{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,29]],"date-time":"2026-10-29T00:46:44Z","timestamp":1793234804286,"version":"4.5.0"},"reference-count":48,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,4,17]],"date-time":"2026-04-17T00:00:00Z","timestamp":1776384000000},"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>Introduction<\/jats:title>\n                    <jats:p>Retinal diseases represent a leading cause of visual impairment and blindness worldwide, with early and accurate diagnosis being crucial for preventing irreversible vision loss. Although deep learning techniques have achieved significant advances in fundus image analysis, existing methods predominantly focus on single tasks, treating vessel segmentation and disease diagnosis as independent problems without fully leveraging their intrinsic relationships. Furthermore, the lack of transparency in deep model predictions limits clinical adoption; while full interpretability remains an open challenge, post-hoc techniques such as Grad-CAM can provide partial transparency by highlighting influential image regions.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>This study presents V-MNet, a vessel-guided multi-task deep learning framework that simultaneously achieves retinal vessel segmentation and multi-disease classification, and provides visual transparency through Grad-CAM-based class activation mapping to support clinical decision-making. The framework comprises three core modules: a shared encoder extracts multi-scale feature representations; a segmentation decoder employs a U-Net-style architecture to generate vessel masks; and a classification decoder incorporates an innovative vessel-guided mechanism that explicitly transfers structural priors from the segmentation branch to the classification task, enabling the model to precisely localize pathological regions. Concurrently, an integrated Grad-CAM module generates post-hoc class activation maps for each disease category, highlighting spatially relevant lesion regions for clinician review. Comprehensive experiments were conducted on four public datasets-RFMiD, ODIR-5K, DRIVE, and EyePACS-light-v2.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>Experiments demonstrate that V-MNet achieves a Dice coefficient of 0.831 and AUC of 0.985 for vessel segmentation tasks, and an average AUC of 0.978 with F1-score of 0.935 for multi-disease classification tasks, significantly outperforming single-task baseline models and existing state-of-the-art methods. Ablation studies systematically quantify the performance contributions of multi-task learning and the vessel-guided mechanism, confirming the effectiveness of the framework's core innovations.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Discussion<\/jats:title>\n                    <jats:p>V-MNet demonstrates broad application potential as a computer-aided diagnostic tool by jointly leveraging vascular structure and disease pathology for superior performance and visual transparency. The vessel-guided multi-task design effectively exploits the intrinsic relationship between vessel segmentation and disease classification, while the integrated Grad-CAM module addresses the lack of model transparency, facilitating clinical adoption and supporting clinical decision-making.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/fmed.2026.1799745","type":"journal-article","created":{"date-parts":[[2026,4,17]],"date-time":"2026-04-17T04:13:04Z","timestamp":1776399184000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":2,"title":["A vessel-guided multi-task deep learning framework with visual interpretability for simultaneous retinal vessel segmentation and multi-disease classification from fundus images"],"prefix":"10.3389","volume":"13","author":[{"given":"Qi","family":"Li","sequence":"first","affiliation":[{"name":"Department of Ophthalmology, The Second Affiliated Hospital of Anhui Medical University","place":["Hefei, China"]},{"name":"Department of Refractive Surgery, Nanhai Aier Eye Hospital","place":["Foshan, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Liming","family":"Tao","sequence":"additional","affiliation":[{"name":"Department of Ophthalmology, The Second Affiliated Hospital of Anhui Medical University","place":["Hefei, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,4,17]]},"reference":[{"key":"B1","volume-title":"Blindness and Vision Impairment","year":"2026"},{"key":"B2","doi-asserted-by":"publisher","first-page":"112346","DOI":"10.1016\/j.diabres.2025.112346","article-title":"IDF diabetes Atlas: a worldwide review of studies utilizing retinal photography to screen for diabetic retinopathy from 2017 to 2024 inclusive","volume":"226","author":"Owens","year":"2025","journal-title":"Diabetes Res Clini Pract"},{"key":"B3","doi-asserted-by":"publisher","first-page":"2081","DOI":"10.1016\/j.ophtha.2014.05.013","article-title":"Global prevalence of glaucoma and projections of glaucoma burden through 2040: a systematic review and meta-analysis","volume":"121","author":"Tham YC Li","year":"2014","journal-title":"Ophthalmology"},{"key":"B4","doi-asserted-by":"publisher","first-page":"169","DOI":"10.1109\/RBME.2010.2084567","article-title":"Retinal imaging and image analysis","volume":"3","author":"Abr\u00e0moff","year":"2010","journal-title":"IEEE Rev Biomed Eng"},{"key":"B5","doi-asserted-by":"publisher","first-page":"170","DOI":"10.1016\/j.diabres.2013.05.006","article-title":"Preventing diabetes blindness: cost effectiveness of a screening programme using digital non-mydriatic fundus photography for diabetic retinopathy in a primary health care setting in South Africa","volume":"101","author":"Khan","year":"2013","journal-title":"Diabetes Res Clin Pract"},{"key":"B6","doi-asserted-by":"publisher","first-page":"59","DOI":"10.1016\/B978-0-12-817440-1.00003-6","article-title":"Fundus retinal image analyses for screening and diagnosing diabetic retinopathy, macular edema, and glaucoma disorders","volume":"1","author":"Chalakkal","year":"2020","journal-title":"Diabet Fundus"},{"key":"B7","doi-asserted-by":"publisher","first-page":"783","DOI":"10.1136\/bjophthalmol-2011-301378","article-title":"The number of ophthalmologists in practice and training worldwide: a growing gap despite more than 200 000 practitioners","volume":"96","author":"Resnikoff","year":"2012","journal-title":"Br J Ophthalmol"},{"key":"B8","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":"B9","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2501.04073","article-title":"Deep learning for ophthalmology: the state-of-the-art and future trends","author":"Nguyen","year":"2025","journal-title":"arXiv"},{"key":"B10","doi-asserted-by":"publisher","first-page":"3108","DOI":"10.3390\/app13053108","article-title":"Classification and segmentation of diabetic retinopathy: a systemic review","volume":"13","author":"Shaukat","year":"2023","journal-title":"Appl Sci"},{"key":"B11","doi-asserted-by":"publisher","first-page":"134","DOI":"10.3390\/diagnostics12010134","article-title":"Review of machine learning applications using retinal fundus images","volume":"12","author":"Jeong","year":"2022","journal-title":"Diagnostics"},{"key":"B12","doi-asserted-by":"publisher","first-page":"102559","DOI":"10.1016\/j.bspc.2021.102559","article-title":"ECNet: An evolutionary convolutional network for automated glaucoma detection using fundus images","volume":"67","author":"Nayak","year":"2021","journal-title":"Biomed Signal Process Cont"},{"key":"B13","doi-asserted-by":"publisher","first-page":"563","DOI":"10.21917\/ijsc.2013.0083","article-title":"Diagnosis of diabetic retinopathy using machine learning techniques","volume":"3","author":"Priya","year":"2013","journal-title":"ICTACT J Soft Comp"},{"key":"B14","doi-asserted-by":"publisher","first-page":"9681034","DOI":"10.1155\/2022\/9681034","article-title":"A classification tree model with optical coherence tomography angiography variables to screen early-stage diabetic retinopathy in diabetic patients","volume":"2022","author":"Yao","year":"2022","journal-title":"J Ophthalmol"},{"key":"B15","doi-asserted-by":"publisher","first-page":"6780","DOI":"10.3390\/s22186780","article-title":"Deep learning for diabetic retinopathy analysis: a review, research challenges, and future directions","volume":"22","author":"Nadeem","year":"2022","journal-title":"Sensors"},{"key":"B16","doi-asserted-by":"publisher","first-page":"1902","DOI":"10.1049\/el.2014.2854","article-title":"Hierarchical local binary pattern for branch retinal vein occlusion recognition with fluorescein angiography images","volume":"50","author":"Zhang","year":"2014","journal-title":"Electron Lett"},{"key":"B17","doi-asserted-by":"publisher","first-page":"184","DOI":"10.1109\/JBHI.2015.2490798","article-title":"Retinal disease screening through local binary patterns","volume":"21","author":"Morales","year":"2015","journal-title":"IEEE J Biomed Health Inform"},{"key":"B18","doi-asserted-by":"publisher","first-page":"14","DOI":"10.1007\/s13755-022-00181-z","article-title":"Features extraction using encoded local binary pattern for detection and grading diabetic retinopathy","volume":"10","author":"Berbar","year":"2022","journal-title":"Health Inform Sci Syst"},{"key":"B19","doi-asserted-by":"publisher","first-page":"58","DOI":"10.1080\/10095020.2020.1843376","article-title":"Feature detection and description for image matching: from hand-crafted design to deep learning","volume":"24","author":"Chen","year":"2021","journal-title":"Geo-Spatial Inform Sci"},{"key":"B20","doi-asserted-by":"publisher","first-page":"380","DOI":"10.2741\/4725","article-title":"State-of-the-art review on deep learning in medical imaging","volume":"24","author":"Biswas","year":"2019","journal-title":"Front Biosci-Landmark"},{"key":"B21","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":"B22","doi-asserted-by":"publisher","first-page":"3883","DOI":"10.3390\/s21113883","article-title":"ResNet based deep features and random forest classifier for diabetic retinopathy detection","volume":"21","author":"Yaqoob","year":"2021","journal-title":"Sensors"},{"key":"B23","doi-asserted-by":"publisher","first-page":"157","DOI":"10.1186\/s12859-023-05293-1","article-title":"Development of revised ResNet-50 for diabetic retinopathy detection","volume":"24","author":"Lin","year":"2023","journal-title":"BMC Bioinformatics"},{"key":"B24","first-page":"770","article-title":"Deep residual learning for image recognition","volume-title":"Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition","author":"He","year":"2016"},{"key":"B25","first-page":"4700","article-title":"Densely connected convolutional networks","volume-title":"Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition","author":"Huang","year":"2017"},{"key":"B26","doi-asserted-by":"publisher","first-page":"1975","DOI":"10.3390\/diagnostics12081975","article-title":"Automated diabetic retinopathy detection using horizontal and vertical patch division-based pre-trained DenseNET with digital fundus images","volume":"12","author":"Kobat","year":"2022","journal-title":"Diagnostics"},{"key":"B27","doi-asserted-by":"publisher","first-page":"14251","DOI":"10.1007\/s00521-024-09820-w","article-title":"Combining EfficientNet with ML-Decoder classification head for multi-label retinal disease classification","volume":"36","author":"Sivaz","year":"2024","journal-title":"Neural Comp Appl"},{"key":"B28","doi-asserted-by":"publisher","first-page":"1290803","DOI":"10.3389\/fnins.2023.1290803","article-title":"Multi-label classification of retinal disease via a novel vision transformer model","volume":"17","author":"Wang","year":"2024","journal-title":"Front Neurosci"},{"key":"B29","doi-asserted-by":"publisher","first-page":"66556","DOI":"10.1109\/ACCESS.2024.3398203","article-title":"Unraveling the black box: a review of explainable deep learning healthcare techniques","volume":"12","author":"Murad","year":"2024","journal-title":"IEEE Access"},{"key":"B30","doi-asserted-by":"publisher","first-page":"364","DOI":"10.4093\/dmj.2018.0182","article-title":"Pathophysiology of diabetic retinopathy: the old and the new","volume":"42","author":"Kusuhara","year":"2018","journal-title":"Diabet Metabol J"},{"key":"B31","first-page":"234","article-title":"U-Net: Convolutional networks for biomedical image segmentation","volume-title":"International Conference on Medical Image Computing and Computer-Assisted Intervention","author":"Ronneberger","year":"2015"},{"key":"B32","doi-asserted-by":"publisher","first-page":"1478","DOI":"10.3390\/mi12121478","article-title":"Blood vessel segmentation of retinal image based on dense-U-Net network","volume":"12","author":"Li","year":"2021","journal-title":"Micromachines"},{"key":"B33","doi-asserted-by":"publisher","first-page":"9948","DOI":"10.3934\/mbe.2022464","article-title":"A generative adversarial network for segmenting retinal vessels","volume":"19","author":"Yue","year":"2022","journal-title":"Math Biosci Eng"},{"key":"B34","doi-asserted-by":"crossref","first-page":"130591","DOI":"10.1016\/j.eswa.2025.130591","article-title":"VasCA-Net: a vascular channel attention network for retinal vessel segmentation","volume":"303","author":"Ma","year":"2026","journal-title":"Expert Syst Appl"},{"key":"B35","doi-asserted-by":"crossref","first-page":"105879","DOI":"10.1016\/j.dsp.2025.105879","article-title":"APU-Net: a U-Net enhanced network with dynamic feature fusion and pyramid cross-attention mechanism for polyp segmentation","volume":"172","author":"Li","year":"2026","journal-title":"Digit Signal Process"},{"key":"B36","doi-asserted-by":"publisher","first-page":"111985","DOI":"10.1109\/ACCESS.2021.3102176","article-title":"Retinal vessel segmentation using deep learning: a review","volume":"9","author":"Chen","year":"2021","journal-title":"IEEE Access"},{"key":"B37","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1145\/3561653","article-title":"multi-label classification with an adversarial-based denoising autoencoder for medical image annotation","volume":"14","author":"Chai","year":"2023","journal-title":"ACM Trans Managem Inform Syst"},{"key":"B38","doi-asserted-by":"publisher","first-page":"e53207","DOI":"10.2196\/53207","article-title":"Melhus \u00c5, McKee M, Stuckler D. How explainable artificial intelligence can increase or decrease clinicians' trust in AI applications in health care: systematic review","volume":"3","author":"Rosenbacke","year":"2024","journal-title":"JMIR AI"},{"key":"B39","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1145\/3510413","article-title":"Avoiding overfitting: A survey on regularization methods for convolutional neural networks","volume":"54","author":"Santos","year":"2022","journal-title":"ACM Comp Surv"},{"key":"B40","doi-asserted-by":"publisher","first-page":"705","DOI":"10.3390\/mi14030705","article-title":"multi-label detection deep learning model with attention-guided image enhancement for retinal images","volume":"14","author":"Li","year":"2023","journal-title":"Micromachines"},{"key":"B41","doi-asserted-by":"publisher","first-page":"4492","DOI":"10.3390\/s25144492","article-title":"Multi-modal AI for multi-label retinal disease prediction using OCT and fundus images: a hybrid approach","volume":"25","author":"Zedadra","year":"2025","journal-title":"Sensors"},{"key":"B42","doi-asserted-by":"publisher","first-page":"106519","DOI":"10.1016\/j.compbiomed.2022.106519","article-title":"Discriminative kernel convolution network for multi-label ophthalmic disease detection on imbalanced fundus image dataset","volume":"153","author":"Bhati","year":"2023","journal-title":"Comput Biol Med"},{"key":"B43","doi-asserted-by":"publisher","first-page":"194551","DOI":"10.1109\/ACCESS.2020.3033273","article-title":"Retinal vessel segmentation combined with generative adversarial networks and dense U-Net","volume":"8","author":"Guo","year":"2020","journal-title":"IEEE Access"},{"key":"B44","first-page":"89","article-title":"Transfer learning based classification of diabetic retinopathy on the Kaggle EyePACS dataset","volume-title":"International Conference on Medical Imaging and Computer-Aided Diagnosis","author":"Tariq","year":"2022"},{"key":"B45","first-page":"618","article-title":"Grad-cam: Visual explanations from deep networks via gradient-based localization","volume-title":"Proceedings of the IEEE International Conference on Computer Vision","author":"Selvaraju","year":"2017"},{"key":"B46","doi-asserted-by":"publisher","first-page":"327","DOI":"10.48550\/arXiv.1512.03385","article-title":"Deep residual learning","volume":"7","author":"He","year":"2015","journal-title":"Image Recognition"},{"key":"B47","doi-asserted-by":"publisher","first-page":"1148","DOI":"10.3390\/e25081148","article-title":"Multi-task segmentation and classification network for artery\/vein classification in retina fundus","volume":"25","author":"Yi","year":"2023","journal-title":"Entropy"},{"key":"B48","doi-asserted-by":"publisher","first-page":"5900","DOI":"10.1109\/ACCESS.2025.3650159","article-title":"TEDR-YNet: a dual-encoder retinal multi-lesion segmentation network","volume":"14","author":"Qi","year":"2026","journal-title":"IEEE Access"}],"container-title":["Frontiers in Medicine"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fmed.2026.1799745\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,4,17]],"date-time":"2026-04-17T04:13:07Z","timestamp":1776399187000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fmed.2026.1799745\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,4,17]]},"references-count":48,"alternative-id":["10.3389\/fmed.2026.1799745"],"URL":"https:\/\/doi.org\/10.3389\/fmed.2026.1799745","relation":{},"ISSN":["2296-858X"],"issn-type":[{"value":"2296-858X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,4,17]]},"article-number":"1799745"}}