A Foundational Generative Model for Breast Ultrasound Image Analysis

A Foundational Generative Model for Breast Ultrasound Image Analysis | 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 A Foundational Generative Model for Breast Ultrasound Image Analysis Liwei Wang, Haojun Yu, Youcheng Li, Nan Zhang, Zihan Niu, Xuantong Gong, and 17 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5794268/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 07 Apr, 2026 Read the published version in Nature Biomedical Engineering → Version 1 posted You are reading this latest preprint version Abstract Foundational models have emerged as powerful tools for addressing various tasks in clinical settings. However, their potential development to breast ultrasound analysis remains untapped. In this paper, we present BUSGen, the first foundational generative model specifically designed for breast ultrasound image analysis. Pretrained on over 3.5 million breast ultrasound images, BUSGen has acquired extensive knowledge of breast structures, pathological features, and clinical variations. With few-shot adaptation, BUSGen can generate repositories of realistic and informative task-specific data, facilitating the development of models for a wide range of downstream tasks. Extensive experiments highlight BUSGen's exceptional adaptability, significantly exceeding real-data-trained foundational models in breast cancer screening, diagnosis, and prognosis. In breast cancer early diagnosis, our approach outperformed all board-certified radiologists (n=9), achieving an average sensitivity improvement of 16.5% (P-value<0.0001). Additionally, we characterized the scaling effect of using generated data which was as effective as the collected real-world data for training diagnostic models. Moreover, extensive experiments demonstrated that our approach improved the generalization ability of downstream models. Importantly, BUSGen protected patient privacy by enabling fully de-identified data sharing, making progress forward in secure medical data utilization. An online demo of BUSGen is available at https://aibus.bio . Biological sciences/Computational biology and bioinformatics/Computational models Biological sciences/Computational biology and bioinformatics/Machine learning Health sciences/Health care/Medical imaging Full Text Additional Declarations There is NO Competing Interest. Supplementary Files supplementarysubmit.pdf Supplementary Information Cite Share Download PDF Status: Published Journal Publication published 07 Apr, 2026 Read the published version in Nature Biomedical Engineering → 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-5794268","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":443914411,"identity":"95b6840b-3ba2-4a68-b4c6-f8d1d45fcf52","order_by":0,"name":"Liwei Wang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAyklEQVRIiWNgGAWjYLCCBww2DBIgBg/RWhIY0kjXcpgELQbHzx5+kdh2Pk9yRgLjg7dtDPLmBLWcyUuzSGy7XSwtkcBsOLeNwXBnAwEtZgdyzAwSt91OnCeRwCbN28aQYHCAkJbzb0BazoG0sP8mTsuNHOMHidsOJM4G2sJMlBb7G2/MGBL/JRdL9jxslpxzTsJwAyEtkv05xh8+nLHLkziefPDDmzIbeYK2AAEbKEoSGBgYG4C0BGH1QMD8AaJlFIyCUTAKRgEOAAAozEI+Da6SxwAAAABJRU5ErkJggg==","orcid":"","institution":"Peking University","correspondingAuthor":true,"prefix":"","firstName":"Liwei","middleName":"","lastName":"Wang","suffix":""},{"id":443914412,"identity":"b9f4a96b-ebae-4624-a2cf-38d4b125c05f","order_by":1,"name":"Haojun Yu","email":"","orcid":"https://orcid.org/0009-0004-5291-8363","institution":"Peking University","correspondingAuthor":false,"prefix":"","firstName":"Haojun","middleName":"","lastName":"Yu","suffix":""},{"id":443914413,"identity":"3579db4b-73ee-4f6c-8883-f32e58083bc7","order_by":2,"name":"Youcheng Li","email":"","orcid":"","institution":"Peking University","correspondingAuthor":false,"prefix":"","firstName":"Youcheng","middleName":"","lastName":"Li","suffix":""},{"id":443914414,"identity":"96d76284-3b8b-4fbd-bead-aecb5bc9b9c3","order_by":3,"name":"Nan Zhang","email":"","orcid":"","institution":"Key laboratory of Carcinogenesis and Translational Research (Ministry of Education/Beijing), Department of Breast","correspondingAuthor":false,"prefix":"","firstName":"Nan","middleName":"","lastName":"Zhang","suffix":""},{"id":443914415,"identity":"6c0733c5-330d-4864-a546-c2353f5f98ae","order_by":4,"name":"Zihan Niu","email":"","orcid":"","institution":"Peking Union Medical College Hospital","correspondingAuthor":false,"prefix":"","firstName":"Zihan","middleName":"","lastName":"Niu","suffix":""},{"id":443914416,"identity":"2d9f7ff2-f14d-46e3-a0f5-7e6689646dc0","order_by":5,"name":"Xuantong Gong","email":"","orcid":"","institution":"Cancer Hospital, Chinese Academy of Medical Sciences","correspondingAuthor":false,"prefix":"","firstName":"Xuantong","middleName":"","lastName":"Gong","suffix":""},{"id":443914417,"identity":"a7426d3a-b39c-41bb-9090-68d72f7bbc6b","order_by":6,"name":"Yanwen Luo","email":"","orcid":"","institution":"Department of Ultrasound, Chinese Academy of Medical Sciences and Peking Union Medical College Hospital","correspondingAuthor":false,"prefix":"","firstName":"Yanwen","middleName":"","lastName":"Luo","suffix":""},{"id":443914418,"identity":"9273d44c-f05a-4e18-998a-2b1de41d61e7","order_by":7,"name":"Haotian Ye","email":"","orcid":"","institution":"Stanford University","correspondingAuthor":false,"prefix":"","firstName":"Haotian","middleName":"","lastName":"Ye","suffix":""},{"id":443914419,"identity":"e4dbf47f-2866-4201-bbfb-97dfda9c382e","order_by":8,"name":"Siyu He","email":"","orcid":"","institution":"Stanford University","correspondingAuthor":false,"prefix":"","firstName":"Siyu","middleName":"","lastName":"He","suffix":""},{"id":443914420,"identity":"782f5d91-d124-4b4b-a938-f99422496b52","order_by":9,"name":"Quanlin Wu","email":"","orcid":"","institution":"Yuanpei College, Peking University","correspondingAuthor":false,"prefix":"","firstName":"Quanlin","middleName":"","lastName":"Wu","suffix":""},{"id":443914421,"identity":"c10a63d3-b1da-4561-b0f8-8fd810037501","order_by":10,"name":"Wangyan Qin","email":"","orcid":"","institution":"Peking University Cancer Hospital \u0026 Institute","correspondingAuthor":false,"prefix":"","firstName":"Wangyan","middleName":"","lastName":"Qin","suffix":""},{"id":443914422,"identity":"52ec174e-403c-4af9-8bc5-f81fd7ffbdac","order_by":11,"name":"Mengyuan Zhou","email":"","orcid":"","institution":"Department of Ultrasound, Chinese Academy of Medical Sciences and Peking Union Medical College Hospital","correspondingAuthor":false,"prefix":"","firstName":"Mengyuan","middleName":"","lastName":"Zhou","suffix":""},{"id":443914423,"identity":"48c4f8e5-0b94-423f-9f61-954a8aac2492","order_by":12,"name":"Jie Han","email":"","orcid":"","institution":"Department of Ultrasound, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College","correspondingAuthor":false,"prefix":"","firstName":"Jie","middleName":"","lastName":"Han","suffix":""},{"id":443914424,"identity":"0df67dbf-4548-4521-b842-b321cde0c578","order_by":13,"name":"Jia Tao","email":"","orcid":"","institution":"Peking Union Medical College Hospital","correspondingAuthor":false,"prefix":"","firstName":"Jia","middleName":"","lastName":"Tao","suffix":""},{"id":443914425,"identity":"1bc7754e-a5ba-49cd-8291-8651737cffdf","order_by":14,"name":"Ziwei Zhao","email":"","orcid":"","institution":"Yizhun Medical AI Co., Ltd.","correspondingAuthor":false,"prefix":"","firstName":"Ziwei","middleName":"","lastName":"Zhao","suffix":""},{"id":443914426,"identity":"039e93f4-30e1-445d-ada0-4e7d6af40cb3","order_by":15,"name":"Di Dai","email":"","orcid":"","institution":"Peking University","correspondingAuthor":false,"prefix":"","firstName":"Di","middleName":"","lastName":"Dai","suffix":""},{"id":443914427,"identity":"0bf6e867-f180-4ea5-ac1a-bfc38c13c2ec","order_by":16,"name":"Di He","email":"","orcid":"","institution":"Peking University","correspondingAuthor":false,"prefix":"","firstName":"Di","middleName":"","lastName":"He","suffix":""},{"id":443914428,"identity":"f90e95d8-af58-4e14-944d-147cd2c79ebe","order_by":17,"name":"Dong Wang","email":"","orcid":"","institution":"Yizhun Medical AI Co., Ltd.","correspondingAuthor":false,"prefix":"","firstName":"Dong","middleName":"","lastName":"Wang","suffix":""},{"id":443914429,"identity":"5268d00e-e4b9-4474-8dcc-a43c48719dd6","order_by":18,"name":"Binghui Tang","email":"","orcid":"","institution":"Nanchang People's Hospital","correspondingAuthor":false,"prefix":"","firstName":"Binghui","middleName":"","lastName":"Tang","suffix":""},{"id":443914430,"identity":"ddb735b6-85ac-4a16-883c-1faf45251dd2","order_by":19,"name":"Ling Huo","email":"","orcid":"","institution":"Key laboratory of Carcinogenesis and Translational Research (Ministry of Education/Beijing), Department of Breast","correspondingAuthor":false,"prefix":"","firstName":"Ling","middleName":"","lastName":"Huo","suffix":""},{"id":443914431,"identity":"76a56254-2907-4470-8326-acf4656029d9","order_by":20,"name":"James Zou","email":"","orcid":"https://orcid.org/0000-0001-8880-4764","institution":"Stanford University","correspondingAuthor":false,"prefix":"","firstName":"James","middleName":"","lastName":"Zou","suffix":""},{"id":443914432,"identity":"6878b616-3cc7-4874-9e22-756826ac76fd","order_by":21,"name":"Qingli Zhu","email":"","orcid":"","institution":"Chinese Academy of Medical Sciences and Peking Union Medical College Hospital","correspondingAuthor":false,"prefix":"","firstName":"Qingli","middleName":"","lastName":"Zhu","suffix":""},{"id":443914433,"identity":"eaaad201-e5e4-45d6-869d-6782f77556f3","order_by":22,"name":"Yong Wang","email":"","orcid":"","institution":"Cancer Hospital, Chinese Academy of Medical Sciences","correspondingAuthor":false,"prefix":"","firstName":"Yong","middleName":"","lastName":"Wang","suffix":""}],"badges":[],"createdAt":"2025-01-09 07:50:12","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5794268/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5794268/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41551-026-01639-1","type":"published","date":"2026-04-07T04:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":106389961,"identity":"ba0a1975-5d87-4f4b-921a-4a3ab70d0829","added_by":"auto","created_at":"2026-04-08 07:08:28","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3051624,"visible":true,"origin":"","legend":"Article File","description":"","filename":"mainpapersubmit.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5794268/v1_covered_8712315a-c3fa-43a9-8cd3-4d9457039dfd.pdf"},{"id":80825238,"identity":"876a1a33-01df-4dd3-9b15-b724927ce9d3","added_by":"auto","created_at":"2025-04-17 13:06:11","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":2386299,"visible":true,"origin":"","legend":"Supplementary Information","description":"","filename":"supplementarysubmit.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5794268/v1/aa96fc0b2315d73f06f5b7d5.pdf"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"A Foundational Generative Model for Breast Ultrasound Image Analysis","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"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-5794268/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5794268/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Foundational models have emerged as powerful tools for addressing various tasks in clinical settings. However, their potential development to breast ultrasound analysis remains untapped. In this paper, we present BUSGen, the first foundational generative model specifically designed for breast ultrasound image analysis. Pretrained on over 3.5 million breast ultrasound images, BUSGen has acquired extensive knowledge of breast structures, pathological features, and clinical variations. With few-shot adaptation, BUSGen can generate repositories of realistic and informative task-specific data, facilitating the development of models for a wide range of downstream tasks. Extensive experiments highlight BUSGen's exceptional adaptability, significantly exceeding real-data-trained foundational models in breast cancer screening, diagnosis, and prognosis. In breast cancer early diagnosis, our approach outperformed all board-certified radiologists (n=9), achieving an average sensitivity improvement of 16.5% (P-value\u003c0.0001). Additionally, we characterized the scaling effect of using generated data which was as effective as the collected real-world data for training diagnostic models. Moreover, extensive experiments demonstrated that our approach improved the generalization ability of downstream models. Importantly, BUSGen protected patient privacy by enabling fully de-identified data sharing, making progress forward in secure medical data utilization. An online demo of BUSGen is available at https://aibus.bio.","manuscriptTitle":"A Foundational Generative Model for Breast Ultrasound Image Analysis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-04-17 13:06:06","doi":"10.21203/rs.3.rs-5794268/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-biomedical-engineering","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"natbiomedeng","sideBox":"Learn more about [Nature Biomedical Engineering](http://www.nature.com/natbiomedeng/)","snPcode":"41551","submissionUrl":"https://mts-natbiomedeng.nature.com/cgi-bin/main.plex","title":"Nature Biomedical Engineering","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Research","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"29ba7242-58e3-45d2-b1c5-de72dc2018b6","owner":[],"postedDate":"April 17th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":47263317,"name":"Biological sciences/Computational biology and bioinformatics/Computational models"},{"id":47263318,"name":"Biological sciences/Computational biology and bioinformatics/Machine learning"},{"id":47263319,"name":"Health sciences/Health care/Medical imaging"}],"tags":[],"updatedAt":"2026-04-08T07:08:10+00:00","versionOfRecord":{"articleIdentity":"rs-5794268","link":"https://doi.org/10.1038/s41551-026-01639-1","journal":{"identity":"nature-biomedical-engineering","isVorOnly":false,"title":"Nature Biomedical Engineering"},"publishedOn":"2026-04-07 04:00:00","publishedOnDateReadable":"April 7th, 2026"},"versionCreatedAt":"2025-04-17 13:06:06","video":"","vorDoi":"10.1038/s41551-026-01639-1","vorDoiUrl":"https://doi.org/10.1038/s41551-026-01639-1","workflowStages":[]},"version":"v1","identity":"rs-5794268","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5794268","identity":"rs-5794268","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}