Dynamic Plaque Evaluation Based on Ultrasound Video Images

Due to aging and lifestyle habits, lesions called "plaques" may form in the carotid arteries. Plaques have different stability depending on their tissue properties, and particularly unstable plaques are known to increase the risk of cerebral infarction. Among these, mobile plaques that exhibit dynamic characteristics are considered to have particularly high risk of rupture, and accurately evaluating these is clinically important. In this research, we are developing systems that more accurately evaluate signs of plaque rupture by analyzing the dynamic characteristics of mobile plaques in carotid ultrasound examinations, aiming to improve the prediction accuracy of cerebral infarction risk. Technical approach: - Automatic plaque detection and segmentation using deep learning - Quantitative analysis of plaque deformation with pulsation - Multimodal machine learning utilizing plaque surface information