{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,3]],"date-time":"2026-08-03T16:53:08Z","timestamp":1785775988140,"version":"3.56.0"},"reference-count":16,"publisher":"Wiley","issue":"8","license":[{"start":{"date-parts":[[2026,6,16]],"date-time":"2026-06-16T00:00:00Z","timestamp":1781568000000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by-nc-nd\/4.0\/"},{"start":{"date-parts":[[2026,6,16]],"date-time":"2026-06-16T00:00:00Z","timestamp":1781568000000},"content-version":"tdm","delay-in-days":0,"URL":"http:\/\/doi.wiley.com\/10.1002\/tdm_license_1.1"}],"funder":[{"DOI":"10.13039\/501100012166","name":"National Key Research and Development Program of China","doi-asserted-by":"publisher","award":["2023YFB4606700"],"award-info":[{"award-number":["2023YFB4606700"]}],"id":[{"id":"10.13039\/501100012166","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Orthopaedic Surgery"],"published-print":{"date-parts":[[2026,8]]},"abstract":"<jats:title>ABSTRACT<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Deep vein thrombosis (DVT) following total knee arthroplasty (TKA) remains a major postoperative complication. Conventional risk assessment tools such as the Caprini score show limited predictive performance (AUC 0.65\u20130.72). We developed a multimodal deep learning framework that integrates early postoperative radiographs with clinical data to improve DVT risk stratification.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>\n                      This single\u2010center retrospective cohort study analyzed 1200 patients undergoing primary unilateral TKA (DVT incidence: 18.0%) from January 2018 to June 2023. A dual\u2010branch architecture was implemented: ViT\u2010B\/16 or ResNet50 processed standardized radiographs (metal artifact suppression; 512\u2009\u00d7\u2009512 resolution), while Clinical\u2010BERT encoded structured electronic medical record (EMR) representations. A dynamic attention fusion module generated sample\u2010specific modality weights (\u03b1) for feature fusion (\u03b1\u00b7v_img\u2009+\u2009(1\u2009\u2212\u2009\u03b1)\u00b7v_txt). Performance was evaluated on a stratified test set (\n                      <jats:italic>n<\/jats:italic>\n                      \u2009=\u2009180) against the Caprini score, unimodal models, and a nonattention concatenation fusion baseline. AUCs were compared using the DeLong test, and sensitivity\/specificity differences were assessed using the McNemar test.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>\n                      The multimodal model achieved an AUC of 0.89 (95% CI: 0.85\u20130.92; DeLong\n                      <jats:italic>p<\/jats:italic>\n                      \u2009&lt;\u20090.001 vs. all comparators), sensitivity 83% (76%\u201389%), specificity 88% (82%\u201393%), and F1\u2010score 0.79. It identified 91.7% (11\/12) of DVT cases labeled low\u2010risk by Caprini. The learned attention coefficient \u03b1 suggested modality reliance patterns associated with specific findings (e.g., higher \u03b1 in cases with prosthesis\u2013bone lucency &gt;\u20092\u2009mm; higher text contribution in cases with D\u2010dimer &gt;\u20091.0\u2009mg\/L), but these associations do not establish causality. Anatomical heatmaps consistently localized high\u2010risk regions, particularly the popliteal vein projection area and tibial interface.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>This attention\u2010based multimodal model integrates biomechanical (radiographic) and biochemical\/clinical (EMR\u2010derived) signals to improve early postoperative DVT risk prediction while providing interpretable anatomical cues.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1111\/os.70335","type":"journal-article","created":{"date-parts":[[2026,6,17]],"date-time":"2026-06-17T05:00:27Z","timestamp":1781672427000},"page":"1574-1580","update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Multimodal Deep Learning for Predicting Deep Vein Thrombosis Risk After Total Knee Arthroplasty: A Clinical Study"],"prefix":"10.1111","volume":"18","author":[{"given":"Kejia","family":"Zhu","sequence":"first","affiliation":[{"name":"Department of Orthopedic Surgery and Orthopedic Research Institute West China Hospital, Sichuan University  Chengdu China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hang","family":"Li","sequence":"additional","affiliation":[{"name":"Department of Orthopedic Surgery and Orthopedic Research Institute West China Hospital, Sichuan University  Chengdu China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hui","family":"Zhang","sequence":"additional","affiliation":[{"name":"Rehabilitation Medicine Center and Institute of Rehabilitation Medicine West China Hospital, Sichuan University  Chengdu China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Biao","family":"Wang","sequence":"additional","affiliation":[{"name":"West China Biomedical Big Data Center West China Hospital, Sichuan University  Chengdu China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-9632-7221","authenticated-orcid":false,"given":"Bin","family":"Shen","sequence":"additional","affiliation":[{"name":"Department of Orthopedic Surgery and Orthopedic Research Institute West China Hospital, Sichuan University  Chengdu China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0005-7813-3764","authenticated-orcid":false,"given":"Yong","family":"Nie","sequence":"additional","affiliation":[{"name":"Department of Orthopedic Surgery and Orthopedic Research Institute West China Hospital, Sichuan University  Chengdu China"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2026,6,16]]},"reference":[{"key":"e_1_2_13_2_1","doi-asserted-by":"publisher","DOI":"10.1378\/chest.11-2404"},{"key":"e_1_2_13_3_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.arth.2023.10.017"},{"key":"e_1_2_13_4_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.thromres.2012.08.001"},{"key":"e_1_2_13_5_1","doi-asserted-by":"publisher","DOI":"10.1177\/17539447221105013"},{"issue":"6","key":"e_1_2_13_6_1","article-title":"Osteochondral Allograft Transplantation for Knee Cartilage and Osteochondral Defects: A Review of Indications, Technique","volume":"7","author":"Cavendish P. A.","year":"2019","journal-title":"Rehabilitation, and Outcomes"},{"key":"e_1_2_13_7_1","doi-asserted-by":"publisher","DOI":"10.4103\/0976-3147.169801"},{"key":"e_1_2_13_8_1","doi-asserted-by":"publisher","DOI":"10.1186\/s12959-015-0062-0"},{"key":"e_1_2_13_9_1","doi-asserted-by":"publisher","DOI":"10.1002\/jbm4.10828"},{"key":"e_1_2_13_10_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijmedinf.2025.105901"},{"key":"e_1_2_13_11_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.13583"},{"key":"e_1_2_13_12_1","doi-asserted-by":"publisher","DOI":"10.3389\/fonc.2019.01333"},{"key":"e_1_2_13_13_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.cell.2017.11.001"},{"key":"e_1_2_13_14_1","doi-asserted-by":"publisher","DOI":"10.3389\/fbioe.2020.00123"},{"key":"e_1_2_13_15_1","doi-asserted-by":"publisher","DOI":"10.1007\/s00068-014-0461-0"},{"key":"e_1_2_13_16_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.arthro.2024.04.035"},{"key":"e_1_2_13_17_1","doi-asserted-by":"publisher","DOI":"10.1007\/s43390-024-00991-z"}],"container-title":["Orthopaedic Surgery"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1111\/os.70335","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/full-xml\/10.1111\/os.70335","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1111\/os.70335","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,3]],"date-time":"2026-08-03T16:26:18Z","timestamp":1785774378000},"score":1,"resource":{"primary":{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/10.1111\/os.70335"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,6,16]]},"references-count":16,"journal-issue":{"issue":"8","published-print":{"date-parts":[[2026,8]]}},"alternative-id":["10.1111\/os.70335"],"URL":"https:\/\/doi.org\/10.1111\/os.70335","archive":["Portico"],"relation":{},"ISSN":["1757-7853","1757-7861"],"issn-type":[{"value":"1757-7853","type":"print"},{"value":"1757-7861","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,6,16]]},"assertion":[{"value":"2025-08-28","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-04-23","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-06-16","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}