{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,2]],"date-time":"2026-07-02T08:43:20Z","timestamp":1782981800841,"version":"3.54.5"},"reference-count":45,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,7,2]],"date-time":"2026-07-02T00:00:00Z","timestamp":1782950400000},"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>\n                    <jats:title>Purpose<\/jats:title>\n                    <jats:p>To develop and validate a deep learning model for automated macular hole (MH) staging using optical coherence tomography (OCT) images.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>This retrospective study included OCT-confirmed MH images collected between July 2019 and October 2024. After quality control and redundancy screening, 6,243 representative OCT images were retained. Highly similar consecutive scans were excluded through combined manual review and program-assisted similarity assessment before model development. Models were developed using standardized preprocessing and five-fold cross-validation. ResNet-based and Swin Transformer\u2013based architectures were constructed and compared. ResNet-18 was further optimized by integrating deformable convolution (DCNv2) and efficient channel attention (ECA).<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>Swin Transformer Tiny achieved high training accuracy but showed unstable performance on the test set. In contrast, the optimized ResNet-18 demonstrated more consistent classification results across validation folds. The optimized ResNet-18 achieved accuracy, precision, recall, F1-score, specificity, and AUC values of 98.23%, 97.17%, 96.90%, 97.00%, 99.44%, and 99.76%, respectively. External validation showed consistent classification accuracy on independent data.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>The optimized ResNet-18 achieved accurate and stable MH staging using OCT images under a standardized training framework. Compared with deeper CNN and Transformer-based architectures, the proposed model demonstrated more reliable performance under limited-data conditions. These findings support the feasibility of automated MH staging using OCT images.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/fcell.2026.1886103","type":"journal-article","created":{"date-parts":[[2026,7,2]],"date-time":"2026-07-02T07:43:03Z","timestamp":1782978183000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Automated macular hole staging on optical coherence tomography using an optimized ResNet-18 framework"],"prefix":"10.3389","volume":"14","author":[{"given":"Zou","family":"Jianjun","sequence":"first","affiliation":[{"name":"Aier Eye Hospital, Jinan University","place":["Guangzhou, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xu","family":"Tao","sequence":"additional","affiliation":[{"name":"Aier Eye Hospital, Jinan University","place":["Guangzhou, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yang","family":"Bo","sequence":"additional","affiliation":[{"name":"Chengdu Aier Eye Hospital","place":["Chengdu, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zeng","family":"Fujiang","sequence":"additional","affiliation":[{"name":"School of Electronics and Information Engineering, Sichuan University","place":["Chengdu, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Huang","family":"Xiaoming","sequence":"additional","affiliation":[{"name":"Aier Sichuan Eye Hospital","place":["Chengdu, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Shi","family":"Saiqing","sequence":"additional","affiliation":[{"name":"Aier Sichuan Eye Hospital","place":["Chengdu, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Liu","family":"Hui","sequence":"additional","affiliation":[{"name":"Aier Sichuan Eye Hospital","place":["Chengdu, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhang","family":"Rongzhu","sequence":"additional","affiliation":[{"name":"School of Electronics and Information Engineering, Sichuan University","place":["Chengdu, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Dan","family":"Xiaofen","sequence":"additional","affiliation":[{"name":"Aier Sichuan Eye Hospital","place":["Chengdu, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Wu","family":"Tong","sequence":"additional","affiliation":[{"name":"Aier Sichuan Eye Hospital","place":["Chengdu, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,7,2]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"1811","DOI":"10.1016\/j.ophtha.2018.04.004","article-title":"Pivotal trial of an autonomous AI-based diagnostic system for detection of diabetic retinopathy in primary care offices","volume":"125","author":"Abr\u00e0moff","year":"2018","journal-title":"Ophthalmology"},{"key":"B2","doi-asserted-by":"publisher","first-page":"86","DOI":"10.3390\/jimaging10040086","article-title":"Denoising of optical coherence tomography images in ophthalmology using deep learning: a systematic review","volume":"10","author":"Ahmed","year":"2024","journal-title":"J. Imaging"},{"key":"B3","doi-asserted-by":"publisher","first-page":"108253","DOI":"10.1016\/j.cmpb.2024.108253","article-title":"Artificial intelligence in retinal screening using OCT images: a review","volume":"254","author":"Akpinar","year":"2024","journal-title":"Comput. Methods Programs Biomed."},{"key":"B4","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1109\/TMI.2020.3035208","article-title":"Utility of deep learning methods for referability classification of age-related macular degeneration from OCT images","volume":"40","author":"Burlina","year":"2021","journal-title":"IEEE Trans. Med. Imaging"},{"key":"B5","doi-asserted-by":"publisher","first-page":"280","DOI":"10.1016\/j.media.2019.03.009","article-title":"Not-so-supervised: a survey of semi-supervised, multi-instance, and transfer learning in medical image analysis","volume":"54","author":"Cheplygina","year":"2019","journal-title":"Med. Image Anal."},{"key":"B6","doi-asserted-by":"publisher","first-page":"764","DOI":"10.1109\/ICCV.2017.89","article-title":"Deformable convolutional networks","author":"Dai","year":"2017","journal-title":"Proc IEEE Int Conf Comput Vis."},{"key":"B7","doi-asserted-by":"publisher","first-page":"1342","DOI":"10.1038\/s41591-018-0107-6","article-title":"Clinically applicable deep learning for diagnosis and referral in retinal disease","volume":"24","author":"De Fauw","year":"2018","journal-title":"Nat. Med."},{"key":"B8","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1167\/iovs.61.7.13","article-title":"OCT biomarkers in retinal disease","volume":"61","author":"de Sisternes","year":"2020","journal-title":"Invest. Ophthalmol. Vis. Sci."},{"key":"B9","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2010.11929","article-title":"An image is worth 16\u00d716 words: transformers for image recognition at scale","author":"Dosovitskiy","year":"2021","journal-title":"Int. Conf. Learn Represent"},{"key":"B10","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1016\/j.ophtha.2020.08.017","article-title":"Vitreomacular interface diseases","author":"Duker","year":"2021","journal-title":"Ophthalmology"},{"key":"B11","doi-asserted-by":"publisher","first-page":"2611","DOI":"10.1016\/j.ophtha.2013.07.042","article-title":"The international vitreomacular traction study group classification of vitreomacular adhesion, traction, and macular hole","volume":"120","author":"Duker","year":"2013","journal-title":"Ophthalmology"},{"key":"B12","doi-asserted-by":"publisher","first-page":"24","DOI":"10.1038\/s41591-018-0316-z","article-title":"A guide to deep learning in healthcare","volume":"25","author":"Esteva","year":"2019","journal-title":"Nat. Med."},{"key":"B13","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1136\/bjophthalmol-2020-317495","article-title":"Idiopathic full-thickness macular hole: natural history and pathogenesis","volume":"105","author":"Ezra","year":"2021","journal-title":"Br. J. Ophthalmol."},{"key":"B14","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1038\/nbt.3590","article-title":"The development of OCT technology","volume":"34","author":"Fujimoto","year":"2016","journal-title":"Nat. Biotechnol."},{"key":"B15","doi-asserted-by":"publisher","first-page":"100823","DOI":"10.1016\/j.preteyeres.2019.100823","article-title":"Insights into the pathogenesis of full-thickness macular holes","volume":"76","author":"Govetto","year":"2020","journal-title":"Prog. Retin Eye Res."},{"key":"B16","doi-asserted-by":"publisher","first-page":"102305","DOI":"10.1016\/j.media.2021.102305","article-title":"UNETR: transformers for 3D medical image segmentation","volume":"76","author":"Hatamizadeh","year":"2022","journal-title":"Med. Image Anal."},{"key":"B17","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":"Proc. IEEE Conf. Comput. Vis. Pattern Recognit."},{"key":"B18","doi-asserted-by":"publisher","first-page":"53","DOI":"10.3390\/jimaging11010053","article-title":"Comparative analysis of deep learning architectures for macular hole segmentation in OCT images","volume":"11","author":"Herath","year":"2025","journal-title":"J. Imaging"},{"key":"B19","doi-asserted-by":"publisher","first-page":"1178","DOI":"10.1126\/science.1957169","article-title":"Optical coherence tomography","volume":"254","author":"Huang","year":"1991","journal-title":"Science"},{"key":"B20","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1097\/IAE.0000000000002649","article-title":"Tractional retinal disorders","volume":"40","author":"Johnson","year":"2020","journal-title":"Retina"},{"key":"B21","doi-asserted-by":"publisher","first-page":"6913","DOI":"10.1038\/s41598-024-57562-5","article-title":"Accuracy of generative deep learning model for macular anatomy prediction from OCT images in macular hole surgery","volume":"14","author":"Kwon","year":"2024","journal-title":"Sci. Rep."},{"key":"B22","doi-asserted-by":"publisher","first-page":"1\u2013e297","DOI":"10.1016\/S2589-7500(19)30126-3","article-title":"Deep learning for detecting retinal diseases from optical coherence tomography images","author":"Liu","year":"2019","journal-title":"Lancet Digit. Health"},{"key":"B23","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":"Proc. IEEE Int. Conf. Comput. Vis."},{"key":"B24","doi-asserted-by":"publisher","first-page":"99\u2013e131","DOI":"10.1111\/aos.14567","article-title":"Vitreomacular traction and macular hole development","author":"Maier","year":"2021","journal-title":"Acta Ophthalmol."},{"key":"B25","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1097\/IAE.0000000000002874","article-title":"Surgical classification of macular holes","volume":"41","author":"Michalewska","year":"2021","journal-title":"Retina"},{"key":"B26","doi-asserted-by":"publisher","first-page":"1345","DOI":"10.1109\/TKDE.2009.191","article-title":"A survey on transfer learning","volume":"22","author":"Pan","year":"2010","journal-title":"IEEE Trans. Knowl. Data Eng."},{"key":"B27","doi-asserted-by":"publisher","first-page":"3347","DOI":"10.48550\/arXiv.1902.07208","article-title":"Transfusion: understanding transfer learning for medical imaging","volume":"32","author":"Raghu","year":"2019","journal-title":"Adv. Neural Inf. Process Syst."},{"key":"B28","doi-asserted-by":"publisher","first-page":"1","DOI":"10.3390\/jcm11030694","article-title":"Macular holes: main clinical presentations, diagnosis, and therapies","volume":"11","author":"Rossi","year":"2022","journal-title":"J. Clin. Med."},{"key":"B29","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1097\/IAE.0000000000002643","article-title":"Morphological characteristics of macular holes","volume":"40","author":"Schumann","year":"2020","journal-title":"Retina"},{"key":"B30","doi-asserted-by":"publisher","first-page":"60","DOI":"10.1186\/s40537-019-0197-0","article-title":"A survey on image data augmentation for deep learning","volume":"6","author":"Shorten","year":"2019","journal-title":"J. Big Data"},{"key":"B31","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1016\/j.preteyeres.2017.11.003","article-title":"Optical coherence tomography angiography","volume":"64","author":"Spaide","year":"2018","journal-title":"Prog. Retin Eye Res."},{"key":"B32","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1038\/s41433-019-0604-5","article-title":"Clinical variability in macular hole progression","volume":"34","author":"Steel","year":"2020","journal-title":"Eye"},{"key":"B33","doi-asserted-by":"publisher","first-page":"248","DOI":"10.1038\/s41433-020-01198-0","article-title":"The evolution of macular hole surgery","volume":"35","author":"Steel","year":"2021","journal-title":"Eye"},{"key":"B34","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1016\/j.survophthal.2019.08.003","article-title":"Optical coherence tomography of macular holes","volume":"65","author":"Tanner","year":"2020","journal-title":"Surv. Ophthalmol."},{"key":"B35","doi-asserted-by":"publisher","first-page":"167","DOI":"10.1136\/bjophthalmol-2018-313173","article-title":"Artificial intelligence and deep learning in ophthalmology","volume":"103","author":"Ting","year":"2019","journal-title":"Br. J. Ophthalmol."},{"key":"B36","doi-asserted-by":"publisher","first-page":"44","DOI":"10.1038\/s41591-018-0300-7","article-title":"High-performance medicine: the convergence of human and artificial intelligence","volume":"25","author":"Topol","year":"2019","journal-title":"Nat. Med."},{"key":"B37","doi-asserted-by":"publisher","first-page":"16652","DOI":"10.1038\/s41598-024-67591-7","article-title":"Deep learning in retinal OCT analysis: current applications","volume":"14","author":"Toto","year":"2024","journal-title":"Sci. Rep."},{"key":"B38","doi-asserted-by":"publisher","first-page":"101832","DOI":"10.1016\/j.media.2020.101832","article-title":"Robust OCT image analysis methods for retinal diseases","volume":"67","author":"Venhuizen","year":"2021","journal-title":"Med. Image Anal."},{"key":"B39","doi-asserted-by":"publisher","first-page":"11534","DOI":"10.1109\/CVPR42600.2020.01155","article-title":"ECA-Net: efficient channel attention for deep convolutional neural networks","author":"Wang","year":"2020","journal-title":"Proc. IEEE Conf. Comput. Vis. Pattern Recognit."},{"key":"B40","doi-asserted-by":"publisher","first-page":"102539","DOI":"10.1016\/j.compmedimag.2025.102539","article-title":"Retinal OCT image segmentation with deep learning: advances and challenges","volume":"123","author":"Wang","year":"2025","journal-title":"Comput. Med. Imaging Graph"},{"key":"B41","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1016\/j.ophtha.2019.10.019","article-title":"Visual outcomes in macular hole surgery","volume":"127","author":"Wong","year":"2020","journal-title":"Ophthalmology"},{"key":"B42","doi-asserted-by":"publisher","first-page":"3","DOI":"10.1007\/978-3-030-01234-2_1","article-title":"CBAM: convolutional block attention module","author":"Woo","year":"2018","journal-title":"Proc. Eur. Conf. Comput. Vis."},{"key":"B43","doi-asserted-by":"publisher","first-page":"2456","DOI":"10.3390\/diagnostics12102456","article-title":"Multimodal deep fusion network for macular hole classification","volume":"12","author":"Xiao","year":"2022","journal-title":"Diagnostics"},{"key":"B44","doi-asserted-by":"publisher","first-page":"103938","DOI":"10.1016\/j.compbiomed.2020.103938","article-title":"Automated analysis of macular hole in OCT images: a review","volume":"123","author":"Xu","year":"2020","journal-title":"Comput. Biol. Med."},{"key":"B45","doi-asserted-by":"publisher","first-page":"261","DOI":"10.1109\/TPAMI.2019.2948742","article-title":"Deformable ConvNets v2: more deformable, better results","volume":"43","author":"Zhu","year":"2021","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."}],"container-title":["Frontiers in Cell and Developmental Biology"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fcell.2026.1886103\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,2]],"date-time":"2026-07-02T07:43:04Z","timestamp":1782978184000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fcell.2026.1886103\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7,2]]},"references-count":45,"alternative-id":["10.3389\/fcell.2026.1886103"],"URL":"https:\/\/doi.org\/10.3389\/fcell.2026.1886103","relation":{},"ISSN":["2296-634X"],"issn-type":[{"value":"2296-634X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,7,2]]},"article-number":"1886103"}}