What question did this study set out to answer?

The aim is to develop a framework for automatically extracting structured clinical variables from neurointerventional reports in acute ischemic stroke patients.

February 22, 2026American Journal of NeuroradiologyOpen Access

Large Language Models for Accurate Medical Chart Abstraction: Enabling Scalable and Secure AI Deployment in Stroke

View Full Paper

Ask AI

Bookmark

Authors

LWLuoyun WangBrown University CPCarl PortoProvidence College DHD Yong HongBrown University

Discussion

Member takes

Implication

This retrospective study evaluates large language models for automated extraction of clinical variables in stroke, suggesting effective AI deployment.

Key Points

The aim is to develop a framework for automatically extracting structured clinical variables from neurointerventional reports in acute ischemic stroke patients.
Retrospective analysis of 2,416 stroke intervention reports from three hospitals
Annotated eight key clinical variables by non-expert staff as a reference standard
Evaluation of 22 instruction-tuned LLMs using two prompting strategies: Quick Response and Chain-of-Thought
Performance assessed against non-expert annotations and expert ratings
LLaMA3.3-70B demonstrated the highest accuracy at 94.8%
Chain-of-Thought prompting improved performance for inferential variables
LLaMA3.3-70B outperformed non-expert annotations in 7 of 8 variables
AI predictions aligned more closely with expert interpretations than those of non-expert staff

Cite This Study

Wang et al. (2026) studied this question.

synapsesocial.com/papers/699a9d65482488d673cd340e https://doi.org/10.3174/ajnr.a9251

View Full Paper

Ask AI

Bookmark

Implication

This retrospective study evaluates large language models for automated extraction of clinical variables in stroke, suggesting effective AI deployment.

Large Language Models for Accurate Medical Chart Abstraction: Enabling Scalable and Secure AI Deployment in Stroke

Discussion

Member takes

Implication

Key Points

Cite This Study

Large Language Models for Accurate Medical Chart Abstraction: Enabling Scalable and Secure AI Deployment in Stroke

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider