Modular Meta-Evolutionary AI Architecture Enables Interpretable Stratification in Heterogeneous Clinical Trials

Modular Meta-Evolutionary AI Architecture Enables Interpretable Stratification in Heterogeneous Clinical Trials | 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 Modular Meta-Evolutionary AI Architecture Enables Interpretable Stratification in Heterogeneous Clinical Trials Joseph Geraci, Bessi Qorri, Christian Cumbaa, Mike Tsay, Paul Leonczyk, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9031851/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 Foundation models are poorly matched to small, heterogeneous clinical datasets where transparency and auditability are required. We describe a modular, meta-evolutionary architecture that pairs an interpretable dynamical-systems learner (NetraAI) that leverages a long-range memory (LRM) mechanism to identify stable, outcome-linked Model-Derived Subgroups (MDS) and abstain when evidence is insufficient, coupled to a literature-grounded large language model (LLM) Strategist used for structured scientific critique and robustness testing. Across three datasets: CATIE schizophrenia (olanzapine vs perphenazine), CAN-BIND depression (escitalopram response), and COMPASS pancreatic cancer (GnP vs FOLFIRINOX), standard classifiers achieved near-chance whole-cohort prediction (AUC 0.51-0.57), whereas NetraAI identified compact 2-4 variable MDS yielding high discrimination within high-confidence patients (AUC up to ~1.0), including a 3-SNV signature with high C-for-benefit (0.92) for regimen selection in PDAC. Combining dynamical-systems subgroup discovery with LLM-guided scientific critique embodies the AI modularity hypothesis – showing how distinct computational models can jointly transform small, heterogeneous clinical datasets into concise, interpretable patient stratifications. Biological sciences/Cancer Biological sciences/Computational biology and bioinformatics artificial intelligence interpretable machine learning clinical trials foundation models selective prediction subgroup discovery large language models Full Text Additional Declarations Competing interest reported. JG, BQ, CC, MT, PL, and AG are employed by NetraMark Corp. JG declares that he owns substantial shares in NetraMark Holdings, which funded a major portion of this study. LP is also a shareholder in this company. LP disclosures (Last 2 years): AbbVie, USA; Acumen, USA; Aicure, USA; Alexion, Italy; BCG, Switzerland; Astra-Zeneca, Italy; Boehringer Ingelheim International GmbH, Germany; EDRA-LSWR Publishing Company, Italy; GH-Pharma, Ireland; GLG-Institute, USA; Immunogen, USA; Johnson & Johnson, USA; LB-Pharmaceuticals, USA; Magdalena BioSciences, USA; MSD, Italy; Sanofi-Aventis-Genzyme, France and USA; Lundbeck, Denmark and Italy; Napo-Pharma, USA and EU; NetraMark, Canada; Pfizer Global, USA; Relmada Therapeutics, USA; Takeda, USA. Shares/Options: Relmada, NetraMark. 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-9031851","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":604816151,"identity":"c9e76f18-1e4e-4e55-a4a6-db53dcb4b7f1","order_by":0,"name":"Joseph Geraci","email":"","orcid":"","institution":"Queen's University","correspondingAuthor":false,"prefix":"","firstName":"Joseph","middleName":"","lastName":"Geraci","suffix":""},{"id":604816152,"identity":"30f42d5a-7810-4e1d-8428-802bd7816b0c","order_by":1,"name":"Bessi Qorri","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAzElEQVRIiWNgGAWjYBACAxCRUAHnJxCr5QwDAw9pWhjbSNFizt7++MPDeXfk7KUPP93wc08aA397N359lj1nzCQStz0z5uFLM7vZ8yyHQeLM2Q34HXYjh40hcdvhxB4eBrMbPAcqGAwkcglouf/88YfEOYfre3jYv938Q5SWG0A1iQ2HE3h4eMxu8xzIIazFsifHTCLh2GHDnjM8ZbdlDqTxEPSLOfvxxx9/1ByWZ+9h33bzzYFkOf72XvxaMAAPacpHwSgYBaNgFGAFAAxISioIFeOAAAAAAElFTkSuQmCC","orcid":"","institution":"NetraMark Corp","correspondingAuthor":true,"prefix":"","firstName":"Bessi","middleName":"","lastName":"Qorri","suffix":""},{"id":604816153,"identity":"fb1fe9cd-6c3f-4a7c-b26d-7be3b6bbfa6c","order_by":2,"name":"Christian Cumbaa","email":"","orcid":"","institution":"NetraMark Corp","correspondingAuthor":false,"prefix":"","firstName":"Christian","middleName":"","lastName":"Cumbaa","suffix":""},{"id":604816154,"identity":"9f7af31f-c78c-4536-ba41-ba630bd84b8b","order_by":3,"name":"Mike Tsay","email":"","orcid":"","institution":"NetraMark Corp","correspondingAuthor":false,"prefix":"","firstName":"Mike","middleName":"","lastName":"Tsay","suffix":""},{"id":604816155,"identity":"d5b91cef-09cd-40d3-904c-1bdb88a18ecd","order_by":4,"name":"Paul Leonczyk","email":"","orcid":"","institution":"NetraMark Corp","correspondingAuthor":false,"prefix":"","firstName":"Paul","middleName":"","lastName":"Leonczyk","suffix":""},{"id":604816156,"identity":"b5cb8b94-d42e-4bdd-bd9b-588b4e6102f7","order_by":5,"name":"Adam Gogacz","email":"","orcid":"","institution":"NetraMark Corp","correspondingAuthor":false,"prefix":"","firstName":"Adam","middleName":"","lastName":"Gogacz","suffix":""},{"id":604816157,"identity":"071beb21-e790-4d15-ab1d-72dc9d4de100","order_by":6,"name":"Luca Pani","email":"","orcid":"","institution":"University of Miami","correspondingAuthor":false,"prefix":"","firstName":"Luca","middleName":"","lastName":"Pani","suffix":""}],"badges":[],"createdAt":"2026-03-04 15:23:05","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9031851/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9031851/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":105728948,"identity":"c248a7ac-b140-4f33-88f1-45ab66999f2c","added_by":"auto","created_at":"2026-03-30 11:13:03","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":953053,"visible":true,"origin":"","legend":"","description":"","filename":"ManuscriptNetraAIMetaEvolutionaryFramework03.04.2026.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9031851/v1_covered_7d3583b2-5323-4e54-b9d2-e1815fe5c5ea.pdf"}],"financialInterests":"Competing interest reported. JG, BQ, CC, MT, PL, and AG are employed by NetraMark Corp. JG declares that he owns substantial shares in NetraMark Holdings, which funded a major portion of this study. LP is also a shareholder in this company. LP disclosures (Last 2 years): AbbVie, USA; Acumen, USA; Aicure, USA; Alexion, Italy; BCG, Switzerland; Astra-Zeneca, Italy; Boehringer Ingelheim International GmbH, Germany; EDRA-LSWR Publishing Company, Italy; GH-Pharma, Ireland; GLG-Institute, USA; Immunogen, USA; Johnson \u0026 Johnson, USA; LB-Pharmaceuticals, USA; Magdalena BioSciences, USA; MSD, Italy; Sanofi-Aventis-Genzyme, France and USA; Lundbeck, Denmark and Italy; Napo-Pharma, USA and EU; NetraMark, Canada; Pfizer Global, USA; Relmada Therapeutics, USA; Takeda, USA. Shares/Options: Relmada, NetraMark.","formattedTitle":"Modular Meta-Evolutionary AI Architecture Enables Interpretable Stratification in Heterogeneous Clinical Trials","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":"artificial intelligence, interpretable machine learning, clinical trials, foundation models, selective prediction, subgroup discovery, large language models","lastPublishedDoi":"10.21203/rs.3.rs-9031851/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9031851/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eFoundation models are poorly matched to small, heterogeneous clinical datasets where transparency and auditability are required. We describe a modular, meta-evolutionary architecture that pairs an interpretable dynamical-systems learner (NetraAI) that leverages a long-range memory (LRM) mechanism to identify stable, outcome-linked Model-Derived Subgroups (MDS) and abstain when evidence is insufficient, coupled to a literature-grounded large language model (LLM) Strategist used for structured scientific critique and robustness testing. Across three datasets: CATIE schizophrenia (olanzapine vs perphenazine), CAN-BIND depression (escitalopram response), and COMPASS pancreatic cancer (GnP vs FOLFIRINOX), standard classifiers achieved near-chance whole-cohort prediction (AUC 0.51-0.57), whereas NetraAI identified compact 2-4 variable MDS \u0026nbsp;yielding high discrimination within high-confidence patients (AUC up to ~1.0), including a 3-SNV signature with high C-for-benefit (0.92) for regimen selection in PDAC. Combining dynamical-systems subgroup discovery with LLM-guided scientific critique embodies the AI modularity hypothesis – showing how distinct computational models can jointly transform small, heterogeneous clinical datasets into concise, interpretable patient stratifications.\u003c/p\u003e","manuscriptTitle":"Modular Meta-Evolutionary AI Architecture Enables Interpretable Stratification in Heterogeneous Clinical Trials","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-13 23:32:48","doi":"10.21203/rs.3.rs-9031851/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":"1f609aab-548c-4803-b0d9-70675f6f15af","owner":[],"postedDate":"March 13th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":64364592,"name":"Biological sciences/Cancer"},{"id":64364593,"name":"Biological sciences/Computational biology and bioinformatics"}],"tags":[],"updatedAt":"2026-03-30T03:39:52+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-13 23:32:48","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9031851","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9031851","identity":"rs-9031851","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}