Related Experiment Video
Updated: Jan 9, 2026
09:10
Determining the Functional Status of the Corticospinal Tract Within One Week of Stroke
Published on: February 22, 2020
9.1K
Development and Validation of a Web-Based Machine Learning Model for Predicting Early Neurological Deterioration
Juan Li1, Huanxian Chang2, Shouyun Du3
1The Neurology Department, Lianyungang Clinical College of Nanjing Medical University, The First People's Hospital of Lianyungang, No. 182 Tongguan North Road, Jiangsu, Lianyungang, 222001, China, 86 18961326515.
Journal of Medical Internet Research
|December 10, 2025
Summary
A new machine learning model, ENDRAS, accurately predicts early neurological deterioration in acute ischemic stroke patients receiving thrombolysis. This tool enables personalized risk stratification and optimized care, potentially improving patient outcomes.
Area of Science:
- Neurology
- Artificial Intelligence in Medicine
- Clinical Decision Support
Background:
- Early neurological deterioration (END) is a significant complication for acute ischemic stroke (AIS) patients treated with intravenous thrombolysis.
- Current methods lack reliability in identifying high-risk patients needing intensified monitoring and interventions.
Purpose of the Study:
- To develop and validate a high-performance machine learning model for predicting END in AIS patients post-thrombolysis.
- To enable personalized, risk-stratified management strategies for improved patient care.
Main Methods:
- A multicenter study involving 1927 AIS patients treated with intravenous thrombolysis.
- Development of the Early Neurological Deterioration Risk Assessment Score (ENDRAS) using 6 clinical variables and XGBoost algorithm.
- Internal and external validation of the model's predictive performance using AUC, accuracy, precision, recall, and F1-score.
Main Results:
- The XGBoost-based ENDRAS demonstrated high predictive performance (AUC=0.988) using readily available clinical parameters.
- A dual-pathway management protocol was established, stratifying patients into low-risk (<29%) and high-risk (≥29%) groups.
- ENDRAS, implemented as a web-based calculator, provides real-time clinical decision support with rapid computation (<0.02 seconds).
Conclusions:
- ENDRAS integrates END prediction into actionable clinical pathways, enhancing post-thrombolysis care through personalized monitoring and interventions.
- The model's robust performance, efficiency, and structured management framework address critical challenges in stroke care and resource utilization.
- Further prospective validation is recommended, but ENDRAS shows potential to significantly improve AIS patient outcomes by enabling early identification of high-risk individuals.
