Generalizable deep learning framework for 3D medical image segmentation using limited training data

Generalizable deep learning framework for 3D medical image segmentation using limited training data | 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 Research Article Generalizable deep learning framework for 3D medical image segmentation using limited training data Tobias Ekman, Arthur Barakat, Einar Heiberg This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4667752/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 06 Mar, 2025 Read the published version in 3D Printing in Medicine → Version 1 posted 9 You are reading this latest preprint version Abstract Medical image segmentation is a critical component in a wide range of clinical applications, enabling the identification and delineation of anatomical structures. This study focuses on segmentation of anatomical structures for 3D printing, virtual surgery planning, and advanced visualization such as virtual or augmented reality. Manual segmentation methods are labor-intensive and can be subjective, leading to inter-observer variability. Machine learning algorithms, particularly deep learning models, have gained traction for automating the process and are now considered state-of-the-art. However, deep-learning methods typically demand large datasets for fine-tuning and powerful graphics cards, limiting their applicability in resource-constrained settings. In this paper we introduce a robust deep learning framework for 3D medical segmentation that achieves high performance across a range of medical segmentation tasks, even when trained on a small number of subjects. This approach overcomes the need for extensive data and heavy GPU resources, facilitating adoption within healthcare systems. The potential is exemplified through six different clinical applications involving orthopedics, orbital segmentation, mandible CT, cardiac CT, fetal MRI and lung CT. Notably, a small number of hyper-parameters and augmentation settings produced >95% segmentation Dice score results for a diverse range of organs and tissues. Segmentation Machine Learning Artificial Intelligence Deep Learning 3D Printing Full Text Additional Declarations Competing interest reported. Einar Heiberg is the founder and owner of Medviso AB, developing the software Segment 3DPrint which was used for manual segmentation as well as for testing and validation. This software is commercially available for clinical use. All other authors report no competing interests. Supplementary Files SupplementalmaterialTestdata.pdf SupplementalmaterialTraindata.pdf Cite Share Download PDF Status: Published Journal Publication published 06 Mar, 2025 Read the published version in 3D Printing in Medicine → Version 1 posted Editorial decision: Revision requested 10 Oct, 2024 Reviews received at journal 05 Sep, 2024 Reviewers agreed at journal 05 Sep, 2024 Reviews received at journal 12 Aug, 2024 Reviewers agreed at journal 11 Jul, 2024 Reviewers invited by journal 10 Jul, 2024 Editor assigned by journal 01 Jul, 2024 Submission checks completed at journal 01 Jul, 2024 First submitted to journal 01 Jul, 2024 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-4667752","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":330003736,"identity":"1723bbac-43a5-4bef-95d2-065518cd76a6","order_by":0,"name":"Tobias Ekman","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABAklEQVRIiWNgGAWjYFACHoYDIEoCRHyosAGSjI0H8GtA0sI440waiGogqIUBpoWZt+0wmINXiz1778EDP/4wyElOO/zs4Ywz5+3Wth8G2lJjE43TFp5zCQd72xiMpaXTzA0+VNxO3nYmEajlWFpuAy4tEjkGB3gbGBLnSSeYSc44czvZ7ABQC2PDYbxaDv75w1A/Tzr9mzRv27lks/MPCWs5zMPGkCAtnWMG1HLAzuwGIVvOnEs4LNsmYThzdk4Z0GHJCWY3gLYk4PELe3vv4Y9v/tjIS9xO3ybxocLO3ux8+sMHH2pscGqBAgk4KxGsMgG/clRgT4riUTAKRsEoGBkAALLQZKVp8yrTAAAAAElFTkSuQmCC","orcid":"","institution":"Lund University","correspondingAuthor":true,"prefix":"","firstName":"Tobias","middleName":"","lastName":"Ekman","suffix":""},{"id":330003737,"identity":"2a7505d3-95a2-4376-a4f2-28e4cdad8fe7","order_by":1,"name":"Arthur Barakat","email":"","orcid":"","institution":"Lund University","correspondingAuthor":false,"prefix":"","firstName":"Arthur","middleName":"","lastName":"Barakat","suffix":""},{"id":330003738,"identity":"cc1d32ce-7f17-43b3-b85a-3a4578836b15","order_by":2,"name":"Einar Heiberg","email":"","orcid":"","institution":"Lund University","correspondingAuthor":false,"prefix":"","firstName":"Einar","middleName":"","lastName":"Heiberg","suffix":""}],"badges":[],"createdAt":"2024-07-01 11:39:53","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4667752/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4667752/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s41205-025-00254-1","type":"published","date":"2025-03-06T15:57:56+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":78191363,"identity":"50ee47b0-044c-43a2-871e-e53ac9542d7e","added_by":"auto","created_at":"2025-03-10 19:57:57","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2234846,"visible":true,"origin":"","legend":"","description":"","filename":"Manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4667752/v1_covered_b54c67be-83d4-4dd4-a45e-d37440c63465.pdf"},{"id":60957648,"identity":"513bec6b-00e0-434d-8ef3-0feb9fbbf6e2","added_by":"auto","created_at":"2024-07-24 03:43:54","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":81472,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementalmaterialTestdata.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4667752/v1/0a696b562a3af3c08ade90aa.pdf"},{"id":60957649,"identity":"60557639-c040-48d1-a28a-c519be5d1199","added_by":"auto","created_at":"2024-07-24 03:43:54","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":85197,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementalmaterialTraindata.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4667752/v1/93e271671f6864e1f832199b.pdf"}],"financialInterests":"Competing interest reported. Einar Heiberg is the founder and owner of Medviso AB, developing the software Segment 3DPrint which was used for manual segmentation as well as for testing and validation. This software is commercially available for clinical use. All other authors report no competing interests.","formattedTitle":"Generalizable deep learning framework for 3D medical image segmentation using limited training data","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"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":"3d-printing-in-medicine","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"tdpm","sideBox":"Learn more about [3D Printing in Medicine](https://threedmedprint.biomedcentral.com/)","snPcode":"41205","submissionUrl":"https://submission.nature.com/new-submission/41205/3","title":"3D Printing in Medicine","twitterHandle":"@BioMedCentral","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Segmentation, Machine Learning, Artificial Intelligence, Deep Learning, 3D Printing","lastPublishedDoi":"10.21203/rs.3.rs-4667752/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4667752/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Medical image segmentation is a critical component in a wide range of clinical applications, enabling the identification and delineation of anatomical structures. This study focuses on segmentation of anatomical structures for 3D printing, virtual surgery planning, and advanced visualization such as virtual or augmented reality. Manual segmentation methods are labor-intensive and can be subjective, leading to inter-observer variability. Machine learning algorithms, particularly deep learning models, have gained traction for automating the process and are now considered state-of-the-art. However, deep-learning methods typically demand large datasets for fine-tuning and powerful graphics cards, limiting their applicability in resource-constrained settings. In this paper we introduce a robust deep learning framework for 3D medical segmentation that achieves high performance across a range of medical segmentation tasks, even when trained on a small number of subjects. This approach overcomes the need for extensive data and heavy GPU resources, facilitating adoption within healthcare systems. The potential is exemplified through six different clinical applications involving orthopedics, orbital segmentation, mandible CT, cardiac CT, fetal MRI and lung CT. Notably, a small number of hyper-parameters and augmentation settings produced \u003e95\\% segmentation Dice score results for a diverse range of organs and tissues. ","manuscriptTitle":"Generalizable deep learning framework for 3D medical image segmentation using limited training data","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-07-24 03:43:49","doi":"10.21203/rs.3.rs-4667752/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2024-10-10T07:11:49+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-09-05T20:36:41+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"35490712179639966893620596905645714234","date":"2024-09-05T20:33:14+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-08-12T04:09:45+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"261573833892626055790783086504641333524","date":"2024-07-11T19:42:58+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-07-10T05:11:05+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-07-01T23:14:22+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-07-01T23:13:35+00:00","index":"","fulltext":""},{"type":"submitted","content":"3D Printing in Medicine","date":"2024-07-01T11:38:39+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"3d-printing-in-medicine","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"tdpm","sideBox":"Learn more about [3D Printing in Medicine](https://threedmedprint.biomedcentral.com/)","snPcode":"41205","submissionUrl":"https://submission.nature.com/new-submission/41205/3","title":"3D Printing in Medicine","twitterHandle":"@BioMedCentral","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"143ba42c-f679-4f32-bfb4-c505d7016465","owner":[],"postedDate":"July 24th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-03-10T19:57:50+00:00","versionOfRecord":{"articleIdentity":"rs-4667752","link":"https://doi.org/10.1186/s41205-025-00254-1","journal":{"identity":"3d-printing-in-medicine","isVorOnly":false,"title":"3D Printing in Medicine"},"publishedOn":"2025-03-06 15:57:56","publishedOnDateReadable":"March 6th, 2025"},"versionCreatedAt":"2024-07-24 03:43:49","video":"","vorDoi":"10.1186/s41205-025-00254-1","vorDoiUrl":"https://doi.org/10.1186/s41205-025-00254-1","workflowStages":[]},"version":"v1","identity":"rs-4667752","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4667752","identity":"rs-4667752","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}