Artificial Intelligence Based Comprehensive Assessment System of Carotid Plaque Stability

Stroke is a common clinical disease with high disability and mortality, which seriously threatens human life and health.Carotid atherosclerotic plaque rupture is an important pathogenic basis of ischemic stroke, so judging the stability of plaque has important clinical significance in preventing ischemic stroke.Ultrasound, as a convenient, rapid, noninvasive, radiation-free auxiliary examination technology, is widely used in carotid plaque stability examination.At present, there are many methods to judge the stability of carotid plaque based on ultrasound, including two-dimensional ultrasound, contrast-enhanced ultrasound, ultrasound elastography and so on. However, the results of plaque stability judgment by various technologies deviate greatly, which is not conducive to the development of standardized diagnosis and treatment strategies by clinicians.Studies have shown that because the neovascular epidermal cells in atherosclerotic plaques are imperfect, they are easy to rupture after stress, and the ruptured neovasculature will lead to intraplaque hemorrhage, thus causing plaque shedding, and eventually obstructing the cerebrovascular cause stroke.Contrast-enhanced ultrasound can sensitively detect the distribution and course of blood vessels.The plaque's softness and hardness determine its stability, while the difference of lipids, fibers and calcium in the plaque determines its softness and hardness.Real-time ultrasound elastography can provide tissue mechanical parameters, express the soft and hard of tissue with strain value, and provide important reference information for judging plaque stability.At present, elastography technology is used to reflect the hardness of plaque, so as to further judge its stability.However, the elastography parameters are prone to deviations due to the influence of the selected section and the selected region of interest.Deep learning is the hottest research method in AI at present. [Deep learning is essentially to construct machine learning models with multiple hidden layers, and use large-scale data to train to obtain a large number of more representative feature information, so as to use these features to classify and predict samples. At present, it is widely used in the field of image analysis and plays an important role in medicine.Such as pathological image detection, regulatory genomics research, diagnosis of retinopathy and quantitative analysis of liver fibrosis, etc.Computational Fluid Dynamics (CFD) is a new discipline developed by the combination of numerical calculation and classical fluid mechanics theory. It can make it convenient for researchers to build a geometric model of cardiovascular system, simulate the real structure of vascular wall and blood, and display the results of "numerical experiments" using visualization technology.More intuitive Comprehensive response to changes in hemodynamic parameters In recent years, its application in cardiovascular hemodynamic research has become increasingly widespread, with a large number of relevant literature reports However, no report has been reported on the study of carotid plaque stability and fluid dynamics using this technology Based on the above reasons,This study attempts to use AI technology to automatically identify and quantitatively evaluate the gray scale differences of plaques, elastic image characteristics of plaques, microvessel density division of plaque contrast-enhanced ultrasound, and velocity vector imaging (VVI) to determine plaque surface stress.To study the effect of hemodynamic parameters on carotid plaque using CFD technology, and to establish a systematic comprehensive evaluation system for carotid plaque stability, which integrates two-dimensional plaque information, texture information, microcirculation perfusion information, biomechanical information and blood flow field information, and combines the results of clinical follow-up and collagen fiber content of surgical specimens, MMP9/CD34 and other examinations.