Related Experiment Video
Updated: May 28, 2025
09:21
Optimized Management of Endovascular Treatment for Acute Ischemic Stroke
Published on: January 18, 2018
11.9K
Large Language Models-Supported Thrombectomy Decision-Making in Acute Ischemic Stroke Based on Radiology Reports:
Jonathan Kottlors1, Robert Hahnfeldt1, Lukas Görtz1
1Institute for Diagnostic and Interventional Radiology, Faculty of Medicine and University Hospital Cologne, University of Cologne, Cologne, Germany.
Journal of Medical Internet Research
|February 13, 2025
Summary
Large language models (LLMs) show promise in aiding medical decisions for acute ischemic stroke (AIS). This AI tool accurately assessed mechanical thrombectomy (MT) eligibility using radiology reports and clinical data.
Area of Science:
- Artificial Intelligence in Medicine
- Natural Language Processing in Healthcare
Background:
- Generative pretrained transformer large language models (LLMs) represent a significant AI advancement, capable of humanlike text responses.
- While LLMs offer potential for medical data integration and clinical decision-making, their specific applications remain under investigation.
Purpose of the Study:
- To evaluate the feasibility of using LLMs for medical decision-making in acute ischemic stroke (AIS).
- To assess the integration of radiology reports and clinical data for mechanical thrombectomy (MT) therapy decisions using LLMs.
Main Methods:
- A retrospective study included 100 AIS patients, with 50% indicated for MT.
- An LLM analyzed computed tomography reports, neurological symptoms, onset information, and patient age.
- AI model performance was benchmarked against expert consensus, calculating sensitivity, specificity, and accuracy.
Main Results:
- The AI model achieved 88% overall accuracy, 96% specificity, and 80% sensitivity for MT indication.
- The area under the curve for the report-based MT decision was 0.92.
Conclusions:
- LLMs demonstrated promising accuracy in determining MT eligibility for AIS patients based on integrated data.
- These findings highlight the potential of LLMs for radiological and medical data integration.
- The study suggests LLMs can serve as augmented decision-support systems in clinical practice.
