基于大语言模型与结构化数据预测卒中预警中急性颅内血管事件的研究
Predicting acute intracranial vascular events in stroke alerts using large-language models and structured data
研究人员基于Mount Sinai健康系统2011至2021年卒中预警记录,开发了结合大语言模型、结构化电子病历与临床时序特征的多模态机器学习模型,用于预测急性颅内血管事件(AIVE)。
- Article
- Open access
- Published: 22 October 2026
- Asala N. Erekat ORCID: orcid.org/0000-0002-8111-46101,2,
- Margaret H. Downes3,
- Laura K. Stein ORCID: orcid.org/0000-0002-8268-616X2,
- Bradley N. Delman4,
- Adam M. Karp2,
- Ankita Tripathi2,
- Girish N. Nadkarni5,
- Mark J. Kupersmith1,6,7 &
- …
- Benjamin R. Kummer1,2,8
npj Cardiovascular Health (2026) Cite this article
We’re sharing this article early to provide faster access to peer-reviewed, accepted research. It is citable and carries a permanent DOI. This version is subject to further edits and will be replaced automatically by the final Version of Record. All legal disclaimers apply.
Abstract
Acute stroke alerts are frequently activated for non-cerebrovascular conditions, creating false positives that strain resources and increase diagnostic uncertainty. We developed machine learning models integrating large-language models, structured electronic health record data, and clinical time-series features to predict acute intracranial vascular event (AIVE) at stroke alert activation. Using retrospective stroke alerts from the Mount Sinai Health System (2011–2021), we extracted demographics, comorbidities, medications, vital signs, laboratory results, and clinical notes documented before alert activation. AIVE was defined using radiographic gold standards as a composite of acute cerebrovascular and selected acute intracranial diagnoses. Notes were modeled using Clinical-Longformer, MxBAI embeddings with a feed-forward neural network, and BioWordVec embeddings with logistic regression; structured-data models used complementary hyperparameter-tuned workflows. Model outputs were combined in a late-fusion stacked ensemble and evaluated on internal and within-system held-out hospital test sets. Among 16,512 stroke alerts, 9,013 (55.6%) were due to AIVE. The multimodal model achieved an AUROC of 0.73 (95% CI, 0.71–0.75), positive predictive value of 0.68, sensitivity of 0.76, F1-score of 0.72, and specificity of 0.56 on internal testing, with similar held-out hospital performance. This approach may support post-activation stroke-alert risk stratification and reduce false-positive evaluations.
Funding
This study was supported by the National Center for Advancing Translational Sciences (CTSA grant UL1TR004419 to B.R.K.), the American Heart Association (grant 857015 to L.K.S. and 26RIRA1642728 to B.R.K.), the Mount Sinai Neuro-Ophthalmology Research Fund (M.J.K.), and the National Institutes of Health (R01DK133539 and R01HL167050 to G.N.N.). The funders had no role in study design, data collection, analysis, interpretation, or manuscript preparation.
Author information
Authors and Affiliations
Clinical Neuro-Informatics Center, Icahn School of Medicine at Mount Sinai, New York, NY, USA
Asala N. Erekat, Mark J. Kupersmith & Benjamin R. Kummer
Department of Neurology, Icahn School of Medicine at Mount Sinai, New York, NY, USA
Asala N. Erekat, Laura K. Stein, Adam M. Karp, Ankita Tripathi & Benjamin R. Kummer
Icahn School of Medicine at Mount Sinai, New York, NY, USA
Margaret H. Downes
Division of Neuroradiology, Department of Radiology, Icahn School of Medicine at Mount Sinai, New York, NY, USA
Bradley N. Delman
Division of Data-Driven Medicine (D3M), Icahn School of Medicine at Mount Sinai, New York, NY, USA
Girish N. Nadkarni
Department of Ophthalmology, Icahn School of Medicine at Mount Sinai and New York Eye and Ear Infirmary, New York, NY, USA
Mark J. Kupersmith
Department of Neurosurgery, Icahn School of Medicine at Mount Sinai, New York, NY, USA
Mark J. Kupersmith
Windreich Department of Artificial Intelligence and Human Health, Icahn School of Medicine at Mount Sinai, New York, NY, USA
Benjamin R. Kummer
Authors
- Asala N. Erekat
- Margaret H. Downes
- Laura K. Stein
- Bradley N. Delman
- Adam M. Karp
- Ankita Tripathi
- Girish N. Nadkarni
- Mark J. Kupersmith
- Benjamin R. Kummer
Corresponding author
Correspondence to Benjamin R. Kummer.
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Erekat, A.N., Downes, M.H., Stein, L.K. et al. Predicting acute intracranial vascular events in stroke alerts using large-language models and structured data. npj Cardiovasc Health (2026). https://doi.org/10.1038/s44325-026-00158-x
Received: 01 July 2026
Accepted: 02 October 2026
Published: 22 October 2026
DOI: https://doi.org/10.1038/s44325-026-00158-x
来源:Nature 健康护理主题 · nature.com