Carotid Atherosclerosis Phenotypic Characteristics Quantification And Analysis In3D Ultrasound Images

Atherosclerosis (AS) and its sequelae, transient ischemic stroke (TIS) and heart attacks are the leading causes of mortality and morbidity in the developed world, and are increasing in the developing countries. Cardiovascular and Cerebrovascular Diseases (CVDs) represent a staggering mortality, morbidity, and economic cost in the worldwide. Therefore, improved methods to identify patients at increased risk of stroke, to recommend advanced techniques to detect and treat the disease, and monitor them will have an enormous impact.Most TIS are due to the blockage of a cerebral artery by a thrombotic embolus generated at the bifurcation (BF) of the common carotid artery (CCA). Most of these strokes can be prevented by lifestyle/dietary changes, medical and surgical treatment, such as drug therapy, carotid artery stenting (CAS) and Carotid Arterial Endarterectomy (CAE) surgeries. Nowadays, researches on new strategies for treating atherosclerosis, such as improving identification of patients, who are at risk for stroke, and developing sensitive techniques for monitoring of carotid plaque response to therapy, will be great helpful for the management of these patients, and decrease the risk of stroke.B-mode ultrasound (US) had been widely used for AS screen examination and only sensitive for those late stage patients in clinic. Three-dimensional (3D) ultrasound not only has the Two-dimensional (2D) US advantages, but also has special capabilities in early stage AS diagnose and vulnerable plaque identification. Therefore, this thesis research orientation was focused on real-time carotid artery segmentation, quantification of atorvastatin effective evaluation and plaque characterize classification.We propose that3D carotid US can provide cost-effective, sensitive and specific measurement tools. The tools are critically needed to define novel risk factors that aggravate atherosclerosis, and to assess efficacy of therapies for carotid atherosclerosis as the research contents. The research means, therefore, of this study is to verify the non-invasive, quantitative three-dimensional carotid ultrasound imaging technology clinical capability, by applying it as one of the new means for monitoring carotid atherosclerosis change based on the carotid atherosclerotic phenotypic characteristics quantification and analysis. And the carotid atherosclerosis phenotype are carotid morphology, function and other aspects of its performance, such as thickness, size, shape, volume, echo features and drug tolerance, and so on. The objective of this paper is to provide a both low cost, high sensitive method for assessment of carotid plaque progression and regression, and some useful information of plaque reaction to stabilization therapies treatment, by segmentation on longitudinal and transverse sections of the carotid artery, identification of the important risk factors that will distinguish patients who have vulnerable plaques.Our purposes are to develop sensitive and reproducible3D US imaging software tools that will allow semi-automated measurement of the following4imaging carotid atherosclerosis phenotypes:local and global vessel wall intima-media thickness (IMT); total plaque thickness, area (TPA) and volume (TPV); and vessel wall volume (VWV); plaque surface morphology and composition, and to provide software imaging tools for monitoring changes in these measures.In order to achieve that, we propose the following3specific technical objectives:(1) In this paper, an integrated segmentation method for3D US carotid artery based on bifurcation and medial axis was proposed.3D US image was sliced into transverse, coronal and sagittal2D images on the carotid bifurcation point. Then, the three images were processed respectively, and the carotid artery contours and thickness were obtained finally. This paper tries to overcome the disadvantages of current computer aided diagnosis method, such as high computational complexity, easily introduced subjective errors et al. The proposed method could get the carotid artery overall information rapidly, accurately and completely. It could be transplanted into clinical usage for atherosclerosis diagnosis and prevention.The following method, Active Contour Model (ACM) with mathematical morphology, Active Shape Models (ASM), and its improved version (Improved Active Shape Models, I-ASM) and Active Appearance Models (AAM), were proposed on carotid artery transverse view for media-adventitia boundary (MAB) and lumen-intima boundary (LIB) segmentations. Different algorithms were compared between and with the manual segmentation results which was the golden standard. The proposed semi-automated segmentation method could outline the carotid wall and lumen boundaries in5minutes, with a variance a factor of2compared with manual segmentation.(3) The computer simulation phantom and agar phantom were used for the carotid vessel Flow Mediated Dilation (FMD) function measurement validation. And features extraction software tool was developed based on those phantom experiments for quantification drug therapy evaluation. The pathophysiology of carotid plaques indicates that morphological and compositional characterization is necessary for identification of vulnerable plaques and monitoring disease progression and regression. Optimization features, extracted from all the morphology, texture and elastic features, were sent to classifier such as:Back Propagation Neural Network (BPNN) and Support Vector Machine (SVM) for classification, as of84%and91%accuracy respectively. Then, Receiver Operating Characteristic (ROC) and Area Under Curve (AUC) were used for the validation and evaluation.All the proposed methods, including IMT measurement, MAB/LIB segmentation, phantom simulations and features extraction and classification, were testified by the real patient data on3D US images. And the experiments results indicate that the proposed methods can promote the carotid3D US usage for a fast, safe and economical monitoring of the atherosclerotic disease progression and regression during therapy. The proposed method would accelerate the translation of3D US to clinical application and be a significant role for CVDs diagnose and treatment.