Applying Machine Learning to Carotid Sonographic Features for Recurrent Stroke in Patients With Acute Stroke
Shih-Yi Lin1,2, Kin-Man Law3,4, Yi-Chun Yeh3
1Graduate Institute of Biomedical Sciences, College of Medicine, China Medical University, Taichung, Taiwan.
Insights
Machine learning effectively predicts recurrent stroke using carotid ultrasound data. The CatBoost model identified anticoagulation, NSAID use, and subclavian artery resistive index as key predictors, showing high accuracy.
Area of Science:
- Medical Imaging
- Machine Learning
- Stroke Prediction
Background:
- Carotid sonographic features are established stroke predictors.
- Large-scale studies on machine learning for recurrent stroke prediction are limited.
Purpose of the Study:
- To investigate the efficacy of machine learning in predicting recurrent stroke using carotid sonographic features.
- To identify key sonographic and clinical predictors for recurrent stroke.
Main Methods:
- Retrospective analysis of 2,411 patients' electronic medical records from a tertiary medical center.
- Carotid ultrasound data and clinical information were collected within 30 days of the first acute stroke.
- Machine learning models, including CatBoost, were developed and validated using Python and scikit-learn.
Main Results:
- The CatBoost model achieved the highest area under the curve (0.844) for recurrent stroke prediction.
- 43 carotid sonographic features were analyzed, with anticoagulation, NSAID use, and left subclavian artery resistive index identified as top predictors.
- Multiple machine learning models were compared, with CatBoost outperforming others.
Conclusions:
- The CatBoost machine learning model demonstrates high efficiency and optimal performance in classifying recurrent stroke.
- Anticoagulation medication use, NSAID use, and left subclavian artery resistive index are significant predictors of recurrent stroke.
- Machine learning analysis of carotid sonographic features offers a promising approach for recurrent stroke prediction.
Background:
Although carotid sonographic features have been used as predictors of recurrent stroke, few large-scale studies have explored the use of machine learning analysis of carotid sonographic features for the prediction of recurrent stroke.
Methods:
We retrospectively collected electronic medical records of enrolled patients from the data warehouse of China Medical University Hospital, a tertiary medical center in central Taiwan, from January 2012 to November 2018. We included patients who underwent a documented carotid ultrasound within 30 days of experiencing an acute first stroke during the study period. We classified these participants into two groups: those with non-recurrent stroke (those who has not been diagnosed with acute stroke again during the study period) and those with recurrent stoke (those who has been diagnosed with acute stroke during the study period). A total of 1,235 carotid sonographic parameters were analyzed. Data on the patients' demographic characteristics and comorbidities were also collected. Python 3.7 was used as the programming language, and the scikit-learn toolkit was used to complete the derivation and verification of the machine learning methods.
Results:
In total, 2,411 patients were enrolled in this study, of whom 1,896 and 515 had non-recurrent and recurrent stroke, respectively. After extraction, 43 features of carotid sonography (36 carotid sonographic parameters and seven transcranial color Doppler sonographic parameter) were analyzed. For predicting recurrent stroke, CatBoost achieved the highest area under the curve (0.844, CIs 95% 0.824-0.868), followed by the Light Gradient Boosting Machine (0.832, CIs 95% 0.813-0.851), random forest (0.819, CIs 95% 0.802-0.846), support-vector machine (0.759, CIs 95% 0.739-0.781), logistic regression (0.781, CIs 95% 0.764-0.800), and decision tree (0.735, CIs 95% 0.717-0.755) models.
Conclusion:
When using the CatBoost model, the top three features for predicting recurrent stroke were determined to be the use of anticoagulation medications, the use of NSAID medications, and the resistive index of the left subclavian artery. The CatBoost model demonstrated efficiency and achieved optimal performance in the predictive classification of non-recurrent and recurrent stroke.
More Related Videos
04:09Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma
Published on: October 10, 2018
06:45Author Spotlight: Integrated Photoacoustic, Ultrasound, and Angiographic Tomography (PAUSAT) for NonInvasive Whole-Brain Imaging of Ischemic Stroke
Published on: June 2, 2023