{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,15]],"date-time":"2026-06-15T04:54:37Z","timestamp":1781499277094,"version":"3.54.1"},"reference-count":43,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,6,15]],"date-time":"2026-06-15T00:00:00Z","timestamp":1781481600000},"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>Accurate radiographic grading of knee osteoarthritis (KOA) using the Kellgren\u2013Lawrence (KL) system is essential for clinical decision-making but suffers from subjective interpretation and inter-observer variability. Deep learning approaches have shown promise, yet existing models often rely on single-scale features, lack effective attention mechanisms, and generalize poorly to external datasets.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>We developed a novel multi-scale attention-guided deep learning framework that synergistically integrates three key components: (1) a Feature Pyramid Network (FPN) for multi-scale feature fusion across different spatial resolutions; (2) a dual-attention mechanism combining a Squeeze-and-Excitation (SE) module (channel-wise recalibration) and a Self-Attention module (long-range spatial dependency modeling); and (3) knowledge distillation to enhance generalization. A total of 5,000 radiographs from the OSAIL KL dataset were used for model development (3,500 training, 1,500 internal validation), with external validation on an independent dataset of 2,000 radiographs. Performance was compared against ResNet50, ResNet18, MobileNet V2, and DenseNet using 5-fold cross-validation.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>On internal validation, our model achieved superior performance across all metrics (F1: 0.726, precision: 0.740, MCC: 0.620, accuracy: 0.726), outperforming all baseline architectures. On external validation, the model maintained robust generalization (F1: 0.656, precision: 0.683, MCC: 0.564, accuracy: 0.685), with only a 4.04% drop in accuracy\u2014demonstrating clinical reliability. Misclassifications were predominantly confined to adjacent KL grades, with no extreme errors (e.g., Grade 0 vs. Grade 4). Gradient-weighted Class Activation Mapping (Grad-CAM) heatmaps revealed that the model\u2019s attention aligns with clinically relevant regions (joint space, osteophyte sites), enhancing interpretability.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>The proposed multi-scale attention-guided framework enables reliable, automated KL grading of KOA from radiographs with strong generalization across independent datasets. By integrating FPN, dual-attention mechanisms, and knowledge distillation, this approach addresses key limitations of prior models and may support standardized, objective, and interpretable clinical assessment of KOA severity.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/fmed.2026.1809768","type":"journal-article","created":{"date-parts":[[2026,6,15]],"date-time":"2026-06-15T04:25:40Z","timestamp":1781497540000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Automated Kellgren\u2013Lawrence grading of knee osteoarthritis using a multi-scale attention-based deep learning framework"],"prefix":"10.3389","volume":"13","author":[{"given":"Henghui","family":"Zhang","sequence":"first","affiliation":[{"name":"Shanghai Key Laboratory of Orthopaedic Implants, Department of Orthopaedic Surgery, Shanghai Jiao Tong University School of Medicine, School of Biomedical Engineering, Shanghai Ninth People\u2019s Hospital, Shanghai Jiao Tong University","place":["Shanghai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Chui","family":"Kong","sequence":"additional","affiliation":[{"name":"School of Information Science and Technology, Fudan University","place":["Shanghai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hanwen","family":"Chang","sequence":"additional","affiliation":[{"name":"Shanghai Key Laboratory of Orthopaedic Implants, Department of Orthopaedic Surgery, Shanghai Jiao Tong University School of Medicine, School of Biomedical Engineering, Shanghai Ninth People\u2019s Hospital, Shanghai Jiao Tong University","place":["Shanghai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yaokai","family":"Gan","sequence":"additional","affiliation":[{"name":"Shanghai Key Laboratory of Orthopaedic Implants, Department of Orthopaedic Surgery, Shanghai Jiao Tong University School of Medicine, School of Biomedical Engineering, Shanghai Ninth People\u2019s Hospital, Shanghai Jiao Tong University","place":["Shanghai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,6,15]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"e0288561","DOI":"10.1371\/journal.pone.0288561","article-title":"Temporal trends in the disease burden of osteoarthritis from 1990 to 2019, and projections until 2030.","volume":"18","author":"Chen","year":"2023","journal-title":"PLoS One."},{"key":"B2","doi-asserted-by":"publisher","first-page":"293","DOI":"10.1016\/j.mcna.2019.10.007","article-title":"Osteoarthritis: pathology, diagnosis, and treatment options.","volume":"104","author":"Abramoff","year":"2020","journal-title":"Med Clin North Am."},{"key":"B3","doi-asserted-by":"publisher","first-page":"199","DOI":"10.1186\/s13075-025-03658-w","article-title":"Epidemiological trends of osteoarthritis at the global, regional, and national levels from 1990 to 2021 and projections to 2050.","volume":"27","author":"Qiao","year":"2025","journal-title":"Arthritis Res Ther."},{"key":"B4","doi-asserted-by":"publisher","first-page":"240","DOI":"10.1016\/j.jphys.2021.08.009","article-title":"Knee osteoarthritis.","volume":"67","author":"Goff","year":"2021","journal-title":"J Physiother."},{"key":"B5","doi-asserted-by":"publisher","first-page":"e20230264","DOI":"10.1590\/1806-9282.20230264","article-title":"Effects of low-level laser therapy on acupuncture points on knee pain and function in knee osteoarthritis.","volume":"70","author":"Yeti\u015fir","year":"2024","journal-title":"Rev Associa\u00e7\u00e3o M\u00e9d Bras."},{"key":"B6","doi-asserted-by":"publisher","first-page":"20","DOI":"10.2174\/1573397117666211116111738","article-title":"Emerging and new treatment options for knee osteoarthritis.","volume":"18","author":"Siddiq","year":"2022","journal-title":"Curr Rheumatol Rev."},{"key":"B7","doi-asserted-by":"publisher","first-page":"675","DOI":"10.1111\/jebm.12627","article-title":"Evaluation and management of knee osteoarthritis.","volume":"17","author":"Zhu","year":"2024","journal-title":"J Evid Based Med."},{"key":"B8","doi-asserted-by":"publisher","first-page":"258","DOI":"10.1186\/s13018-022-03142-0","article-title":"The effects of total knee arthroplasty on knee proprioception of patients with knee osteoarthritis: a meta-analysis.","volume":"17","author":"Xue","year":"2022","journal-title":"J Orthop Surg Res."},{"key":"B9","doi-asserted-by":"publisher","first-page":"805","DOI":"10.2106\/jbjs.21.00892","article-title":"Osseous morphological differences in knee osteoarthritis.","volume":"104","author":"Siddiqi","year":"2022","journal-title":"J Bone Joint Surg Am."},{"key":"B10","doi-asserted-by":"publisher","first-page":"1530","DOI":"10.1016\/j.joca.2021.07.013","article-title":"Performance of knee image digital analysis of radiographs of patients with end-stage knee osteoarthritis.","volume":"29","author":"Jansen","year":"2021","journal-title":"Osteoarthritis Cartilage."},{"key":"B11","doi-asserted-by":"publisher","first-page":"120454","DOI":"10.1016\/j.jns.2022.120454","article-title":"Deep learning in acute vertigo diagnosis.","volume":"443","author":"Rastall","year":"2022","journal-title":"J Neurol Sci."},{"key":"B12","doi-asserted-by":"publisher","first-page":"1179","DOI":"10.1093\/europace\/euaa377","article-title":"Deep learning and the electrocardiogram: review of the current state-of-the-art.","volume":"23","author":"Somani","year":"2021","journal-title":"Europace."},{"key":"B13","doi-asserted-by":"publisher","first-page":"152","DOI":"10.1186\/s13073-021-00968-x","article-title":"Deep learning in cancer diagnosis, prognosis and treatment selection.","volume":"13","author":"Tran","year":"2021","journal-title":"Genome Med."},{"key":"B14","doi-asserted-by":"publisher","first-page":"844","DOI":"10.1016\/j.ard.2025.01.012","article-title":"Deep learning-based clustering for endotyping and post-arthroplasty response classification using knee osteoarthritis multiomic data.","volume":"84","author":"Rockel","year":"2025","journal-title":"Ann Rheum Dis."},{"key":"B15","doi-asserted-by":"publisher","first-page":"102840","DOI":"10.1016\/j.arr.2025.102840","article-title":"Advancements in deep learning-based image screening for orthopedic conditions: emphasis on osteoporosis, osteoarthritis, and bone tumors.","volume":"111","author":"Guo","year":"2025","journal-title":"Ageing Res Rev."},{"key":"B16","doi-asserted-by":"publisher","first-page":"75","DOI":"10.1016\/j.semcancer.2023.07.003","article-title":"Artificial intelligence-based radiomics in bone tumors: technical advances and clinical application.","volume":"95","author":"Meng","year":"2023","journal-title":"Semin Cancer Biol."},{"key":"B17","doi-asserted-by":"publisher","first-page":"530","DOI":"10.1109\/tbme.2025.3587003","article-title":"Confidence-driven deep learning framework for early detection of knee osteoarthritis.","volume":"73","author":"Wang","year":"2025","journal-title":"IEEE Trans Biomed Eng."},{"key":"B18","doi-asserted-by":"publisher","first-page":"1549","DOI":"10.26355\/eurrev_202203_28220","article-title":"Artificial intelligence assistance in radiographic detection and classification of knee osteoarthritis and its severity: a cross-sectional diagnostic study.","volume":"26","author":"Pongsakonpruttikul","year":"2022","journal-title":"Eur Rev Med Pharmacol Sci."},{"key":"B19","doi-asserted-by":"publisher","first-page":"100161","DOI":"10.1016\/j.ostima.2023.100161","article-title":"Artificial intelligence in knee osteoarthritis: a comprehensive review for 2022.","volume":"3","author":"Cigdem","year":"2023","journal-title":"Osteoarthr Imaging."},{"key":"B20","doi-asserted-by":"publisher","first-page":"110249","DOI":"10.1016\/j.ejrad.2022.110249","article-title":"External validation of an artificial intelligence tool for radiographic knee osteoarthritis severity classification.","volume":"150","author":"Brejneb\u00f8l","year":"2022","journal-title":"Eur J Radiol."},{"key":"B21","doi-asserted-by":"publisher","first-page":"398","DOI":"10.1007\/s11547-022-01476-7","article-title":"Automatic detection and classification of knee osteoarthritis using deep learning approach.","volume":"127","author":"Abdullah","year":"2022","journal-title":"Radiol Med."},{"key":"B22","doi-asserted-by":"publisher","first-page":"e21703","DOI":"10.1016\/j.heliyon.2023.e21703","article-title":"Development of an automated optimal distance feature-based decision system for diagnosing knee osteoarthritis using segmented X-ray images.","volume":"9","author":"Fatema","year":"2023","journal-title":"Heliyon."},{"key":"B23","doi-asserted-by":"publisher","first-page":"bbab503","DOI":"10.1093\/bib\/bbab503","article-title":"ResNet18DNN: prediction approach of drug-induced liver injury by deep neural network with ResNet18.","volume":"23","author":"Chen","year":"2022","journal-title":"Brief Bioinform."},{"key":"B24","doi-asserted-by":"publisher","first-page":"740","DOI":"10.1186\/s12888-025-07194-4","article-title":"MentalAId: an improved DenseNet model to assist scalable psychosis assessment.","volume":"25","author":"Li","year":"2025","journal-title":"BMC Psychiatry."},{"key":"B25","doi-asserted-by":"publisher","first-page":"20220390","DOI":"10.1259\/dmfr.20220390","article-title":"A support vector machine-based algorithm to identify bisphosphonate-related osteonecrosis throughout the mandibular bone by using cone beam computerized tomography images.","volume":"52","author":"G\u00fcrses","year":"2023","journal-title":"Dentomaxillofac Radiol."},{"key":"B26","doi-asserted-by":"publisher","first-page":"107291","DOI":"10.1016\/j.aap.2023.107291","article-title":"Computed tomography assessment of pelvic bone density: associations with age and pelvic fracture in motor vehicle crashes.","volume":"193","author":"Weaver","year":"2023","journal-title":"Accid Anal Prev."},{"key":"B27","doi-asserted-by":"publisher","first-page":"7225","DOI":"10.1038\/s41598-024-57002-4","article-title":"Gaussian Aquila optimizer based dual convolutional neural networks for identification and grading of osteoarthritis using knee joint images.","volume":"14","author":"Subha","year":"2024","journal-title":"Sci Rep."},{"key":"B28","unstructured":"Nejat\n              P.\n            \n          \n          OSAIL Knee Osteoarthritis KL Scoring Dataset.\n          \n          2022"},{"key":"B29","volume-title":"Knee Osteoarthritis Dataset with Severity Grading.","author":"Tiwari","year":"2021"},{"key":"B30","doi-asserted-by":"publisher","first-page":"11314","DOI":"10.1038\/s41598-023-38109-6","article-title":"Region of interest (ROI) selection using vision transformer for automatic analysis using whole slide images.","volume":"13","author":"Hossain","year":"2023","journal-title":"Sci Rep."},{"key":"B31","doi-asserted-by":"publisher","first-page":"6934","DOI":"10.1109\/tcyb.2024.3424858","article-title":"Gaussian particle filtering for nonlinear systems with heavy-tailed noises: a progressive transform-based approach.","volume":"54","author":"Zhang","year":"2024","journal-title":"IEEE Trans Cybern."},{"key":"B32","doi-asserted-by":"publisher","first-page":"280","DOI":"10.3390\/s24010280","article-title":"An image histogram equalization acceleration method for field-programmable gate arrays based on a two-dimensional configurable pipeline.","volume":"24","author":"Wang","year":"2024","journal-title":"Sensors."},{"key":"B33","doi-asserted-by":"publisher","first-page":"2011","DOI":"10.1109\/TPAMI.2019.2913372","article-title":"Squeeze-and-excitation networks.","volume":"42","author":"Hu","year":"2020","journal-title":"IEEE Transact Pattern Anal Mach Intell."},{"key":"B34","first-page":"30","article-title":"Attention is all you need.","volume-title":"Proceedings of the Advances in Neural Information Processing Systems.","author":"Vaswani","year":"2017"},{"key":"B35","article-title":"Efficientnet: rethinking model scaling for convolutional neural networks.","volume-title":"Paper Presented at the International Conference on Machine Learning.","author":"Tan","year":"2019"},{"key":"B36","doi-asserted-by":"publisher","first-page":"1437188","DOI":"10.3389\/fbioe.2024.1437188","article-title":"OA-MEN: a fusion deep learning approach for enhanced accuracy in knee osteoarthritis detection and classification using X-Ray imaging.","volume":"12","author":"Ren","year":"2024","journal-title":"Front Bioeng Biotechnol."},{"key":"B37","doi-asserted-by":"publisher","first-page":"2107","DOI":"10.1002\/ksa.12221","article-title":"A multiview deep learning-based prediction pipeline augmented with confident learning can improve performance in determining knee arthroplasty candidates.","volume":"32","author":"Liu","year":"2024","journal-title":"Knee Surg Sports Traumatol Arthrosc."},{"key":"B38","doi-asserted-by":"publisher","first-page":"e29839","DOI":"10.2196\/29839","article-title":"Application of artificial intelligence in community-based primary health care: systematic scoping review and critical appraisal.","volume":"23","author":"Abbasgholizadeh Rahimi","year":"2021","journal-title":"J Med Internet Res."},{"key":"B39","doi-asserted-by":"publisher","first-page":"618496","DOI":"10.3389\/fonc.2021.618496","article-title":"3D IFPN: improved feature pyramid network for automatic segmentation of gastric tumor.","volume":"11","author":"Li","year":"2021","journal-title":"Front Oncol."},{"key":"B40","doi-asserted-by":"publisher","first-page":"83","DOI":"10.1109\/tnsre.2023.3336897","article-title":"A connectivity-aware graph neural network for real-time drowsiness classification.","volume":"32","author":"Zhuang","year":"2024","journal-title":"IEEE Trans Neural Syst Rehabil Eng."},{"key":"B41","doi-asserted-by":"publisher","first-page":"186","DOI":"10.1504\/IJSISE.2024.143822","article-title":"Automotive knee osteoarthritis severity assessment by adaptive multiscale deep learning techniques.","volume":"13","author":"Devarapaga","year":"2024","journal-title":"Int J Signal Imaging Syst Eng."},{"key":"B42","doi-asserted-by":"publisher","first-page":"1563","DOI":"10.1007\/s10278-024-01306-4","article-title":"Knee Osteoarthritis SCAENet: adaptive knee osteoarthritis severity assessment using spatial separable convolution with attention-based ensemble networks with hybrid optimization strategy.","volume":"38","author":"Devarapaga","year":"2024","journal-title":"J Imaging Informat Med."},{"key":"B43","doi-asserted-by":"publisher","first-page":"966","DOI":"10.1016\/j.joca.2023.03.011","article-title":"Relation of pain sensitization to self-reported and performance-based measures of physical functioning: the Multicenter Osteoarthritis (MOST) study.","volume":"31","author":"Corrigan","year":"2023","journal-title":"Osteoarthritis Cartilage."}],"container-title":["Frontiers in Medicine"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fmed.2026.1809768\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,6,15]],"date-time":"2026-06-15T04:25:48Z","timestamp":1781497548000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fmed.2026.1809768\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,6,15]]},"references-count":43,"alternative-id":["10.3389\/fmed.2026.1809768"],"URL":"https:\/\/doi.org\/10.3389\/fmed.2026.1809768","relation":{},"ISSN":["2296-858X"],"issn-type":[{"value":"2296-858X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,6,15]]},"article-number":"1809768"}}