Deep Learning of Suboptimal Spirometry to Predict Respiratory Outcomes and Mortality

Deep Learning of Suboptimal Spirometry to Predict Respiratory Outcomes and Mortality | 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 Deep Learning of Suboptimal Spirometry to Predict Respiratory Outcomes and Mortality Michael Cho, Davin Hill, Max Torop, Aria Masoomi, Peter Castaldi, and 8 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6296752/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 Importance: Obtaining spirometry requires repeated testing and using the maximal values based on quality control criteria. Whether the suboptimal efforts are useful for the prediction of respiratory outcomes is not clear. Objective: To determine whether a machine learning model could predict respiratory outcomes and mortality based on suboptimal spirometry. Design: Observational cohorts (UK Biobank and COPDGene). Setting: Multi-center; population, and disease-enriched. Participants: UK aged 40-69; US aged 45-80, >10 pack-years smoking, without respiratory diseases other than COPD or asthma. Exposures: Raw spirograms (volume-time). Main outcomes and measures: To create a combined representation of lung function we implemented a contrastive learning approach, Spirogram-based Contrastive Learning Framework (Spiro-CLF), which utilized all recorded volume-time curves per participant and applied different transformations (e.g. flow-volume, flow-time). We defined “maximal” efforts as those passing quality control (QC) with the maximum FVC; all other efforts, including submaximal and QC-failing efforts, were defined as “suboptimal”. We trained the Spiro-CLF model using both maximal and suboptimal efforts from the UK Biobank. We tested the model in a held-out 20% testing UK Biobank subset and COPDGene, on 1) binary predictions of FEV1/FVC <0.7, and FEV1 Percent Predicted (FEV1PP) <80%, 2) Cox regression for all-cause mortality, and 3) prediction of respiratory phenotypes. Results: We trained Spiro-CLF on 940,705 volume-time curves from 352,684 UKB participants with 2-3 spirometry efforts per individual (66.7% with 3 efforts) and at least one QC-passing spirometry effort. Of all spirometry efforts, 61.6% were suboptimal (37.5% submaximal and 24.1% QC-failing). In the UK Biobank, Spiro-CLF using QC-failing and submaximal efforts predicted FEV1/FVC < 0.7 with an Area under the Receiver Operating Characteristics (AUROC) of 0.956, mortality with a concordance index of 0.647, and asthma with a 9-42% improvement versus baseline models. In COPDGene (n=10,110 participants), adding QC-passing, submaximal efforts did not improve the prediction of lung function or mortality; however, Spiro-CLF representations predicted asthma and respiratory phenotypes (joint test P ≤ 2 × 10−3). Conclusions and Relevance: A machine-learning model can predict respiratory phenotypes using suboptimal spirometry; results from all spirometry efforts may contain valuable data. Additional studies are required to determine performance and utility in specific clinical scenarios. Health sciences/Diseases/Respiratory tract diseases/Chronic obstructive pulmonary disease Health sciences/Biomarkers/Diagnostic markers Full Text Additional Declarations Yes there is potential Competing Interest. BDH received grant support from Bayer. MHC has received grant support from GlaxoSmithKline and Bayer, consulting fees from Genentech and AstraZeneca, and speaking fees from Illumina. EKS has received grant support from and Bayer and Northpond Laboratories. PJC has received grant support from Bayer. SPB has received consulting fees from Sanofi/Regeneron and Boehringer Ingelheim, and CME fees from IntegrityCE. His institute has received funds from Sanofi and Nuvaira for the conduct of clinical trials. TY, FH, and CYM are employees of Google LLC and own Alphabet stock. 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-6296752","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":477417503,"identity":"83f8f5e4-770d-4cd8-b476-9d837872c9bf","order_by":0,"name":"Michael Cho","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABCklEQVRIiWNgGAWjYHACNgaGCoYEEOsAECfABcFcnFrOkKyFsQ2mkhgt5u3daQ9+zrPL45/de/BwwS8GIOOM4eeKMgY5vhsJWLXInDm73bB3W3KxxJ1zCYdn9jEUS5zLMZY8c47BWBKHFgmJ3G3SjNsOJDbcyDE4zNvDkNhwhi1BsrGNIXEDLi3yb4Fa5hxInA/TMv8MW/JPoJZ6nFokeIFaGg4AzQRq4fkBNPwM8zGQLQkGuLTw5G6T7DmWnLgRbEuDROJGoBbLhnMShjPPPMCuhf3sNokfNXaJ827kGH/m+WOTOO8MY/PNhjIbeb7j2G1BBYxtEnCziFAOBn+IVTgKRsEoGAUjCQAAzH9nZDkmxA0AAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0002-4907-1657","institution":"Brigham and Women's Hospital","correspondingAuthor":true,"prefix":"","firstName":"Michael","middleName":"","lastName":"Cho","suffix":""},{"id":477417504,"identity":"6c8a4b2d-ae7b-42ce-9dd2-a00547a2554c","order_by":1,"name":"Davin Hill","email":"","orcid":"","institution":"Northeastern University","correspondingAuthor":false,"prefix":"","firstName":"Davin","middleName":"","lastName":"Hill","suffix":""},{"id":477417505,"identity":"c982c9e8-0ddd-49af-99ff-5d23fee768bf","order_by":2,"name":"Max Torop","email":"","orcid":"","institution":"Northeastern University","correspondingAuthor":false,"prefix":"","firstName":"Max","middleName":"","lastName":"Torop","suffix":""},{"id":477417506,"identity":"c9a99ddb-efdb-4fe1-949b-9d437cfd0851","order_by":3,"name":"Aria Masoomi","email":"","orcid":"","institution":"Northeastern University","correspondingAuthor":false,"prefix":"","firstName":"Aria","middleName":"","lastName":"Masoomi","suffix":""},{"id":477417507,"identity":"5aa6b3b9-7abc-4554-a963-a35ddec45b68","order_by":4,"name":"Peter Castaldi","email":"","orcid":"","institution":"Brigham and Women's Hospital","correspondingAuthor":false,"prefix":"","firstName":"Peter","middleName":"","lastName":"Castaldi","suffix":""},{"id":477417508,"identity":"f87d3edd-4f23-4740-9ec4-6312ac5ddfec","order_by":5,"name":"Edwin Silverman","email":"","orcid":"","institution":"Brigham and Women's Hospital, Harvard Medical School","correspondingAuthor":false,"prefix":"","firstName":"Edwin","middleName":"","lastName":"Silverman","suffix":""},{"id":477417509,"identity":"057d49a8-8ccf-43a9-9eb2-6e64c90b50e1","order_by":6,"name":"Sandeep Bodduluri","email":"","orcid":"","institution":"University of Alabama at Birmingham","correspondingAuthor":false,"prefix":"","firstName":"Sandeep","middleName":"","lastName":"Bodduluri","suffix":""},{"id":477417510,"identity":"eca1f0a9-f3b0-4ed5-85e5-7eab69c19fb1","order_by":7,"name":"Surya Bhatt","email":"","orcid":"https://orcid.org/0000-0002-8418-4497","institution":"University of Alabama at Birmingham","correspondingAuthor":false,"prefix":"","firstName":"Surya","middleName":"","lastName":"Bhatt","suffix":""},{"id":477417511,"identity":"9915ebce-4ca4-450b-b606-d3b30319a9fb","order_by":8,"name":"Taedong Yun","email":"","orcid":"https://orcid.org/0000-0002-6242-5536","institution":"Google","correspondingAuthor":false,"prefix":"","firstName":"Taedong","middleName":"","lastName":"Yun","suffix":""},{"id":477417512,"identity":"65cfe0ec-3078-4531-a0a5-dfb0d753ad84","order_by":9,"name":"Cory McLean","email":"","orcid":"https://orcid.org/0000-0001-9928-8216","institution":"Google (United States)","correspondingAuthor":false,"prefix":"","firstName":"Cory","middleName":"","lastName":"McLean","suffix":""},{"id":477417513,"identity":"5517b718-d8fa-432d-a43e-0aa01adb9bbd","order_by":10,"name":"Farhad Hormozdiari","email":"","orcid":"https://orcid.org/0000-0002-5617-6174","institution":"Google","correspondingAuthor":false,"prefix":"","firstName":"Farhad","middleName":"","lastName":"Hormozdiari","suffix":""},{"id":477417514,"identity":"99c7ad39-4e24-4400-871b-dee7e52145f4","order_by":11,"name":"Jennifer Dy","email":"","orcid":"","institution":"Northeastern University","correspondingAuthor":false,"prefix":"","firstName":"Jennifer","middleName":"","lastName":"Dy","suffix":""},{"id":477417515,"identity":"3fe8be86-54f2-433a-a359-271494f0d607","order_by":12,"name":"Brian Hobbs","email":"","orcid":"https://orcid.org/0000-0001-9564-0745","institution":"Brigham and Women's Hospital","correspondingAuthor":false,"prefix":"","firstName":"Brian","middleName":"","lastName":"Hobbs","suffix":""}],"badges":[],"createdAt":"2025-03-24 15:21:46","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6296752/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6296752/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":85638304,"identity":"1b9b44ff-259c-4407-9325-d37e3898f0ba","added_by":"auto","created_at":"2025-06-30 06:39:11","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3037738,"visible":true,"origin":"","legend":"Article File","description":"","filename":"DeepLearningUtilizingSuboptimalSpirometry3.20.25.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6296752/v1_covered_3bd3a3b3-d656-4c3a-917b-e8258a4472ff.pdf"}],"financialInterests":"\u003cb\u003eYes\u003c/b\u003e there is potential Competing Interest.\nBDH received grant support from Bayer.\r\nMHC has received grant support from GlaxoSmithKline and Bayer, consulting fees from Genentech and\r\nAstraZeneca, and speaking fees from Illumina.\r\nEKS has received grant support from and Bayer and Northpond Laboratories.\r\nPJC has received grant support from Bayer.\r\nSPB has received consulting fees from Sanofi/Regeneron and Boehringer Ingelheim, and CME fees from\r\nIntegrityCE. His institute has received funds from Sanofi and Nuvaira for the conduct of clinical trials.\r\nTY, FH, and CYM are employees of Google LLC and own Alphabet stock.","formattedTitle":"Deep Learning of Suboptimal Spirometry to Predict Respiratory Outcomes and Mortality","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-6296752/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6296752/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Importance: Obtaining spirometry requires repeated testing and using the maximal values based\r\non quality control criteria. Whether the suboptimal efforts are useful for the prediction of respiratory\r\noutcomes is not clear.\r\n\r\nObjective: To determine whether a machine learning model could predict respiratory outcomes and\r\nmortality based on suboptimal spirometry.\r\n\r\nDesign: Observational cohorts (UK Biobank and COPDGene).\r\n\r\nSetting: Multi-center; population, and disease-enriched.\r\n\r\nParticipants: UK aged 40-69; US aged 45-80, \u003e10 pack-years smoking, without respiratory diseases\r\nother than COPD or asthma.\r\n\r\nExposures: Raw spirograms (volume-time).\r\n\r\nMain outcomes and measures: To create a combined representation of lung function we\r\nimplemented a contrastive learning approach, Spirogram-based Contrastive Learning Framework\r\n(Spiro-CLF), which utilized all recorded volume-time curves per participant and applied different\r\ntransformations (e.g. flow-volume, flow-time). We defined “maximal” efforts as those passing quality\r\ncontrol (QC) with the maximum FVC; all other efforts, including submaximal and QC-failing efforts,\r\nwere defined as “suboptimal”. We trained the Spiro-CLF model using both maximal and suboptimal\r\nefforts from the UK Biobank. We tested the model in a held-out 20% testing UK Biobank subset and\r\nCOPDGene, on 1) binary predictions of FEV1/FVC \u003c0.7, and FEV1 Percent Predicted (FEV1PP)\r\n\u003c80%, 2) Cox regression for all-cause mortality, and 3) prediction of respiratory phenotypes.\r\n\r\nResults: We trained Spiro-CLF on 940,705 volume-time curves from 352,684 UKB participants with\r\n2-3 spirometry efforts per individual (66.7% with 3 efforts) and at least one QC-passing spirometry\r\neffort. Of all spirometry efforts, 61.6% were suboptimal (37.5% submaximal and 24.1% QC-failing).\r\nIn the UK Biobank, Spiro-CLF using QC-failing and submaximal efforts predicted FEV1/FVC \u003c\r\n0.7 with an Area under the Receiver Operating Characteristics (AUROC) of 0.956, mortality with\r\na concordance index of 0.647, and asthma with a 9-42% improvement versus baseline models. In\r\nCOPDGene (n=10,110 participants), adding QC-passing, submaximal efforts did not improve the\r\nprediction of lung function or mortality; however, Spiro-CLF representations predicted asthma and\r\nrespiratory phenotypes (joint test P ≤ 2 × 10−3).\r\n\r\nConclusions and Relevance: A machine-learning model can predict respiratory phenotypes using\r\nsuboptimal spirometry; results from all spirometry efforts may contain valuable data. Additional\r\nstudies are required to determine performance and utility in specific clinical scenarios.","manuscriptTitle":"Deep Learning of Suboptimal Spirometry to Predict Respiratory Outcomes and Mortality","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-06-30 06:31:02","doi":"10.21203/rs.3.rs-6296752/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":"1c903040-a189-474a-9ccd-c672060e7ec8","owner":[],"postedDate":"June 30th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":50698798,"name":"Health sciences/Diseases/Respiratory tract diseases/Chronic obstructive pulmonary disease"},{"id":50698799,"name":"Health sciences/Biomarkers/Diagnostic markers"}],"tags":[],"updatedAt":"2026-04-15T14:12:45+00:00","versionOfRecord":[],"versionCreatedAt":"2025-06-30 06:31:02","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6296752","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6296752","identity":"rs-6296752","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}