Related Experiment Video
Updated: Jul 18, 2025
07:31
Implementation of a Real-Time Psychosis Risk Detection and Alerting System Based on Electronic Health Records using CogStack
Published on: May 15, 2020
7.1K
Development and validation of a machine learning-based prognostic risk stratification model for acute ischemic
Kai Wang1,2, Tao Hong3,4,5, Wencai Liu6
1Department of Neurology, The Second Affiliated Hospital of Xuzhou Medical University, Xuzhou, Jiangsu, China.
Scientific Reports
|August 23, 2023
Summary
Machine learning accurately predicts acute ischemic stroke (AIS) prognosis using neuron specific enolase (NSE), homocysteine (HCY), and other factors. The Random Forest model offers a reliable tool for assessing patient outcomes and guiding treatment strategies.
Area of Science:
- Neurology
- Biomedical Informatics
- Data Science
Background:
- Acute ischemic stroke (AIS) is a leading cause of long-term disability globally.
- Accurate prognosis prediction is crucial for effective AIS intervention and treatment.
Purpose of the Study:
- To develop a reliable machine learning (ML)-based model for predicting prognosis in AIS patients.
- To identify key prognostic factors influencing AIS outcomes.
Main Methods:
- Retrospective data collection from 677 AIS patients.
- Identification of prognostic factors using logistic analysis.
- Development and evaluation of ML models, including Random Forest (RF).
- Feature importance assessed using Shapley Additive explanations (SHAP).
Main Results:
- Poor prognosis was observed in 30.9% of patients.
- Six key variables identified: neuron specific enolase (NSE), homocysteine (HCY), S-100β, dysphagia, C-reactive protein (CRP), and anticoagulation.
- The RF model achieved the highest AUC of 0.908.
- NSE was the most impactful predictor, followed by HCY, S-100β, dysphagia, CRP, and anticoagulation.
Conclusions:
- The developed RF model provides accurate AIS prognosis prediction.
- NSE, HCY, CRP, S-100β, anticoagulation, and dysphagia are significant predictors of poor prognosis.
- An online tool based on the RF model can assist clinicians in optimizing AIS patient treatment.
