DentVLM: A Multimodal Vision-Language Model for Comprehensive Dental Diagnosis and Enhanced Clinical Practice

DentVLM: A Multimodal Vision-Language Model for Comprehensive Dental Diagnosis and Enhanced Clinical Practice | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article DentVLM: A Multimodal Vision-Language Model for Comprehensive Dental Diagnosis and Enhanced Clinical Practice Zuozhu Liu, Zijie Meng, Jin Hao, Xiwei Dai, Yang Feng, Jiaxiang Liu, and 17 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7403627/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract Diagnosing and managing oral diseases necessitate advanced visual interpretation across diverse imaging modalities and integrated information synthesis. While current AI models excel at isolated tasks, they often fall short in addressing the complex, multimodal requirements of comprehensive clinical dental practice. Here we introduce DentVLM, a multimodal vision-language model engineered for expert-level oral disease diagnosis. DentVLM was developed using a comprehensive, large-scale, bilingual dataset of 110,447 images and 2.46 million visual question-answering (VQA) pairs. The model is capable of interpreting seven 2D oral imaging modalities across 36 diagnostic tasks, significantly outperforming leading proprietary and open-source models by 19.6% higher accuracy for oral diseases and 27.9% for malocclusions. In a clinical study involving 25 dentists, evaluating 1,946 patients and encompassing 3,105 QA pairs, DentVLM surpassed the diagnostic performance of 13 junior dentists on 21 of 36 tasks and exceeded that of 12 senior dentists on 12 of 36 tasks. When integrated into a collaborative workflow, DentVLM elevated junior dentists’ performance to senior levels and reduced diagnostic time for all practitioners by 15-22%. Furthermore, DentVLM exhibited promising performance across three practical utility scenarios, including home-based dental health management, hospital-based intelligent diagnosis and multi-agent collaborative interaction. These findings establish DentVLM as a robust clinical decision support tool, poised to enhance primary dental care, mitigate provider-patient imbalances, and democratize access to specialized medical expertise within the field of dentistry. Health sciences/Health care/Dentistry Health sciences/Diseases/Dental diseases Full Text Additional Declarations There is NO Competing Interest. Supplementary Files SupplementaryNotes.pdf Note1, ..., Note 5 SupplementaryTable.pdf Extended Table1, ..., Extented Table 14 DentVLMrepository.zip The repository for code review. Cite Share Download PDF Status: Under Review Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7403627","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":508602030,"identity":"c31d2658-fa7c-4c6d-a34a-7fefc77a5cf0","order_by":0,"name":"Zuozhu Liu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABBUlEQVRIie3RsUvDQBTH8XccpMtr51cU8i+8w6EOgfwrJwGnDrqUjpHCZekfEMH+D13dGjp0KZ0tLkqhk0LB0cXndaxJ7SZ4XwJZ7sPvSABCoT+aBrhEaN3lADcASl6AxwkhYJXDjE8g8ljwxNdE4jKbf9w6Ouf1xm12DKkurXp5dxD3aogqr+3FvSPk56uCZUWr0mozcWAe85pLUZ+z9p44EhIJic7aDizPfiaRkLkn68oTFNL6bCJIfTPy5El5Qt8ruokQbjONK8LuUi62ZGI1fh11H1ZkpjUkLuSL4SBJO4vFlobDJDVFVu3eBklct3KwCib3P5N+d36/e8LZUCgU+h99AUp8Txd/2R/4AAAAAElFTkSuQmCC","orcid":"","institution":"Zhejiang University","correspondingAuthor":true,"prefix":"","firstName":"Zuozhu","middleName":"","lastName":"Liu","suffix":""},{"id":508602031,"identity":"bfb9e882-fdeb-4556-bd78-d0e9ca82401d","order_by":1,"name":"Zijie Meng","email":"","orcid":"","institution":"Zhejiang University","correspondingAuthor":false,"prefix":"","firstName":"Zijie","middleName":"","lastName":"Meng","suffix":""},{"id":508602032,"identity":"dfdab3ac-6d6d-4e29-af04-74e50cd2da0d","order_by":2,"name":"Jin Hao","email":"","orcid":"https://orcid.org/0000-0002-6685-2017","institution":"Shanghai Jiao Tong University","correspondingAuthor":false,"prefix":"","firstName":"Jin","middleName":"","lastName":"Hao","suffix":""},{"id":508602033,"identity":"553abfea-296b-46a1-b88f-50d6a0656c21","order_by":3,"name":"Xiwei Dai","email":"","orcid":"","institution":"Zhejiang University","correspondingAuthor":false,"prefix":"","firstName":"Xiwei","middleName":"","lastName":"Dai","suffix":""},{"id":508602034,"identity":"e1fffbfb-9fe8-4175-a9df-35b925fe22be","order_by":4,"name":"Yang Feng","email":"","orcid":"","institution":"Angel Align Inc.","correspondingAuthor":false,"prefix":"","firstName":"Yang","middleName":"","lastName":"Feng","suffix":""},{"id":508602035,"identity":"86efa7f4-9926-49f9-9c11-2209796d6b6c","order_by":5,"name":"Jiaxiang Liu","email":"","orcid":"","institution":"Zhejiang University-University of Illinois at Urbana-Champaign Institute, ZJU-Angelalign R\u0026D Center for Intelligent Healthcare, Zhejiang University","correspondingAuthor":false,"prefix":"","firstName":"Jiaxiang","middleName":"","lastName":"Liu","suffix":""},{"id":508602036,"identity":"5dbfb99b-f33f-4051-8ae3-828956c2d107","order_by":6,"name":"Bin Feng","email":"","orcid":"","institution":"Zhejiang University","correspondingAuthor":false,"prefix":"","firstName":"Bin","middleName":"","lastName":"Feng","suffix":""},{"id":508602037,"identity":"b594a5a3-5eb9-443e-b375-1e6ba15c2193","order_by":7,"name":"Huikai Wu","email":"","orcid":"https://orcid.org/0000-0001-9233-6350","institution":"Zhejiang University-University of Illinois at Urbana-Champaign Institute, ZJU-Angelalign R\u0026D Center for Intelligent Healthcare, Zhejiang University","correspondingAuthor":false,"prefix":"","firstName":"Huikai","middleName":"","lastName":"Wu","suffix":""},{"id":508602038,"identity":"c1eee5de-14a5-4f35-acf2-1fb0b9c06dff","order_by":8,"name":"Xiaotang Gai","email":"","orcid":"","institution":"Zhejiang University","correspondingAuthor":false,"prefix":"","firstName":"Xiaotang","middleName":"","lastName":"Gai","suffix":""},{"id":508602039,"identity":"96326a90-bee0-4ee5-bce4-7ade1c3b65e2","order_by":9,"name":"Hengchuan Zhu","email":"","orcid":"","institution":"Zhejiang University","correspondingAuthor":false,"prefix":"","firstName":"Hengchuan","middleName":"","lastName":"Zhu","suffix":""},{"id":508602040,"identity":"d8d2e3cf-309c-4f0f-90f9-96ca0f8e781c","order_by":10,"name":"Tianxiang Hu","email":"","orcid":"","institution":"Zhejiang University","correspondingAuthor":false,"prefix":"","firstName":"Tianxiang","middleName":"","lastName":"Hu","suffix":""},{"id":508602041,"identity":"d487d253-309d-449d-bf40-73b688b60ff7","order_by":11,"name":"Yangyang Wu","email":"","orcid":"","institution":"Zhejiang University","correspondingAuthor":false,"prefix":"","firstName":"Yangyang","middleName":"","lastName":"Wu","suffix":""},{"id":508602042,"identity":"72217a97-f566-4352-b914-4635e96303d6","order_by":12,"name":"Hongxia Xu","email":"","orcid":"","institution":"Zhe Jiang University","correspondingAuthor":false,"prefix":"","firstName":"Hongxia","middleName":"","lastName":"Xu","suffix":""},{"id":508602043,"identity":"4673834b-e1a9-4fa5-88b2-a12cdedc80b0","order_by":13,"name":"Jin Li","email":"","orcid":"","institution":"Zhejiang University-University of Illinois at Urbana-Champaign Institute, ZJU-Angelalign R\u0026D Center for Intelligent Healthcare, Zhejiang University","correspondingAuthor":false,"prefix":"","firstName":"Jin","middleName":"","lastName":"Li","suffix":""},{"id":508602044,"identity":"568a9e13-bb11-44a7-b5f9-cb2e931a9c3b","order_by":14,"name":"Jun Xiao","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Jun","middleName":"","lastName":"Xiao","suffix":""},{"id":508602045,"identity":"06343c36-5309-45bb-aeec-fdf4e455bd2b","order_by":15,"name":"Xiaoqiang Liu","email":"","orcid":"","institution":"Peking University School and Hospital of Stomatology","correspondingAuthor":false,"prefix":"","firstName":"Xiaoqiang","middleName":"","lastName":"Liu","suffix":""},{"id":508602046,"identity":"2febc656-eac3-49bb-8c68-ea6815afb1f2","order_by":16,"name":"Joey Tianyi Zhou","email":"","orcid":"https://orcid.org/0000-0002-4675-7055","institution":"Technology and Research (A*STAR)","correspondingAuthor":false,"prefix":"","firstName":"Joey","middleName":"Tianyi","lastName":"Zhou","suffix":""},{"id":508602047,"identity":"e0fe11d5-734c-409e-a6e2-1e75272c2ea0","order_by":17,"name":"Fudong Zhu","email":"","orcid":"","institution":"Zhejiang University","correspondingAuthor":false,"prefix":"","firstName":"Fudong","middleName":"","lastName":"Zhu","suffix":""},{"id":508602048,"identity":"3075a082-0b47-49c3-9078-277156b262cc","order_by":18,"name":"Zhihe Zhao","email":"","orcid":"https://orcid.org/0000-0003-2955-1706","institution":"Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Zhihe","middleName":"","lastName":"Zhao","suffix":""},{"id":508602049,"identity":"b937f641-f7bf-407e-bdc7-148a4ec4d738","order_by":19,"name":"Lunguo Xia","email":"","orcid":"","institution":"Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Lunguo","middleName":"","lastName":"Xia","suffix":""},{"id":508602050,"identity":"f55c7f82-6f2d-4673-95e6-b9eedede403a","order_by":20,"name":"Bing Fang","email":"","orcid":"","institution":"Shanghai Ninth People's hospital","correspondingAuthor":false,"prefix":"","firstName":"Bing","middleName":"","lastName":"Fang","suffix":""},{"id":508602051,"identity":"52d03a94-c54b-4434-8914-c4592f5efdb5","order_by":21,"name":"Jimeng Sun","email":"","orcid":"https://orcid.org/0000-0003-1512-6426","institution":"University of Illinois at Urbana Champaign","correspondingAuthor":false,"prefix":"","firstName":"Jimeng","middleName":"","lastName":"Sun","suffix":""},{"id":508602052,"identity":"9d9c844f-47bf-49f0-9af7-0366a1e598b2","order_by":22,"name":"jian wu","email":"","orcid":"","institution":"Zhejiang University","correspondingAuthor":false,"prefix":"","firstName":"jian","middleName":"","lastName":"wu","suffix":""}],"badges":[],"createdAt":"2025-08-19 02:15:08","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7403627/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7403627/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":90461085,"identity":"6c1e6dc0-7963-44a1-8354-5a16cf08cbbd","added_by":"auto","created_at":"2025-09-03 04:00:23","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2926570,"visible":true,"origin":"","legend":"Article File","description":"","filename":"DentVLMManuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7403627/v1_covered_19d58a0e-215e-4d4a-897f-f31ceb5bd3f3.pdf"},{"id":90460321,"identity":"274459c7-687e-49df-92db-d0c5b847de46","added_by":"auto","created_at":"2025-09-03 03:44:15","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":235997,"visible":true,"origin":"","legend":"Note1, ..., Note 5","description":"","filename":"SupplementaryNotes.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7403627/v1/5d117d32577e8b22bf8ce1ba.pdf"},{"id":90460642,"identity":"a5e33a1d-2840-4787-a82f-8c516db32bb8","added_by":"auto","created_at":"2025-09-03 03:52:15","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":548980,"visible":true,"origin":"","legend":"Extended Table1, ..., Extented Table 14","description":"","filename":"SupplementaryTable.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7403627/v1/f08fa2736f9e18412360cc5b.pdf"},{"id":90460328,"identity":"b7d83bbc-f212-461e-a6ee-fd06192cbba9","added_by":"auto","created_at":"2025-09-03 03:44:16","extension":"zip","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":35497375,"visible":true,"origin":"","legend":"The repository for code review.","description":"","filename":"DentVLMrepository.zip","url":"https://assets-eu.researchsquare.com/files/rs-7403627/v1/b17d43869162ada35d04cb31.zip"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"DentVLM: A Multimodal Vision-Language Model for Comprehensive Dental Diagnosis and Enhanced Clinical Practice","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-7403627/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7403627/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Diagnosing and managing oral diseases necessitate advanced visual interpretation across diverse imaging modalities and integrated information synthesis. While current AI models excel at isolated tasks, they often fall short in addressing the complex, multimodal requirements of comprehensive clinical dental practice. Here we introduce DentVLM, a multimodal vision-language model engineered for expert-level oral disease diagnosis. DentVLM was developed using a comprehensive, large-scale, bilingual dataset of 110,447 images and 2.46 million visual question-answering (VQA) pairs. The model is capable of interpreting seven 2D oral imaging modalities across 36 diagnostic tasks, significantly outperforming leading proprietary and open-source models by 19.6% higher accuracy for oral diseases and 27.9% for malocclusions. In a clinical study involving 25 dentists, evaluating 1,946 patients and encompassing 3,105 QA pairs, DentVLM surpassed the diagnostic performance of 13 junior dentists on 21 of 36 tasks and exceeded that of 12 senior dentists on 12 of 36 tasks. When integrated into a collaborative workflow, DentVLM elevated junior dentists’ performance to senior levels and reduced diagnostic time for all practitioners by 15-22%. Furthermore, DentVLM exhibited promising performance across three practical utility scenarios, including home-based dental health management, hospital-based intelligent diagnosis and multi-agent collaborative interaction. These findings establish DentVLM as a robust clinical decision support tool, poised to enhance primary dental care, mitigate provider-patient imbalances, and democratize access to specialized medical expertise within the field of dentistry.","manuscriptTitle":"DentVLM: A Multimodal Vision-Language Model for Comprehensive Dental Diagnosis and Enhanced Clinical Practice","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-03 03:44:10","doi":"10.21203/rs.3.rs-7403627/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-communications","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"NCOMMS","sideBox":"Learn more about [Nature Communications](http://www.nature.com/ncomms/)","snPcode":"","submissionUrl":"https://mts-ncomms.nature.com/","title":"Nature Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Communications","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"65884acc-2bfe-4693-8f2e-7e1a6babcdee","owner":[],"postedDate":"September 3rd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":54015816,"name":"Health sciences/Health care/Dentistry"},{"id":54015817,"name":"Health sciences/Diseases/Dental diseases"}],"tags":[],"updatedAt":"2026-02-25T04:25:29+00:00","versionOfRecord":[],"versionCreatedAt":"2025-09-03 03:44:10","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7403627","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7403627","identity":"rs-7403627","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}