Identification of high-risk basilar artery plaque with HR-VWI-based radiomics and machine learning

The stability of plaques is the key cause of cerebral ischemic in patients with basilar atherosclerosis. High-resolution vessel wall imaging (HR-VWI) is a useful technique for study of plaques. Radiomics are widely used in recently years to discover the radiographic information of diseases and to improve diagnostic performance. In this study, we constructed six machine learning models based on radiomics features from HR-VWI to predict the stability of plaques of the basilar artery, finding support vector machine has the best performance.