Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Video

Updated: Jan 11, 2026

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

Predicting Acute Cerebrovascular Events in Stroke Alerts Using Large-Language Models and Structured Data.

Asala N Erekat^1,2, Margaret H Downes³, Laura K Stein²

¹Clinical Neuro-Informatics Center, Icahn School of Medicine at Mount Sinai, New York, NY.

Medrxiv : the Preprint Server for Health Sciences

|November 19, 2025

Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Real-World Associations of KidneyIntelX Risk Stratification With Guideline-Directed Therapy, Kidney Outcomes, and Metabolic Trajectories in Early Diabetic Kidney Disease.

Diabetes, obesity & metabolism·2026

Same author

Education Research: Quality of Narrative Feedback Generated by a Large Language Model Compared With Expert Faculty for Case-Based Learning in Neurology Education.

Neurology. Education·2026

Same author

Identification of a robust QTL hotspot across environments governing drought tolerance in lentil (Lens culinaris Medikus).

BMC genomics·2026

Same author

Genomic analyses implicate hormonal and metabolic dysregulation in polycystic ovary syndrome.

Nature genetics·2026

Same author

How to meaningfully evaluate AI in clinical medicine.

Nature medicine·2026

Same author

Large Language Models in Cardiology: Systematic Review.

JMIR cardio·2026

Same journal

Predicting Chemotherapy Response from Staging Laparoscopy Images.

medRxiv : the preprint server for health sciences·2026

Same journal

Development and External Validation of a Machine Learning Model for 10-Year Ischemic Stroke Risk Prediction in Diverse Populations.

medRxiv : the preprint server for health sciences·2026

Same journal

MCH-Guard: Multimodal Machine Learning Framework for Risk Stratification of Cerebral Microhemorrhage Risk in the Alzheimer's Disease Neuroimaging Initiative.

medRxiv : the preprint server for health sciences·2026

Same journal

Genetic and maternal environmental contributions to estimated fetal weight at 20 weeks gestation compared with birthweight.

medRxiv : the preprint server for health sciences·2026

Same journal

Better immediate declarative memory is associated with forgetting during locomotor adaptation in chronic stroke and in older adults.

medRxiv : the preprint server for health sciences·2026

Same journal

An empirical Bayes framework for burden and dispersion association tests helps prioritize rare variants associated with Alzheimer's disease.

medRxiv : the preprint server for health sciences·2026

See all related articles

This summary is machine-generated.

Machine learning models integrating electronic health records and large-language models can predict acute cerebrovascular disease (ACD) at stroke alert activation, aiming to reduce false positives and optimize stroke triage.

Area of Science:

Neurology
Medical Informatics
Machine Learning

Background:

Acute stroke alerts are frequently triggered by non-cerebrovascular conditions, causing false positives.
These false alarms strain clinical resources and increase diagnostic uncertainty.

Purpose of the Study:

To develop machine learning (ML) models for predicting acute cerebrovascular disease (ACD) at stroke alert activation.
Integrate large-language models (LLMs), structured EHR data, and clinical time series data for improved prediction.

Main Methods:

Retrospective analysis of stroke alerts from 2011-2021 at Mount Sinai Health System.
Extracted structured EHR data and unstructured clinical notes; processed notes using embeddings (word, BioWordVec, LLM).
Trained individual ML models (unstructured, structured, multimodal ensemble) using auto-ML and evaluated using AUROC, PPV, sensitivity, and F1-score.

Keywords:

Acute cerebrovascular disease electronic health records large language models machine learning natural language processing predictive modeling stroke alerts stroke care

More Related Videos

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Published on: December 15, 2023

Related Experiment Videos

Last Updated: Jan 11, 2026

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Published on: December 15, 2023

Main Results:

The study analyzed 16,512 stroke alerts, with 54.6% confirmed as ACD.
The multimodal model achieved an AUROC of 0.72, outperforming individual models.
A structured model focusing on demographics, comorbidities, and medications showed the highest sensitivity (0.95).

Conclusions:

A multimodal ML model was developed to predict ACD during stroke alerts.
This approach shows potential for optimizing stroke triage and reducing unnecessary alerts.