好色先生

好色先生

Explore the latest content from across our publications
Join The AAN

Log In

Forgot Password?
Create New Account

Loading... please wait

Abstract Details

Artificial Intelligence in Acute Stroke Diagnostics: Application in Large Vessel Occlusions
Cerebrovascular Disease and Interventional Neurology
P2 - Poster Session 2 (5:30 PM-6:30 PM)
1-006
To systematically review the literature of Artificial Intelligence (AI) use in stroke diagnostics and characterize the use of AI products currently available for detecting a Large Vessel Occlusion (LVO) and expediting its treatment. 
Acute stroke caused by LVOs require emergent detection and, if the patient is a suitable candidate, rapid intervention via thrombectomy.  LVO detection is technically challenging, subject to time delays, and inconsistent between hospitals. Use of AI in stroke provides a method to standardize rapid, sensitive LVO detection and is now available for clinical use.  Stroke AI may provide a critical tool to increase accuracy of LVO detection and automatically activate treatment teams.  

A systematic review was performed using PubMed for all articles in English using predetermined search terms, including “artificial intelligence” and “stroke”.  A total of 16 studies met the inclusion criteria.  Characterization of AI LVO detection technology currently used in stroke was also performed by review of FDA documents, manufacturer websites, and personal communication.    

16 studies were identified that reviewed AI in acute stroke diagnostics.  Stroke AI literature and applications included: technical principles of AI machine learning, studies of stroke imaging (e.g. CTA AI to detect LVOs), and use of “big data”.   Stroke onset time, clinical examination, core and penumbral volume may be combined with AI for highly accurate LVO detection rates to significantly expedite patient selection for rapid thrombectomy.  We also provide the first description of image processing used by AI for LVO detection, including performance measures for such. 

Artificial intelligence is a critical tool increasingly used for optimal and rapid LVO stroke diagnostics.  AI used to assist interpretation of traditional stroke imaging technology may reduce inefficiency and expedite stroke treatment, therefore minimizing long-term morbidity and mortality.

Authors/Disclosures

PRESENTER 		No disclosure on file