{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"institution":[{"name":"medRxiv"}],"indexed":{"date-parts":[[2026,3,13]],"date-time":"2026-03-13T23:28:52Z","timestamp":1773444532977,"version":"3.50.1"},"posted":{"date-parts":[[2025,9,19]]},"group-title":"Oncology","reference-count":34,"publisher":"openRxiv","license":[{"start":{"date-parts":[[2025,9,19]],"date-time":"2025-09-19T00:00:00Z","timestamp":1758240000000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by-nc-nd\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"accepted":{"date-parts":[[2025,9,19]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                <jats:p>Cancer cachexia, a multifactorial metabolic syndrome characterized by severe muscle wasting and weight loss, contributes to poor outcomes across various cancer types but lacks a standardized, generalizable biomarker for early detection. We present a multimodal AI-based biomarker trained on real-world clinical, radiologic, laboratory, and unstructured clinical note data, leveraging foundation models and large language models (LLMs) to identify cachexia at the time of cancer diagnosis. Prediction accuracy improved with each added modality: 77% using clinical variables alone, 81% with added laboratory data, and 85% with structured symptom features extracted from clinical notes. Incorporating embeddings from clinical text and CT images further improved accuracy to 92%. The framework also demonstrated prognostic utility, improving survival prediction as data modalities were integrated. Designed for real-world clinical deployment, the framework accommodates missing modalities without requiring imputation or case exclusion, supporting scalability across diverse oncology settings. Unlike prior models trained on curated datasets, our approach utilizes standard-of-care clinical data, facilitating integration into oncology workflows. In contrast to fixed-threshold composite indices such as the cachexia index (CXI), the model generates patient-specific predictions, enabling adaptable, cancer-agnostic performance. To enhance clinical reliability and safety, the framework incorporates uncertainty estimation to flag low-confidence cases for expert review. This work advances a clinically applicable, scalable, and trustworthy AI-driven decision support tool for early cachexia detection and personalized oncology care.<\/jats:p>","DOI":"10.1101\/2025.09.17.25336028","type":"posted-content","created":{"date-parts":[[2025,9,19]],"date-time":"2025-09-19T15:00:13Z","timestamp":1758294013000},"source":"Crossref","is-referenced-by-count":2,"title":["Multimodal AI-driven Biomarker for Early Detection of Cancer Cachexia"],"prefix":"10.64898","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-0456-3073","authenticated-orcid":false,"given":"Sabeen","family":"Ahmed","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Nathan","family":"Parker","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Margaret","family":"Park","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Evan W.","family":"Davis","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Daniel","family":"Jeong","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Jennifer B.","family":"Permuth","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Matthew B.","family":"Schabath","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yasin","family":"Yilmaz","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Ghulam","family":"Rasool","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"54368","reference":[{"key":"2025092206350646000_2025.09.17.25336028v1.1","doi-asserted-by":"publisher","DOI":"10.1186\/s13550-020-00737-8"},{"key":"2025092206350646000_2025.09.17.25336028v1.2","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijcard.2009.07.042"},{"key":"2025092206350646000_2025.09.17.25336028v1.3","doi-asserted-by":"publisher","DOI":"10.1016\/S1470-2045(10)70218-7"},{"key":"2025092206350646000_2025.09.17.25336028v1.4","doi-asserted-by":"crossref","unstructured":"Jafri, S.H.R. , et al., Cachexia index in advanced non-small-cell lung cancer patients. Clinical Medicine Insights: Oncology, 2015. 9: p. CMO. S30891.","DOI":"10.4137\/CMO.S30891"},{"issue":"6","key":"2025092206350646000_2025.09.17.25336028v1.5","doi-asserted-by":"crossref","first-page":"2211","DOI":"10.1002\/jcsm.12837","article-title":"Cachexia index as a potential biomarker for cancer cachexia and a prognostic indicator in diffuse large B-cell lymphoma","volume":"12","year":"2021","journal-title":"Journal of cachexia, sarcopenia and muscle"},{"issue":"10","key":"2025092206350646000_2025.09.17.25336028v1.6","doi-asserted-by":"crossref","first-page":"880","DOI":"10.1038\/s41430-024-01469-x","article-title":"Developing and validating a modified cachexia index to predict the outcomes for colorectal cancer after radical surgery","volume":"78","year":"2024","journal-title":"European Journal of Clinical Nutrition"},{"issue":"2","key":"2025092206350646000_2025.09.17.25336028v1.7","doi-asserted-by":"crossref","first-page":"891","DOI":"10.1002\/jcsm.13207","article-title":"Development and validation of a cancer cachexia risk score for digestive tract cancer patients before abdominal surgery","volume":"14","year":"2023","journal-title":"Journal of cachexia, sarcopenia and muscle"},{"key":"2025092206350646000_2025.09.17.25336028v1.8","doi-asserted-by":"crossref","first-page":"87","DOI":"10.1007\/s13539-011-0027-5","article-title":"The cachexia score (CASCO): a new tool for staging cachectic cancer patients","volume":"2","year":"2011","journal-title":"Journal of cachexia, sarcopenia and muscle"},{"issue":"1","key":"2025092206350646000_2025.09.17.25336028v1.9","doi-asserted-by":"crossref","first-page":"136","DOI":"10.1038\/s41746-020-00341-z","article-title":"Fusion of medical imaging and electronic health records using deep learning: a systematic review and implementation guidelines","volume":"3","year":"2020","journal-title":"NPJ digital medicine"},{"key":"2025092206350646000_2025.09.17.25336028v1.10","unstructured":"Waqas, A. , et al., Embedding-based Multimodal Learning on Pan-Squamous Cell Carcinomas for Improved Survival Outcomes. arXiv preprint arXiv:2406.08521, 2024."},{"key":"2025092206350646000_2025.09.17.25336028v1.11","doi-asserted-by":"crossref","first-page":"1408843","DOI":"10.3389\/frai.2024.1408843","article-title":"Multimodal data integration for oncology in the era of deep neural networks: a review","volume":"7","year":"2024","journal-title":"Frontiers in Artificial Intelligence"},{"key":"2025092206350646000_2025.09.17.25336028v1.12","doi-asserted-by":"crossref","unstructured":"Waqas, A. , et al., Revolutionizing digital pathology with the power of generative artificial intelligence and foundation models. Laboratory Investigation, 2023: p. 100255.","DOI":"10.1016\/j.labinv.2023.100255"},{"issue":"10","key":"2025092206350646000_2025.09.17.25336028v1.13","doi-asserted-by":"crossref","first-page":"630","DOI":"10.1007\/s00520-024-08833-4","article-title":"Machine learning to identify precachexia and cachexia: a multicenter, retrospective cohort study","volume":"32","year":"2024","journal-title":"Supportive Care in Cancer"},{"key":"2025092206350646000_2025.09.17.25336028v1.14","doi-asserted-by":"crossref","first-page":"919046","DOI":"10.3389\/fmedt.2022.919046","article-title":"Failure detection in deep neural networks for medical imaging","volume":"4","year":"2022","journal-title":"Frontiers in Medical Technology"},{"issue":"1","key":"2025092206350646000_2025.09.17.25336028v1.15","doi-asserted-by":"crossref","first-page":"26","DOI":"10.1038\/s41746-024-01007-w","article-title":"Uncertainty-aware deep-learning model for prediction of supratentorial hematoma expansion from admission non-contrast head computed tomography scan","volume":"7","year":"2024","journal-title":"NPJ Digital Medicine"},{"issue":"1","key":"2025092206350646000_2025.09.17.25336028v1.16","doi-asserted-by":"crossref","first-page":"4","DOI":"10.1038\/s41746-020-00367-3","article-title":"Second opinion needed: communicating uncertainty in medical machine learning","volume":"4","year":"2021","journal-title":"NPJ Digital Medicine"},{"issue":"4","key":"2025092206350646000_2025.09.17.25336028v1.17","doi-asserted-by":"crossref","first-page":"809","DOI":"10.3390\/cancers13040809","article-title":"The Florida pancreas collaborative next-generation biobank: infrastructure to reduce disparities and improve survival for a diverse cohort of patients with pancreatic cancer","volume":"13","year":"2021","journal-title":"Cancers"},{"issue":"5","key":"2025092206350646000_2025.09.17.25336028v1.18","doi-asserted-by":"crossref","first-page":"1378","DOI":"10.1016\/j.clnu.2016.09.008","article-title":"Use of routinely available clinical, nutritional, and functional criteria to classify cachexia in advanced cancer patients","volume":"36","year":"2017","journal-title":"Clinical nutrition"},{"issue":"1","key":"2025092206350646000_2025.09.17.25336028v1.19","doi-asserted-by":"crossref","first-page":"194","DOI":"10.1038\/s41746-022-00742-2","article-title":"A large language model for electronic health records","volume":"5","year":"2022","journal-title":"NPJ digital medicine"},{"key":"2025092206350646000_2025.09.17.25336028v1.20","unstructured":"Grattafiori, A. , et al., The llama 3 herd of models. arXiv preprint arXiv:2407.21783, 2024."},{"key":"2025092206350646000_2025.09.17.25336028v1.21","unstructured":"Bai, J. , et al., Qwen technical report. arXiv preprint arXiv:2309.16609, 2023."},{"key":"2025092206350646000_2025.09.17.25336028v1.22","unstructured":"Liu, A. , et al., Deepseek-v3 technical report. arXiv preprint arXiv:2412.19437, 2024."},{"key":"2025092206350646000_2025.09.17.25336028v1.23","doi-asserted-by":"crossref","unstructured":"Ahmed, S. , et al., Reliable Radiologic Skeletal Muscle Area Assessment-A Biomarker for Cancer Cachexia Diagnosis. medRxiv, 2025: p. 2025.04. 21.25326162.","DOI":"10.1101\/2025.04.21.25326162"},{"key":"2025092206350646000_2025.09.17.25336028v1.24","unstructured":"Beenish Zia, S.A. , Steve Johnson , Jared Pager , S Antoine Aliotti , Yingpo Huang , Jerome Knoplioch , Yannick Leberre , Lionel Marmonier , Florentin Toulemon , White Paper: Accelerate Your Visualization Experience. 2020, Intel Corporation and GE Healthcare: Santa Clara, CA."},{"key":"2025092206350646000_2025.09.17.25336028v1.25","doi-asserted-by":"crossref","unstructured":"Tripathi, A. , et al., HONeYBEE: Enabling Scalable Multimodal AI in Oncology Through Foundation Model\u2013Driven Embeddings. medRxiv, 2025: p. 2025.04. 22.25326222.","DOI":"10.1101\/2025.04.22.25326222"},{"issue":"5","key":"2025092206350646000_2025.09.17.25336028v1.26","first-page":"e210315","article-title":"RadImageNet: an open radiologic deep learning research dataset for effective transfer learning","volume":"4","year":"2022","journal-title":"Radiology: Artificial Intelligence"},{"key":"2025092206350646000_2025.09.17.25336028v1.27","unstructured":"Biewald, L ., Experiment Tracking with Weights and Biases. 2020."},{"issue":"129","key":"2025092206350646000_2025.09.17.25336028v1.28","first-page":"1","article-title":"Time-to-event prediction with neural networks and Cox regression","volume":"20","year":"2019","journal-title":"Journal of machine learning research"},{"issue":"1","key":"2025092206350646000_2025.09.17.25336028v1.29","doi-asserted-by":"crossref","first-page":"192","DOI":"10.1186\/s12859-021-04103-w","article-title":"Survival time prediction by integrating cox proportional hazards network and distribution function network","volume":"22","year":"2021","journal-title":"BMC bioinformatics"},{"key":"2025092206350646000_2025.09.17.25336028v1.30","doi-asserted-by":"publisher","DOI":"10.1001\/jama.1982.03320430047030"},{"key":"2025092206350646000_2025.09.17.25336028v1.31","doi-asserted-by":"publisher","DOI":"10.1002\/(SICI)1097-0258(19960229)15:4<361::AID-SIM168>3.0.CO;2-4"},{"issue":"4","key":"2025092206350646000_2025.09.17.25336028v1.32","doi-asserted-by":"crossref","first-page":"286","DOI":"10.1097\/CCO.0000000000000517","article-title":"Cancer cachexia, a clinical challenge","volume":"31","year":"2019","journal-title":"Current opinion in oncology"},{"key":"2025092206350646000_2025.09.17.25336028v1.33","doi-asserted-by":"publisher","DOI":"10.1200\/JCO.20.00611"},{"key":"2025092206350646000_2025.09.17.25336028v1.34","doi-asserted-by":"publisher","DOI":"10.1038\/s41572-018-0025-4"}],"container-title":[],"original-title":[],"link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.1101\/2025.09.17.25336028","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,1,15]],"date-time":"2026-01-15T20:46:17Z","timestamp":1768509977000},"score":1,"resource":{"primary":{"URL":"http:\/\/medrxiv.org\/lookup\/doi\/10.1101\/2025.09.17.25336028"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,9,19]]},"references-count":34,"URL":"https:\/\/doi.org\/10.1101\/2025.09.17.25336028","relation":{},"subject":[],"published":{"date-parts":[[2025,9,19]]},"subtype":"preprint"}}