Contralateral artery enlargement predicts carotid plaque progression based on machine learning algorithm models in...

Area of Science:

• Cardiovascular Research
• Medical Imaging
• Machine Learning in Medicine

Background:

• Atherosclerotic plaque vulnerability is a key concern in cardiovascular disease.
• Low shear stress is implicated in the development of atherosclerotic plaques.
• Accurate characterization and prediction of vulnerable plaques are crucial for effective treatment.

Purpose of the Study:

• To characterize low shear stress-induced atherosclerotic plaque in mice using anatomical MRI.
• To develop predictive models for vulnerable plaque using machine learning algorithms.
• To identify key predictive factors for vulnerable plaque progression.

Main Methods:

• A partial carotid artery ligation model was created in apoE-/- and C57BL/6J mice.
• Serial in vivo MRI and ultrasonography were used to assess carotid artery diameter and blood flow.
• Support Vector Machine (SVM) and Decision Tree (DT) algorithms were employed for feature selection and predictive modeling.
• Histological analysis was performed on harvested carotid arteries.

Main Results:

• Neointima formation was observed in both mouse models within 7 days post-surgery.
• ApoE-/- mice developed vulnerable plaque, while C57BL/6J mice showed arterial stenosis after 4 weeks.
• Contralateral carotid artery diameter at 7 days post-surgery was a reliable predictor of plaque progression.
• SVM achieved >87.5% accuracy, 80% sensitivity, and 95% specificity; DT achieved 90% accuracy, sensitivity, and specificity.

Conclusions:

• SVM and DT models are suitable for identifying vulnerable plaque progression in mice.
• Early contralateral artery enlargement predicts vulnerable plaque development.
• These findings may assist in optimizing clinical workflows for patient selection in treatment.