Related Experiment Video
Updated: Jun 10, 2025
13:53
The Helsinki Rat Microsurgical Sidewall Aneurysm Model
Published on: October 12, 2014
12.2K
Artificial intelligence applied to development of predictive stability model for intracranial aneurysms
Junmin Tao1,2, Wei Wei2,3, Meiying Song1
1Department of Epidemiology, School of Public Health, Dalian Medical University, No. 9, West Section of Lvshun South Road, Lvshunkou District, Dalian, Liaoning Province, China.
European Journal of Medical Research
|October 18, 2024
Summary
Machine learning models accurately predict intracranial aneurysm (IA) rupture risk. Age, white blood cell count (WBC), and uric acid (UA) are key predictors of IA stability.
Area of Science:
- Neurosurgery
- Medical Imaging
- Machine Learning
Background:
- Intracranial aneurysms (IAs) pose a significant risk of rupture.
- Predicting IA rupture is crucial for timely intervention and patient management.
Purpose of the Study:
- To develop and evaluate machine learning models for predicting early intracranial aneurysm rupture risk.
- To compare the performance of different artificial intelligence models in IA stability prediction.
Main Methods:
- Collected data from 989 IA patients (CT angiography) including clinical, blood, and morphological parameters.
- Defined IA instability as rupture or growth ≥0.5 mm within 1 month.
- Developed and validated Random Forest (RF), Support Vector Machine (SVM), and Artificial Neural Network (ANN) models.
Main Results:
- RF models achieved AUCs of 0.734-0.809 on the test set.
- SVM models showed variable performance, with test set AUCs ranging from 0.699-0.806.
- ANN models demonstrated test set AUCs from 0.680-0.860.
- Age, white blood cell count (WBC), and uric acid (UA) were identified as significant predictors of IA stability.
Conclusions:
- AI-based models for IA stability prediction show good clinical applicability.
- Age, WBC, and UA are important predictors for assessing IA stability risk.
