Multi-scale ensemble model for dMMR prediction from histopathological images of colorectal cancer

Multi-scale ensemble model for dMMR prediction from histopathological images of colorectal cancer | 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 Multi-scale ensemble model for dMMR prediction from histopathological images of colorectal cancer Liisa Petäinen, Juha P. Väyrynen, Jan Böhm, Pekka Ruusuvuori, and 12 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5677679/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Colorectal cancer, the second most fatal malignancy globally, burdens public healthcare systems. AI-assisted cancer diagnostics could enable significant cost savings. This study presents a multi-scale ensemble model for DNA mismatch repair deficiency (dMMR) detection from Whole Slide Images (WSIs). dMMR is a clinically important feature, traditionally identified through labor- and time-intensive DNA analysis. The dMMR prediction capability of non-tumorous regions was also evaluated, but it showed limited potential. Therefore, tumorous regions were utilized. The model, comprising two convolutional neural network (CNN) branches and an XGBoost layer, was trained on 1,228 WSIs. It achieved an F 1 score of 0.863 (sensitivity 0.852) on internal testing, and F 1 scores of 0.770 (sensitivity 0.868) and 0.743 (sensitivity 0.951) on external test sets of 1,010 and 457 WSIs, respectively. The results indicate that a multi-scale approach can be an effective strategy when developing digital pathology algorithms. Biological sciences/Computational biology and bioinformatics/Machine learning Health sciences/Health care/Diagnosis/Pathology Biological sciences/Cancer/Gastrointestinal cancer/Colorectal cancer colorectal cancer DNA mismatch-repair deficiency microsatellite instability multi-scale ensemble model digital pathology artificial intelligence deep learning histopathology Full Text Additional Declarations No competing interests reported. Supplementary Files graphicalabstract.png SupplementalInformation.pdf Cite Share Download PDF Status: Posted 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-5677679","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":396294810,"identity":"c144a943-a8d6-4346-9d69-4e129384f3dc","order_by":0,"name":"Liisa Petäinen","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAtUlEQVRIiWNgGAWjYBACPmYeEGXBwM9wAEjbSBDWwgbRIsEg2QDSkkaMFgaoFgOQDoY0IhzGxs578OGPGgk544NnzCQYEiyIcRhfsjHPMQljswNgLcT5xUyagU0icRtIC+MP4rSY//zxTyJxcwMptjDwtkkkbmAgQYuxNG+fhLHEgWPFFgnEaOHnP2P48cc3Gzn+GYc33viQUEdYCwJInDBgSCBFA9C+9gekaRgFo2AUjIIRAwDygC5Y4dT1VwAAAABJRU5ErkJggg==","orcid":"","institution":"Faculty of Information Technology, University of Jyväskylä","correspondingAuthor":true,"prefix":"","firstName":"Liisa","middleName":"","lastName":"Petäinen","suffix":""},{"id":396294811,"identity":"bb167c4c-3491-4da0-881c-297d11ae82ad","order_by":1,"name":"Juha P. Väyrynen","email":"","orcid":"","institution":"Translational Medicine Research Unit, Medical Research Center, University of Oulu","correspondingAuthor":false,"prefix":"","firstName":"Juha","middleName":"P.","lastName":"Väyrynen","suffix":""},{"id":396294812,"identity":"1e6fca22-cbff-4acb-a0ca-51f78c02075d","order_by":2,"name":"Jan Böhm","email":"","orcid":"","institution":"Department of Pathology, Wellbeing Services County of Central Finland","correspondingAuthor":false,"prefix":"","firstName":"Jan","middleName":"","lastName":"Böhm","suffix":""},{"id":396294813,"identity":"6afba413-dd7c-42cf-93d3-1199883ca5fa","order_by":3,"name":"Pekka Ruusuvuori","email":"","orcid":"","institution":"Cancer Research Unit, Institute of Biomedicine, University of Turku","correspondingAuthor":false,"prefix":"","firstName":"Pekka","middleName":"","lastName":"Ruusuvuori","suffix":""},{"id":396294814,"identity":"4eeb744c-3034-4ad7-8cd9-7b68e8878e96","order_by":4,"name":"Maarit Ahtiainen","email":"","orcid":"","institution":"Central Finland Biobank, Wellbeing Services County of Central Finland","correspondingAuthor":false,"prefix":"","firstName":"Maarit","middleName":"","lastName":"Ahtiainen","suffix":""},{"id":396294815,"identity":"83ad0d28-6881-4323-a61a-9dbb9d74e97e","order_by":5,"name":"Hanna Elomaa","email":"","orcid":"","institution":"Department of Education and Research, Wellbeing Services County of Central Finland","correspondingAuthor":false,"prefix":"","firstName":"Hanna","middleName":"","lastName":"Elomaa","suffix":""},{"id":396294816,"identity":"2a8388f8-f15c-45c4-84b7-0dce8b072af9","order_by":6,"name":"Henna Karjalainen","email":"","orcid":"","institution":"Translational Medicine Research Unit, Medical Research Center, University of Oulu","correspondingAuthor":false,"prefix":"","firstName":"Henna","middleName":"","lastName":"Karjalainen","suffix":""},{"id":396294817,"identity":"fc302dd7-9478-42ae-8298-70d69ee1b359","order_by":7,"name":"Meeri Kastinen","email":"","orcid":"","institution":"Translational Medicine Research Unit, Medical Research Center, University of Oulu","correspondingAuthor":false,"prefix":"","firstName":"Meeri","middleName":"","lastName":"Kastinen","suffix":""},{"id":396294818,"identity":"6db9a26e-0e0c-4e0f-85c2-42ad62a98940","order_by":8,"name":"Vilja V. Tapiainen","email":"","orcid":"","institution":"Translational Medicine Research Unit, Medical Research Center, University of Oulu","correspondingAuthor":false,"prefix":"","firstName":"Vilja","middleName":"V.","lastName":"Tapiainen","suffix":""},{"id":396294820,"identity":"e286a44b-0cac-48f8-b152-0cfa1abbfc44","order_by":9,"name":"Ville K. Äijälä","email":"","orcid":"","institution":"Translational Medicine Research Unit, Medical Research Center, University of Oulu","correspondingAuthor":false,"prefix":"","firstName":"Ville","middleName":"K.","lastName":"Äijälä","suffix":""},{"id":396294821,"identity":"82f2d69e-3b27-4df4-b52f-c092425eeb3d","order_by":10,"name":"Päivi Sirniö","email":"","orcid":"","institution":"Translational Medicine Research Unit, Medical Research Center, University of Oulu","correspondingAuthor":false,"prefix":"","firstName":"Päivi","middleName":"","lastName":"Sirniö","suffix":""},{"id":396294823,"identity":"3745dce5-1789-406e-8c82-82f1a8272d6a","order_by":11,"name":"Anne Tuomisto","email":"","orcid":"","institution":"Translational Medicine Research Unit, Medical Research Center, University of Oulu","correspondingAuthor":false,"prefix":"","firstName":"Anne","middleName":"","lastName":"Tuomisto","suffix":""},{"id":396294825,"identity":"86917e69-93fd-4b9e-bc2f-4bb0a0a9a499","order_by":12,"name":"Markus J. Mäkinen","email":"","orcid":"","institution":"Translational Medicine Research Unit, Medical Research Center, University of Oulu","correspondingAuthor":false,"prefix":"","firstName":"Markus","middleName":"J.","lastName":"Mäkinen","suffix":""},{"id":396294826,"identity":"4012979c-d9f3-457c-be95-8aa923281df3","order_by":13,"name":"Jukka-Pekka Mecklin","email":"","orcid":"","institution":"Faculty of Sport and Health Sciences, University of Jyväskylä","correspondingAuthor":false,"prefix":"","firstName":"Jukka-Pekka","middleName":"","lastName":"Mecklin","suffix":""},{"id":396294827,"identity":"12667820-39b2-4f4b-afa5-f0fde827f734","order_by":14,"name":"Ilkka Pölönen","email":"","orcid":"","institution":"Faculty of Information Technology, University of Jyväskylä","correspondingAuthor":false,"prefix":"","firstName":"Ilkka","middleName":"","lastName":"Pölönen","suffix":""},{"id":396294828,"identity":"771d94f1-ce9e-479e-a6af-fea3aa9b204f","order_by":15,"name":"Sami Äyrämö","email":"","orcid":"","institution":"Faculty of Information Technology, University of Jyväskylä","correspondingAuthor":false,"prefix":"","firstName":"Sami","middleName":"","lastName":"Äyrämö","suffix":""}],"badges":[],"createdAt":"2024-12-19 14:38:27","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5677679/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5677679/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":74351964,"identity":"0990c80c-6682-4c71-abd8-89724e5c34b3","added_by":"auto","created_at":"2025-01-21 11:02:10","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2790365,"visible":true,"origin":"","legend":"","description":"","filename":"manuscriptpetainen01.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5677679/v1_covered_d2fe9004-d564-4697-8cdf-096b72728b25.pdf"},{"id":72766484,"identity":"355457cb-71a7-4036-9ec9-2910bb49ef8e","added_by":"auto","created_at":"2025-01-02 02:15:05","extension":"png","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":3022907,"visible":true,"origin":"","legend":"","description":"","filename":"graphicalabstract.png","url":"https://assets-eu.researchsquare.com/files/rs-5677679/v1/b0b25e790c7a9fe6fc53a15d.png"},{"id":72766478,"identity":"e52c85c7-6dac-4348-8b15-745b2b717904","added_by":"auto","created_at":"2025-01-02 02:15:05","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":665306,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementalInformation.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5677679/v1/6fec8f1efd88a7f915dfd0ed.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Multi-scale ensemble model for dMMR prediction from histopathological images of colorectal cancer","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"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":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"colorectal cancer, DNA mismatch-repair deficiency, microsatellite instability, multi-scale, ensemble model, digital pathology, artificial intelligence, deep learning, histopathology","lastPublishedDoi":"10.21203/rs.3.rs-5677679/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5677679/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eColorectal cancer, the second most fatal malignancy globally, burdens public healthcare systems. AI-assisted cancer diagnostics could enable significant cost savings. This study presents a multi-scale ensemble model for DNA mismatch repair deficiency (dMMR) detection from Whole Slide Images (WSIs). dMMR is a clinically important feature, traditionally identified through labor- and time-intensive DNA analysis. The dMMR prediction capability of non-tumorous regions was also evaluated, but it showed limited potential. Therefore, tumorous regions were utilized. The model, comprising two convolutional neural network (CNN) branches and an XGBoost layer, was trained on 1,228 WSIs. It achieved an \u003cem\u003e\u003cstrong\u003eF\u003c/strong\u003e\u003c/em\u003e\u003csub\u003e\u003cstrong\u003e1\u003c/strong\u003e\u003c/sub\u003e score of 0.863 (sensitivity 0.852) on internal testing, and \u003cem\u003e\u003cstrong\u003eF\u003c/strong\u003e\u003c/em\u003e\u003csub\u003e\u003cstrong\u003e1\u003c/strong\u003e\u003c/sub\u003e \u0026nbsp;scores of 0.770 (sensitivity 0.868) and 0.743 (sensitivity 0.951) on external test sets of 1,010 and 457 WSIs, respectively. The results indicate that a multi-scale approach can be an effective strategy when developing digital pathology algorithms.\u003c/p\u003e","manuscriptTitle":"Multi-scale ensemble model for dMMR prediction from histopathological images of colorectal cancer","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-01-02 02:15:00","doi":"10.21203/rs.3.rs-5677679/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"276db835-be7a-40d6-9d3e-2017c8f56777","owner":[],"postedDate":"January 2nd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":42221074,"name":"Biological sciences/Computational biology and bioinformatics/Machine learning"},{"id":42221075,"name":"Health sciences/Health care/Diagnosis/Pathology"},{"id":42221076,"name":"Biological sciences/Cancer/Gastrointestinal cancer/Colorectal cancer"}],"tags":[],"updatedAt":"2025-01-21T10:53:59+00:00","versionOfRecord":[],"versionCreatedAt":"2025-01-02 02:15:00","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-5677679","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5677679","identity":"rs-5677679","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}