"article-number":"1630667"}}