AI-driven predictive biomarker discovery with contrastive learning to improve clinical trial outcomes

IF 48.8 1区医学 Q1 CELL BIOLOGY

Cancer Cell Pub Date : 2025-04-17 DOI:10.1016/j.ccell.2025.03.029

Gustavo Arango-Argoty, Damian E. Bikiel, Gerald J. Sun, Elly Kipkogei, Kaitlin M. Smith, Sebastian Carrasco Pro, Elizabeth Y. Choe, Etai Jacob

{"title":"AI-driven predictive biomarker discovery with contrastive learning to improve clinical trial outcomes","authors":"Gustavo Arango-Argoty, Damian E. Bikiel, Gerald J. Sun, Elly Kipkogei, Kaitlin M. Smith, Sebastian Carrasco Pro, Elizabeth Y. Choe, Etai Jacob","doi":"10.1016/j.ccell.2025.03.029","DOIUrl":null,"url":null,"abstract":"Modern clinical trials can capture tens of thousands of clinicogenomic measurements per individual. Discovering predictive biomarkers, as opposed to prognostic markers, remains challenging. To address this, we present a neural network framework based on contrastive learning—the Predictive Biomarker Modeling Framework (PBMF)—that explores potential predictive biomarkers in an automated, systematic, and unbiased manner. Applied retrospectively to real clinicogenomic datasets, particularly for immuno-oncology (IO) trials, our algorithm identifies biomarkers of IO-treated individuals who survive longer than those treated with other therapies. We demonstrate how our framework retrospectively contributes to a phase 3 clinical trial by uncovering a predictive, interpretable biomarker based solely on early study data. Patients identified with this predictive biomarker show a 15% improvement in survival risk compared to those in the original trial. The PBMF offers a general-purpose, rapid, and robust approach to inform biomarker strategy, providing actionable outcomes for clinical decision-making.","PeriodicalId":9670,"journal":{"name":"Cancer Cell","volume":"263 1","pages":""},"PeriodicalIF":48.8000,"publicationDate":"2025-04-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Cancer Cell","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.1016/j.ccell.2025.03.029","RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"CELL BIOLOGY","Score":null,"Total":0}

引用次数: 0

Abstract

Modern clinical trials can capture tens of thousands of clinicogenomic measurements per individual. Discovering predictive biomarkers, as opposed to prognostic markers, remains challenging. To address this, we present a neural network framework based on contrastive learning—the Predictive Biomarker Modeling Framework (PBMF)—that explores potential predictive biomarkers in an automated, systematic, and unbiased manner. Applied retrospectively to real clinicogenomic datasets, particularly for immuno-oncology (IO) trials, our algorithm identifies biomarkers of IO-treated individuals who survive longer than those treated with other therapies. We demonstrate how our framework retrospectively contributes to a phase 3 clinical trial by uncovering a predictive, interpretable biomarker based solely on early study data. Patients identified with this predictive biomarker show a 15% improvement in survival risk compared to those in the original trial. The PBMF offers a general-purpose, rapid, and robust approach to inform biomarker strategy, providing actionable outcomes for clinical decision-making.

Abstract Image

查看原文本刊更多论文

求助全文

约1分钟内获得全文求助全文

来源期刊

Cancer Cell 医学-肿瘤学

CiteScore

55.20

自引率

1.20%

发文量

179

审稿时长

4-8 weeks

期刊介绍： Cancer Cell is a journal that focuses on promoting major advances in cancer research and oncology. The primary criteria for considering manuscripts are as follows: Major advances: Manuscripts should provide significant advancements in answering important questions related to naturally occurring cancers. Translational research: The journal welcomes translational research, which involves the application of basic scientific findings to human health and clinical practice. Clinical investigations: Cancer Cell is interested in publishing clinical investigations that contribute to establishing new paradigms in the treatment, diagnosis, or prevention of cancers. Insights into cancer biology: The journal values clinical investigations that provide important insights into cancer biology beyond what has been revealed by preclinical studies. Mechanism-based proof-of-principle studies: Cancer Cell encourages the publication of mechanism-based proof-of-principle clinical studies, which demonstrate the feasibility of a specific therapeutic approach or diagnostic test.