ZAP-2.5DSAM: Zero Additional Parameters Advancing 2.5D SAM Adaptation to 3D Tumor Segmentation

ZAP-2.5DSAM: Zero Additional Parameters Advancing 2.5D SAM Adaptation to 3D Tumor Segmentation | 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 ZAP-2.5DSAM: Zero Additional Parameters Advancing 2.5D SAM Adaptation to 3D Tumor Segmentation Cai Guo, Yuxi Jin, Bishenghui Tao, Jianzhong Li, Hong-Ning Dai, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6916638/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 14 Jul, 2025 Read the published version in The Visual Computer → Version 1 posted 8 You are reading this latest preprint version Abstract The Segment Anything Model (SAM) demonstrated outstanding performance in 2D segmentation tasks, exhibiting robust generalization to natural images through its prompt-driven design. However, due to the lack of volumetric spatial information modeling and the domain gap between nature and medical images, its direct application to 3D medical image segmentation is suboptimal. Existing approaches to adapting SAM for 3D segmentation typically involve architectural adjustments by integrating additional components, thereby increasing trainable parameters and requiring higher GPU memory during fine-tuning. Moreover, retraining the prompt encoder may result in degraded spatial localization, especially when annotated data is scarce. To address these limitations, we propose ZAP-2.5DSAM, a parameter-efficient fine-tuning framework, which effectively extends the segmentation capacity of SAM to 3D medical images through a 2.5D decomposition scheme without introducing any additional adapter modules. Our method fine-tunes only 3.51M parameters from the original SAM, significantly reducing GPU memory requirements during training. Extensive experiments on multiple 3D tumor segmentation benchmarks demonstrate that ZAP-2.5DSAM achieves superior segmentation accuracy compared to conventional fine-tuning methods. Our code and models are available at: https://github.com/CaiGuoHS/ZAP-2.5DSAM.git. Segment Anything Model (SAM) Tumor Segmentation Zero Additional Parameters (ZAP) 2.5D Fine-tuning Framework Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 14 Jul, 2025 Read the published version in The Visual Computer → Version 1 posted Editorial decision: Accepted 24 Jun, 2025 Reviews received at journal 23 Jun, 2025 Reviewers agreed at journal 20 Jun, 2025 Reviewers agreed at journal 20 Jun, 2025 Reviewers invited by journal 20 Jun, 2025 Editor assigned by journal 18 Jun, 2025 Submission checks completed at journal 18 Jun, 2025 First submitted to journal 17 Jun, 2025 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-6916638","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":474990954,"identity":"eb9d6575-f83d-4773-bd83-9c5641dcfb9d","order_by":0,"name":"Cai Guo","email":"","orcid":"","institution":"Hanshan Normal University","correspondingAuthor":false,"prefix":"","firstName":"Cai","middleName":"","lastName":"Guo","suffix":""},{"id":474990955,"identity":"9ab8b0a7-7506-411d-8686-b696758297ce","order_by":1,"name":"Yuxi Jin","email":"","orcid":"","institution":"Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences","correspondingAuthor":false,"prefix":"","firstName":"Yuxi","middleName":"","lastName":"Jin","suffix":""},{"id":474990960,"identity":"6f305c89-e401-47b4-8732-222d8bbef638","order_by":2,"name":"Bishenghui Tao","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAuklEQVRIiWNgGAWjYBACAxDxsSEBzJFgYDhgQFgLGwMD40yStTDzkqTFXL754WPbHWnRBgeYD97mYbhjTFCLZRubsXHumZzcDQfYkq15GJ6ZEXbYMQYz6dy2CqAWHjNpHobDNkRoYf8mbQnWwv+NWC1AwxnbQA7jYQNpIewwy7acYsPetrTcmYfZjC3nGBwm7H1z5uMbH/xsS87tO9788MabisOGDQT1wAEz2J3Eqx8Fo2AUjIJRgAcAAGSfO8HCWI5rAAAAAElFTkSuQmCC","orcid":"","institution":"Hong Kong Metropolitan University","correspondingAuthor":true,"prefix":"","firstName":"Bishenghui","middleName":"","lastName":"Tao","suffix":""},{"id":474990962,"identity":"2e0d6f49-162e-48e6-899d-9f6d834e4ab7","order_by":3,"name":"Jianzhong Li","email":"","orcid":"","institution":"Hanshan Normal University","correspondingAuthor":false,"prefix":"","firstName":"Jianzhong","middleName":"","lastName":"Li","suffix":""},{"id":474990964,"identity":"8a5c3c4a-375f-46b5-9a5a-b605896c2216","order_by":4,"name":"Hong-Ning Dai","email":"","orcid":"","institution":"Hong Kong Baptist University","correspondingAuthor":false,"prefix":"","firstName":"Hong-Ning","middleName":"","lastName":"Dai","suffix":""},{"id":474990965,"identity":"2a2b7e8d-d96e-4560-aea3-a981bbdcb9af","order_by":5,"name":"Ping Li","email":"","orcid":"","institution":"The Hong Kong Polytechnic University","correspondingAuthor":false,"prefix":"","firstName":"Ping","middleName":"","lastName":"Li","suffix":""}],"badges":[],"createdAt":"2025-06-17 17:23:29","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6916638/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6916638/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s00371-025-04092-4","type":"published","date":"2025-07-14T16:05:29+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":87220863,"identity":"d02af1e4-8972-48c1-b760-bb2e4bd36cec","added_by":"auto","created_at":"2025-07-21 16:13:06","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2175909,"visible":true,"origin":"","legend":"","description":"","filename":"CGI2025358.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6916638/v1_covered_16e40c00-6e86-4b1d-a9cf-12a88e8a9fed.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"ZAP-2.5DSAM: Zero Additional Parameters Advancing 2.5D SAM Adaptation to 3D Tumor Segmentation","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":"the-visual-computer","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"tvcj","sideBox":"Learn more about [The Visual Computer](http://link.springer.com/journal/371)","snPcode":"371","submissionUrl":"https://submission.nature.com/new-submission/371/3","title":"The Visual Computer","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Segment Anything Model (SAM), Tumor Segmentation, Zero Additional Parameters (ZAP), 2.5D Fine-tuning Framework","lastPublishedDoi":"10.21203/rs.3.rs-6916638/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6916638/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe Segment Anything Model (SAM) demonstrated outstanding performance in 2D segmentation tasks, exhibiting robust generalization to natural images through its prompt-driven design. However, due to the lack of volumetric spatial information modeling and the domain gap between nature and medical images, its direct application to 3D medical image segmentation is suboptimal. Existing approaches to adapting SAM for 3D segmentation typically involve architectural adjustments by integrating additional components, thereby increasing trainable parameters and requiring higher GPU memory during fine-tuning. Moreover, retraining the prompt encoder may result in degraded spatial localization, especially when annotated data is scarce. To address these limitations, we propose ZAP-2.5DSAM, a parameter-efficient fine-tuning framework, which effectively extends the segmentation capacity of SAM to 3D medical images through a 2.5D decomposition scheme without introducing any additional adapter modules. Our method fine-tunes only 3.51M parameters from the original SAM, significantly reducing GPU memory requirements during training. Extensive experiments on multiple 3D tumor segmentation benchmarks demonstrate that ZAP-2.5DSAM achieves superior segmentation accuracy compared to conventional fine-tuning methods. Our code and models are available at: https://github.com/CaiGuoHS/ZAP-2.5DSAM.git.\u003c/p\u003e","manuscriptTitle":"ZAP-2.5DSAM: Zero Additional Parameters Advancing 2.5D SAM Adaptation to 3D Tumor Segmentation","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-06-24 15:49:04","doi":"10.21203/rs.3.rs-6916638/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Accepted","date":"2025-06-24T13:33:07+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-06-23T07:06:34+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"145265853100474433555540547617253020200","date":"2025-06-20T17:39:24+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"258548363704710763703584073440500676798","date":"2025-06-20T06:43:03+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-06-20T06:32:43+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-06-19T03:11:01+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-06-18T05:17:38+00:00","index":"","fulltext":""},{"type":"submitted","content":"The Visual Computer","date":"2025-06-17T17:07:04+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"the-visual-computer","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"tvcj","sideBox":"Learn more about [The Visual Computer](http://link.springer.com/journal/371)","snPcode":"371","submissionUrl":"https://submission.nature.com/new-submission/371/3","title":"The Visual Computer","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"6b2f3c78-dc01-4517-92da-d11767763219","owner":[],"postedDate":"June 24th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-07-21T16:12:23+00:00","versionOfRecord":{"articleIdentity":"rs-6916638","link":"https://doi.org/10.1007/s00371-025-04092-4","journal":{"identity":"the-visual-computer","isVorOnly":false,"title":"The Visual Computer"},"publishedOn":"2025-07-14 16:05:29","publishedOnDateReadable":"July 14th, 2025"},"versionCreatedAt":"2025-06-24 15:49:04","video":"","vorDoi":"10.1007/s00371-025-04092-4","vorDoiUrl":"https://doi.org/10.1007/s00371-025-04092-4","workflowStages":[]},"version":"v1","identity":"rs-6916638","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6916638","identity":"rs-6916638","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}